Modelling non-industrial private forest management in Galicia: a new approach for forest research

MODELO DE GESTIQUE PARA LA INVESTIGACIÓN FORESTAL MODELLING NON-INDUSTRIAL PRIVATE FOREST MANAGEMENT IN GALICIA: A NEW APPROACH FOR FOREST RESEARCH Verónica Rodríguez Vicente Departamento de Enxeñería Agroforestal TESIS DOCTORAL Universidade de Santiago de Compostela Lugo 2010 Prof. Dr. Manuel Fco. Marey Pérez Tutor (co-autor) MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research Verónica Rodríguez-Vicente Departamento de Enxeñería Agroforestal TESIS DOCTORAL Universidade de Santiago de Compostela Lugo 2010 Prof. Dr. Manuel Fco. Marey-Pérez Tutor (co-autor) MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research UNIVERSIDAD DE SANTIAGO DE COMPOSTELA ESCUELA POLITÉCNICA SUPERIOR DE LUGO DEPARTAMENTO DE ENXEÑARÍA AGROFORESTAL MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research MEMORIA QUE PARA OPTAR AL GRADO DE DOCTOR PRESENTA LA INGENIERA VERÓNICA RODRÍGUEZ VICENTE. REALIZADA BAJO LA DIRECCIÓN DEL DOCTOR MANUEL FRANCISCO MAREY PÉREZ 2011 ISBN 978-84-9887-725-0 (Edición digital PDF) MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research MOTIVACIÓN En 1994 inicio la titulación de Ingeniería Técnica en Explotaciones Forestales por la Universidad de Santiago de Compostela en la Escuela Politécnica Superior de Lugo, para continuar, en el mismo centro, la titulación de Ingeniería de Montes en 1998. Realizar el proyecto fin de carrera en el Departamento de Ingeniería Agroforestal de la Escuela Politécnica Superior de Lugo en el período de 2000-2001, me permitió completar mi formación académica como becaria de dicho departamento hasta el año 2005, determinando además el inicio de mis Estudios de Tercer Ciclo entre 2002-2003 con la presentación del Trabajo de Iniciación a la Investigación titulado La propiedad forestal en Europa: situación y problemática actual. A finales del año 2005, inicio mi actividad profesional en la Asociación Sectorial Forestal Galega (ASEFOGA), sectorial de Unións AgrariasUPA (Unión de Pequeños Agricultores), donde tuve y tengo la oportunidad de desarrollarme como ingeniera de montes. El trabajo diario en esta asociación me ha permitido asesorar y ayudar en materia de gestión y planificación a un gran número de propietarios agroforestales asociados (cifra que casi asciende a 10.000), así como colaborar y participar, junto con otras entidades y asociaciones profesionales, con la administración pública en aquellos puntos de relevancia para este sector estratégico de Galicia. Así, la actividad técnica desarrollada en la anterior asociación de propietarios forestales me ha permitido y me permite actualmente ser miembro de los siguientes órganos consultivos:  Entidade Galega Solicitante da Certificación Forestal Rexional PEFC (presidencia)  Asociación Galega Promotora da Certificación Forestal (vocal)  Consello Forestal de Galicia (vocal)  Consello Galego de Medio Ambiente e Desenvolvemento Sostible (vocal)  Comisión Técnica de Prezos e Valores do Banco de Terras de GaliciaBANTEGAL (suplente)  Xurado provincial de clasificación de montes veciñais en man común de A Coruña (vocal). Mi labor como técnico de extensión forestal, junto con la representatividad que la propiedad privada individual tiene en Galicia, determinó que la línea de investigación, reflejada en la presente tesis doctoral, se centrara en el estudio empírico de la conducta de gestión forestal desempeñada por este tipo de propiedad en la región, generando un análisis socioeconómico pormenorizado de dicha actividad y articulando ejes clave de mejora. El fin de la presente tesis doctoral es, en definitiva, desarrollar una línea de investigación estratégica para el medio agrario de Galicia, ser el modesto inicio de un ciclo de mejora continua. Aportar paulatina e incrementalmente conocimientos y experiencia que permitan comprender las pautas de gestión forestal desempeñadas por este tipo de propiedad privada en la región a partir del conocimiento de sus objetivos y motivaciones para con la tierra. INVESTIGAR E INNOVAR COLABORAR Y PARTICIPAR CAPACITAR Y MOTIVAR DEMANDAR Y VALORAR APORTAR APORTAR CONSENSUAR CONSENSUAR PROFESIONALIZAR PROFESIONALIZAR MEJORAR MEJORAR Sólo así se podrá diseñar y trabajar en iniciativas públicas que promuevan una actividad forestal profesional acorde a los principios de gestión forestal sostenible. A mi familia MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research ÍNDICE DE CONTENIDOS Resumen 1 Introducción 3 Objetivo 6 Material y métodos 8 Resultados y discusión 11 Conclusiones 15 Bibliografía 16 ANEXOS 19 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research ANEXOS ARTÍCULOS I-V La presente tesis se basa en los siguientes artículos científicos relacionados a continuación en números romanos: I. Rodríguez-Vicente V., Marey-Pérez M.F., 2008. Sistemas de apoio á propiedade privada forestal e a súa aplicación en Galicia. Revista Galega de Economía 17(1), 111-130 II. Rodríguez-Vicente V., Marey-Pérez M.F., 2009. Characterization of nonindustrial private forest owners and their influence on forest management aims and practices in Northern Spain. Small-Scale Forestry 8 (4), 479513 III. Rodríguez-Vicente V., Marey-Pérez M.F., 2008. Assessing the role of the family unit in individual private forestry in northern Spain. Scandinavian Journal of Forest Research 23 (1), 53-77. IV. Rodríguez-Vicente V., Marey-Pérez M.F., 2009. Land-use and land-base patterns in non-industrial private forests: Factors affecting forest management in Northern Spain. Forest Policy and Economics 11 (7), 475-490. V. Rodríguez-Vicente V., Marey-Pérez M.F., 2010. Analysis of individual private forestry in northern Spain according to economic factors related to management. Journal of Forest Economics 16 (4), 269-295. Los artículos publicados son reproducidos con el permiso de las respectivas revistas. 1 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research RESUMEN El sector forestal gallego ha experimentado en las últimas décadas una fuerte expansión en superficie sobre tierras antiguamente dedicadas a la agricultura y a la ganadería, incremento superficial que no ido acompañado de una repercusión económica y social similar. Partiendo de la importante representación de la propiedad privada de la tierra en la región, y en relación con el concepto de gestión forestal sostenible, se revisan en primer lugar diferentes metodologías de planificación y seguimiento forestal implementadas en otras regiones europeas atendiendo a su potencialidad como alternativas de avance y de desarrollo del sector forestal gallego. Estas líneas - redes de contabilidad, modelos de cooperación, programas de educación y asesoramiento y medidas públicas de apoyo económico - son analizadas con el fin último de generar criterios para su adaptación a la comunidad gallega. En segundo lugar, teniendo en cuenta el papel clave de la propiedad privada individual en la gestión de las tierras forestales de numerosas áreas rurales del mundo (non-industrial private forest ownership, NIPF ownership), los siguientes cuatro artículos de la presente tesis doctoral se centran particularmente en explicar y modelar la gestión NIPF - plantación, selvicultura y corta forestalmediante el análisis empírico de atributos vinculados al perfil del propietario, unidad familiar, dinámica en los usos agroforestales de la tierra y características de la propiedad, y factores económicos. Así, un total de 103 NIPF de la región Mariña Oriental (noroeste de Galicia) fueron entrevistados personalmente en marzo de 2004 con el objetivo de conocer sus prácticas de gestión forestal durante el período 1999-2003. Los resultados sugieren que: (i) la ocupación profesional es el principal factor que, directa o indirectamente, influye en la conducta de gestión forestal, en concreto los antecedentes agrícolas del propietario; (ii) el patrón de adquisición de las tierras, los reempleos forestales, la disponibilidad de maquinaria agroforestal en la explotación, junto con la mano de obra familiar y el asesoramiento técnico en la actividad, son también factores significativos en las prácticas forestales analizadas; (iii) la gestión forestal responde principalmente al principio de capitalización e incremento de la productividad de la tierra como capital activo, siendo determinante el tamaño y grado de parcelación de la propiedad, así como el propio interés del propietario en la producción maderera; (iv) unos ingresos forestales atractivos y unas condiciones favorables para el mercado de madera son ítems clave en la iniciación y continuidad de la actividad forestal, siendo relevante en el tipo de gestión desarrollada las circunstancias personales y familiares. Los resultados pueden ser de interés para el diseño, planificación e implementación de medidas públicas de investigación y promoción que incentiven una gestión forestal sostenible al amparo del desarrollo rural entre la población NIPF. En la región, la actividad forestal podría ser una actividad económica valiosa, si bien actualmente no es valorada como tal. Palabras clave: agricultura; dinámica en los usos de la tierra; prácticas de plantación, selvicultura y corta forestal; propietario privado forestal no-industrial (propietario NIPF); rentabilidad forestal. 2 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research ABSTRACT The Galician forestry sector has strongly expanded during the last few decades over lands formerly devoted to agrarian and livestock activities, increment in land which has not been accompanied by a similar socioeconomic repercussion. Given the important representation of the private land ownership in the region, and dealing with the concept of sustainable forest management, different potential methodologies for forest planning and monitoring implemented in other European regions are firstly reviewed as potential alternatives to advance and development of the Galician forestry sector. These guidelines - accountancy data network, co-operation models, education and advice programmes and public measures of economic supportare analysed with the final aim of proposing criteria for their adaptation to the Galician community. Secondly, taking into account non-industrial private forest (NIPF) ownership as a key component in most rural areas worldwide, the following four articles are specifically centred on explaining and predicting NIPF owner land management - planting, silvicultural and harvesting practices - by analysing attributes of landowner profile, family unit, dynamics in farming and forestry practices and landholding characteristics, and forest economics. In March 2004, 103 forest landowners were personally interviewed about their commitment to and involvement in land management during 1999-2003, considering a forest region in northern Galicia, the Mariña Oriental. Results suggest that: (i) professional occupation, particularly farming background, is the main factor affecting, either directly or indirectly, the forest management behaviour; (ii) pattern of land acquisition, household dependence on forest products for self-consumption, the availability of machinery, in addition to family labour force and technical guidance in forestry, are all significantly related to the ability to manage and use forestland; (iii) forest management mainly responds to investment and increasing the productivity of the land as a capital asset, which is directly influenced by the size and degree of parcellation of the holding, and directly or indirectly related to the owner's interest in timber production; (iv) attractive forest returns and favourable market conditions for timber production are significant factors for investment in and development of forestry, with personal and family conditions also being important factors in explaining the type of land management carried out. These findings may be of interest in designing, planning and implementing research and policy measures that allow NIPF landowners to promote sustainable forestry for rural development. In the region, forestry could be a valuable economic activity but it is not considered as such today. Keywords: farming; forest profitability; land-use change; non-industrial private forest (NIPF) owner; planting, silviculture and harvesting practices. 3 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research INTRODUCCIÓN EL RECURSO MONTE Se entiende por monte todo terreno en el que vegetan especies forestales arbóreas, arbustivas, de matorral o herbáceas, sea espontáneamente o procedan de siembra o plantación, que cumplan o puedan cumplir funciones ambientales, protectoras, productoras, culturales, paisajísticas o recreativas. Concepto de monte, artículo 5 de la Ley 43/ 2003, de 21 de noviembre, de montes. Se define monte arbolado aquel monte poblado con especies forestales arbóreas como manifestación vegetal dominante y con una fracción de cabida cubierta (fcc) igual o superior al 20%. Este concepto incluiría las dehesas de base cultivo o pastizal con labores siempre que la fcc arbolada sea igual o superior al 20%. También comprendería los terrenos con plantaciones monoespecíficas o poco diversificadas de especies forestales arbóreas, sean autóctonas o alóctonas, siempre que la intervención humana sea débil y discontinua. Se entiende por monte arbolado ralo el terreno de monte poblado con especies arbóreas como manifestación botánica dominante y con una fcc comprendida entre el 1020%; también el terreno con especies de matorral o pastizal natural como manifestación vegetal dominante, pero con una presencia de árboles forestales importante cuantificada por una fcc arbórea igual o superior al 10% e inferior al 20%, incluyéndose aquí las dehesas de base cultivo cuando la fcc forestal esté entre el 1020%. Por su parte, se define monte arbolado disperso como el terreno ocupado por especies arbóreas como presencia vegetal dominante y con una fcc entre el 510%; igualmente espacio de tierra conteniendo matas, malezas y herbazales naturales como fenómeno botánico preponderante, pero con una manifestación de árboles forestales que cubran una fcc sobre el suelo igual o superior al 5% y menor del 10%. Las dehesas con base cultivo no se clasificarían dentro de este grupo aunque la fcc de los árboles esté entre el 510%, pues la importancia del uso agrícola anularía prácticamente a los demás. Finalmente se entiende por monte desarbolado el terreno poblado con especies de matorral o/y pastizal natural o con débil intervención humana como manifestación vegetal dominante con presencia o no de árboles forestales, pero en todo caso con la fcc inferior al 5%. Definiciones tomadas del Plan Forestal Español del MMA (2002), a partir del Inventario Forestal Nacional (IFN). En torno al 69% del territorio de Galicia (Figura 1) es clasificado actualmente como monte (2.039.574,11 ha), siendo éste un recurso clave no sólo en el paisaje de la región, sino también en su economía e identidad cultural (MMA 1998). El monte gallego representa casi el 8% del monte estatal, cuando esta Comunidad Autónoma (CCAA) no llega a alcanzar el 6% de la superficie geográfica nacional; en 2001, la superficie forestal gallega ascendía a 688,1 ha por cada 1.000 habitantes, un 71,9% más que la media española (Xunta de Galicia 2005). Con esta superficie forestal, Galicia sería la sexta CCAA en contribuir al monte estatal, por detrás de Castilla y León (17,2% del monte estatal), Andalucía (16,5%), CastillaLa Mancha (13,2%), Aragón (9,4%) y Extremadura (8,7%). Figura 1. Situación de la CCAA de Galicia, noroeste de España Península Ibérica GALICIA Según los datos del III Inventario Forestal Nacional (IFN) del MMA (1998), los montes arbolados en Galicia abarcan 1.276.651,64 ha (62,6% del terreno forestal gallego y en torno al 11% de la superficie forestal arbolada de España). Mientras, las zonas de arbolado ralo y disperso se extienden por 82.140,92 y 23.864,22 ha, respectivamente, es decir, un 4,0 y 1,2%, 10 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research residentes permanentemente en A Mariña Oriental, el tamaño muestral se diseñó finalmente con un total de 103 propietarios, organizados en cuatro estratos, a contactar y entrevistar en persona. Dicha muestra de propietarios NIPF gestionaba el 12% del monte y el 13% del monte arbolado productivo en el área de estudio. Diseño del cuestionario y realización de encuestas Con el objetivo de obtener la mayor calidad y cantidad posible de información al menor coste fue necesario ensayar previamente varios cuestionarios personales antes de decidir una versión final y definitiva que recopilase toda aquella información de interés relacionada con las prácticas de gestión forestal (plantación, selvicultura y corta), el perfil del propietario, su unidad familiar, explotación y usos de la tierra, y economía para el período de 1999 a 2003. Las entrevistas fueron realizadas en dos fases en el mes de marzo de 2004. La primera fase consistía en una entrevista telefónica entre las 20:00 y 22:00 horas, preguntando por la posibilidad de participar en el estudio. Si el propietario accedía, el entrevistador fijaba la realización de la entrevista personal entre 12 días después (segunda fase). Cada entrevista personal supuso una duración media de 36 minutos. Las variables de estudio para la presente investigación se basaron en la información obtenida en las entrevistas personales, añadiéndose datos complementarios del Catastro. Mediante software SAS/ STATTM (versión 9.1), dicha información fue redefinida y codificada en variables nominales, ordinales y binarias con el fin de resumir los datos del estudio y cumplir posteriormente las premisas de los análisis estadísticos a emplear. Análisis estadísticos Puesto que la población de estudio no atendía las premisas establecidas por los tests de bondad-de-ajuste de KolmogorovSmirnov ni de normalidad K-S o de homogeneidad de varianzas de Levene, los artículos IIV incluyen análisis estadísticos no-paramétricos en función del tipo de variable analizada (continua, nominal u ordinal y binaria) con el objetivo de explicar estadísticamente, de forma completa y fiable, la conducta de gestión forestal realizada por los propietarios NIPF de estudio en A Mariña Oriental en base a cuatro puntos clave: el perfil del propietario, la unidad familiar, la explotación y usos de la tierra, y la economía. En primer, se comprobaba la fuerza y significación de correlación lineal entre variables para, a continuación, testar la existencia de diferencias significativas para un nivel de confianza del 95% y un nivel mínimo de significación estadística del 5%. Mediante el uso de tablas de contingencia, la relación estadística entre variables nominales, ordinales y/ o binarias se fundamentó en el coeficiente D de Somers, comprobar posteriormente la existencia de diferencias significativas en la distribución de frecuencias entre variables mediante tabulación cruzada chi-cuadrado χ2de Pearson. Para los análisis estadísticos entre variables continuas con respecto a variables nominales, ordinales y/ o binarias, se usó el coeficiente ρ de Spearman para niveles de significación estadística del 1% y 5%. La existencia de diferencias significativas en la distribución de medias entre variables se basó en el test H de Kruskal-Wallis, realizando además comparaciones-poriguales mediante el test T3 de Dunnett; tras chequear la existencia de diferencias significativas, el estudio se completaba con análisis post-hoc del test HSD de Tukey, el cual permitiría definir subgrupos homogéneos de variables nominales/ ordinales que mostrasen una conducta estadística similar con respecto a la variable continua. Otros métodos El artículo nº II de la presente investigación completó el estudio ajustando estadísticamente la participación del propietario NIPF en un grupo profesional de asesoramiento (asociación, cooperativa o sindicato, entre otros) con respecto a una combinación de variables explicativas (otros atributos del perfil del propietario, unidad familiar, explotación y usos de la tierra, y economía) para un nivel de significación estadística del 5%. Del mismo modo, el artículo nº IV modeliza la relación entre la transformación pasada de prado a monte arbolado entre 1999-2003, e intenciones futuras de incrementar el terreno forestal productivo a corto/ medio plazo y de cambiar la actual especie forestal arbolada en el próximo ciclo. Dada la naturaleza binaria de las variables dependientes, se usó regresión logística mediante selección por pasos hacia adelante, método basado en una función de probabilidad acumulativa cuyo principal objetivo es modelizar cómo la presencia de diversos factores y el valor o nivel de los mismos, afecta la probabilidad de ocurrencia (Ryan 1997). 11 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research Finalmente, el artículo nº V de la presente investigación incluyó una ecuación de regresión para el ratio anual de plantación entre 1999-2003 con respecto a una combinación lineal de variables continuas de explicación relacionadas con factores del perfil del propietario, unidad familiar, explotación y usos de la tierra, y economía. Atendiendo a la naturaleza cuantitativa continua de la variable dependiente, se optó por la regresión lineal múltiple por pasos, modelo predictivo cuyo principal objetivo es modelizar cómo la presencia de diversos factores y el valor o nivel de los mismos, afecta la proporción de la variable dependiente, de acuerdo con los criterios de selección estadística de minimización del sesgo y maximización del ajuste R2-ajustado (Neter et al. 1996). Al igual que en los análisis no-paramétricos, se empleó software SAS/ STATTM (versión 9.1). RESULTADOS Y DISCUSIÓN ARTÍCULO I. GENERALIDADES DE LA SELVICULTURA FAMILIAR Como principales resultados se citan:  A pesar de la relativa importancia en el conjunto económico de la región y de su interrelación con otros sectores, la actividad forestal en Galicia ha sido sistemáticamente marginada de las reflexiones y negociaciones políticas en materia económica, presentando un desfase temporal en contabilidad forestal con respecto a otros países más dinámicos en la materia. Así, los estudios realizados hasta el momento en la región se han limitado al campo de la economía aplicada o bien se han resumido en estadísticas de diferentes anuarios e informes de la administración pública, sin desarrollar un plan contable que permita determinar la progresión socioeconómica de una muestra de explotaciones forestales tipo agrupadas a partir de factores productivos. Dado que la selvicultura computa una pequeña parte de todas las actividades agrícolas y no-agrícolas de una explotación, sería una tarea complicada establecer una red pura de contabilidad y rentabilidad forestal, citándose la posible adaptación y proyección de las estadísticas económicas agrícolas ya existentes.  A los problemas económicos citados anteriormente se unen los inconvenientes asociados a la parcelación territorial de Galicia, obstáculo de vital importancia para desarrollar una actividad competitiva y rentable. En este sentido, los modelos de gestión forestal conjunta han demostrado ser motores de dinamización socioeconómica en numerosas áreas rurales, incentivando la interacción y cooperación entre propietarios forestales cara un fin común viable. La aplicación a Galicia de este modelo de gestión forestal en común ya existe y se lleva a cabo en los denominados Montes Veciñais en Man Común, descritos anteriormente. Sin embargo, la mayoría de estas comunidades son deficitarias en capital humano, estando principalmente integradas por miembros de edad avanzada, retirados de la actividad agroganadera o profesionales activos no-agrarios, sin información ni formación forestal, que participarían en tales colectivos sin percibir contraprestaciones económicas individuales. Todos estos factores pueden determinar y determinan que, no depender de los recursos colectivos y no obtener beneficios de la participación, desencadene en absentismo o ausencia de acción para la gestión común.  Otro factor de relevancia que es necesario estudiar en la eficiencia forestal es el capital humano, esto es, las habilidades, cualidades y saber-hacer (experiencia) de los propietarios forestales. A diferencia de otras regiones con amplia tradición forestal, los propietarios forestales gallegos no disponen servicios de extensión forestal como tal, modelos de asistencia técnica que han demostrado ser claramente efectivos a la hora de dinamizar y profesionalizar la actividad. Las principales fuentes públicas de información y asesoramiento forestal actualmente existentes en la región son los servicios administrativos de la Xunta de Galicia a través de las cámaras de extensión agraria y los distritos forestales. Estos últimos (distritos forestales) conformarían la unidad pública básica de asesoramiento, gestión y ejecución forestal más directa y práctica. Con todo, la extensión forestal en la comunidad es asumida y desempeñada principalmente por asociaciones privadas sin ánimo de lucro de propietarios de montes, grupos 12 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research profesionales que trabajan y cooperan estrechamente con sus asociados en todas aquellas materias relacionadas con el monte.  En la región gallega, la dimensión política del monte se ha venido centrando, de forma generalizada, en líneas de apoyo a la producción y en la innovación tecnológica de plantaciones monocultivo de escasas especies comerciales de crecimiento rápido, así como en la actuación directa en materia de incendios forestales, dejándose la inversión forestal a la iniciativa privada mediante incentivos económicos. El contexto forestal actual demanda, sin embargo, nuevos enfoques económicos en el diseño de programas o de medidas públicas para el sector en un intento de abarcar o de cubrir la multiplicidad de objetivos relacionados con el desarrollo rural en general y con la gestión forestal sostenible en particular. Así, se está asistiendo a un cambio en la valoración de los recursos naturales, especialmente de los espacios rurales, donde los valores y las actividades tradicionales deben y deberán combinarse con los ambientales y con los culturales, buscando la profesionalización y viabilidad en la actividad forestal. ARTÍCULO II. PERFIL DEL PROPIETARIO Como principales resultados se citan:  El nivel de educación reglada del propietario NIPF y su participación en grupos profesionales se asociaban significativamente con el ratio anual de plantación. Este grupo de propietarios NIPF se correspondía principalmente con profesionales activos fuera de la agricultura, de mediana edad e ingresos familiares altos, gestores además de grandes superficies de monte arbolado productor.  La ocupación primaria del propietario NIPF, en concreto su condición como agricultor activo, y su participación en grupos profesionales de asesoramiento se relacionaban significativamente con el ratio anual de tratamientos selvícolas. Nuevamente, estos propietarios NIPF eran mayoritariamente profesionales activos no-agrarios que apostarían por la selvicultura mediante la contratación de asistencia técnica profesionalizada. Fuera de las ocupaciones no-agrarias, otro perfil de propietario NIPF a destacar en cuanto a su contribución a las actividades de plantación y selvicultura en el área de estudio fueron los agricultores retirados y activos, respectivamente. En este caso, la gestión forestal se basaba en la propia formación del propietario en la materia, su participación en grupos profesionales y la disponibilidad de maquinaria agroforestal en la explotación, además de apoyarse en importantes fracciones de trabajo propio y familiar.  El ratio anual de corta se relacionaba significativamente con la edad del propietario NIPF, la disponibilidad de información sobre el mercado forestal (madera) y el conocimiento y aplicación de criterios técnicoproductivos en materia de corta. Estos propietarios de mediana edad eran profesionales fuera de la agricultura o agricultores en activo que, posiblemente gracias a sus conocimientos y experiencia en la materia, habían fijado los mejores precios de madera en cortas previas, además de obtener los mayores ingresos en dichas transacciones. El modelo de regresión logística desarrollado para la participación del propietario NIPF en grupos profesionales de asesoramiento (asociaciones) reveló que dicho atributo estaba significativa y positivamente vinculado a aspectos propios de la actividad agrícola (el estatus del propietario como agricultor en activo y la disponibilidad de maquinaria agroforestal), así como con la intención futura de aumentar la superficie forestal productora a corto/ medio plazo en la explotación. ARTÍCULO III. UNIDAD FAMILIAR Como principales resultados se citan:  El patrón de adquisición de las tierras y la fracción anual de trabajo familiar en la explotación eran factores positiva y significativamente relacionados con el ratio anual de plantación.  El ratio anual de tratamientos selvícolas se vinculaba positiva y significativamente con la disponibilidad de maquinaria agroforestal en la explotación. 13 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research  La fracción anual de reempleos forestales dentro de la unidad familiar se asociaba positiva y significativamente con el ratio anual de corta. Además, un factor clave en la conducta de gestión forestal de los propietarios NIPF del área de estudio fue la fracción anual de trabajo contratado, variable positiva y significativamente relacionada con los ratios anuales de plantación, selvicultura y corta. Los propietarios NIPF que gestionaban bases territoriales adquiridas mediante herencia y compra eran significativamente más proclives a plantar sus terrenos de monte que los restantes propietarios analizados, identificándose como agricultores retirados que habían transformado anteriormente terrenos agroganaderos abandonados en monte arbolado mediante incentivos públicos. Dichos propietarios se caracterizaban por dedicar una importante fracción de trabajo propio a la actividad forestal, apoyándose en su formación y saberhacer en la materia, así como en ayuda y dedicación familiar. La inversión y gestión forestal respondería básicamente a valores emocionales, esto es, capitalizar la tierra heredada para ser transmitida. Por su parte, aquellos propietarios NIPF con maquinaria agroforestal disponible en la explotación y aquellos con una importante tasa de reempleo forestal en la unidad familiar eran significativamente más proclives a realizar tratamientos selvícolas y cortas de madera, respectivamente, que la restante población de estudio. Dichos propietarios se perfilaban principalmente como agricultores activos con formación forestal miembros de grupos profesionales de asesoramiento (asociaciones). El mayor vínculo con la tierra, asociado a su ocupación primaria en la agricultura, determinaba que estos propietarios dedicasen una importante fracción de trabajo propio a la actividad forestal, uniéndose además una importante ayuda familiar. Sin embargo, es necesario distinguir un grupo de propietarios NIPF claramente diferenciado de los anteriores perfiles. Esta tipología de propietarios se ajustaba a profesionales activos no-agrarios que, del mismo modo que los agricultores retirados, capitalizarían las tierras mediante la inversión y gestión forestal. Estos propietarios eran gestores de grandes y poco parceladas superficies de monte arbolado que, no dedicando importantes fracciones de trabajo propio y familiar en la explotación, trabajaban estrechamente con técnicos profesionales. Mediante esta contratación de trabajo forestal, este grupo de propietarios NIPF contribuía significativamente a la actividad forestal desarrollada en A Mariña Oriental. ARTÍCULO IV. EXPLOTACIÓN AGROFORESTAL Como principales resultados se citan:  La transformación pasada de terreno forestal a prado respondía a la demanda de superficie agroganadera en la base territorial de la explotación, estando positiva y significativamente vinculada con el estatus del propietario NIPF como agricultor activo. Así, en contraste con otras ocupaciones profesionales, los propietarios NIPF vinculados activamente en la agricultura se caracterizaban, de forma generalizada, por gestionar menores áreas de uso forestal, posiblemente por su completa dedicación o implicación en la agricultura, así como más parceladas, probablemente para mejorar e incrementar la productividad de la tierra.  Tanto la transformación pasada de terreno agrícola abandonado a monte arbolado como la intención futura de incrementar el terreno forestal productivo a corto/ medio plazo dependían claramente de experiencias previas en materia de corta y venta de madera. Los propietarios NIPF con mayores beneficios económicos de la actividad forestal (mejores precios unitarios y mayores ingresos por cortas previas de madera) se caracterizaban por haber aumentando recientemente el monte arbolado en su explotación o por pretender incrementarlo en un futuro cercano. El perfil de este tipo de propietario NIPF se correspondía con el de agricultor jubilado o profesional no-agrario en activo, propietarios que capitalizarían sus tierras mediante el cultivo forestal. Los resultados del modelo de regresión logística para estimación de la transformación pasada de terrenos agroganaderos abandonados a monte arbolado indicaron que el perfil del propietario NIPF, representado por su condición de agricultor en activo y su participación en organizaciones profesionales, junto con la actividad anual en tratamientos selvícolas, eran los factores 14 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research determinantes de haber practicado dicho cambio de uso en la explotación. Por su parte, el modelo de regresión logística ajustado para la intención futura de incrementar el terreno forestal productivo a corto/ medio plazo dentro de la explotación atendió nuevamente a la participación del propietario NIPF en asociaciones profesionales, añadiéndose la fracción anual de trabajo contratado en la actividad forestal y los ingresos anuales por venta de madera.  La intención futura de cambiar la especie forestal productiva en el siguiente turno o ciclo se asociaba positiva y significativamente con las fracciones anuales de trabajo propio y contratado en la explotación, así como con los gastos anuales en plantación y selvicultura. Este perfil de propietario NIPF se correspondía principalmente con un agricultor activo que, no disponiendo de más base territorial para la actividad forestal, dada la necesidad de terrenos para la actividad agroganadera, pretendía mejorar la rentabilidad del monte arbolado mediante dicho cambio. El modelo de regresión logística desarrollado para esta intención futura fue el más complejo en cuanto a variables explicativas incluidas. Así, además de vincularse con el ratio anual en tratamientos selvícolas, este propósito se completaba nuevamente con la participación del propietario NIPF en asociaciones profesionales, el conocimiento en y aplicación de criterios técnicoproductivos en materia de corta forestal, la fracción anual de trabajo propio en la explotación, la superficie forestal productiva y, finalmente, el ingreso anual por venta de madera.  Finalmente, los análisis de correlación y dependencia mostraron que los propietarios NIPF más eficientes y dinámicos en materia de gestión forestal, esto es, aquellos con los mayores ratios anuales de plantación, selvicultura y corta en el área de estudio, eran los propietarios y gestores de mayores superficies de monte arbolado dentro sus explotaciones, siendo además, terrenos forestales menos parcelados (menor número de parcelas por unidad de superficie de monte arbolado). ARTÍCULO V. ACTIVIDAD ECONÓMICA Como principales resultados se citan:  El ratio anual de plantación forestal se relacionaba positiva y significativamente con la inversión anual en mejora de la explotación (equipamientos e infraestructuras), los gastos anuales en plantación y selvicultura, la solicitud de subvenciones públicas, la cuantía anual de ayuda finalmente concedida, y el ingreso anual y precio unitario por venta de madera.  El ratio anual de tratamientos selvícolas se asociaba positiva y significativamente con la inversión anual en mejora de la explotación, los gastos anuales en plantación y selvicultura, y la solicitud de subvención pública.  El ratio anual de corta de madera se vinculaba positiva y significativamente con los gastos anuales en plantación y selvicultura, el ingreso anual y precio unitario por venta de madera, y los ingresos anuales no-madereros (venta de tierras). El modelo de regresión múltiple ajustado para el ratio anual de plantación indicaba que dicha práctica en la región se explicaba casi exclusivamente por atributos de carácter económico (inversión anual en mejoras de la explotación, gastos anuales en plantación y selvicultura, e ingresos anuales por venta de madera), añadiéndose la superficie forestal productora dentro de la explotación como otra variable de importancia en el ajuste. Sin restar importancia a la situación personal y familiar del propietario NIPF, así como a las propias características de la explotación agroforestal, a la hora describir y explicar el patrón de gestión forestal en el área, los principales resultados obtenidos sugirieron que unos ingresos forestales atractivos y unas condiciones de mercado favorables para la producción de madera parecían influenciar, directa o indirectamente, la inversión en el monte y la continuidad de su manejo como capital activo. Así, los gestores forestales más dinámicos de A Mariña Oriental se caracterizaban por beneficiarse ampliamente de la actividad forestal en cuanto a ingresos económicos, pero igualmente invertían cumplidamente en mejoras de la explotación, y en la plantación y mejora 15 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research selvícola de sus terrenos forestales con el fin de mantener su productividad y aumentar su rentabilidad. CONCLUSIONES Analizar el crecimiento del sector forestal en Galicia mediante indicadores de superficie puede apuntar, en primer término, hacia una apuesta decidida por parte de los propietarios de tierras por este subsector agrario. Sin embargo, estos índices no reflejan realmente la situación del monte gallego, puesto que este incremento se asienta principalmente en la crisis de los subsectores agrícola y ganadero, así como en la falta de alternativas para la tierra que no pasen en muchos casos por su forestación. En la actualidad, casi un 70% de la superficie de la comunidad gallega es terreno forestal, lo que supone un enorme potencial socioeconómico que aún no se ha visto refrendado por las características productivas de este sector, principalmente por sus deficiencias estructurales y por la falta de definición de objetivos y de modelos de desarrollo forestal. La selvicultura, como práctica de la tierra claramente diferente con respecto a otros usos del territorio, parecer estar en una fase de iniciación en la economía rural, compartiendo numerosos objetivos y prácticas de gestión con la agricultura, pero no al mismo nivel económico ni formativo. El punto de partida para el desarrollo del sector forestal en Galicia estará en ajustar e implementar procedimientos o metodologías de rigor científico, testados exitosamente en otras áreas europeas con amplia tradición forestal, que permitan caracterizar exhaustivamente su situación actual y estudiar su progresión espaciotemporal. No se trata de romper la lógica existente en el medio rural gallego, sino de complementarlo mediante su adaptación y modernización a las nuevas condiciones y demandas en materia de producción y organización forestal: aprovechar la oportunidad productiva del monte en Galicia sin abandonarlo ante el atraso en formación y en gestión. Los resultados de la presente investigación indican que los distintos programas o líneas públicas de apoyo a la gestión forestal de propietarios NIPF deben ser afines y coherentes a la existencia de diferentes perfiles de propietarios y unidades familiares y, por tanto, a la existencia de diferentes explotaciones y prácticas de gestión de la tierra. En este sentido, es particularmente importante diferenciar los objetivos y motivaciones de gestión forestal de los agricultores, que muestran un fuerte vínculo emocional con la tierra y que dedican una importante fuerza de trabajo a la actividad, de los propietarios no-agrarios, que estarían menos arraigados a la propiedad de la tierra, pero que también contribuirían al subsector forestal mediante la contratación de trabajo profesional. Independientemente del perfil de propietario NIPF, los terrenos forestales muestran ser, indiscutiblemente, una importante parte de la base territorial de las explotaciones rurales, considerando la selvicultura como una opción al abandono de terrenos agroganaderos descapitalizados y un complemento económico en la unidad familiar. Respondiendo a señales atractivas en cuanto al mercado de la madera, los propietarios NIPF atienden a la responsabilidad moral de cuidar y mantener la productividad de sus tierras. Como resultado, el interés de los propietarios NIPF en la selvicultura no podría expresarse explícitamente en términos económicos (recurso monte como medio para generar ingresos económicos a partir de la producción de madera) o en términos sociológicos (recurso monte como capital a transmitir en herencia a las generaciones futuras), sino en una combinación de ambos. Sin embargo, y sin desestimar el peso de factores de índole social, geográfico o político, los atributos económicos son claves y determinantes del desarrollo e intensidad de la gestión forestal desarrollada por propietarios NIPF, indicando que promover una selvicultura social y ambientalmente sostenibles implicará primeramente promover una selvicultura rentable económicamente. Así, unas prácticas forestales responsables, dentro del contexto de desarrollo rural y de protección ambiental, dependerán de la existencia de una red social de propietarios que perciban ingresos y contraprestaciones económicas por conservar, mejorar y gestionar monte como fuente de bienes y servicios para la sociedad. Considerando la actual dirección de las políticas forestales, los programas e incentivos públicos relacionados con la certificación forestal serán particularmente importantes, puesto que, dicha herramienta permite combinar las opciones rentabilidad y utilidad aportadas por el monte. Por tanto, será esencial motivar y compensar a los propietarios NIPF por gestionar recursos forestales que, entre otros múltiples puntos, mejoran y mantienen la riqueza natural y vitalidad rural, contribuyen al ciclo global del carbono, proporcionan numerosos productos y servicios, e incrementan el producto doméstico bruto y ratio de empleo de la región. 16 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research Los recursos forestales son la esperanza social, económica y ambiental para la recuperación y renovación de las áreas rurales, reorientándose las prácticas y valores tradicionales para con el monte hacia los principios y criterios de gestión forestal sostenible mediante un proceso participativo de toma de decisiones de todos los actores implicados. BIBLIOGRAFÍA Arano KG, Munn IA, Gunter JE, Bullard SH, Doolitle ML, 2004. Comparison between regenerators and non-regenerators in Mississippi: a discriminant analysis. Southern Journal of Applied Forestry 22 (2): 132-138. Arano KG, Munn IA, 2006. Evaluating forest management intensity: a comparison among major forest landowner types. Forest Policy and Economics 9, 237-248. Bergseng E, Vatn A, 2009. Why protection of biodiversity creates conflictSome evidence from the Nordic countries. Journal of Forest Economics 15, 147-165. Bolkesjø TF, Baardsen S, 2002. Roundwood supply in Norway: micro-level analysis of self-employed forest owners. Forest Policy and Economics 4: 55-64. Bolkesjø TF, Solberg B, Wangen KR, 2007. Heterogeneity in nonindustrial private roundwood supply: lessons from a large panel of forest owners. Journal of Forest Economics 13: 7-28. Conway MC, Amacher GS, Sullivan BJ, 2003. Decisions non-industrial forest landowners make: an empirical examination. Journal of Forest Economics 9: 181-203. Dennis D, 1990. A profit analysis of the harvest decision using pooled time-series and cross-sectional data. Journal of Environmental Economics and Management 18: 176-187. Doolittle L, Straka TJ, 1987. Regeneration following harvest on nonindustrial private lands in the south: a diffusion of innovations perspective. Southern Journal of Applied Forestry 11 (1): 37-41. FAO 2006. Evaluación de los recursos forestales mundiales 2005. Hacia la ordenación forestall sostenible. Organización de las Naciones Unidades para la Agricultura y la Alimentación (FAO). Roma, Italia. Fernández X, López E, Jordán M, Besteiro B, Viso P, Balboa XL, Fernández L, Soto D, 2006. Os montes veciñais en man común: o patrimonio silente. Natureza, economía, identidade e democracia na Galicia rural. Vigo, España. Gunter JE, Bullard SH, Doolitle ML, Arano KG, 2001. Reforestation of harvested timberlands in Mississippi: behaviour and attitudes of nonindustrial private forest landowners. Mississippi State University. Mississippi, Estados Unidos. 25 p. Hardie IW, Parks PJ, 1996. Program enrolment and acreage response to reforestation cost-sharing programs. Land Economics 72: 248-260. Hyberg B, Holthausen D, 1989. The behavior of nonindustrial private forest landowners. Canadian Journal of Forest Research 19:1014-1023. Joshi S, Arano KG, 2009. Determinants of private forest management decisions: a study on West Virginia NIPF landowners. Forest Policy and Economics 11, 118-125. Kant S, 2003. Extending the boundaries of forest economics. Forest Policy and Economics 5, 39-56. Karppinen H, 1998. Values and objectives of nonindustrial private forest owners in Finland. Silva Fennica 32 (1), 43-59. Kline JD, Butler BJ, Alig RJ, 2002. Tree planting in the South: what does the future hold? Southern Journal of Applied Forestry 26 (2), 99-107. Kuuluvainen J, Salo J, 1991. Timber supply and life cycle harvest of non-industrial private forest owners: 17 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research an empirical analysis of the Finnish case. Forest Science 37: 1011-1029. Kuuluvainen J, Karppinen H, Ovaskainen V, 1996. Landowner objectives and non-industrial private timber supply. Forest Science 4: 300308. Löyland K, Kringstad V, Öy H, 1995. Determinants of forest activitiesa study of private nonindustrial forestry in Norway. Journal of Forest Economics 1 (2): 219-237. Marey-Pérez MF, 2003. Tenencia de la tierra en Galicia: modelo para la caracterización de los propietarios forestales. Tesis doctoral, Universidade de Santiago de Compostela. Santiago de Compostela, España. 633 p. Marey-Pérez MF, Rodríguez-Vicente V, CrecenteMaseda R, 2006. Using GIS to measure changes in the temporal and spatial dynamics of forestland: experiences from north-west Spain. Forestry 79 (4): 409-423. MMA, 1998. III Inventario Forestal de España. Dirección General de Conservación de la Naturaleza, Ministerio de Medio Ambiente. Madrid, España. MMA, 1987. II Inventario Forestal de España. Dirección General de Conservación de la Naturaleza, Ministerio de Medio Ambiente. Madrid, España. MMA, 2002. Plan Forestal Español. Dirección General de Conservación de la Naturaleza. Ministerio de Medio Ambiente. Madrid, España. Munton R, 2009. Rural land ownership in the United Kingdom: Changing patterns and future possibilities for land use. Land Use Policy 26S, S54-S61. Neter J, Kutner MH, Nachtsheim CJ, Wasserman W, 1996. Applied linear statistical models. McGrawHill. Nueva York, Estados Unidos. Parker DC, Hessl A, Davis SC, 2008. Complexity, landuse modeling, and the human dimension: Fundamental challenges for mapping unknown outcome spaces. Geoforum 39, 789-804. Potter-Witter K, 2005. A cross-sectional analysis of Michigan non-industrial private forest landowners. Southern Journal of Applied Forestry 22 (2): 132-138. Prada A, Vázquez, MX, Soliño M, 2005. Beneficios y costes sociales en la conservación de la Red Natura 2000. A Coruña, España. Prestemon J, Wear D, 2000. Linking harvest choices to timber supply. Forest Science 46 (3): 377-389. Ross-Davis AL, Broussard SR, Jacobs DF, Davis AS, 2005. Afforestation motivations of private landowners: and examination of hardwood tree plantings in Indiana. Northern Journal of Applied Forestry 22 (3): 149-153. Ryan TP, 1997. Modern regression methods. John Wiley. Nueva York, Estados Unidos. Schelhas J, Zabawa R, Molnar JJ, 2003. New opportunities for social research on forest landowners in the South. Soufhern Rural Sociology 19 (2), 60-69. Størdal S, Lien G, Baardsen S, 2008. Analyzing determinants of forest owners’ decision-making using a sample selection framework. Journal of Forest Economics 14: 159-176. Straka TJ, Doolittle S, 1988. Propensity of nonindustrial private forest landowners to regenerate following harvest: relationship to socioeconomic characteristics, including innovativeness. Resource Management and Optimization 6 (2): 121-128. Sukhatme PU, 1953. Sampling theory of surveys. FAO. Roma, Italia. Vokoun M, Amacher GS, Sullivan J, Wear D, 2010. Examining incentives for adjacent nonindustrial private forest landowners to cooperate. Forest Policy and Economics 12, 104-110. 18 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research Xunta de Galicia, 2001. O monte galego en cifras. Consellería de Medio Ambiente, Xunta de Galicia. Santiago de Compostela, España. Xunta de Galicia, 2005. Análise, balance e propostas sobre incendios forestais. Medidas preventivas, aproveitamento da biomasa residual e outras. Consello Económico e Social de Galicia, nº 2/ 05. Santiago de Compostela, España. Zhang D, Flick W, 2001. Sticks, carrots and reforestation investment. Land Economics 77 (3): 443-456. Zhang D, Mehmood SR, 2001. Predicting non-industrial private forest landowners’ choice of a forester for harvesting and tree planting assistance in Alabama. Southern Journal of Applied Forestry 25 (3): 101-107. MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research ANEXOS Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 114 o minifundismo da terra xunto coa propia estrutura social do ámbito rural galego, caracterizada por un escaso investimento tecnolóxico e por unha insuficiente man de obra na actividade agraria, imposibilitan que este sector poida garantir o benestar económico de numerosas familias (Marey et al., 2004). Na figura 1 preséntanse as liñas de traballo e os retos que cómpre desenvolver para o conxunto do sector forestal en Galicia a partir dunha análise cuantitativa de anuarios públicos, así como de datos cuantitativos e cualitativos obtidos a partir de enquisas a propietarios forestais da rexión (Marey, 2003). Figura 1.- Planificación estratéxica do sector forestal en Galicia CORRIXIR Ausencia de experiencia forestal Estrutura territorial de minifundio Produtos forestais de escasa calidade Escasa dimensión da industria forestal Descoñecemento da realidade económica PALIAR Paulatino éxodo rural Avellentamento da poboación Deterioración da explotación familiar Alto nivel de desemprego Escaso nivel formativo POTENCIAR Aptitude agroforestal do territorio Gran superficie forestal arborada Gran volume madeirable en existencias Motor de materia prima para España Vínculo emocional poboación-monte APROVEITAR Aproveitamento múltiple do monte Concienciación polos recursos forestais Medidas de incentivación económica Demanda de materia prima de calidade Asociacionismo e/ou cooperación forestal FONTE: Marey et al. (2006). Mediante a coherencia nas actuacións que cómpre desenvolver, recuperar a confianza nun sistema forestal de enorme capacidade produtiva, pero xestionado coma un investimento de escaso interese, requirirá potenciar e aproveitar de forma sostible aquelas fortalezas e oportunidades do monte galego a fin de corrixir e paliar as principais debilidades e ameazas do noso medio rural (Marey et al., 2006). Así, podemos dicir que o actual subsector forestal galego é o resultado da crise dos subsectores agrícola e gandeiro, máis ca unha aposta decidida polo monte como actividade economicamente rendible. Dese modo, é habitual atoparnos nesta rexión con sistemas forestais infraexplotados tecnolóxica e economicamente (agás casos excepcionais), asentados máis na súa capacidade produtiva natural ca no desenvolvemento de modelos de actuación silvícola que, partindo do estudo da realidade produtiva, xeren respostas eficientes que maximicen o rendemento. Como sinalan Prada et al. (2005), en Galicia poderíase asumir que existe riqueza forestal no que respecta á cantidade pero non así á calidade. Na figura 2 preséntanse as diferentes etapas no ciclo de mellora da planificación forestal, análise fundamentada en tres piares básicos para garantir unha xestión forestal eficiente: o estudo da realidade socioeconómica da propiedade privada forestal, a existencia de servizos de extensión e innovación forestal, e un apoio legal e Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 115 político coherente. Este proceso levarase a cabo a través dun plan estratéxico acorde coa situación forestal existente, cunha espiral de mellora continua como eixe central que permita e que alcance o equilibrio entre recursos forestais existentes e demandas da sociedade en xeral e das industrias forestais de transformación en particular. Dese modo, xérase un sector económico dinámico e equilibrado, onde unha xestión forestal eficiente parte de criterios de experiencia e innovación forestal. Figura 2.- Ciclo de mellora continua no desenvolvemento forestal Visión estratéxica REALIDADE FORESTAL Caracterización Información Acción e implementación Visión estratéxica INNOVACIÓN FORESTAL Caracterización e seguimento NOVA REALIDADE FORESTAL Novas experiencias Acción e implementación REVITALIZACIÓN MELLORA CONTINUA Información Experiencia FONTE: Adaptado a partir de Local Development Process, de Amdam (2001). Conforme ao primeiro dos criterios, as experiencias que ata agora se viñeron desenvolvendo en Galicia trataron principalmente con sistemas de prevención e extinción de incendios, aínda que campos como a mellora xenética e silvícola tamén adquiriron relevancia, dada a importancia das plantacións monoespecíficas para a produción de madeira (Chas et al., 2002). Porén, o estudo da explotación forestal desde a súa perspectiva socioeconómica non está suficientemente documentado e son estas liñas, dinámicas e adaptadas ás necesidades da explotación forestal galega, as que permitirán superar unha das súas principais carencias: a falta de veracidade na información por descoñecemento do valor económico real da actividade forestal. En Galicia, a gran maioría dos propietarios forestais, especialmente individuais, xestionan as súas masas atendendo aos seus propios obxectivos, xeralmente asociados ao investimento do monte no curto-medio prazo sen criterios de eficiencia. Este comportamento ‘individualista’ do xestor forestal maniféstase na falta Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 116 de traballo con experiencia en modelos de xestión que melloren a capacidade para superar atrancos e a habilidade para aproveitar oportunidades. Consonte co segundo dos criterios, a innovación partirá da acumulación previa de experiencias en xestión, do coñecemento das posibilidades que pode ofrecer o noso ámbito territorial e da actitude e da aptitude dos verdadeiros actores: os propietarios forestais. E todo isto baixo a supervisión e o apoio da Administración Pública, que debe velar por que se reduzan as dificultades que presenta o investimento nun modelo económico como é o forestal, cunha alta incerteza e con proxección no longo prazo (figura 3). Figura 3.- Esquema metodolóxico da innovación forestal INNOVACIÓN Posibilidades dos recursos forestais OPORTUNIDADES USO Experiencia CAUSA Conduta do propietario-xestor OBXECTIVOS COÑECER TOMA DE DECISIÓNS HABILIDADE DESENVOLVEMENTO FORESTAL Incerteza-Vulnerabilidade REDUCIR APOIO POLÍTICO En definitiva, o sector forestal en Galicia ten que retomar e completar o seu crecemento con modelos de xestión silvícola adoptados en países que, cunhas certas similitudes, teñen un maior desenvolvemento en investigación e en innovación forestal. Para iso, seleccionáronse catro puntos básicos de estudo en distintos ámbitos europeos para a súa posible adaptación a Galicia: as redes de contabilidade, os modelos de cooperación, os programas de educación e asesoramento, e as medidas públicas de apoio económico. Todos eles serán analizados en tres fases: a descrición das experiencias que desta materia se dispón actualmente en Galicia, base para a posible aplicación ou adaptación de modelos europeos; unha posterior análise dos principais programas ou medidas desenvolvidos en diferentes rexións europeas e, por último, a xeración de criterios para a súa adaptación á Comunidade. 2.1. REDES DE CONTABILIDADE FORESTAL: ANÁLISE DA RENDIBILIDADE ECONÓMICA FORESTAL Malia a relativa importancia no conxunto económico da rexión e da interrelación con outros sectores, a silvicultura foi sistematicamente marxinada das re- Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 117 flexións e negociacións políticas en materia económica. Os estudos realizados ata o momento en Galicia limitáronse ao campo da economía aplicada ou ben se resumiron en estatísticas de diferentes anuarios e informes da Administración Pública, sen desenvolver un plan contable que determinara a progresión socioeconómica dunha mostra de explotacións forestais tipo. Neste sentido, caracterizar polo miúdo a propiedade forestal constitúe a base para a aplicación da rama contable á explotación forestal, sendo esencial avanzar nesta liña para concretar claramente o papel económico do monte en Galicia. Como consecuencia da falta de referencias bibliográficas, a situación económica da explotación forestal galega comeza a atoparse cunha serie de dificultades para a súa continuidade dentro dunha economía de mercado. Revitalizar un sector forestal familiar implica novas fórmulas metodolóxicas de estudo que desenvolvan unha base de datos completa en canto a esta actividade, especialmente en contabilidade de custos e de ingresos da explotación. En Europa, a pesar da importancia desta actividade, tamén as súas condicións económicas están pobremente documentadas, identificándose repetidamente a ausencia deste tipo de información coma un atranco para o seu perfeccionamento (Hyder et al., 1994; Harrison, 2001). Porén, Galicia presenta un desfasamento temporal en contabilidade forestal con respecto a outros países máis dinámicos nesta materia, atranco que orixinou que os principios económicos do sistema forestal galego non respondan ás características particulares da súa propiedade e industria. Desde esta perspectiva, a dispoñibilidade de información forestal contable é unha referencia de grande utilidade para os propios xestores. Coñecendo todas as clases de actividade propias dunha explotación proporciónase a estrutura ideal para unha valoración completa do papel e significación da silvicultura (Sekot, 2001), desenvolvendo metodoloxías públicas ou programas lexislativos efectivos que resolvan problemas específicos na toma de decisións e na xestión (figura 4). Pero ademais destes factores intrínsecos á explotación forestal, para un completo estudo económico é necesario involucrar axentes externos a esta, como son os provedores de materia prima, as industrias do sector e os servizos de seguimento e administración (Harrison e Qureshi, 2000). Polo tanto, e dado que a silvicultura computa unha pequena parte de todas as actividades agrícolas e non agrícolas dunha explotación, establecer unha rede pura de contabilidade forestal para o seguimento da súa rendibilidade é unha tarefa complexa e, neste sentido, menciónase a posible adaptación e proxección á silvicultura das estatísticas económicas agrícolas existentes. En Europa, a socioeconomía da silvicultura a pequena escala foi −e é− amplamente estudada en países cunha gran tradición en anuarios forestais de contabilidade ou con redes permanentes de seguimento silvícola (Niskanen e Sekot, 2001; Brandl, 2002). Así, en Finlandia destaca, entre outros, o Statistical Yearbook of Forestry (Anuario Estatístico de Silvicultura) do Finnish Forest Research Institute (Metsäntutkimuslaitos-METLA); no Reino Unido, a Forestry Commission recompi- Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 118 la anualmente información estatística relativa a múltiples aspectos da actividade forestal de Inglaterra, Gales, Escocia e norte de Irlanda no anuario Forestry Facts & Figures (Información Detallada sobre Silvicultura); en Noruega, o servizo Statistics Norway (Statistisk Sentralbyrå), servizo oficial de estatísticas do país e que depende administrativamente do Ministerio de Finanzas, realiza anualmente un estudo exhaustivo do subsector forestal en colaboración con outras institucións; e en Suecia, a Swedish Forest Agency (Skogsstyrelsen), institución pública responsable en materia forestal e con competencia en estatística oficial do monte, presenta anualmente o Swedish Statistical Yearbook of Forestry (Anuario Estatístico Sueco de Silvicultura). Como sinala Sekot (2001) na súa revisión de experiencias e resultados de redes contables forestais en Austria, estas metodoloxías demostran mellorar a xestión da explotación forestal. Figura 4.- Esquema dunha rede de contabilidade forestal SISTEMA DE APOIO Á XESTIÓN FORESTAL Manter Crear Incentivos lexislación Intercambio de experiencias Intercambio de experiencias Administración forestal Rendibilidade da explotación forestal Institucións forestais Mellorar Caracterizar Investigacións secundarias PREDICIR DEFINIR POTENCIAR CORRIXIR Recursos forestais Medios técnicos Medios humanos Custos- -Beneficios Aspectos sociais Vantaxes Desvantaxes APRENDIZAXE CONTINUA DO XESTOR FORESTAL Así e todo, a información contable existente na actualidade mostra unha serie de deficiencias que é necesario mellorar tanto na fase de creación coma na de mantemento deste tipo de redes. De feito, as redes contables desenvolvidas ata o momento, como submostras dunha rede de contabilidade agrícola, non revelan a significación ou a representatividade da silvicultura na explotación agroforestal. A este problema de definición únense os inconvenientes asociados á inexistencia dun plan contable na maioría das explotacións que, xunto coa gran variabilidade en terminoloxía forestal entre países, dificulta a súa análise e comparativa. Outro problema na formulación dunha rede contable é a selección e a representatividade da mostra que Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 119 reflicta de forma obxectiva a evolución do modelo de explotación analizada e que permita obter conclusións estatisticamente fiables. Coñecida a mostra motivo de análise, é necesario reducir a incerteza propia da toma de datos, considerando o grao de fiabilidade dos resultados achegados polo xestor forestal. Ademais, ás dificultades propias do establecemento dunha rede uniranse os potenciais inconvenientes do seu seguimento. Entre estes cítanse a falta de cooperación e de continuidade na rede por parte dos seus participantes, a fiabilidade dos datos achegados polos propietarios integrantes e o elevado custo de seguimento. Polo tanto, é preciso establecer un amplo rango de aspectos metodolóxicos dentro dunha estrutura analítica detallada para alcanzar de forma eficiente uns resultados estándares imparciais. Cambiar esta situación require activar ou mobilizar as redes xa existentes e establecer outras novas que proporcionen un maior coñecemento da economía da explotación forestal. A aplicación de redes contables en Galicia debe partir dunha análise previa das explotacións forestais existentes e a súa agrupación posterior a partir de factores produtivos. Esta rede proporcionará a información contable necesaria para ser utilizada de forma directa na medición de indicadores socioeconómicos de sostibilidade forestal propios da certificación. 2.2. MODELOS DE COOPERACIÓN: PARTICIPACIÓN E XESTIÓN EN COMUNIDADES FORESTAIS Aos problemas socioeconómicos citados anteriormente únense as desvantaxes propias da fragmentación territorial de Galicia, atranco importante para un manexo silvícola eficiente. Partindo desta base, asegurar a competitividade e a rendibilidade desta actividade require desenvolver novas formas de cooperación e melloras loxísticas e informativas que atenúen este inconveniente (Mitchell-Banks, 2001; Uusivuori e Kuuluvainen, 2001; Marey et al., 2004). Neste sentido, as comunidades, agrupacións, asociacións e/ou cooperativas forestais demostraron un maior grao de dinamismo e de flexibilidade fronte aos novos desafíos na xestión forestal ca outras formas de posesión da terra (Bollin e Eklkofer, 2000; Frank, 2001). Como modelo operativo de xestión, este tipo de comunidades estimulan a continua interacción e cooperación entre propietarios forestais de cara a un fin común. Cuestións como, entre outras, de que medios humanos e técnicos se dispón?, cales son as metas de xestión que hai que alcanzar? ou, que procedemento se ten que seguir?, son materializadas nun plan estratéxico de actuación silvícola e comercial de maior eficacia e adaptabilidade, que permite a incorporación de novas tecnoloxías e a cualificación de recursos humanos. A figura 5 mostra como conseguir unha administración e xestión forestal responsable partindo de dous piares: o interese común dos propietarios e a flexibilidade e o dinamismo das solucións achegadas. A aplicación a Galicia deste modelo de xestión forestal comunitaria xa existe e lévase a cabo mediante os montes veciñais en man común (MVMC) no 33% da súa Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 120 superficie forestal. Como xa se dixo, nun número superior a 2.800 comunidades este tipo de “grupos forestais” xestionan máis de 670.000 hectáreas na rexión. Con diferentes teorías en canto á súa orixe, neste tipo de propiedade privada colectiva ser “veciño comuneiro” implica legalmente ter dereito a participar nun proceso democrático-asembleario de decisión colectiva sobre unha serie de asuntos relacionados co monte: o dereito de acceso, o dereito de uso e o dereito de participación no rendemento económico da explotación forestal (Fernández et al., 2006). Figura 5.- Principios operativos da cooperación forestal INTERCAMBIAR PRIORIZAR ANALIZAR INTERESE COMÚN COLABORACIÓN PARTICIPACIÓN XESTIÓN FORESTAL SOSTIBLE FLEXIBILIDADE DINAMISMO Obxectivos de xestión Condutas de xestión Debilidades Ameazas Fortalezas Oportunidades Administración forestal Institucións forestais Propietario forestal Xestor forestal Idoneidade en operacións silvícolas Uso eficiente de recursos-medios Rendibilidade en operacións comerciais Xeración de información-formación Estas comunidades caracterízanse na súa gran maioría por ser deficitarias en capital humano, e están principalmente integradas por comuneiros de avanzada idade, retirados da actividade agrogandeira, sen información nin formación forestal e que participan nestes colectivos sen percibir contraprestacións económicas individuais, factores que poden aumentar as posibilidades de abandono ou de infrautilización das terras (Fernández et al., 2006). Se a este problema de xestión colectiva lle sumamos que cada vez con maior frecuencia unha importante fracción destas comunidades veciñais está representada por profesionais de fóra da agricultura ou da gandería, non depender dos recursos colectivos e non obter beneficios da participación poden determinar tamén o absentismo ou a falta de acción na xestión (Ostrom, 1990). Así, no rural galego a inexistencia dunha base social que permita desenvol- Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 121 ver ou continuar proxectos de aproveitamento colectivo nos MVMC conforma unha importante debilidade (Fernández et al., 2006), a pesar de seren espazos clave para o sector forestal galego e fonte de renda para as comunidades rurais (Prada et al., 2005). En Europa, atopámonos con múltiples exemplos de cooperación ou de asociacionismo forestal, destacando especialmente a Unión de Silvicultores do Sur de Europa, que agrupa xestores franceses, españois, gregos, portugueses e italianos. En Austria, destacan as cooperativas locais de propietarios forestais Waldwirtschaftsgemeinschaft, a miúdo integradas en asociacións federais de propietarios. En Francia existen unhas 11.000 comunidades forestais (Communes Forestières), a maioría en zonas de montaña, que agrupan preto do 20% da superficie forestal total do país. En Italia, a Magnifica Comunitá, no val de Fíeme, é un exemplo de asociación forestal de montaña que proporciona un modelo bottom-up (de abaixo a arriba) de sostibilidade no uso dos recursos (Merlo, 1995), sendo menos coñecida a Comunaliae Pamensi no norte dos Apeninos. Xa no ámbito europeo, a European Federation of Municipal and Local Community Forests é unha alianza de propietarios e de comunidades de montes, e nalgúns casos de xestores, cuxo obxectivo é a xestión dos intereses e motivacións forestais dos distintos países membros. Esta alianza, que representa uns mil propietarios forestais, engloba unha superficie de 25 a 30 millóns de hectáreas arboradas. O proceso democrático-asembleario dunha comunidade forestal como organización colectiva ten que resolver, ademais, outro inconveniente de carácter interno. Como a propia palabra indica, “participar” implica un conxunto de individuos, de diferentes opinións, perspectivas e intereses con respecto ao uso e á xestión dos recursos da comunidade. Ante este feito xorden conflitos internos na comunidade que poden determinar a non participación dos integrantes e a paralización das actividades. No seu estudo sobre conflitos de participación e xestión en comunidades forestais, Skutsch (2000) sinala que non recoñecer a existencia destes conflitos internos pode ter desencadeado e seguir desencadeando o fracaso −colapso− de numerosos proxectos en comunidades forestais: se non se identifica a existencia de conflito, non se poderá entender e analizar a súa natureza e, polo tanto, non se poderá resolver. En definitiva, a “participación” é o mellor camiño, a forma máis eficiente para alcanzar obxectivos en xestión forestal e non un tópico relacionado “cos dereitos ou coas autorizacións políticas” (Skutsch, 2000). 2.3. PROGRAMAS DE EDUCACIÓN E ASESORAMENTO: EXTENSIÓN FORESTAL Outro factor de relevancia que é necesario investigar na eficiencia forestal é o capital humano. Así, Stefanou e Saxena (1988) verifican a importancia da educación e da formación dos propietarios forestais para unha explotación eficiente, é dicir, as habilidades e as calidades, sendo tamén especialmente importante a propia Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 122 experiencia adquirida, o saber facer (Evans, 1987; Marey et al., 2006). Noutras palabras, a carencia de educación formal entorpece, aínda que non impide, a mobilización produtiva do monte (Fernández et al., 2006). En Galicia non existe un servizo de extensión forestal como tal, polo que un dos problemas máis significativos aos que se enfronta o propietario galego para o seu efectivo desenvolvemento e completa implicación na súa explotación é a ausencia de formación. Por iso, a ordenación de montes, a pesar de ser Galicia unha das rexións españolas de maior vocación forestal e de capacidade produtiva, non tivo unha implantación adecuada. A pequena superficie de xestión, o escaso peso dos ingresos silvícolas na economía da explotación e, en definitiva, o paulatino desarraigamento pola propiedade maniféstanse na falta de formación e de interese forestal por parte do propietario. Pero a falta de información e de formación do xestor forestal tamén se asocian á descoordinación entre departamentos administrativos como, por exemplo, entre agricultura e montes, ou outros servizos públicos de asesoría afíns. Estas deficiencias provocan a necesidade de xerar, mediante a participación activa de todas as partes interesadas, un sistema de información transparente, que se plasme en iniciativas forestais coherentes coa realidade existente. Neste sentido, Hermelin (2001) sostén que son tres os puntos que hai que desenvolver para motivar o xestor forestal de cara a unha explotación racional dos recursos dentro dunhas condicións de mercado determinadas: a dispoñibilidade de servizos de información forestal, a preparación técnica e o asesoramento profesional. Polo tanto, a educación e a formación dos propietarios de terras, xunto coa interacción destes con outros propietarios e profesionais forestais, xera experiencia e mellora as habilidades de xestión (Mahapatra e Mitchell, 2001). Deste modo, é posible recuperar a confianza pola actividade forestal como fonte de ingresos tanxibles e beneficios non comerciais, reducíndose a incerteza á hora de innovar ou de adoptar novas tecnoloxías na explotación (Pattanayak et al., 2003). Os sistemas de asesoría −extensión− forestal, medios habituais de consulta en rexións de ampla tradición forestal, demostraron ser o medio máis efectivo para promover unha xestión forestal sostible e revitalizar a silvicultura familiar. Dese modo, asegúrase a integridade dos recursos forestais no tempo e refórzase o papel dos ingresos forestais na economía familiar. E dado que os propietarios non só adoptan medidas en función dos seus intereses senón que nalgúns casos seguen tendencias conxuntas con outros membros da súa comunidade rural (Mahapatra e Mitchell, 2001), é necesario que esta información alcance o seu equilibrio a través dos organismos competentes. Por tradición e por coñecementos, os profesionais forestais han de iniciar unha xestión adaptativa do monte, minimizando as consecuencias do conflito xerado entre diferentes intereses e valores en canto ao uso e ao manexo dos recursos forestais. E este proceso implica que a Administración desenvolva unha información e consulta pública continua e transparente, que baixo os criterios dunha comunicación efectiva leve a un entendemento común (figura 6). En definitiva, o obxectivo estratéxico da extensión forestal ha ser a promoción Rodríguez, V.; Marey, M.F. Sistemas de apoio á propiedade privada forestal... Revista Galega de Economía, vol. 17, núm. 1 (2008), pp. 111-130 ISSN 1132-2799 123 dunha información e formación continua e renovada, facilitando a toma de decisións de éxito nesta actividade e que proporcione a habilidade e os recursos necesarios para iso. Hoxe, as principais fontes de información forestal dispoñibles en Galicia son os servizos administrativos da Xunta de Galicia, a través dos servizos de extensión agraria ou a través dos distritos forestais, e as organizacións e asociacións forestais. A Administración Pública destaca especialmente polo seu particular compromiso na coordinación de información e formación profesional en materia de montes, incentivando liñas de educación, investigación e innovación forestal a través de institutos ou de escolas profesionais e fomentando a sensibilización por unha xestión sostible dos recursos forestais. Dentro dos servizos administrativos, os “distritos forestais” conforman a unidade pública básica de asesoramento, xestión e execución forestal máis directa e práctica, especialmente importante naqueles montes consorciados ou conveniados coa Administración. Este tipo de asistencia pública constituiría o modelo de extensión forestal desexable para Galicia, aínda que as súas funcións informativas e formativas teñen que mellorar de forma continuada mediante unha auténtica coordinación e avaliación ex-ante, mid-term e ex-post daquelas prácticas forestais, comerciais e industriais realizadas no territorio obxecto de intervención do distrito. Figura 6.- Promoción da educación forestal Administración Pública Centros de Investigación Forestal Centros de Formación Forestal Asociacións, federacións e outros grupos EDUCACIÓN FORESTAL CONCIENCIACIÓN-CAPACITACIÓN FORESTAL APOIO ASESORAMENTO EDUCACIÓN INFORMACIÓN Capacidade para a toma de decisións Con todo, a extensión forestal en Galicia é principalmente asumida e desempeñada por asociacións sen ánimo de lucro de propietarios de montes, conformando actualmente un eixe clave no asesoramento e profesionalización forestal na rexión. Como ben definen Fernández et al. (2006), estas entidades nacen ante a necesidade MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research II MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I RESEARCH PAPER Characterization of nonindustrial private forest owners and their influence on forest management aims and practices in Northern Spain Vero ´nica Rodrı ´guez-Vicente ÆManuel F. Marey-Pe ´rez Accepted: 3 August 2009 / Published online: 22 August 2009 ÓSteve Harrison, John Herbohn 2009 Abstract Explaining and predicting nonindustrial private forest (NIPF) owner land management based on social, economic, and environmental factors is an increasingly important issue in policy arenas and academic research on rural development and planning. This study empirically explores and assesses management behavior by NIPF owners by analyzing attributes of landowner profile (age, educational level, primary occupation, engagement in farming, membership of professional groups, training in forestry, availability of market information, and specific knowledge and use of production criteria for timber harvesting). With the aim of predicting outcomes, a multiple regression model was constructed to investigate and quantify the probabilities of and factors influencing the participation of owners in agricultural and forestry associations. In March 2004, 103 resident forest landowners were interviewed about their commitment to and involvement in land management during 1999–2003 in Marin ˜a Oriental, a forest region of Galicia, Northern Spain. Results suggest that professional occupation, particularly farming background, is the main factor affecting, either directly or indirectly, the forest management behavior of NIPF owners in the area. In particular, our logistic regression model for landowner membership of professional groups explained 77.9% of the variability observed in the study population, which suggests that the agricultural background of NIPF owners and their expectations from forests, represented by their future intention to enlarge the forestland base, play an important role in membership. In the region, forestry could be a valuable economic activity but it is not considered as such today. Findings could be used as a guide for design, V. Rodrı ´guez-Vicente (&) Galician Sectorial Forestry Association (ASEFOGA), Doutor Maceira 13-baixo, 15706 Santiago de Compostela, Spain e-mail: [email protected] M. F. Marey-Pe ´rez Department of Agroforestry Engineering, University of Santiago de Compostela, Campus Universitario s/n, 27002 Lugo, Spain 123 Small-scale Forestry (2009) 8:479–513 DOI 10.1007/s11842-009-9097-z planning, and implementation of research and policy measures that allow NIPF landowners to promote sustainable forestry for rural development. Keywords Farming Forestry Nonindustrial private forest (NIPF) owner  Planting Silviculture and harvesting practices Introduction Explaining and predicting nonindustrial private forest (NIPF) owner land management based on social, economic, and environmental factors is an increasingly important issue in policy arenas and academic research on rural development and planning, particularly because of the large diversity of factors and the complexity of characterizing and providing a framework for the management objectives and goals of NIPF owners. Because the nature of NIPF ownership differs notably from country to country, defining the meaning of this term is difficult; this type of individual forest ownership consists of a single or small number of planting blocks, nonprofessional management, and often a lack of silvicultural skills, with little planning for future marketing (Herbohn 2001). NIPF owners’ commitment to and involvement in land-use and management stems from a dynamic environment in which personal and family decisions and/or needs, geographic context, and policy guidelines are closely interrelated. According to Beach et al. (2005), forest owners produce a variety of forest products and benefits using several types of inputs including forestland, growing timber stock, labor, technical assistance, materials, and machinery to perform various land management activities. In particular, forests owned by NIPF owners are a part of the total land-use system of the holding, and management goals are not solely focused on industrial timber production; NIPF owners own and manage land for a wide variety of purposes, and thus their practices and values are equally diverse. Therefore, understanding and modeling the determinants of NIPF owners’ land-use and management behavior is a complex task, insofar as any analysis must consider the interaction between many decision-making and motivational factors. More specifically, understanding the land management behavior of NIPF owners requires knowledge of the agroforestry system and of the landowners’ personal goals and circumstances. Because NIPF owners are key actors in sustainable development and welfare in rural areas, their land management behavior has been studied extensively by researchers from different disciplines, such as sociology, economics, ecology or land planning. Accordingly, there is a growing descriptive, theoretical and empirical literature on the land management behavior of NIPF owners. In this way, understanding why NIPF owners engage in forestland management has been seen as key by many authors concerned with the choice and level of investment in forest management practices such as tree planting and forest stand improvement (Doolittle and Straka 1987; Straka and Doolittle 1988; Hyberg and Holthausen 1989;Lo ¨yland et al. 1995; Hardie and Parks 1996; Gunter et al. 2001; Zhang and Flick 2001; Zhang and Mehmood 2001; Kline et al. 2002; Arano et al. 2004; Ross-Davis et al. 2005), timber harvesting (Hyberg and Holthausen 1989; 480 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 Dennis 1990; Kuuluvainen and Salo 1991;Lo ¨yland et al. 1995; Kuuluvainen et al. 1996; Prestemon and Wear 2000; Zhang and Mehmood 2001; Bolkesjø and Baardsen 2002; Conway et al. 2003; Størdal et al. 2008), and nontimber forest activities (Pattanayak et al. 2002; Conway et al. 2003; Kittredge 2005; PotterWitter 2005; Van Gossum et al. 2005; Van Gossum and De Maeyer 2007; Boon and Meilby 2007). The choice and level of investment in different forest management practices has been analyzed as a function of different variables that may affect land management decisions and practices. Pattanayak et al. (2002), Amacher et al. (2003), and Beach et al. (2005) concisely and systematically review some empirical studies focused on land management decisions made by NIPF owners. The Autonomous Community of Galicia, a region in Northern Spain, provides an interesting setting for the analysis of NIPF owners’ management. During the last 50 years, forestry in Northern Spain has undergone significant changes brought about by Spain’s entry into a highly competitive market. Such changes have altered the landscape patterns of the region. Fewer farms, decreasing rural populations, and increasing fragmentation of forest property have led to significant alterations in the traditional farming system (Marey-Pe ´rez et al. 2006; Marey-Pe ´rez and Rodrı ´guezVicente 2008). These changes have also affected forests. After a period of neglect, the current approach to forest management emphasizes timber production using fast-growing species. Galicia has over 2 million ha of forestland, which accounts for 50% of the total land area of this region and 8% of the total land area of Spain. The Galician forestry sector has grown slowly but steadily over the last 15 years, and is currently one of the economic mainstays of the region. According to the 1998 Spanish forestry survey (MMA 1998), almost 50% of the timber produced in Spain comes from Galicia, which is the Spanish region with the highest standing volume and growing stock. Land-use changes have affected the social and economic structure of land management in Galicia. Until the mid 19th century, the typical land manager profile in Galicia was a crop or livestock farmer clearly linked to family knowledge and needs. In other words, the Galician rural economy was largely based on subsistence farming, with a large share of forestland. From that moment, the diversification and specialization of the industrial sector and the development of a range of policy tools for encouraging the capitalization of marginal agricultural lands have brought about important socioeconomic changes in many Galician rural areas, which have affected land management and planning (Marey-Pe ´rez and Rodrı ´guez-Vicente 2008). From the mid 20th century, a new profile of forest landowner and/or manager has emerged. New forest landowners and/or managers are different from traditional crop or livestock farmers, and have different goals and decision-making processes about their commitment to and involvement with the land. Today, Galician forests are mostly managed by private owners, with about 425,000–673,000 NIPF owners managing over two-thirds of the forest area (Marey-Pe ´rez 2003). The rapid expansion of forest area has generated many studies and statistical reports concerned with the current state and evolution of forestry resources in Galicia. However, none of these studies has focused empirically on the social and structural factors affecting forest decision-making and management. Despite the long agroforestry tradition of Galicia, few evaluations of its current state and future Characterization of NIPF owners and their influence on forest management 481 123 prospects have been conducted, which has led research institutions to investigate individual private property management and planning. Small-scale forestry must be recognized as a productive activity in order to invigorate the economy, to make the sector socially attractive, and to maintain the environmental integrity of rural areas (Marey-Pe ´rez et al. 2004). Knowing and understanding the factors that may influence individual forest management decisions and practices is essential to improve the forestry sector and to shape it as a social, economic, and environmental mainstay in Galicia, and consequently in Spain. As reported by Karppinen (1998), forest management, as a voluntary action, is primarily driven by the motivations of the landowners, i.e., their values and goals. As assumptions about landowner objectives have evolved, so has our understanding of the decisions they make (Amacher et al. 2003). Therefore, the starting point for understanding why NIPF owners engage in land management is to analyze and explain their motivations and objectives with regard to forestlands. Our aim is to explore and assess NIPF owners’ management behavior based on an empirical analysis of attributes of landowner profile, focusing on a sample of forest holdings surveyed in Northern Spain. Thus, based on a thorough analysis of individual landowner characteristics pertaining to age, educational level, primary occupation, engagement in farming, membership of professional groups, training in forestry, availability of market information, and specific knowledge and use of production criteria for timber harvesting, this paper aims to analyze in detail and better understand forest management behavior in the region. Accordingly, we search for a possible statistical relationship or difference between these variables and other factors related to family unit, forest property and land-use changes, or forest economics. In order to complete the results, we have included three practices traditionally used in the relevant literature to predict forest management behavior of NIPF owners, i.e., planting and silviculture on forestland, and timber harvesting on woodland (Lo ¨yland et al. 1995; Hardie and Parks 1996; Kuuluvainen et al. 1996; Prestemon and Wear 2000; Zhang and Flick 2001; Kline et al. 2002; Conway et al. 2003; Arano et al. 2004; Potter-Witter 2005; Ross-Davis et al. 2005; Størdal et al. 2008). In addition, these variables are statistically analyzed with respect to the characteristics pertaining to the profile of forest owners. The characterization of landowners and the identification of patterns in their forest management practices may allow policy-makers to improve existing policies and develop further public measures based on different landowner profiles in terms of their motivations and needs in forestry that encourage landowners to adopt sustainable land management. Materials and methods Identification of the population and questionnaire design This study updates and expands an earlier analysis by Marey-Pe ´rez (2003) exploring private ownership and forest management in the Autonomous Community of Galicia. Data was collected from face-to-face interviews with randomly selected 482 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 NIPF owners in Marin ˜a Oriental pilot area, located in Northeast Galicia, Spain (Fig. 1). The study area is representative of much of Northern Spain, where forests cover most of the land (53%) and forestry is an increasing activity (over 46% of Fig. 1 Location of Marin ˜a Oriental in Galicia, Northern Spain Characterization of NIPF owners and their influence on forest management 483 123 forests are woodlands). Forestlands are owned by 3,043 NIPF owners, who manage more than 90% of the forested area in the region. In Marin ˜a Oriental, forestland increased by 8.5% between 1957 and 2001 because of the conversion of agricultural land and shrubland into forests, such that the agricultural and farmland area in the region currently accounts for 27% of the total surface area. According to INE (1999), the number of farm holdings declined by 33.7% in Marin ˜a Oriental between 1962 and 1999. Such a decline in the number of farm holdings was caused mainly by social and production deficiencies in rural areas. However, the most important land-use change in the study area and in the Galician region was the composition of woodlands. Monospecific stands increased considerably between 1962 and 1999, about 300%, probably because of the spread of the forest species Eucalyptus globulus Labill. (blue gum). Plantations of blue gum increased by more than 63% in Marin ˜a Oriental between 1957 and 2001 (Marey-Pe ´rez 2003). Today, Eucalyptus globulus Labill. and Pinus pinaster Ait. spp. atlantica (maritime pine) are the main productive forest species in the area and cover more than 71% of the forestland, with a mean timber yield of 26.4 m 3 /ha per year, at a rotation age of 12 and 25 years, respectively, for blue gum and maritime pine. The strong increase in timber-producing forests in the study region, in Galicia, and in many other areas of Northern Spain was due to significant changes in the traditional agricultural and forestry system during the 1950s. As a result of increasing urban and industrial development during the mid 19th century, traditional agricultural and forestry activities were no longer economically viable for many rural communities, and active population largely shifted to urban agglomerations, similarly as in other European countries (Marey-Pe ´rez and Rodrı ´guez-Vicente 2008). Changes in land use at farm level are widely acknowledged as a response to decreased agricultural economic viability (Marey-Pe ´rez et al. 2004). The lack of labor in rural areas brought about an increase in abandoned agricultural land, which was gradually occupied by shrub or planted with native trees. From the late 19th century, the Spanish government attempted to tackle the situation by promoting forest plantations primarily used for further processing in the fiber and chipboard industries. The measure reached its peak in the mid 20th century and was aimed at improving the low productivity, low profitability, and deforestation of Spanish forests (Marey-Pe ´rez and Rodrı ´guez-Vicente 2008). Later, European Council Regulation (EEC) no. 2080/1992 of 30 June 1992, instituting a community aid scheme for forestry measures in agriculture, favored the establishment of large forest plantations in Galicia, especially during the period 1993–1997, when grantaided afforestation constituted an important choice for private landowners who wished to improve the productivity of marginal lands through tree plantations, mainly with Pinus spp. and Eucalyptus spp. The addresses of NIPF owners and the attributes of their holdings, such as location, land uses, and size, were identified from the Land Register. In agreement with European Council Regulation (EEC) no. 571/1988 of 29 February 1988, on the organization of community surveys on the structure of agricultural holdings, the study population was composed of individual forest owners who owned at least 1 ha of productive forestland in Marin ˜a Oriental, which totaled 750 forest owners and covered 2,009 ha of productive forestland. The 1-ha threshold value was selected based on the 484 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 assumption that only forest owners with a certain land area would have the information necessary to account for their management goals and practices. A high percentage of the selected owners did not live, however, in the region where their land was located (Marey-Pe ´rez et al. 2004) because of the 20th century migration patterns in Northern Spain, mainly migration from rural to urban areas (Beiras-Torrado 1975). More specifically, over two-thirds of the population of nonresidents lived in other areas of Galicia, and the rest of nonresidents lived in other Spanish regions or in other countries. Consequently, we decided not to include owners who were emigrants or descendants of emigrants because they did not directly manage or monitor their land properties in the study area, and they did not have more or less continuous contact with the land. Moreover, including nonresident owners in the analysis would not guarantee the possibility of obtaining useful information for this research. Accordingly, we used population census data to restrict the study population to registered individual landowners who lived in the area. From among the original population of 750 NIPF owners, 376 were registered as residents, but only 333 actually lived in Marin ˜a Oriental. The remaining 43 NIPF owners were living in the house of a recently deceased relative. Consequently, we had access to around 50% of the owners who met the first condition for the study population. This group of landowners owned 1,154 ha of productive forestland, which accounts for 42% of all productive forestland in the region, including plots smaller than 1 ha. Land Register data proved rather divergent from population census: while according to the Land Register almost all the 750 cadastral landowners lived in Marin ˜a Oriental, the population census suggested that only 45% of landowners actually lived in the region. Because the large number of variables included in the Land Register was highly heterogeneous, stratification was a key factor in the characterization and subsequent validation of results. According to the Land Register database, the variable ‘‘productive forest area per landowner’’ was the most suitable variable to determine the minimum threshold of NIPF owners that should be interviewed and to stratify landowners. In order to determine the number of strata and the cutoff points required, NIPF owners were classified based on data pertaining to timber harvesting in Marin ˜a Oriental. Landowner stratification was defined by the size of productive forestland that enabled NIPF owners to fell the equivalent of the mean annual harvest per plot in the region (Marey-Pe ´rez 2003), set at 3.5 ha of productive forestland, considering a weighted rotation age of 15 years for the two main forest species in the region, Eucalyptus globulus Labill. and Pinus pinaster Ait. spp. atlantica. NIPF owners were classified into four groups according to this value (Table 1). We designed a questionnaire based on the subjective method of sample selection. Sample size was designed to achieve a 5% sampling error at the 95% confidence level. A priori, the error level was set at 3% for quantitative answers (mean estimation) and 6% for qualitative answers (proportion estimation). In order to obtain complete and reliable results, we enlarged as much as possible the interviewable landowner sample and minimized the economic costs of the interview during the design of the sampling size. A self-weighting design was used; sample size was determined accordingly and allocated using Neyman allocation (Sukhatme Characterization of NIPF owners and their influence on forest management 485 123 statistically confirmed this relationship. Educational level is usually interpreted as the owner’s ability to manage the existing resources and value them as new opportunities or management challenges, that is, formal education trains and encourages professionalism in certain management skills. In our analysis, landowners with secondary education were the most active planters in the area, with an annual rate of planting significantly different from the rate for uneducated landowners, as shown in Table 3(H=6.685). According to this table, owners without formal education and owners with secondary education are characterized by clearly different rates of planting. Considering the mean value obtained for this forest practice, these two education-level groups were termed ‘‘family planter’’ and ‘‘new planter,’’ respectively. The other groups of owners (primary education and university education) showed an annual rate of planting that could define them as either ‘‘family planter’’ or ‘‘new planter.’’ Mean annual expenditures on planting also varied significantly according to the EDUC group; once again, significant differences were between uneducated and secondary-education groups (H=4.713). On average, landowners with secondary education invested annually twice the amount invested by the uneducated group in planting. The owner profile allowed us to clarify the reason behind this forest management trend in the area. Landowner formal education was moderately and negatively correlated with age, as suggested in the above section (q=-0.550; P\0.01). As described in Table 3, the tertiary-education group included significantly younger landowners as compared with owners who were uneducated or had completed primary education; furthermore, the mean landowner age for uneducated and primary-education groups differed significantly (H=26.173). Considering the importance of the owner age in his/her educational level and labor situation, we Table 3 Mean annual rates of planting (%), owner age (year), family income (€/year), and size of productive forest holding (ha) for homogeneous EDUC subgroups EDUC No education Primary education Secondary education Tertiary education Pvalue Percentage of interviewed landowners 37.2 47.7 9.3 5.8 PLANT Family planter 0.69 1.76 1.54 0.765 New planter 1.76 4.17 1.54 0.088 AGE New forester 51.00 48.00 0.931 Farmer 62.24 51.00 0.116 Retiree 72.66 62.24 0.163 HOUSEHOLD Farmer traditionalist 13,137.75 18,935.55 18,875.54 0.535 New professional 18,935.55 18,875.54 26,714.99 0.270 SIZE Small landowner 3.26 5.15 6.68 0.150 Large landowner 5.15 6.68 8.14 0.248 492 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 additionally verified that the owner’s level of formal education and his/her primary occupation were positively correlated (D=0.406). All owners with tertiary education and half of the owners with secondary education were active workers not linked to agriculture, while 75% and 43.9% of owners who were uneducated or had completed primary education, respectively, were retired farmers; the largest share of active farmers, over 30%, was found in primary-education and secondary-education groups (v 2 =39.148). Landowner level of formal education was also positively correlated with training in forestry, and particularly with the knowledge of timber market conditions (D=0.145 and 0.188, respectively). Of landowners with secondary education 37.5% were trained in forestry, as compared with 5.2% of landowners who were uneducated or had completed primary education; none of the owners with tertiary education was trained in forestry (v 2 =10.780). With regard to the availability of timber market information, primary-education and secondary-education groups included the largest share of landowners with this type of knowledge, near 81%; 56.6% of uneducated owners and owners with tertiary education had information related to timber market. Furthermore, none of the owners with tertiary education had the machinery necessary to carry out forest activities; the best equipped holdings were managed by uneducated owners and owners with primary education, 75.6% and 87.5% of whom, respectively, owned some type of machinery (v 2 =14.839). As pointed out in the sections below, engagement in farming could be the reason why owners who had primary or secondary formal education were trained in forestry and had timber market knowledge and logistic resources. As suggested above, a better level of education would generally improve prospects and success in the labor market. The annual income per family unit weakly increased with the landowner’s level of education (q=0.336; P\0.01). Significant differences were found between uneducated and tertiary-education groups, with a mean difference in income of more than 13,550 €/year (H=10.396), as shown in Table 3. The positive significant correlations found between landowner’s educational level and annual family income or annual rate of planting could indirectly support the findings of many authors who have suggested that high plantation investments are made by landowners with greater household incomes (Doolittle and Straka 1987; Straka and Doolittle 1988; Hardie and Parks 1996; Gunter et al. 2001; Mahapatra and Mitchell 2001; Arano et al. 2004). For example, Beach et al. (2005) mentioned that landowner income may be normally used as a measure of their available resources for forest investment and, hence, better income may imply better access to the capital necessary for planting. However, annual income per family unit was not a significant factor in planting in the region, which is in agreement with Zhang and Flick (2001). We could only verify that the annual expenditure on forest planting increased slightly in relation to annual household income (q=0.220; P\0.05), and that there were significant differences in annual household income between the owners who did not invest in planting and the owners who spent between 127.0 and 250.0 €/ha annually (H=9.038). As described by Dewees (1992), Kurttila et al. (2001), and Marey-Pe ´rez et al. (2004), while some owners would be less dependent on forestry because of the increased proportion of other incomes, farmers as land managers would clearly Characterization of NIPF owners and their influence on forest management 493 123 seem to be more dependent on forestry as a source of revenue because of a reduced development of the agricultural income. This fact would explain why the annual fraction of forest reinvestments, i.e., the annual rate of forest products that are used within the family unit for self-consumption, increased weakly with the landowner’s level of education in the study area (q=-0.215; P\0.05). The owners with primary and secondary education, which were the groups with the largest share of active farmers, actively benefited from forests for self-consumption, with reinvestments averaging 60.4 €/ha per year. Uneducated landowners, mainly retired farmers, had a mean forest reinvestment valued at 90.5 €/ha per year; this mean value was four times the mean fraction for the tertiary-education group, composed of professionals outside agriculture, in which over two-thirds of them did not exploit forest for self-consumption. The size of productive forestland and the landowner’s level of formal education were barely and positively correlated (q=0.294; P\0.01), while the degree of parcellation of the productive forestland was barely and negatively correlated with the educational level (q=-0.236; P\0.05). Table 3shows that the mean values of the area of productive forestland in ownership were significantly different between educational groups. The productive forest holdings of uneducated owners were significantly smaller than the holdings belonging to owners with primary education (H=7.721). As regards the number of plots per unit of productive forestland, the largest rate of forest parcellation was observed in uneducated and secondary-education groups, where 26.6% of landowners managed more than 5.29 plots per hectare of productive forestland. As expected, 40% of landowners with tertiary education had \1.52 productive forest plots per hectare. Based on the characterization of landowner’s main occupation according to his/her level of formal education, we could verify that there might be a close link between land parcellation and farming activity. Primary occupation Karppinen (1998) affirmed that the most significant characteristic of the structural change among NIPF owners was the transfer of forest ownership from farmers to nonfarmers through inheritance, which should be reflected in forestry practices. As a measure of professional occupation, Gunter et al. (2001), Arano et al. (2004), and Potter-Witter (2005) confirmed that the place of residence of the landowner was significantly related to the planting decision. In the region, annual rates of silviculture and timber harvesting showed significant mean differences according to the owner’s occupation. As detailed in Table 4, entrepreneurs were the most active silviculturists in the area, differing significantly from the remaining groups, excluding selfemployees (H=5.155). Significant differences were found between the mean rates of timber harvesting for retirees and self-employees (H=3.053); retirees annually harvested seven times more woodland than self-employees, who showed the lowest annual rate of timber harvesting in the region. Authors such as Kuuluvainen and Salo (1991) and Kuuluvainen et al. (1996) found a negative statistical association between the intensity of timber harvesting and the landowner’s professional occupation. In Central Virginia, Conway et al. (2003) tested that absentee landowners, whose 494 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 residences were located at least 50 miles from their properties, were less likely to harvest. The strong and negative relationship between landowner age and his/her professional category (q=-0.772; P\0.01), and the positive relationship between primary occupational and level of formal education (D=0.406) allowed us to characterize landowners who annually carried out considerable forest improvement and timber harvesting in Marin ˜a Oriental. Table 4shows the means for the owner age that showed significant differences for the study population when such a population was classified according to professional occupation. Particularly, retired farmers were significantly older than the landowners in the other groups, excluding self-employees (H=58.282). In relation to landowner education level, over 55% of retired farmers did not have formal studies, as compared with 37.5% and 31.3% of active farmers and entrepreneurs, respectively, who had tertiary education (v 2 =39.148). Based on the landowner profile described here, it would not be surprising that annual income per family unit was slightly and positively correlated with landowner’s primary occupation (q=0.379; P\0.01). Retired farmers annually received a household income that was significantly lower than the family income of active farmers, with a mean difference of almost 11,500 €/year (H=21.396). Engagement in farming might probably account for the significant differences observed in the level of membership of professional groups and in the availability of machinery according to the OCCUP group (v 2 =24.687 and 12.935, respectively). Of retired and active farmers 49.5% were members of a professional organization, while 29.2% of hired workers and self-employees and none of the entrepreneurs and other professionals was associated. According to these results, the condition of the landowner as an absentee did not seem to be an added barrier to organizing professional groups, contrary to Finley’s (2002) report. The best equipped landowners were active farmers and self-employees, 95.3% of whom owned some kind of agroforestry machinery; logistic resources were available for almost twothirds of retirees and hired workers. Accordingly, the annual rate of silviculture seemed to be weakly and positively correlated with machinery availability (q=0.373; P\0.01), and the family labor-force devoted annually to the holding was based on accessibility to this resource (v 2 =11.158). Some of the machinery Table 4 Mean annual rates of silviculture (%) and owner age (years) for homogeneous OCCUP subgroups OCCUP Retiree Farmer Hired worker Self-employee Entrepreneur Other P-value Percentage of interviewed landowners 51.2 24.4 9.3 3.5 4.6 7.0 TREAT Nonsilviculturist 0.66 1.01 0.65 0.53 0.90 1.000 Silviculturist 6.29 1.000 AGE Active 54.90 48.67 50.88 55.25 53.33 0.763 Retiree 74.50 1.000 Characterization of NIPF owners and their influence on forest management 495 123 available, such as chain-saws, pruning-saws or hand-weeders, could be used for forest improvement activities by family labor. Therefore, the annual amounts of personal and family labor spent within the holding were interrelated (v 2 =101.468). The landowners who devoted between 11 and 100 personal labor-days per year to forestry annually benefited from a family labor-force ranging from 11 to 50 labor-days for forestry, while owners who annually spent \5 personal labor-days on the holding received \10 family labordays per year. In addition, professional assistance in forestry proved an important factor in keeping and making forestry viable in the area, as suggested by other studies about NIPF management (Lo ¨yland et al. 1995; Hardie and Parks 1996; Zhang and Flick 2001; Zhang and Mehmood 2001). Thus, the annual amount of professional labor-force on the holding differed significantly according to landowner primary occupation (v 2 =36.601). In fact, the annual rates of planted and improved forestlands increased moderately with the number of professional labor-days annually hired in forestry (q=0.408; P\0.01 and 0.251; P\0.05, respectively). More specifically, 67% of retired and active farmers hired professional labor amounting to\5 days per year, in comparison with 44.5% of the other owner groups, who annually hired professional labor amounting to 51–100 labordays. These results could clarify the noticeable commitment of professionals outside agriculture to forest improvements, who seemed to hire technical guidance to compensate for their absenteeism in forestry. The land acquisition pattern varied significantly based on the main occupation of owners (v 2 =22.191). Retired farmers represented a specific group, with holdings that were mainly acquired by purchasing or inheriting. In brief, retired farmers would be more likely to make land transactions, probably to improve and increase their former agricultural productivity, and hence show a more significant land mobility (Marey-Pe ´rez et al. 2004). Thus, whereas 59.1% of retirees managed inherited and purchased lands, purchase was the sole pattern of land acquisition for 22.9% of hired workers and self-employees; inherited and purchased lands were combined in 25.8% of the remaining occupational groups. This orientation towards farming could again clarify the significant differences found in the annual rate of forest reinvestment between retired or active farmers and self-employees or entrepreneurs (H=15.833). Over 70.5% of retired and active farmers took advantage of forest reinvestments between 71 and 233.7 €/ha per year, whereas 66.7% of self-employees owners and none of the entrepreneurs obtained forest products for self-consumption. Annual expenditure on forest plantation and landowner’s professional category were weakly and positively correlated (q=0.235; P\0.05). None of the active farmers allocated more than 250 €/ha per year on planting forestlands, probably associated with a greater likelihood of managing agricultural land, but large expenditures were exclusively incurred by retired farmers or professionals unrelated to agriculture. Thus, self-employees and entrepreneurs annually invested in planting the highest amounts in the region, over 355 €/ha per year. This annual amount exceeded the mean expenditure on planting by retired and active farmers, almost 165 €/ha per year. In keeping with previous studies (Karppinen 1998; Gunter et al. 2001; Marey-Pe ´rez et al. 2004), the owners that were more likely to invest in lands 496 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 to keep them productive seemed to be the owners outside agriculture who worked part-time at their property, and also retired and older landowners. The key role of public subsidies in forest involvement in Marin ˜a Oriental was again evidenced by the large share of owners who were financially compensated by means of public funding in forestry. Thus, the annual amount of public subsidies and the landowner’s professional occupation were barely and negatively correlated (q=-0.236; P\0.05). None of the landowners unrelated to agriculture, that is, hired workers, self-employees, entrepreneurs, and other professionals, received public financial aid for forestry. On the contrary, the greatest beneficiaries of this type of measure were farmers; 76.9% of retired owners and all of the active farmers who asked for public incentives for forestry finally received the subsidies, with a mean payment of 54.7 €/ha per year. The degree of parcellation of productive forestland and the landowner’s occupational group were barely and negatively correlated (q=-0.265; P\0.05), which confirms that agricultural productivity could be an important factor in land parcellation in the study area. The largest number of plots per unit of productive forestland corresponded to retired and active farmers, with an average of 3.5 plots per hectare of productive forestland. Conversely, hired workers and entrepreneurs managed the least parcelled holdings, with over 2.4 plots per unit of productive forestland. Condition as an active farmer Literature concerned with the land management behavior of NIPF owners suggests that farmers, as land managers, are distinct from other landowners in terms of their commitment to and involvement in forestry. Thus, Hardie and Parks (1996) found that the level of planting was negatively associated with the landowner’s condition as a farmer. However, Hyberg and Holthausen (1989) observed that the choice of timber harvesting was positively related to the landowner’s profile as a farmer. Similarly, Kuuluvainen and Salo (1991) reported that farmers were characterized by significantly lower harvest volumes than other types of NIPF owners. In Marin ˜a Oriental, only a weak positive correlation was found between the annual rate of silviculture and the landowner’s condition as an active farmer (q=0.212; P\0.05). The fact that active farmers were slightly more active with regard to silviculture could be explained by considering that these landowners were occupied in agriculture full-time (H=3.829). In agreement with Lo ¨yland et al. (1995), landowner’s occupation outside his/her property may mean having less time available for working the land, and therefore the landowner is less likely to carry out forest practices him/herself. Because of their close association with the land, farmers would actively manage their property themselves and, particularly, they would generally have more time for forest management (Zhang and Mehmood 2001; Lindroos et al. 2005). The moderate and negative relationship between landowner age and his/her condition as an active farmer (q=-0.451; P\0.01), and the small and positive relationship between the annual family income and the owner’s occupational profile (q=0.343; P\0.01) allowed us to characterize active farmers in the area. Post hoc analyses revealed that the mean age of active farmers was significantly Characterization of NIPF owners and their influence on forest management 497 123 lower than the mean age of nonfarmers, with a difference of almost 15 years (H=17.297). With regard to the landowner’s household income, active farmers annually received an average family income of 23,855 €, exceeding by more than 8,800 €the income of the remainder (H=10.023). Such distributions would be probably associated with the inclusion of all the retirees in the nonfarmer group (which accounts for 51.2% of the whole study population in Marin ˜a Oriental), which would increase owner age and decrease the landowner’s mean earning per family unit in this group. As suggested in preceding sections, the professional associations in the area could be mainly agricultural groups, especially cooperatives or trade unions. The equipment available on the majority of holdings could be a proof of the nature of the associations. The assumptions were confirmed by the positive correlations observed between these two factors and the owner’s condition as an active farmer (D=0.508 and 0.312, respectively). Only 20% of nonfarmers were members of a professional organization, as compared with 76.2% of active farmers (v 2 =22.422). With regard to the availability of agroforestry machinery, active farmers were characterized by a large personal labor-force annually devoted to forest management, which would explain why 90.5% of them had equipment on their holding, as compared with 55.4% of the other landowners (v 2 =8.478). In fact, the agricultural link might also explain the low increase in the annual amount of forest reinvestments when the landowner was actively related to agriculture (q=0.304; P\0.01). The active farmers annually took advantage of forest reinvestments twice as often as the nonfarmer group (H=7.840). Moreover, the profile of owners as active farmers and their production requirements and goals for land management would explain why such landowners could consider the option of using part of their forestlands for agricultural production. Consequently, significant differences in the likelihood of converting forestlands into meadows according to the FARM group were tested (v 2 =6.338). As expected, none of the nonfarmers adopted this productive orientation for their holding, and only 9.5% of active farmers considered such an option. Furthermore, we could verify that professional engagement in agriculture and the pattern of land transmission were negatively correlated, for the reasons already explained (D=- 0.255). Of the nonfarmer group 53.8% were characterized by managing inherited and purchased lands, as compared with 19% of active farmers; for each nonfarmer who owned a fully inherited landholding, there were almost two active farmers who owned this type of holding (v 2 =7.769). Patterns of land acquisition and future transmission may be important indicators of NIPF owner land decisions; thus, Hardie and Parks (1996) and Ross-Davis et al. (2005) showed that maintaining the viability of the property by land management by future landowners through inheritance may be a key in securing continuity of the forestry sector. The following results verified this hypothesis about land capitalization for the study area: first, the pattern of land acquisition in the region significantly influenced the annual rate of planting forestlands (H=5.973), and second, annual expenditure in planting weakly increased in forest holdings that combined inherited and purchased lands (q=0.224; P\0.05). Landowners whose holdings combined inherited and purchased lands, which was a group largely represented by retirees, 498 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 annually planted over seven times more forestland than owners whose holdings were acquired solely through purchasing, and almost twice the forestland planted by owners whose holdings were acquired solely through inheritance. In addition, owners with inherited and purchased lands annually spent almost ten times more in planting than the others (H=6.732). A negative association was found between the conversion of marginal meadows into woodlands and the condition of the landowner as an active farmer (D=- 0.233). The farming group hardly considered such a shift in land use for their holding (9.5%), as compared with 33.8% of nonfarmers who put it into practice in order to keep their land productive (v 2 =4.667). Therefore, land capitalization might largely respond to the owner’s willingness to improve land productivity and to ensure a complementary source of household income by means of forest investment. Landowners who implemented such a land-use change declared having done so as a way of ceasing farming (67%) and because of the higher profitability of forestry (21%). These findings would support the statement by Beach et al. (2005) that the increase in forest income with respect to agricultural income may tend to increase forest management. Membership of agricultural and forestry groups Landowners who are members of cooperative organizations share information, techniques, experiences, and advice with one another (Kittredge 2005; Van Gossum and De Maeyer 2007). The most representative model of professional associations in the study area was a private professional group partially supported by public funds. Participating as a member in this type of landowner organization appeared to be a significant determinant in the annual planting and silviculture management behaviors in Marin ˜a Oriental (H=3.648 and 3.376, respectively). The associated landowners were slightly more active with regard to planting and silviculture than the nonassociated group. The biggest impact on uptake of planting by the segment of landowners who were members of an organization was also tested by Doolittle and Straka (1987), Straka and Doolittle (1988), and Mahapatra and Mitchell (2001). Such active forest management behavior of associated landowners could be due to the fact that professional groups offer information sources and technical advice for members. However, we were unable to associate this finding with the access of landowners to professional services from agricultural and forestry groups, as Mahapatra and Mitchell (2001), Kittredge (2005) or Van Gossum et al. (2005) suggested. The important role of professional groups in planting and silviculture management would seem to be related to the owner’s engagement in farming. As cited above, the level of participation in a professional body was positively correlated with the landowner’s occupation in agriculture (D=0.508). Only 8.8% of nonassociated owners were active farmers, as opposed to almost 55.2% of associated owners (v 2 =22.422). More specifically, 59.6% of nonassociated owners were retired farmers, as compared with more than 34.5% of associated owners (v 2 =24.687). This result could explain why participation in such groups was weakly and negatively correlated with owner age (q=-0.271; P\0.05), but weakly and Characterization of NIPF owners and their influence on forest management 499 123 positively correlated with annual family income (q=0.259; P\0.05). Associated owners were almost 10 years younger than nonassociated owners (H=6.249). With regard to the annual household income, associated owners earned almost 7,000 €/ year more than nonassociated owners (H=5.691). The owner profile, and particularly his/her occupation as an active farmer, could justify that the level of participation in professional groups and the availability of equipment on the holding were positively correlated (D=0.433). There were two associated owners for every nonassociated owner who had suitable agroforestry machinery to support land management (v 2 =16.128). Technology users would seem to be more likely to participate in social groups and to share experience, which verifies the findings reported by Hodges and Cubbage (1990). Taking into consideration the statistical relationship between the availability of machinery and the annual rate of silviculture, and the significant differences in this forest practice depending on the landowner’s level of participation in professional groups, it would not be surprising to find that associated owners were more interested in changing their current productive forest species in the short/mid-term. Thus, the intention of replacing the current productive forest species in the future increased with the level of participation of landowners in professional groups (D=0.207). The number of associated owners who intended to change the main productive forest species in the near future was matched by the number of nonassociated owners, possibly due to the larger labor-force devoted to silviculture (v 2 =3.902). This statement was supported by the finding that 80% of owners intended to replace the current productive forest species with Eucalyptus globulus Labill. The weak positive correlations found between this future land-use intention and the annual rate of and expenditure on silviculture confirmed this hypothesis (q=0.238 and 0.216, respectively; P\0.05). Landowners who intended to replace the main productive forest species annually treated four times more forestland (H=4.809), and invested in silviculture almost twice as much as those who did not (H=3.833), hence their interest in eucalyptus, a fast-growing forest species that is highly productive and easy to manage. Other future intentions, such as increasing woodlands on the holding, increased with the participation of landowners in professional groups (D=0.211). Over 58% of associated owners mentioned their intention of increasing their productive forestlands in the near future, as compared with 37% of nonassociated owners (v 2 =3.697). The major reason for this behavior could be attributed to previous harvests and timber sales, that is, to the interest in timber production. In fact, the annual rate of harvesting woodlands in the region increased strongly in proportion to annual income from timber sales and stumpage price per unit (q=0.809 and 0.781, respectively; P\0.01). Kuuluvainen and Salo (1991), Bolkesjø and Baardsen (2002), and Bolkesjø et al. (2007) statistically proved that timber price (roundwood, pulpwood or sawtimber) positively affects harvest choice and intensity or volume, and timber supply. The future intention of enlarging woodlands weakly increased with the annual rate of timber harvesting (q=0.242; P\0.05). Landowners who had the intention of enlarging woodlands harvested annually twice as much woodland as the remaining owner population (H=4.972), with an annual timber income of 100 €/ha more (H=3.118) at almost twice the stumpage price per unit 500 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 (H=3.104). On average, associated landowners annually benefited from a 1.5 times higher timber income, with a stumpage price per unit 0.61 €/T higher than the price for the nonassociated group. In addition to the purpose of changing the current productive forest species, the future increase in woodlands seemed to correspond to landowners with an extensive surface area. In keeping with Hodges and Cubbage (1990) and Van Gossum et al. (2005), we observed that the productive forest holdings managed by owners who were members of a professional group were somewhat larger than the holdings of owners uninterested in participating in such groups. As mean values, associated owners had almost 1 ha more than the nonassociated group (H=3.768). Table 5shows the results of the logistic regression model for estimating landowner membership of agricultural and forestry associations. At 1% statistical significance, the fitted model correctly predicted 77.9% of the overall observations. The binary variables FARM and MACHINERY, represented as the status of the owner as an active farmer and his/her availability of agroforestry machinery within the holding, respectively, and the binary variable IFOREST, represented as the owner’s intention of enlarging the productive forestland in the short/mid-term, proved to have a significant positive effect on explaining the landowners’ condition as a member of a professional group in the study region, as previously suggested and analyzed: PðASSOCÞ¼ 1 1þeð3:873 þ1:425IFOREST þ2:380MACHINERY þ2:460FARMÞ These results show that active farmers were almost 12 times more likely to be included in an agroforestry association than were retired farmers or other professionals not related to agriculture. Those landowners who had agroforestry machinery within their holding as support to land management were almost 11 times more likely to partake in an association related to the sector than landowners without any type of logistic resources. Finally, those owners who had the future aim Table 5 Parameter estimates of the logistic regression model that examines the factors affecting NIPF landowner membership of agroforestry associations Variable Coefficient Wald P-value Standard error FARM 2.460 12.661 0.000 0.691 MACHINERY 2.380 7.750 0.005 0.855 IFOREST 1.425 5.167 0.023 0.627 Constant -3.873 17.012 0.000 0.939 -2 Log likelihood 71.637 Model v2 38.300* Nagelkerke R 2 0.498 Obs. with ASSOC =1 51.7 Obs. with ASSOC =0 91.2 Overall % correct 77.9 *PB0.01 Characterization of NIPF owners and their influence on forest management 501 123 Table 7 continued Variable Code Definition No. of interviewed NIPF owners OCCUP Nominal Main primary occupation of the owner 1 If owner was a retired owner 53 2 If owner was an active farmer 24 3 If owner was a hired worker 10 4 If owner was a self-employee 4 5 If owner was an entrepreneur 5 6 If other 7 FARM Binary Condition of the owner as an active farmer 1 If owner was an active farmer 24 0 If otherwise 79 ASSOC Binary Participation of the owner in professional associations 1 If owner was a member of a professional association 35 0 If otherwise 68 TRAINING Binary Forestry training of the owner 1 If owner participated in a forestry course 8 0 If otherwise 95 IMARKET Binary Specific training in market (timber prices, supplydemand, etc.) 1 If owner had market information 35 0 If otherwise 68 TECHNIC Nominal Knowledge and use of production criteria for timber harvesting 1 If owner did not know about and did not take into account rotation age 8 2 If owner knew about and did not take into account rotation age 84 3 If owner knew about and took into account rotation age 11 Family unit INHERIT Nominal Acquisition of the forest holding 1 If owner inherited lands 53 2 If owner inherited and bought lands 46 3 If owner bought lands 4 BEQUEST Nominal Intention of bequeathing the forest holding 1 If owner intended to bequeath lands to heirs 97 2 If owner intended to bequeath some lands to heirs and sell the remainder 2 3 If owner intended to sell lands 4 HOUSEHOLD Ordinal Annual net family income in euros during 1999–2003 (x=17,224.94; r=10,339.69) 1 If net household income was \6,000 11 2 If net household income was between 6,000 and 9,000 18 3 If net household income was between 9,001 and 18,000 37 508 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 Table 7 continued Variable Code Definition No. of interviewed NIPF owners 4 If net household income was between 18,001 and 30,000 20 5 If net household income was more than 30,000 17 REINVEST Ordinal Annual reinvestment for household consumption in euros during 1999–2003, per unit of forest area (x=71.01; r=81.32) 0 If owner did not obtain reinvestments 29 1 If reinvestment was \71.0 33 2 If reinvestment was between 71.0 and 152.4 24 3 If reinvestment was between 152.5 and 233.7 11 4 If reinvestment was between 233.8 and 315.0 4 5 If reinvestment was more than 315.0 2 PERSONAL Ordinal Personal labor-days spent annually on forestry during 1999–2003 1 If personal labor was \231 2 If personal labor was between 2 and 5 11 3 If personal labor was between 6 and 10 13 4 If personal labor was between 11 and 50 32 5 If personal labor was between 51 and 100 12 6 If personal labor was more than 100 4 FAMILY Ordinal Family labor-days spent annually on forestry during 1999–2003 1 If family labor was \244 2 If family labor was between 2 and 5 12 3 If family labor was between 6 and 10 8 4 If family labor was between 11 and 50 15 5 If family labor was between 51 and 100 18 6 If family labor was more than 100 6 MACHINERY Binary Logistic resources available for forestry activities 1 If owner had agricultural and forestry machinery 66 0 If otherwise 37 PROFESS Ordinal Professional labor-days spent annually on forestry during 1999–2003 1 If professional labor was \241 2 If professional labor was between 2 and 5 6 3 If professional labor was between 6 and 10 23 4 If professional labor was between 11 and 50 12 5 If professional labor was between 51 and 100 19 6 If professional labor was more than 100 2 Forest property and land-use changes FMEADOW Binary Past conversion of forestland into meadow during 1999–2003 Characterization of NIPF owners and their influence on forest management 509 123 Table 7 continued Variable Code Definition No. of interviewed NIPF owners 1 If owner made this land-use change 2 0 If otherwise 101 MWOOD Binary Past conversion of marginal meadow into woodland during 1999–2003 1 If owner made this land-use change 29 0 If otherwise 74 CSPECIE Binary Future intention of changing the current productive forest species 1 If owner had this future purpose 23 0 If otherwise 80 IFOREST Binary Future intention of increasing the productive forestland 1 If owner had this future purpose 46 0 If otherwise 57 PLOT Ordinal Number of plots per hectare of productive forestland in ownership (x=3.40; r=1.88) 1 If fragmentation degree was smaller than 1.52 14 2 If fragmentation degree ranged between 1.52 and 3.40 41 3 If fragmentation degree ranged between 3.41 and 5.29 29 4 If fragmentation degree was larger than 5.29 19 SIZE Ordinal Area of productive forestland in ownership, in hectares (x=4.76; r=3.86) 1 If ownership sized between 1.00 and 1.70 22 2 If ownership sized between 1.71 and 3.50 24 3 If ownership sized between 3.51 and 7.00 32 4 If ownership sized more than 7.00 25 Forest economics INVEST Ordinal Annual investment in holding improvement in euros during 1999–2003, per unit of forest area (x=10.35; r=24.56) 0 If owner did not invest in the forest holding 76 1 If owner invested \39.7 15 2 If owner invested more than 39.7 12 PEXP Ordinal Annual expenditure on plantation in euros during 1999– 2003, per unit of forest area (x=193.48; r=176.28) 0 If owner did not spend on planting forestlands 5 1 If owner spent \127.0 61 2 If owner spent between 127.0 and 250.0 25 3 If owner spent between 250.1 and 400.0 7 4 If owner spent between more than 400.0 5 TEXP Ordinal Annual expenditure on silviculture treatments in euros during 1999–2003, per unit of forest area (x=71.83; r=81.76) 510 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 References Amacher GS, Conway MC, Sullivan J (2003) Econometric analyses of nonindustrial forest landowners: is there anything left to study? J For Econ 9:137–164 Arano KG, Munn IA (2006) Evaluating forest management intensity: a comparison among major forestland owner types. For Policy Econ 9:237–248 Table 7 continued Variable Code Definition No. of interviewed NIPF owners 0 If owner did not spend on forestland improvement 12 1 If owner spent \90.7 63 2 If owner spent between 90.7 and 221.7 20 3 If owner spent more than 221.7 8 REQUEST Binary Formal application for a forest management subsidy during 1999–2003 1 If owner applied for economic aid 22 0 If otherwise 81 SUB Ordinal Annual forest subsidy in euros during 1999–2003, per unit of forest area (x=4.92; r=20.09) 0 If owner did not apply for economic aid or was not finally compensated 95 1 If owner received \25.0 2 2 If owner received between 25.0 and 50.1 2 3 If owner received between 50.2 and 75.2 2 4 If owner received more than 75.2 2 TINCOME Ordinal Annual income from timber sales in euros during 1999– 2003, per unit of forest area (x=166.62; r=193.39) 0 If owner did not receive timber income 46 1 If owner received \195.1 18 2 If owner received between 195.1 and 425.2 19 3 If owner received more than 425.2 20 TPRICE Ordinal Stumpage price in euros during 1999–2003, per ton (x=3.99; r=4.25) 0 If owner did not sell timber 46 1 If stumpage price per unit was \48.9 40 2 If stumpage price per unit was between 48.9 and 74.3 11 3 If stumpage price per unit was more than 74.3 6 NTINCOME Ordinal Annual income from land sales in euros during 1999– 2003, per unit of forest area (x=170.79; r=281.77) 0 If owner did not receive nontimber income 35 1 If owner received \381.8 52 2 If owner received more than 381.8 16 Characterization of NIPF owners and their influence on forest management 511 123 Arano KG, Munn IA, Gunter JE, Bullard SH, Doolitle ML (2004) Comparison between regenerators and non-regenerators in Mississippi: a discriminant analysis. South J Appl For 22(2):132–138 Beach RH, Pattanayak SK, Yang J, Murray BC, Abt RC (2005) Econometric studies of non-industrial private forest management a review and synthesis. For Policy Econ 7:261–281 Beiras-Torrado XM (1975) A emigracio ´n: o seu papel na dina ´mica da formacio ´n social. In: Garcı ´a-Sabell D (ed) A Galicia rural na encrucillada. Vigo, Spain, pp 39–73 Bolkesjø TF, Baardsen S (2002) Roundwood supply in Norway: micro-level analysis of self-employed forest owners. For Policy Econ 4:55–64 Bolkesjø TF, Solberg B, Wangen KR (2007) Heterogeneity in nonindustrial private roundwood supply: lessons from a large panel of forest owners. J For Econ 13:7–28 Boon TE, Meilby H (2007) Describing management attitudes to guide forest policy implementation. Small-Scale For 6(1):79–92 Butler BJ, Swenson JJ, Alig RJ (2004) Forest fragmentation in the Pacific Northwest: quantification and correlations. For Ecol Manag 189:363–373 Cao-Abad R (2002) Ana ´lisis multivariante. Curso de postgrado en estadı ´stica aplicada. University of A Corun ˜a, A Corun ˜a, Spain Conway MC, Amacher GS, Sullivan BJ (2003) Decisions non-industrial forest landowners make: an empirical examination. J For Econ 9:181–203 Dennis D (1990) A profit analysis of the harvest decision using pooled time-series and cross-sectional data. J Environ Econ Manag 18:176–187 Dewees PA (1992) Social and economic incentives for smallholder tree growing: a case study from Muranga District, Kenya. Community forestry case study series 5, FAO, Rome Doolittle L, Straka TJ (1987) Regeneration following harvest on nonindustrial private lands in the south: a diffusion of innovations perspective. South J Appl For 11(1):37–41 Finley AO (2002) Assessing private forest landowners’ attitudes towards, and ideas for, cross-boundary cooperation in western Massachusetts. PhD thesis, University of Massachusetts, Amherst, 28pp Gunter JE, Bullard SH, Doolitle ML, Arano KG (2001) Reforestation of harvested timberlands in Mississippi: behaviour and attitudes of non-industrial private forest landowners. Mississippi State University, Mississippi, 25 pp Hardie IW, Parks PJ (1996) Program enrolment and acreage response to reforestation cost-sharing programs. Land Econ 72:248–260 Herbohn J (2001) Prospects for small-scale forestry in Australia. In: Niskanen A, Va ¨yrynen J (eds) Economic sustainability of small-scale forestry. European Forest Institute, Finland, EFI Proceedings No. 36, pp 9–20 Hodges DG, Cubbage FW (1990) Adoption behaviour of technical assistance foresters in the Southern Pine Region. For Sci 36(3):516–530 Hosmer DW, Lemeshow S (2000) Applied logistic regression, 2nd edn. Wiley, New York, USA Hugosson M, Ingemarson F (2004) Objectives and motivations of small-scale forest owners: theoretical modeling and qualitative assessment. Silva Fennica 38(3):217–231 Hyberg B, Holthausen D (1989) The behavior of non-industrial private forest landowners. Can J For Res 19:1014–1023 INE (1999) Agrarian census. Instituto Nacional de Estadı ´stica, Madrid, Spain INE (2004) Consumer Price Index. Instituto Nacional de Estadı ´stica, Madrid, Spain. DIALOG, http://www.ine.es/daco/ipc.htm. Accessed 6 Dec 2004 Ingemarson F, Lindhagen A, Eriksson L (2006) A typology of small-scale private forest owners in Sweden. Scand J For Res 21(3):249–259 Karppinen H (1998) Values and objectives of non-industrial private forest owners in Finland. Silva Fennica 32(1):43–59 Kittredge DB (2005) The cooperation of private forest owners on scales larger than one individual property: international examples and potential application in the United States. For Policy Econ 7:671–688 Kline JD, Butler BJ, Alig RJ (2002) Tree planting in the South: what does the future hold? South J Appl For 26(2):99–107 Kurttila M, Hamalainen K, Kajanus M, Pesonen M (2001) Non-industrial private forest owners’ attitudes towards the operational environment of forestry - a multinomial logit model analysis. For Policy Econ 2(1):13–28 Kuuluvainen J, Salo J (1991) Timber supply and life cycle harvest of non-industrial private forest owners: an empirical analysis of the Finnish case. For Sci 37:1011–1029 512 V. Rodrı ´guez-Vicente, M. F. Marey-Pe ´rez 123 Kuuluvainen J, Karppinen H, Ovaskainen V (1996) Landowner objectives and non-industrial private timber supply. For Sci 4:300–308 Lindroos O, Lidestav G, Nordfjell T (2005) Swedish non-industrial private forest owners: a survey of self-employment and equipment investments. Small-Scale For 4(4):409–425 Lo ¨yland K, Kringstad V, O ¨y H (1995) Determinants of forest activities - a study of private non-industrial forestry in Norway. J For Econ 1(2):219–237 Mahapatra AK, Mitchell CP (2001) Classifying tree planters and non planters in a subsistence farming system using a discriminant analytical approach. Agrofor Syst 54:41–52 Marey-Pe ´rez MF (2003) Tenencia de la tierra en Galicia: modelo para la caracterizacio ´n de los propietarios forestales. PhD thesis, University of Santiago de Compostela, Spain, 633 pp Marey-Pe ´rez MF, Rodrı ´guez-Vicente V (2008) Forest transition in Northern Spain: local responses on large-scale programmes of field-afforestation. Land Use Policy 26(1):139–156 Marey-Pe ´rez MF, Rodrı ´guez-Vicente V, Crecente-Maseda R (2004) El monte en Galicia en el siglo XXI: Balance evolutivo y consideraciones para el futuro. In: Maya-Frades A (ed) >Que ´futuro para los espacios rurales?. University of Leo ´n, Leo ´n, Spain, pp 117–125 Marey-Pe ´rez MF, Rodrı ´guez-Vicente V, Crecente-Maseda R (2006) Using GIS to measure changes in the temporal and spatial dynamics of forestland: experiences from north-west Spain. Forestry 79(4):409–423 MMA (1998) III Inventario Forestal de Espan ˜a. Direccio ´n General de Conservacio ´n de la Naturaleza, Ministerio de Medio Ambiente. Madrid, Spain Niskanen A, Pettenella D, Slee B (2007) Barriers and opportunities for the development of small-scale forest enterprises in Europe. Small-Scale For 6(4):331–345 Pattanayak SK, Murray BC, Abt R (2002) How joint in joint forest production: an econometric analysis of timber supply conditional on endogenous amenity values. For Sci 48(3):479–491 Potter-Witter K (2005) A cross-sectional analysis of Michigan non-industrial private forest landowners. South J Appl For 22(2):132–138 Prestemon J, Wear D (2000) Linking harvest choices to timber supply. For Sci 46(3):377–389 Ross-Davis AL, Broussard SR, Jacobs DF, Davis AS (2005) Afforestation motivations of private landowners: and examination of hardwood tree plantings in Indiana. North J Appl For 22(3):149– 153 Ryan TP (1997) Modern regression methods. Wiley, New York, USA Størdal S, Lien G, Baardsen S (2008) Analyzing determinants of forest owners’ decision-making using a sample selection framework. J For Econ 14:159–176 Straka TJ, Doolittle S (1988) Propensity of nonindustrial private forest landowners to regenerate following harvest: relationship to socioeconomic characteristics, including innovativeness. Resour Manag Optim 6(2):121–128 Sukhatme PU (1953) Sampling theory of surveys. FAO, Rome Van Gossum P, De Maeyer W (2007) Performance of forest groups in achieving multifunctional forestry in Flanders. Small-Scale For 5(1):19–36 Van Gossum P, Luyssaert S, Serbruyns I, Mortier F (2005) Forest groups as support to private forest owners in developing close-to-nature management. For Policy Econ 7:589–601 Wear D, Parks P (1994) The economics of timber supply: an analytical synthesis of modelling approaches. Nat Resour Modell 8(3):199–223 Zhang D, Flick W (2001) Sticks, carrots and reforestation investment. Land Econ 77(3):443–456 Zhang D, Mehmood SR (2001) Predicting non-industrial private forest landowners’ choice of a forester for harvesting and tree planting assistance in Alabama. South J Appl For 25(3):101–107 Characterization of NIPF owners and their influence on forest management 513 123 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research III MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I ORIGINAL ARTICLE Assessing the role of the family unit in individual private forestry in northern Spain VERO ´NICA RODRI ´GUEZ VICENTE 1 & MANUEL FRANCISCO MAREY PE ´REZ 2 1 Galician Sectorial Forestry Association (ASEFOGA), Santiago de Compostela, Spain, and 2 Department of Agroforestry Engineering, University of Santiago de Compostela, Lugo, Spain Abstract Farm forestry has been often linked to family knowledge and needs, and even to local expertise through several generations. Among the several factors that may influence farm forestry, family welfare and support on forest decision making and management are nowadays key arguments to provide a richer and better understanding of the land behaviour of non-industrial private forest owners (NIPFOs). This paper empirically explores and assesses the potential direct effects of the characteristics of the family unit (bequests, household income, forest reinvestments, personal and family labour, logistic resources and professional assistance) on individual forest management in terms of planting, silvicultural and harvesting practices. In March 2004, 103 forest landowners were personally interviewed about their commitment to and involvement in land management during 19992003, considering a forest region in northern Spain. The pattern of land acquisition, household dependence on forest products for self-consumption, the availability of machinery, in addition to family labour force and technical guidance in forestry, are all significantly related to the ability to manage and use forestland as a capital asset. These issues may be essential for advancing research in individual forestry and for improving policy objectives and programmes on forest planning and management within the increasing demands for sustainable forestry and rural development. Keywords: Bequest, forest labour, forest reinvestment, household income, logistic support, non-industrial private forest owner. Introduction Nowadays, the territorial system is characterized worldwide by urban development. Populations tend to concentrate in large cities, where the supply of services and infrastructures available is greater and better than in rural areas. Such a polarization of population has brought about patterns of gradual and increasing socioeconomic recession in rural areas, changing land ownership and reducing land management as the requirements and prospects of land managers adapt to new economic opportunities. In the face of such a process of loss of values and rural heritage, agricultural and forestry activities are no longer economically viable for many rural communities, and many rural landscapes have begun to be abandoned because of the lack of an active population and the ageing residents who maintain and conserve them (Karppinen, 2005; Kittredge, 2005; Marey et al., 2006). Consequently, as rural landscapes become increasingly unmaintained, many approaches to land management become more relevant at different scales. In the past few decades, the unsustainable growth of a large part of the urban areas and the new needs of the current population have promoted the consideration of forests as the social, economic and environmental hope for a future recovery and renewal of rural areas, aimed at reorientating their traditional values and activities (Marey et al., 2004). Therefore, scientific and public organizations, as well as policy statements of many countries, have begun to focus on rural development issues, especially those centred on forest sustainability, and try to resolve questions such as which style of forestry can generate the best bundle of benefits by means of land-use decision making among diverse stakeholders (Slee & Wiersum, 2001). This issue is particularly worrying in regions where forestland is largely owned and managed by non-industrial private forest owners (NIPFOs), and Correspondence: V. Rodrı´guez Vicente, Doutor Maceira 13 baixo ES-15706, Santiago de Compostela, Spain. E-mail: consulto[email protected] Scandinavian Journal of Forest Research, 2008; 23: 5377 (Received 7 January 2007; accepted 24 August 2007) ISSN 0282-7581 print/ISSN 1651-1891 online #2008 Taylor & Francis DOI: 10.1080/02827580701672212 percentage of interviewed owners for each category, as well as the mean and SD parameters observed for planting, silvicultural and harvesting practices during the 5 year study period. Finally, all the economic variables analysed in the present study were adjusted to constant euros for 2004 to control the inflation rate, and summarized in mean annual eco amounts per hectare owned, with the exception of family households. The information source was the Spanish consumer price index of the National Statistics Institute (INE, 2004). Statistical analyses Given that the study population was not adjusted to the KolmogorovSmirnov goodness-of-fit test for the normality K-S test, or to the Levene test for homogeneity of variances, the statistical analyses conducted in this study were based on distributionfree tests, that is, non-parametric tests. The nonparametric procedure took into account the typology of variables measured in the study, i.e. continuous, nominal/ordinal and binary variables, and was defined to explain statistically the planting, silvicultural and harvesting management practices in Marin ˜a Oriental in relation to family unit characteristics. In addition, these explanatory variables were associated with other attributes linked to landowner profile, forest property, land-use changes and forest economics. Thus, first, the strength and significance of the linear correlation among the variables were tested, and then the significant differences among them contrasted at a 95% confidence limit and a minimum 0.05 level of statistical significance. Using Somers’Dcoefficient and its critical significance level, the statistical relationship among nominal, ordinal and/or binary variables was measured using contingency tables. Table III details the degree of correlation found for this type of explanatory factor considered in the study. Significant differences in the frequency distribution across nominal, ordinal and/or binary variables of the cross-tabulation were computed and detected using Pearson’s chi-square statistic (x 2 ) and the two-tailed asymptotic significance (Table IV). Spearman’s rho coefficient (r) was used to estimate the statistical association between the continuous variables and these measures versus nominal, ordinal and/or binary attributes at 0.01 and 0.05 significance levels. Table V shows the Spearman’s correlation coefficients for the three forest management practices of reference and the family attributes. This table also shows the values of the correlation parameter between the group of characteristics selected to describe the family unit and the remaining explanatory variables. Later, the mean distribution Table II (Continued) PLANT TREAT HARV Variable Code Definition % of interviewed NIPFOs Mean SD Mean SD Mean SD T I0 0 if owner did not receive timber income 45.1 0.61 0.99 0.56 1.02 0.00 0.00 T I1 1 if owner receivedB195.1 17.1 2.03 4.25 2.61 6.90 7.34 7.09 T I2 2 if owner received 195.1425.2 18.3 3.20 3.54 0.78 0.95 3.64 3.34 T I3 3 if owner received425.2 19.5 1.76 2.24 0.77 0.96 8.18 9.46 TPRICE Ordinal Stumpage price per unit (tt 1 ) during 19992003 T P0 0 if owner did not sell timber 45.1 0.61 0.99 0.56 1.02 0.00 0.00 T P1 1 if stumpage price per unit wasB48.9 39.0 2.42 3.64 1.73 4.61 5.90 5.56 T P2 2 if stumpage price per unit was 48.974.3 11.0 1.83 2.53 0.39 0.56 6.46 7.04 T P3 3 if stumpage price per unit was74.3 4.9 2.61 3.54 0.44 0.40 8.17 12.17 NTINCOME Ordinal Annual income from land sales (t) during 19992003, per unit of forest area nT I0 0 if owner did not receive non-timber income 33.7 1.69 2.73 0.57 0.91 0.81 0.92 nT I1 1 if owner receivedB381.8 50.6 1.37 2.75 1.33 4.07 2.14 3.13 nT I2 2 if owner received381.8 15.7 2.11 2.64 1.06 1.25 8.86 5.79 Note: NIPFOnon-industrial private forest owner. 60 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez Table III. Correlations among nominal, ordinal and/or binary explanatory variables. INHERIT (N) BEQUEST (N) PERSONAL (O) FAMILY (O) MACHINERY (B) PROFESS (O) Landowner profile EDUC (O) D0.043 0.092 0.100 0.001 0.073 0.152 p0.659 0.182 0.332 0.990 0.506 0.101 OCCUP (N) D0.092 0.041 0.063 0.049 0.091 0.015 p0.349 0.580 0.480 0.583 0.387 0.882 FARM (B) D0.255* 0.025 0.038 0.149 0.312* 0.058 p0.013* 0.766 0.663 0.117 0.000* 0.520 ASSOC (B) D0.077 0.060 0.061 0.053 0.433* 0.086 p0.455 0.428 0.522 0.591 0.000* 0.380 TRAINING (B) D0.195* 0.100 0.209* 0.024 0.192* 0.055 p0.041* 0.515 0.006* 0.756 0.005* 0.456 IMARKET (B) D0.169* 0.055 0.045 0.063 0.023 0.052 p0.050* 0.476 0.638 0.515 0.829 0.574 TECHNIC (N) D0.123 0.013 0.114 0.035 0.039 0.107 p0.244 0.624 0.151 0.667 0.711 0.249 Family unit INHERIT (N) D0.135 0.296* 0.006 0.076 0.107 p0.139 0.000* 0.952 0.460 0.256 BEQUEST (N) D0.135 0.009 0.074 0.020 0.025 p0.139 0.899 0.184 0.821 0.692 PERSONAL (O) D0.296* 0.009 0.079 0.135 0.116 p0.000* 0.899 0.438 0.152 0.207 FAMILY (O) D0.006 0.074 0.079 0.063 0.021 p0.952 0.184 0.438 0.490 0.822 MACHINERY (B) D0.076 0.020 0.135 0.063 0.019 p0.460 0.821 0.152 0.490 0.836 PROFESS (O) D0.107 0.025 0.116 0.021 0.019 p0.256 0.692 0.207 0.822 0.836 Forest property and land-use changes FMEADOW (B) D0.001 0.035 0.018 0.010 0.066 0.010 p0.986 0.216 0.690 0.800 0.151 0.866 MWOOD (B) D0.114 0.146 0.116 0.001 0.019 0.088 p0.279 0.195 0.263 0.989 0.862 0.365 CSPECIE (B) D0.097 0.008 0.163* 0.038 0.222* 0.127 p0.354 0.927 0.050* 0.689 0.018* 0.165 IFOREST (B) D0.187* 0.062 0.185* 0.024 0.034 0.308* p0.050* 0.471 0.045* 0.806 0.752 0.001* Forest economics REQUEST (B) D0.203* 0.002 0.087 0.026 0.089 0.147 p0.039* 0.979 0.336 0.781 0.419 0.134 Note: variable subscripts indicate the type of variable used in the statistical analysis (Nnominal; Oordinal; Bbinary). *Statistically significant (pB0.05). Role of the family in private forestry 61 across variables was analysed using the Kruskal Wallis’Htest, a non-parametric test of variance homogeneity equivalent to one-way anova (Table VI). Confirming this premise, pairwise comparisons were conducted using the Dunnett’s T3 test to determine which categories (levels) of nominal/ordinal (independent) variables showed behaviours (means) that were significantly different from the Table IV. Significant differences among nominal, ordinal and/or binary explanatory variables. INHERIT (N) BEQUEST (N) HOUSEHOLD (O) REINVEST (O) PERSONAL (O) FAMILY (O) MACHINERY (B) PROFESS (O) Landowner profile EDUC (O) x 2 5.659 1.769 16.890 11.182 22.500 16.081 14.839* 17.140 p0.462 0.940 0.154 0.740 0.095 0.377 0.002* 0.311 OCCUP (N) x 2 22.191* 12.432 34.207* 27.150 33.709 27.819 12.935* 36.601* p0.014* 0.257 0.025* 0.348 0.114 0.316 0.024* 0.050* FARM (B) x 2 7.769* 1.682 12.208* 9.404* 1.570 6.947 8.478* 2.705 p0.021* 0.431 0.016* 0.050* 0.905 0.225 0.003* 0.745 ASSOC (B) x 2 1.666 1.789 12.321* 2.762 1.958 5.030 16.128* 6.750 p0.435 0.409 0.015* 0.737 0.855 0.412 0.000* 0.240 TRAINING (B) x 2 5.715* 5.011* 8.803* 1.062 10.113* 2.844 4.295* 1.450 p0.050* 0.050* 0.050* 0.957 0.050* 0.724 0.038* 0.919 IMARKET (B) x 2 3.074 1.044 6.992 2.700 1.207 3.157 0.046 3.285 p0.215 0.593 0.136 0.746 0.944 0.676 0.512 0.656 TECHNIC (N) x 2 10.352* 1.213 4.230 10.310 13.919 8.217 0.173 9.366 p0.035* 0.876 0.836 0.414 0.177 0.608 0.917 0.498 Family unit INHERIT (N) x 2 14.342* 5.966 6.879 16.724* 6.737 1.819 11.002 p0.006* 0.651 0.737 0.050* 0.750 0.403 0.357 BEQUEST (N) x 2 14.342* 4.801 10.754 18.287* 5.856 2.350 7.047 p0.006* 0.779 0.377 0.050* 0.827 0.309 0.721 PERSONAL (O) x 2 16.724* 18.287* 15.895 27.385 101.468* 5.183 19.216 p0.050* 0.050* 0.723 0.337 0.000* 0.394 0.787 FAMILY (O) x 2 6.737 5.856 22.997 25.662 101.468* 11.158* 28.486 p0.750 0.827 0.289 0.426 0.000* 0.048* 0.286 MACHINERY (B) x 2 1.819 2.350 3.929 19.003* 5.183 11.158* 2.179 p0.403 0.309 0.416 0.002* 0.394 0.048* 0.824 PROFESS (O) x 2 11.002 7.047 12.485 23.268 19.216 28.486 2.179 p0.357 0.721 0.898 0.562 0.787 0.286 0.824 Forest property and land-use changes FMEADOW (B) x 2 0.082 0.126 3.687 22.165* 1.276 2.811 1.154 1.999 p0.960 0.939 0.450 0.000* 0.937 0.729 0.406 0.849 MWOOD (B) x 2 1.201 2.883 4.716 0.783 11.833* 6.019 0.031 3.233 p0.548 0.237 0.318 0.978 0.037* 0.304 0.526 0.664 CSPECIE (B) x 2 0.897 0.788 1.842 5.757 11.377* 8.154 4.341* 9.847* p0.639 0.674 0.765 0.331 0.044* 0.148 0.031* 0.050* IFOREST (B) x 2 4.325 0.673 12.004* 1.241 6.522 8.995 0.100 12.387* p0.115 0.714 0.017* 0.941 0.259 0.109 0.465 0.030* Forest economics REQUEST (B) x 2 6.853* 0.808 4.692 2.846 3.997 2.548 0.696 8.849 p0.033* 0.667 0.320 0.724 0.550 0.769 0.285 0.115 Note: variable subscripts indicate the type of variable used in the statistical analysis (Nnominal; Oordinal; Bbinary). *Statistically significant (pB0.05). 62 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez Table V. Correlations among continuous and/or nominal, ordinal and binary variables. INHERIT (N) BEQUEST (N) HOUSEHOLD (C) REINVEST (C) PERSONAL (O) FAMILY (O) MACHINERY (B) PROFESS (O) Owner land management practices PLANT (C) r0.124 0.020 0.208 0.124 0.130 0.095 0.183 0.408** p0.256 0.853 0.055 0.257 0.233 0.385 0.092 0.000** TREAT (C) r0.010 0.013 0.149 0.079 0.014 0.056 0.373** 0.251* p0.926 0.903 0.170 0.470 0.899 0.608 0.000** 0.020* HARV (C) r0.157 0.014 0.190 0.032 0.095 0.002 0.133 0.302** p0.148 0.901 0.079 0.768 0.382 0.989 0.223 0.005** Landowner profile AGE (C) r0.146 0.032 0.416** 0.015 0.018 0.035 0.203 0.062 p0.179 0.773 0.000** 0.892 0.867 0.752 0.061 0.569 EDUC (O) r0.336** 0.215* p0.002** 0.047* OCCUP (N) r0.379** 0.121 p0.000** 0.269 FARM (B) r0.343** 0.304** p0.001** 0.004** ASSOC (B) r0.259* 0.151 p0.016* 0.164 TRAINING (B) r0.084 0.079 p0.441 0.471 IMARKET (B) r0.122 0.004 p0.261 0.974 TECHNIC (N) r0.058 0.016 p0.599 0.887 Family unit INHERIT (N) r0.009 0.034 p0.937 0.757 BEQUEST (N) r0.095 0.056 p0.382 0.608 HOUSEHOLD (C) r0.009 0.095 0.077 0.086 0.107 0.197 0.166 p0.937 0.382 0.481 0.432 0.327 0.069 0.127 REINVEST (C) r0.034 0.056 0.077 0.021 0.099 0.400** 0.289** p0.757 0.608 0.481 0.846 0.363 0.000** 0.007** PERSONAL (O) r0.086 0.021 p0.432 0.846 FAMILY (O) r0.107 0.099 p0.327 0.363 MACHINERY (B) r0.197 0.400** p0.069 0.000** PROFESS (O) r0.166 0.289** p0.127 0.007** Forest property and land-use changes FMEADOW (B) r0.182 0.214* p0.094 0.048* MWOOD (B) r0.117 0.003 p0.283 0.981 Role of the family in private forestry 63 continuous (dependent) variables. Homogeneous subgroups of similar statistical behaviour were finally defined by Tukey’s HSD procedure, after the significant differences had been tested. Results and discussion Evolution of rural property Patterns of land acquisition and future transmission may be important indicators of NIPFOs’land decisions. In this sense, Hardie and Parks (1996), Conway et al. (2003) and Ross-Davis et al. (2005) have documented that maintaining the viability of the property by land management by future landowners through inheritance may be a key to securing the continuity of the forest sector. In Marin ˜a Oriental, this study only verified that the INHERIT group significantly affected the annual plantation behaviour (H5.973). The I H2 owners annually planted over 2% of their forestlands, which differed significantly from the owners whose holdings were acquired solely through purchasing (0.3%). A similar trend was observed for the annual investment in this activity; the annual expenditure on planting and the pattern of land acquisition were weakly and positively correlated (r0.224). In this case, significant differences were observed in the mean investment in forest plantation for the I H1 and I H2 groups (H6.732). Again, I H2 owners developed the most active forest activity, and invested annually in planting almost twice as much as the I H1 group. Furthermore, the main occupation of owners significantly varied based on the pattern of land acquisition (x 2 22.191). Retired farmers represented a specific group with holdings that were mainly acquired by purchasing and inheriting. In brief, retired farmers would be more likely to make land transactions for improving and increasing their former agricultural productivity, and hence show a more significant land mobility (Marey et al., 2004). The large fraction of retired farmers in the I H2 group might Table V (Continued) INHERIT (N) BEQUEST (N) HOUSEHOLD (C) REINVEST (C) PERSONAL (O) FAMILY (O) MACHINERY (B) PROFESS (O) CSPECIE (B) r0.006 0.161 p0.953 0.139 IFOREST (B) r0.087 0.081 p0.424 0.458 PLOT (C) r0.044 0.129 0.250* 0.102 0.058 0.014 0.119 0.246* p0.685 0.237 0.020* 0.349 0.595 0.897 0.277 0.022* SIZE (C) r0.039 0.050 0.085 0.226* 0.035 0.109 0.089 0.277** p0.721 0.647 0.437 0.036* 0.749 0.317 0.416 0.010** Forest economics INVEST (C) r0.246* 0.082 0.007 0.024 0.094 0.058 0.145 0.308** p0.023* 0.455 0.946 0.826 0.394 0.600 0.186 0.004** PEXP (C) r0.224* 0.001 0.220* 0.031 0.225* 0.109 0.174 0.402** p0.039* 0.995 0.043* 0.780 0.038* 0.321 0.111 0.000** TEXP (C) r0.013 0.054 0.166 0.097 0.009 0.037 0.073 0.313** p0.911 0.627 0.134 0.385 0.939 0.738 0.513 0.004** REQUEST (B) r0.084 0.153 p0.442 0.159 SUB (C) r0.073 0.079 0.022 0.005 0.005 0.022 0.070 0.189 p0.503 0.467 0.840 0.966 0.964 0.841 0.524 0.082 TINCOME (C) r0.097 0.009 0.151 0.042 0.177 0.002 0.172 0.207 p0.387 0.939 0.174 0.707 0.111 0.986 0.123 0.062 TPRICE (C) r0.081 0.056 0.120 0.042 0.225* 0.114 0.162 0.206 p0.471 0.617 0.284 0.711 0.042* 0.309 0.145 0.063 NTINCOME (C) r0.034 0.045 0.105 0.502** 0.079 0.004 0.203 0.048 p0.761 0.689 0.346 0.000** 0.480 0.973 0.066 0.664 Note: variable subscripts indicates which typology of variable was used in the statistical analysis (Ccontinuous; Nnominal; O ordinal; Bbinary). Statistically significant coefficients: **pB0.01, *pB0.05. 64 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez account for the negative association observed between the pattern of land acquisition and the landowner’s condition as an active farmer (D0.255). None of the I H3 owners was a retired farmer, in contrast to the I H1 and I H2 groups, where retired farmers amounted to 41% and 67%, respectively (x 2 7.769). Condition of the owner as a retired farmer might moreover explain the fact that the mean age of owners in the I H2 group was significantly higher than the mean age of the owners included in the I H3 group (H7.607). More than 61% of the I H2 owners were more than 65 years old, as opposed to 75% of the owners in the I H1 and I H3 groups, who were 4065 years old. The important fraction of retired farmers included in the I H2 group and the noticeable planting behaviour of this group could suggest that these landowners invested in agricultural marginal land as capital, as reported by several authors (Karppinen, 1998; Gunter et al., 2001; Kurttila et al., 2001; Arano et al., 2004; Marey et al., 2004). Table VI. Significant differences in continuous variables based on nominal, ordinal and/or binary explanatory variables. INHERIT (N) BEQUEST (N) HOUSEHOLD (O) REINVEST (O) PERSONAL (O) FAMILY (O) MACHINERY (B) PROFESS (O) Owner land management practices PLANT (C) H5.973* 1.104 5.352 7.720 4.646 9.527* 2.846 23.670* p0.050* 0.576 0.253 0.172 0.461 0.050* 0.092 0.000* TREAT (C) H0.102 0.320 2.062 5.917 4.241 3.371 11.854* 11.541* p0.950 0.852 0.724 0.314 0.515 0.643 0.001* 0.042* HARV (C) H2.308 0.441 4.734 7.087* 7.182 2.795 1.497 9.499* p0.315 0.802 0.316 0.050* 0.207 0.732 0.221 0.050* Landowner profile AGE (C) H7.607* 3.085 18.940* 2.629 2.685 4.971 3.487* 4.128 p0.022* 0.214 0.001* 0.757 0.748 0.419 0.041* 0.531 Family unit HOUSEHOLD (C) H0.288 0.774 6.965 2.750 3.996 3.291* 4.917 P0.866 0.679 0.223 0.738 0.550 0.050* 0.426 REINVEST (C) H0.249 0.304 2.842 7.298 7.024 13.631* 9.191 p0.883 0.859 0.585 0.199 0.219 0.000* 0.102 Forest property and land-use changes PLOT (C) H0.168 1.733 0.079* 4.643 1.245 2.485 1.194 10.939* p0.920 0.420 0.050* 0.461 0.941 0.779 0.274 0.050* SIZE (C) H2.909 0.599 9.159* 21.459* 1.189 4.919 0.670 13.131* p0.234 0.741 0.050* 0.001* 0.946 0.426 0.413 0.022* Forest economics INVEST (C) H8.975* 0.566 0.697 5.573 13.278* 0.631 1.764 19.907* p0.011* 0.754 0.952 0.350 0.021* 0.631 0.184 0.001* PEXP (C) H6.732* 0.699 6.059 2.923 6.568 9.003* 2.553 23.701* p0.035* 0.705 0.195 0.712 0.255 0.050* 0.110 0.000* TEXP (C) H0.487 0.980 3.768 6.191 1.434 4.891 0.435 13.599* p0.784 0.613 0.438 0.288 0.921 0.429 0.509 0.018* SUB (C) H1.200 0.536 6.348 0.632 4.523 2.124 0.413 6.806* p0.549 0.765 0.175 0.986 0.477 0.832 0.521 0.050* TINCOME (C) H1.512 2.288 2.254 7.936 11.509* 8.356* 2.393 7.036* p0.470 0.319 0.689 0.160 0.042* 0.050* 0.122 0.050* TPRICE (C) H1.602 1.600 3.475 10.313* 16.877* 10.017* 2.135 9.739* p0.449 0.449 0.482 0.050* 0.005* 0.050* 0.144 0.039* NTINCOME (C) H0.256 2.161 2.089 21.207* 2.914 15.083* 3.373* 2.588 p0.880 0.339 0.719 0.001* 0.713 0.010* 0.066* 0.763 Note: variable subscripts indicate the type of variable used in the statistical analysis (Ccontinuous; Nnominal; Oordinal; Bbinary). *Statistically significant (pB0.05). Role of the family in private forestry 65 In addition, the strong forest involvement of the I H2 group was evidenced by the large fraction of owners who were interested in increasing their woodland in the near future. In the study area, the intention of further enlarging woodlands on the holding increased in those landowner groups who had inherited and purchased land (D0.187). In particular, 56% of the I H2 owners and 34% of the owners in the remaining groups wanted to extend their productive woodland in the short to mid-term. In keeping with other studies, the major cause of this behaviour could be attributed to previous harvests and resulting timber sales, i.e. to the interest in timber production (Hardie & Parks, 1996; Bolkesjø & Baardsen, 2002; Kline et al., 2002; Li & Zhang, 2004). The annual ratio of woodland harvesting increased very strongly in proportion to the annual timber income and stumpage price per unit from previous harvests (r0.809 and 0.781, respectively, at pB0.01). Moreover, this forest practice grew slightly when the landowner intended to enlarge woodlands in the future (r0.242 at pB0.05). Landowners who had the intention of enlarging woodlands harvested twice as much woodland per year as the remaining population of owners (H 4.972), with a timber income of t100 per year more (H3.118) at almost double unitary price (H3.104). Verifying the previous hypothesis about land capitalization, the annual investment in holding improvement increased weakly in forest holdings that combined inherited and purchased land (r0.246). The I H2 owners spent six times more per year on new forest buildings or infrastructure than the I H3 group, a statistically significant mean difference (H8.975). Taking into account that converting former meadows into woodlands as a production alternative for the holding was negatively related to the landowner’s condition as an active farmer, because such a land practice is clearly linked to retired farmers (D 0.233), it was shown that such landowners were effectively capitalizing on their marginal land by plantation. In addition, the great investment in annually improving and planting the holding rose weakly in relation to the payment of public subsidies received annually in forestry (r0.297 at pB0.01 and 0.246 at pB0.05, respectively). This finding would prove the significant role of public measures in NIPFOs’land management decisions (Hodges & Cubbage, 1990; Gunter et al., 2001; Zhang & Flick, 2001; Kline et al., 2002; Arano et al., 2004). Under these circumstances, the likelihood of applying for public subsidies increased when the owner managed both types of land (D0.203). The proportion of I H2 owners who applied for economic subsidies was three times the ratio observed for the I H1 group; conversely, none of the I H2 landowners applied for forest subsidy (x 2 6.853). The status of farmers, both current and former, might justify the personal labour force used for forestry in Marin ˜a Oriental. The surveyed farmers generally managed their land themselves, and in particular, they had more time for forest management, as Zhang and Mehmood (2001) and Karppinen (2005) have pointed out. Thus, the personal time annually devoted to forestry was positively associated with the pattern of land acquisition (D0.296). Landowners who had both inherited and purchased land spent annually a higher amount of personal labour-days in forestry on their holdings. Even though 47% of the I H1 and I H2 owners spent annually more than 11 personal labour-days on forestry, none of the I H3 owners worked on the holding above this threshold (x 2 16.724). Spending more time working on the property may result in better forestry training, which would qualify landowners for managing land (Dennis, 1990; Lo ¨yland et al., 1995; Mahapatra & Mitchell, 2001; Arano et al., 2004). In the study region, the way in which forestland was acquired was positively correlated with the landowners’forest training (D 0.195) and, specifically, with their knowledge about timber market conditions (D0.169). Again, owners who managed inherited and purchased land were generally better trained in these subjects. Over 30% of the I H2 owners were trained in forestry, compared with 8% in the remaining groups (x 2 5.715). In particular, 21% of the I H2 owners knew and applied the suitable rotation age in previous harvests, compared with less than 3% of the I H1 owners and none of the I H3 group who considered these requirements (x 2 10.352). This result, together with the result obtained for personal time used for forestry, confirmed that forest training may enable landowners to overcome some management difficulties and to take a more active role in forestry, as suggested by Dole (1995), Zhang and Mehmood (2001) and Kittredge (2005), among others. Furthermore, the results obtained for the future prospects for forestlands based on INHERIT group were significantly different (x 2 14.342). Over 97% of the I H1 and I H2 owners wanted to pass the holding on to their heirs, independently of holding profitability, while less than 67% of the I H3 owners expected to do so in the near future. As a first conclusion, it may be suggested that land transmission governed by emotional values (inheritance-tobequeath) seems to be a clear motivation for taking care of the holding and passing it on to future generations as best landowners can. Literature on the relationships between land tenure and forest management has generally recognized legacy as key 66 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez to land investment (Conway et al., 2003; Ross-Davis et al., 2005; Marey et al., 2006). Such a strong emotional relationship would additionally explain why landowners in the region continued to work their land, although they did not receive regular income from it. To complete this hypothesis, the authors tested whether significant differences were observed for the personal labour force used for forestry annually according to the future plans for the property (x 2 18.287). All B E2 owners spent more than 11 labour-days per year on their holding, and more specifically, half of the owners in this group worked more than 100 labour-days per year. In contrast, 67% of the B E3 owners devoted less than 2 labourdays per year to forestry. In line with Karppinen (1998) and Ingemarson et al. (2006), inheritance-tobequeath, and not only economic criteria, may be a significant reason for securing the continuity of forest management. As mentioned above, this apparent commitment to land management was linked to the owner’s training in forestry, which may reduce some difficulties in forest management. Half of the B E2 owners, none of the B E3 group and less than 8% of the B E1 owners were trained in forestry (x 2  5.011). Again, the important role of forest knowledge in identifying and understanding management opportunities and constraints was proven. Support to the family economy Household income. The HOUSEHOLD group was not a significant factor in the three forest practices analysed in the study region, in agreement with Bolkesjøand Baardsen (2002), Pattanayak et al. (2002) and Potter-Witter (2005), but contrary to the results reported by other authors (Dennis, 1990; Kuuluvainen et al., 1996; Gunter et al., 2001; Zhang & Flick, 2001; Arano et al., 2004). This result may be clarified by the fact that none of the interviewed landowners mentioned forestry as his or her main occupation, but rather as a complementary activity that supported the family’s well-being. Nevertheless, landowners with a higher annual income per household were more likely to plant and improve their forestlands, and to harvest their woodlands annually. In light of these results, even without statistical evidence, large forest investments may be made by owners with higher family income, in keeping with other studies focused on the positive effects of household earnings on forest involvement (Gunter et al., 2001; Mahapatra & Mitchell, 2001; Arano et al., 2004; Ross-Davis et al., 2005). Timber harvesting may not work as a supplementary source of income for the family unit in times of economic difficulty. Dennis (1990), Kuuluvainen and Salo, (1991) and Kuuluvainen et al. (1996), among other authors, have statistically confirmed that family earnings negatively affect timber harvesting activity. To verify the previous statement about forest investment intensity based on landowners’income, it was confirmed statistically that annual expenditure on forest plantation weakly increased in relation to annual household income (r0.220). Mean values were not statistically significant; landowners whose income exceeded t9000 per year invested almost t100 ha 1 per year more in planting their forestlands than those groups with household earnings not reaching this economic figure. This premise was completed by the assumption that the landowner’s intention of further enlarging woodlands was significantly different according to the group of family earnings considered (x 2 12.004). The H H5 group included the largest fraction of owners interested in enlarging woodlands (76.5%), whereas this ratio did not reach 43% in the intermediate income groups (H H2 and H H3 landowners). In contrast, the study population with annual earnings per family unit below t6000 or above t30,000 was clearly the least interested in the measure (21.8%). The moderate decrease in the annual income earned per family unit with the increase in the landowner’s age (r0.416), and the moderate increase in the annual income with the increase in the owner’s educational level (r0.336) allowed landowners who received a high annual household income to be characterized. Pairwise comparisons revealed that the mean age of owners in the H H2 group was significantly higher than the mean age of the owners included in the H H4 and H H5 groups (H18.940). While 67% of the H H2 owners were more than 65 years old, 75.2% of the owners included in the H H4 and H H5 groups were 4065 years old. The significant differences observed in landowners’mean age according to HOUSEHOLD group classified the study population as shown in Table VII. With regard to landowners’educational level, there were no significant differences. Over 63% of the study population with an annual family income above t9000 had completed primary or secondary education and, in particular, 9.8% of them had completed a university degree. Conversely, none of the owners included in the H H1 and H H2 groups had an educational level higher than secondary studies, and more than 60% had received no formal education. The moderate positive correlations found between the annual family income and the owner’s primary occupation and his or her specific condition as an active farmer defined the previous landowner profile (r0.379 and 0.343, respectively). Over 71% of the H H1 ,H H2 and H H3 owners were retired farmers, an Role of the family in private forestry 67 occupational group which diminished in the H H4 and H H5 groups, where more than 56% were professionals not linked to agriculture and 25% were active farmers, respectively (x 2 34.207). The fraction of active farmers was significantly different based on the group of family earnings (x 2 12.208). Therefore, not only were retired farmers more likely to invest in land to keep them productive, as suggested earlier, but also owners who did not work actively on their property were also actively involved in forestry. These results are in line with those reported by Karppinen (1998), Gunter et al. (2001), Marey et al. (2004) and Rickenbach et al. (2005). Furthermore, the fraction of owners who were trained in forestry was significantly different according to the annual income per family unit (x 2  8.803). For the same reasons discussed in the previous section, the status of the landowner as an active or a retired farmer may justify that all the owners in the study area, except for the H H4 and H H5 groups, were trained in forestry. The owner’s occupation within the farming activity may account for the weak increase in the likelihood of participating in agricultural and forestry bodies observed with the increase in the annual income per family unit (r0.259). Professional associations in Marin ˜a Oriental were mainly agricultural groups, more specifically co-operatives or trade unions, in which 35% and 55% of the participants were active farmers and retired farmers, respectively. On average, for each H H1 and H H2 owner who was involved in an association, two H H4 owners and three H H5 owners participated in the association (x 2 12.321). Finally, the lowest degree of land fragmentation was found in productive forestlands belonging to owners who earned the highest annual household income (r0.250). Post hoc analyses showed that H H5 landholdings were significantly less fragmented than H H3 and H H4 productive forestlands (H 0.079). The most prevalent primary occupations in the H H1 ,H H2 and H H3 groups were retired (70.4%) and active (18.6%) farmers, respectively, which may explain why the most fragmented plots in the region belonged to these landowners (four plots per unit of productive forestland). As shown by Butler et al. (2004) and Marey et al. (2006), land fragmentation may be motivated mainly by the demand for improving and increasing agricultural productivity. Such a landowner categorization may also clarify the significant differences observed in the production size of the forest holding based on the group of annual household considered (H9.159). Pairwise comparison analyses revealed that the holdings of the H H5 owners were significantly different in size from the holdings of the H H1 and H H4 owners, and differentiated two significant landowner subgroups in the area (Table VII). Therefore, these results agree with those of Nagubadi et al. (1996): landowners with a higher income were more likely to manage larger forest holdings effectively. Forest reinvestments. The behaviour of the annual ratio of timber harvesting was significantly different according to the annual rate of dependency on forest products for personal use. In this sense, R E5 owners were the most active harvesters in the region, and harvested annually almost three times more woodland than the owners in the R E1 group (H7.087). Table VIII shows two significant forest-consumer profiles in Marin ˜a Oriental. Such a classification would be largely associated with the landowner’s primary occupation, more specifically with his or her condition as an active farmer. As described by Dewees (1992), Kurttila et al. (2001) and Marey et al. (2004), while some owners would be less dependent on forestry because of the increased proportion of other incomes, farmers might become more dependent on forestry as a source of revenue because of a reduced development of the agricultural income. Hence, farmers would be great consumers of forest products for self-consumption. The annual amount of forest products for personal use increased with the owner’s involvement in agriculture (r0.304). For each R E0 and R E1 owner who was an active farmer, two R E4 and R E5 owners were active farmers (x 2 9.404). More than half of Table VII. Homogeneous HOUSEHOLD subgroups according to landowner’s age (years) and size of productive forest holding (ha). HOUSEHOLD H H1 H H2 H H3 H H4 H H5 p % of interviewed landowners 10.5 17.4 36.0 19.8 16.3 AGE New landowner 67.56 68.19 55.65 56.43 0.055 Retired landowner 67.56 71.13 68.19 0.935 SIZE Small landowner 2.92 4.59 4.93 3.33 0.575 Large landowner 4.59 4.93 7.51 0.209 68 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez the landowners included in the R E0 and R E1 groups were retired farmers. Educational level was another landowner attribute that was statistically associated with annual dependency on forestry. The annual fraction of forest products for self-consumption weakly declined in relation to the owner’s formal education (r0.215). The descriptive statistics showed that owners with primary studies, a group comprising mainly active farmers, actively benefited from forests for self-consumption, with an average reinvestment of t65 ha 1 per year, which was three times the mean value obtained for the owners with tertiary education. Even though this result was not statistically proven, the profile of the owner as an active farmer and his or her forest management guidelines suggest that these landowners were active harvesters, in line with the findings reported by Mahapatra and Mitchell (2001) and Pattanayak et al. (2002). Thus, the annual ratio of timber harvesting by farmers reached twice the fraction observed for other landowners in the study area. In Marin ˜a Oriental, only the annual ratio of silviculture was weakly correlated with the owner’s condition as an active farmer (r0.212 at pB0.05). As previously suggested for the level of participation of landowners in professional bodies, farming occupation and, therefore, a stronger relationship with land management, may also justify the moderate increase in availability of equipment according to the annual rate of forest reinvestment (r0.400). In fact, both the level of participation in associations and the availability of machinery increased with the owner’s status as an active farmer (D0.508 and 0.312, respectively). All R E5 owners had machinery on the holding, while less than half the R E0 and R E1 owners had these resources for forest management (x 2 19.003). As suggested by Hodges and Cubbage (1990), it may be concluded that technology users seemed to be more likely to participate in social groups and to share experience. The owner’s main occupation may also account for the slight increase in the annual amount of professional forest labour on the holding with the decrease in the annual amount of forest products for self-consumption (r0.289). According to the literature reviewed (Lo ¨yland et al., 1995; Zhang & Mehmood, 2001; Finley, 2002; Conway et al., 2003; Arano et al., 2004; Kittredge, 2005; Potter-Witter, 2005), forestry would generally be less feasible for managers with occupations outside the property. The absence of the landowner would consequently force him or her to hire an important fraction of professional assistance in forestry. Hardie and Parks (1996), Gunter et al. (2001) and Zhang and Flick (2001) analysed the role of technical guidance in forest adoption and management. Landowner absence from the holding may explain the fact that R E4 owners hired workers for less than 5 labour-days per year, whereas the R E0 group spent 1150 labourdays annually on professional labour. This technical advice from professional foresters may cause the significant mean differences in the stumpage price per unit between R E5 landowners and landowners in the R E0 ,R E1 and R E2 groups (H10.313). The R E1 owners were characterized by having sold timber at the best stumpage price per unit, a mean of t5.70 t 1 . Thus, R E0 ,R E1 and R E2 landowners sold timber from previous harvests at a unitary price that was twice as high as the economic value fixed by the R E5 group. The change from forestland into meadow was weakly and positively correlated with the annual rate of forest reinvestment (r0.214). All R E5 owners were active farmers and, as expected, all of them made this land-use change; conversely, none of the R E0 and R E1 owners had converted forestlands into meadows, but 29% of them had made the opposite land-use change, from marginal meadow into woodland (x 2 22.165). This finding finally supported that both retired farmers and professionals not related to farming initiated forestry over marginal land, following the production trend observed for the great majority of NIPFOs (Karppinen, 1998; Gunter et al., 2001; Kurttila et al., 2001; Arano et al., 2004; Marey et al., 2004). Land sale also seemed to be clearly dependent on the landowner’s Table VIII. Homogeneous REINVEST subgroups with regard to annual ratio of harvesting (%) and income from land sales (tha 1 per year). REINVEST R E0 R E1 R E2 R E3 R E4 R E5 p % of interviewed landowners 27.9 32.5 23.3 10.5 3.5 2.3 HARV Non-harvester 4.42 3.32 2.34 2.23 4.20 0.816 Harvester 4.42 3.32 2.34 2.23 4.20 8.40 0.051 NTINCOME Non-forest consumer 78.44 141.44 154.87 338.64 426.94 0.260 Forest consumer 141.44 154.87 338.64 548.51 426.94 0.124 Role of the family in private forestry 69 not receive regular income from it. Landowners who had agricultural and forestry machinery on the holding and landowners who clearly depended on forest products for self-consumption stood out by being far more active silviculturists and harvesters, respectively. They were mainly profiled as active farmers who were trained in forestry and organized into professional groups. Because of their status as farmers, these landowners were more likely to manage the land themselves, and devoted an important fraction of their labour to forestry. As in the case of retired farmers, self-employed silviculturists and harvesters received important family aid in the accomplishment of the different forest practices on the holding. However, it was necessary to distinguish a group of landowners who were clearly differentiated from the previous two profiles. Such owners did not earn their living from agriculture; rather, they were also seeking other sources of family income through forest investment and management on the basis of forest subsidies. These owners were managers of large and less fragmented forest holdings, who did not use forest products for self-consumption and did not spend much personal and family time on their land. These landowners worked closely with professional foresters to continue with their holdings and make land management viable, contributing significantly to the planting, silvicultural and harvesting patterns of the region. The findings of this survey may provide important guidance on how and where to focus policy efforts on NIPFOs’land decision making and management. In accordance with the results, public programmes must be adapted to different landowner profiles and family prospects. Therefore, it would be particularly important to differentiate between farmers who display a remarkable emotional link to land and devote more time to forestry, and professionals outside agriculture, who are less attached to land ownership and generally turn to forest assistance. Nowadays, bridging the gap between agriculture and forestry is clearly a key step towards promoting a suitable service of forest assistance that improves farm forestry by means of information, training, technology and market access, and specialization. By means of close forest extension, such forest services could guide NIPFOs’land practices towards economically viable, socially attractive and environmentally sustainable forest management within the framework of rural development. Acknowledgements The authors are grateful to the Forest Statistics Administration for making statistical data available and to the Galician Environmental Council for supporting this study through the formal agreement entitled ‘‘Methodology for surveying and characterizing private forest property in Galicia’’. References Arano, K. G., Munn, I. A., Gunter, J. E., Bullard, S. H. & Doolitle, M. L. (2004). Comparison between regenerators and non-regenerators in Mississippi: A discriminant analysis. Southern Journal of Applied Forestry,22, 132138. Beiras, X. M. (1975). A emigracio´n: O seu papel na dina´mica da formacio´n social [Emigration: Its role in social formation dynamics]. In D. Garcı´a-Sabell (Ed.),, A Galicia rural na encrucillada [Rural Galicia at a crossroad] (pp. 3973). Vigo, Spain: Galaxia. Bolkesjø, T. F. & Baardsen, S. (2002). Roundwood supply in Norway: Micro-level analysis of self-employed forest owners. Forest Policy and Economics,4,5564. Butler, B. J., Swenson, J. J. & Alig, R. J. (2004). Forest fragmentation in the Pacific Northwest: Quantification and correlations. Forest Ecology and Management,189, 363373. Cao, R. (2002). Ana ´lisis multivariante. Curso de postgrado en estadı ´stica aplicada [Multivariate analysis. A postgraduate course on applied statistics]. A Corun ˜a, Spain: University of A Corun ˜a. Conway, M. C., Amacher, G. S. & Sullivan, B. J. (2003). Decisions nonindustrial forest landowners make: An empirical examination. Journal of Forest Economics,9, 181203. Dennis, D. (1990). A probit analysis of the harvest decision using pooled time-series and cross-sectional data. Journal of Environmental Economics and Management,18, 176187. Dewees, P. A. (1992). Social and economic incentives for smallholder tree growing: A case study from Muranga District, Kenya. Community forestry case study series,5. Rome: FAO. Dole, A. (1995). The economics of private forest management unifying the Faustmann model and nonindustrial private forest management models. Nedlands: University of Western Australia. Finley, A. O. (2002). Assessing private forest landowners’attitudes towards, and ideas for, cross-boundary cooperation in western Massachusetts. Amherst, USA: Doctoral dissertation, University of Massachusetts. Gunter, J. E., Bullard, S. H., Doolitle, M. L. & Arano, K. G. (2001). Reforestation of harvested timberlands in Mississippi: Behaviour and attitudes of non-industrial private forest landowners. Mississippi: Mississippi State University. Hardie, I. W. & Parks, P. J. (1996). Program enrollment and acreage response to reforestation cost-sharing programs. Land Economics,72, 248260. Hodges, D. G. & Cubbage, F. W. (1990). Adoption behaviour of technical assistance foresters in the southern pine region. Forest Science,36, 516530. INE (2004). Consumer price index. Instituto Nacional de Estadı´stica. Retrieved December 6, 2004, from http://www.ine.es/ daco/ipc.htm Ingemarson, F., Lindhagen, A. & Eriksson, L. (2006). A typology of small-scale private forest owners in Sweden. Scandinavian Journal of Forest Research,21, 249259. Karppinen, H. (1998). Values and objectives of non-industrial private forest owners in Finland. Silva Fennica,32,4359. Karppinen, H. (2005). Forest owners’choice of reforestation method: An application of the theory of planned behaviour. Forest Policy and Economics,7, 393409. 76 V. Rodrı ´guez Vicente & M. F. Marey Pe ´rez Kittredge, D. B. (2005). The cooperation of private forest owners on scales larger than one individual property: International examples and potential application in the United States. Forest Policy and Economics,7, 671688. Kline, J. D., Butler, B. J. & Alig, R. J. (2002). Tree planting in the south: What does the future hold? Southern Journal of Applied Forestry,26,99107. Kurttila, M., Hamalainen, K., Kajanus, M. & Pesonen, M. (2001). Non-industrial private forest owners’attitudes towards the operational environment of forestry*A multinominal logit model analysis. Forest Policy and Economics,2, 1328. Kuuluvainen, J. & Salo, J. (1991). Timber supply and life cycle harvest of non-industrial private forest owners: An empirical analysis of the Finnish case. Forest Science,37, 10111029. Kuuluvainen, J., Karppinen, H. & Ovaskainen, V. (1996). Landowner objectives and nonindustrial private timber supply. Forest Science,42, 300308. Li, Y. & Zhang, D. (2004). Tree planting in the US south: A panel data analysis. Washington, DC: Auburn University. Lo ¨yland, K., Kringstad, V. & .O ¨y, H. (1995). Determinants of forest activities*A study of private non-industrial forestry in Norway. Journal of Forest Economics,1, 219237. Mahapatra, A. K. & Mitchell, C. P. (2001). Classifying tree planters and non planters in a subsistence farming system using a discriminant analytical approach. Agroforestry Systems,54,4152. Marey, M. F. (2003). Tenencia de la tierra en Galicia: Modelo para la caracterizacio ´n de los propietarios forestales [Land tenure in Galicia: A model for forest landowner characterization]. Doctoral dissertation, University of Santiago de Compostela, Santiago de Compostela, Spain. Marey, M. F., Rodrı´guez, V. & Crecente, R. (2004). El monte en Galicia en el siglo XXI: Balance evolutivo y consideraciones para el futuro [Galician forests during the 21st century: Evolution and considerations for the future]. In A. Maya (Ed.), ¿ Que ´futuro para los espacios rurales? (pp. 117125). Leo´n, Spain: Universidad de Leo´n. Marey, M. F., Rodrı´guez, V. & Crecente, R. (2006). Using GIS to measure changes in the temporal and spatial dynamics of forestland: Experiences from north-west Spain. Forestry,79, 409423. Munn, I. & Rucker, R. (1994). The value of information in a market for factors of production with multiple attributes: The role of consultants in private timber sales. Forest Science, 40, 474486. Nagubadi, V., McNamara, K. T., Hoover, W. L. & Mills, W. L., Jr (1996). Program participation behaviour of nonindustrial forest landowners: A probit analysis. Journal of Agricultural and Applied Economics,28, 323336. Pattanayak, S. K., Murray, B. C. & Abt, R. (2002). How joint is joint forest production: An econometric analysis of timber supply conditional on endogenous amenity values. Forest Science,48, 479491. Potter-Witter, K. (2005). A cross-sectional analysis of Michigan nonindustrial private forest landowners. Southern Journal of Applied Forestry,22, 132138. Prestemon, J. & Wear, D. (2000). Linking harvest choices to timber supply. Forest Science,46, 377389. Rickenbach, M., Zeuli, K. & Sturgess-Cleek, E. (2005). Despite failure: The emergence of ‘‘new’’ forest owners in private forest policy in Wisconsin, USA. Scandinavian Journal of Forest Research,20, 503513. Ross-Davis, A. L., Broussard, S. R., Jacobs, D. F. & Davis, A. S. (2005). Afforestation motivations of private landowners: An examination of hardwood tree plantings in Indiana. Northern Journal of Applied Forestry,22, 149153. Slee, B. & Wiesum, F. (2001). New opportunities for forestrelated rural development in industrialized countries. Forest Policy and Economics,3,14. Sukhatme, P. U. (1953). Sampling theory of surveys. Rome: FAO. Zhang, D. & Flick, W. (2001). Sticks, carrots and reforestation investment. Land Economics,77, 443456. Zhang, D. & Mehmood, S. R. (2001). Predicting nonindustrial private forest landowners’choice of a forester for harvesting and tree planting assistance in Alabama. Southern Journal of Applied Forestry,25, 101107. Role of the family in private forestry 77 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I V MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I Land-use and land-base patterns in non-industrial private forests: Factors affecting forest management in Northern Spain Verónica Rodríguez-Vicente a,1 , Manuel F. Marey-Pérez b, ⁎ a Galician Sectorial Forestry Association (ASEFOGA), Rúa Doutor Maceira 13-Baixo, 15706, Santiago de Compostela, Spain b Department of Agroforestry Engineering, University of Santiago de Compostela, Spain, Campus Universitario s/n, 27002, Lugo, Spain abstractarticle info Article history: Received 6 November 2008 Received in revised form 30 March 2009 Accepted 25 May 2009 Keywords: Holding size Land parcellation Non-industrial private forest (NIPF) owner Land-use change Planting, silviculture and harvesting practices There is increasing worldwide interest in land-use allocation and management within the sphere of rural planning and development. The study of land-use patterns mainly focuses on understanding the practices and values of individuals involved, and no debate of this issue would be complete without taking into account non-industrial private forest (NIPF) ownership as a key component in most rural areas worldwide. This paper empirically explores and assesses NIPF owners' management in terms of analysing dynamics in farming and forestry practices (past conversions from forestland to meadow and from marginal meadow to woodland, and intentions to change the current productive forest species and to extend the area of woodland) and landholding attributes (size and degree of parcellation in productive forestland). Logistic regression models were also used to investigate the probabilities and influencing factors involved in transforming marginal meadows to woodland, and attempts on the part of NIPF owners to change the current productive forest species and increase productive forestland. For this, a total of 103 NIPF owners in Northern Spain were interviewed in person, in March 2004, about their commitment to and involvement in land management during 1999–2003. The models correctly explained 73.3%, 83.7% and 73.3% of the variability in having converted marginal meadow in woodland and of future intentions to change the productive forest species and increase the area of productive forestland, respectively. The results of the study indicate that forest management mainly responds to investment and increasing the productivity of the land as a capital asset, which is directly influenced by the size and degree of parcellation of the holding, and directly or indirectly related to the owner's interest in timber production. The results may be used by forest professionals, researchers and policymakers in order to design and execute successful forest policies related to land management and planning. © 2009 Elsevier B.V. All rights reserved. 1. Introduction The study of global land-use changes cuts across several academic disciplines and fields of inquiry, with the overall aim of achieving sustainable development (Biĉík et al., 2001). Having seen the consequences of centuries of forest overexploitation and clear changes in the socio-economic structure of many rural areas, land-use allocation and management have recently become of concern within the sphere of rural planning and development worldwide. Thus, as Wiersum et al. (2002) have stated, land-use choices can have a profound influence upon the development of rural areas, and landowners' strategies and preferred rural development options vary under different rural conditions. Regional land-use studies have therefore made possible to compare the importance and the structure of the ‘driving forces’of land-use changes at different spatial levels (Biĉík et al., 2001), by modelling and predicting changes in land use and management. According to Madsen and Andriansen (2004), the study of land-use patternsand dynamics mainly focus on understanding the practices and values of individuals involved, where ‘practices’are understood as actions related to land use carried out by individuals, and ‘values’as traditions, thoughts, and beliefs. Understanding land use therefore requires an understanding of landowners and how they make decisions (Koontz, 2001). Given a set of biophysical parameters, individuals choose a specific type of land use on the basis of the potential benefits, either tangible or non-tangible, while also taking into account personal aims and prospects, available means and resources, and possible constraints (Jansen and Di Gregorio, 2003). In addition, institutional, legal, socio-economic and cultural settings also influence an individual's decisions as regards land-use and management (Cihlar and Jansen, 2001). Changes in land use are clearly influenced by land tenure regimes, and hence, by the type of land management carried out (Marey-Pérez et al., 2006). Forest Policy and Economics 11 (2009) 475–490 ⁎Corresponding author. Tel.: +34 982 252 303x23292; fax: +34 982 285 926. E-mail addresses: [email protected] (M.F. Marey-Pérez), [email protected] (V. Rodríguez-Vicente). 1 Tel.: +34 981 530 500; fax: +34 981 531 724. 1389-9341/$ –see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.forpol.2009.05.008 Contents lists available at ScienceDirect Forest Policy and Economics journal homepage: www.elsevier.com/locate/forpol Of particular concern within the wide range of literature related to land-use dynamics are changes in the use and management of forestland, because of theimportantconsequences for the futureavailabilityof timber, wildlife habitats, and other benefits provided by forests (Kline and Alig, 2001; Marey-Pérez and Rodríguez-Vicente, 2008). Considering forestland tenure regimes, no debate of land-use dynamics would be complete without taking into account non-industrial private forest ownership as a key component in most rural areas worldwide. The nature of non-industrial private forest ownership differs notably from country to country and it is difficult to define the meaning of this term (Herbohn, 2001). According to the latter author, this type of forest ownership consists of a single or small number of planting blocks, nonprofessional management and often a lack of silvicultural skills, with little planning for future marketing. In other words, non-industrial private forest (NIPF) owners are individual landowners, not juristic landowners, whose forests constitute a part of their total land-use system of the holding and whose management aims are not solely centred on industrial timber production. NIPF owners own and manage land for a wide variety of purposes, and thus the ‘practices’and ‘values’ are equally as diverse. They are heterogeneous by nature; some NIPF owners carry out intensive management, like industrial owners, while others may have a complete disregard for forest management for any purpose, productive or protective (Siry et al., 2005; Arano and Munn, 2006). As a result, collaborative and coordinated approaches among researchers and policymakers are essential, firstly, to improve knowledge about which and how holding characteristics and land-use patterns affect the decisions made by NIPF owners and the intensity of land investments, and secondly, to develop and execute successful forest policies in relation to land management and planning. To helpto understand NIPF owners' managementobjectives, motivations and behaviour, this study was specifically aimed at the analysis of NIPFowners' management in terms of empirical analysis of dynamics in farming and forestry practices, represented as past conversions from forestland to meadow and from marginal meadow to woodland, intentions to extend the area of woodland and to change the current productive forest species, and landholding attributes, represented as the size anddegreeofparcellation ofproductiveforestland,by interviewingNIPF landowners in an area in Northern Spain. Thus, we examined possible statistical relationships or differencesbetween thesevariables and other factors related to the landowner profile, family unit, and forest economics. In order to complete the results, we moreover included the three practices traditionally used to predict forest management behaviour of NIPF owners in the relevant literature, i.e. planting and silviculture in forestlands (Hyberg and Holthausen, 1989; Löyland et al., 1995; Hardie and Parks,1996; Gunter et al., 2001; Zhang and Flick, 2001; Zhang and Mehmood, 2001; Kline et al., 2002; Arano et al., 2004; Ross-Davis et al., 2005), and timber harvesting in woodlands (Hyberg and Holthausen, 1989;Kuuluvainen and Salo, 1991; Löyland et al., 1995; Kuuluvainen et al., 1996; Prestemon and Wear, 2000; Zhang and Mehmood, 2001; Bølskejo and Baardsen, 2002; Conway et al., 2003; Størdal et al., 2008), and empiricallyexamined these as regards the dynamicsof farmingand forestry practices and the characteristics linked to the landholding. Within the extensive literature on land management behaviour of NIPF owners, we focused on past conversion of forestland to meadow and marginal meadow to woodland, given the agricultural background of the great majority of these landowners (Marey-Pérez and Rodríguez-Vicente, 2008). The abandonment of farms due to lack of generational replacement, together with favourable market conditions for timber production, may be important factors in the conversion of agrarian land to forestland for maintaining the productivity of the land and increasing the returns to complement household income (Gunter et al., 2001; Marey-Pérez et al., 2004). Furthermore, we consider proposals to replace the current productive forest species and to increase the extent of woodland within holdings. Favourable market conditions for certain types of industrial timber, the greater productivity or minor silvicultural requirements of particular forest species, profitable previous harvests and resulting timber sales may motivate NIPF owners to break with the usual patterns of forest management (Binkley, 1981; Boyd, 1984; Kuuluvainen and Salo, 1991; Bolkesjø and Baardsen, 2002), and to replace the current forest species or increase the area of productive woodland. As regards land holding attributes, most of the existing literature identifies land size and the degree of parcellation as the most relevant factors affecting forest management. Owners who manage larger areas of forestland are more likely to invest in land with forest plantations, as capital (Boyd,1984; Löyland et al.,1995; Hardie and Parks,1996; Zhang and Pearse, 1997; Arano et al., 2004). Furthermore, larger productive forest holdings are also more suitable for carrying out forest improvement treatments and timber harvesting (Binkley, 1981; Hyberg and Holthausen, 1989; Kuuluvainen and Salo, 1991; Löyland et al., 1995; Kuuluvainen et al.,1996). Land parcellation may lead to some structural difficulties for forest investment and management, due to the small size of tracts, and hence, an increase in management costs and a decline in productive yields may be expected (Healy, 1985; Conway et al., 2003; Potter-Witter, 2005). Characterization of land decision-making and management by NIPF owners in this way will provide more information about why certain land-use practices are carried out, and which can then be used to restructure and adjust territorial land-base. Forest professionals, researchers and policy makers may therefore benefitfromfurther efforts as regards land management and planning by taking advantage of opportunities to overcome the obstacles that most NIPF landowners face, in order to satisfy their interests and prospects as regards land, in accordance with current demands for sustainable development. 2. Materials and methods 2.1. Study area and data collection In an update and expansion of an earlier study by Marey-Pérez (2003), which explored individual private ownership and forest management in the Autonomous Community of Galicia (Northern Spain), data for the present study was collected in personal interviews with randomly selected NIPF landowners in the Mariña Oriental region, in Northeast Galicia, Spain (Fig. 1). In order to understand the selection of the area of study, it is first necessary to characterize clearly and concisely the characteristics of Galician forests. The particular climatic conditions in Galicia enable the region to be a leader in terms of the Spanishforestry sector. Some 9.5% of woodland in Spain is in Galicia, and this land holds 19.7% of the total volume of Spanish timber (Prada et al., 2005). According to the III National Forest Inventory (MMA, 1998), the total forest area in Galicia is more than 2 million hectares, equivalent to 69% of the territory; more than 1,405,000 ha of this is woodland, i.e.48% of the Galician territory or almost 69% of the Galician forestland. However, the main attribute that characterizes forestry in Galicia is without a doubt, the ownership system. Private ownership, whether individual orcollective, constitutes the main type of forestownership in Galicia, and corresponds to more than 1,994,000 ha, i.e. almost 98% of Galician forest; only 2.2%of Galician forestis publicly owned, the state or autonomous governments or other public bodies. In general, individual private property in the region occupies the most productive land, so that around 81.5% of the woodland in Galicia is under this type of regime (Marey-Pérez, 2003). Thus, 67.9% of private forest in Galicia is managed by more than 672,000 individual private owners, who each own a mean forested area of less than 2 ha, which may be typically subdivided into 10 plots. The remaining private forests (29.9%) correspond to communal forests –in Galician called Montes Veciñais en Man Común,acommunal form of private land tenure (Marey-Pérez and Rodríguez-Vicente, 2008) –which are currently managed by 2835 collectives of private owners. The Mariña Oriental region was therefore chosen for study since it is aforest regiontypicalofmuchofNorthernSpain,andparticularlywithin Galicia, where forest covers mostof the land (53%), and forest activity is 476 V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 increasing (over 46% of the forest is woodland managed for timber production). Thus, land use for forestry purposes increased by 8.5% between 1957 and 2001, mainly to the detriment of agrarian and scrub land, although the most important changes in land use involve forest stand composition (Marey-Pérez, 2003). Pure forest stands have increased considerably during this period (by about 300%), probably Fig. 1. Location of Mariña Oriental region in Galicia, Northern Spain. 477V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 because of the singular expansion of the forest species Eucalyptus globulus Labill. in theregion; coverof the latter species isthought to have increased to more than 63% in the area between 1957 and 2001 (MareyPérez, 2003). Thelarge expansion of productive forest stands (for timber production) in the Mariña Oriental and throughout the rest of Galicia, originates from the important changes that took place in the agroforestrysystem duringthe 1950s, as mass rural emigration hindered the development of a competitive and intensive agricultural sector, causing instability at socio-economic and land use levels (Etxezarreta,1979). As in other European countries, production ceased on a large number of farms, thus transforming land management and land patterns themselves: farm-level changes in land use are widely acknowledged as a response to decreasing agricultural economic viability (Marey-Pérez et al., 2004). The lack of farm labour led to much land becoming abandoned and progressively occupied by scrubland and native woodland; in order to counteract this tendency, the Spanish government's forest policy –begunin the late nineteenth century and reaching a peak in the mid-twentieth century –encouraged an increase in forest plantations destined principally for the fibre and chipboard industries, in an attempt to tackle the problem of low production, scarce profitability and deforestation of state woodlands (Marey-Pérez et al., 2004; Marey-Pérez and Rodríguez-Vicente, 2008). Posteriorly, European Council Regulation no. 2080/1992 of 30 June 1992, instituting a community aid scheme for forestry measures in agriculture, again favoured theestablishment of large forest plantationsin Galicia, especially during the period 1993 to 1997, when grant-aided afforestation constituted an important choice for private landowners who wished to improve the productivity of marginal lands through tree plantations, mainly with Pinus spp. and Eucalyptus spp. At present in the Mariña Oriental 3043 NIPF owners manage more than90% of the forested area in the region. The addresses of NIPF owners and attributes of their holdings,such as location, land-use and size, were identified from the Land Register (Cadastre). In accordance with Council Regulation (CEE) no. 571/88 of 29 February 1988, on organizing European Union surveys of farm structure, the sampling framework firstly consisted of all individual forest landowners living in the region who owned at least 1 ha of productive forestland, a sampling scheme proposed by Marey-Pérez (2003). This excluded many forest owners who probably would not have the necessary information to account for their management goals and practices. As a result of migratory phenomena in Galicia during the 20th century (Beiras-Torrado, 1975), a high percentage of the NIPF owners did not actually reside in MariñaOriental. This forced us to review the population census and to reclassify landowners into non-residentand resident. From 750 NIPF owners who manage more than 1 ha of productive forestland in the region,333 were classified as permanent residents, who together own 1154 ha of woodland, i.e. 42% of all productive forestland in the area, including plots smaller than 1 ha. The large number of variables included in the cadastral database revealed a high degree of heterogeneity in the study population, which is why stratification was considered a key factor in characterizing and subsequently validating the results. The information in the cadastral database indicated that the variable designated ‘productive forest area per landowner’was the most suitable for determining the minimal number of NIPF owners to interview, and their subsequent stratification. The classification of NIPF owners for determining the number of strata and the cut-off points was based on data on timber harvesting in Mariña Oriental. The size of productive forestland that enabled NIPF owners to fell the equivalent of the mean annual harvest per plot in the region defines the landowner stratification (Marey-Pérez, 2003). This was established as 3.5 ha of productive forestland, considering a weighted rotation age of 15 years for the two main forest species in the region, E. globulus Labill. and Pinus pinaster Ait. ssp. atlantica. We opted to use a questionnaire and statistical sampling within the subjective methodologyof analysis, inwhich the sample size was designed to achieve a 5% sampling error at the 95% confidence level. A priori, the error level was set at 3% for quantitative answers (mean estimation) and at 6% for qualitative answers (proportion estimation). In order to obtain comprehensive, reliable results, we attempted to enlarge the landowner sample in order to interview as many owners as possible, while minimizing the economic costs involved. Within the stratified methods, a self-weighting sample size was accordingly determined and allocated by means of Neyman's formula of minimum variance (Sukhatme, 1953). Theresultsshowedthataninitialestimationof 3%(mean estimation) wouldrequire atotalof 101questionnaires,whereasan initial estimation of 4% (proportion estimation) would require 99 questionnaires. From 333 NIPF owners, each responsible for more than 1 ha of productive forestland and permanently resident in Mariña Oriental, the selfweighting sample size was finally formed by 103 NIPF owners who had to be contacted and interviewed in person (Table 1). The NIPF landowners who were interviewed owned 12% of the forestland and 13% of the woodland (i.e., productive forestland) in the region.Thus, the error level was finally set at 4% for quantitative answers (mean estimation) and at 8% for qualitative answers (proportion estimation). We considered a number of questionnaires before deciding on a final version thatwasdividedintofoursections,whichsoughtinformationon owner profile, family unit, forest property and land-use changes, and forest economics during the period from 1999 to 2003. The definitive questionnaire was completed in two stages in March 2004. The first stage consisted of a telephone interview –between 20:00 and 22:00 h – inquiring as to the owner's willingness to participate in the study. If confirmed, the interviewer arranged for an interview in person within one or two days (the second stage). If the owner declined to be interviewed in person, the interviewer posed the questions included in the owner profile section. Each interview lasted an average of 36 min. The information was finally completed from the official data contained in the Land Register. The analytical variables for the present study are based on the information obtained in the previous personal interviews, information that was redefined and coded in nominal, ordinal or binary variables that summarized the surveyed data and met the assumptions of the statistical analyses. For the formulation of ordinal variables, we applied simple statistical criteria by using SAS/STAT™and STAT-GRAPHICS™ software. Firstly, we produced descriptive statistics and frequency histograms for the variables and then selected statistics of location and dispersion, i.e. the mean xand standard deviation σ,respectively(CaoAbad, 2002). We then used the location measurement (x) as the centre of the variables considered for calculating the class intervals from the dispersion statistics (σ). The variables considered in the study are listed and described in Table 2, along with the percentage of interviewed owners for each category. Given that forest decisions and/or practices of a representative NIPF owner are the result of a combination of individual decisions and/or practices regarding planting and silviculture treatments on forestland, as well as harvesting of woodland, these three individual forest practices were also included in the personal interviews and later defined as the following continuous variables in the present study: 1. PLANT. Forest planting, measured as the proportion of the entire forest area planted annually. 2. TREAT. Stand improvement treatments, measured as the proportion of the entire forest area in which silvicultural treatments are annually carried out. This includes activities such as the use of Table 1 Classification of interviewed NIPF owners per stratum. Stratum Productive forestland (ha) No. % A 1.00–1.70 22 21.4 B 1.71–3.50 28 27.2 C 3.50–7.00 31 30.1 DN7.01 22 21.4 103 100 478 V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 fertilizers, application of insecticides, pesticides or herbicides, thinning of competing vegetation, and other management or improvement activities. 3. HARV. Timber harvesting, measured as the proportion of the entire wooded area harvested annually. Table 2 also lists and describes the mean and standard deviation parameters observed for planting, silvicultural and harvesting practices during the five-year study period. Finally, all of the economic variables analysed in the present study wereadjustedtoconstantEurosfor2004 tocontroltheinflation rate, and were summarized as mean annual euro amounts per-hectare-owned, with the exception of family incomes and the unit stumpage timber price. The relevant information was obtained from the Spanish consumer price index of the National Statistics Institute (INE, 2004). 2.2. Statistical analyses Given that the study population was not suited to the Kolmogorov– Smirnov goodness-of-fit test for the normality K–S test or to the Levene test for homogeneity of variances, the statistical analyses conducted in this study were based on distribution-free tests, that is, non-parametric tests. The non-parametric procedure took into account the type of variables measured in the study, i.e. continuous, nominal/ordinal and binary variables, and was defined in order to provide a statistical explanation of the NIPF owners' management in terms of analysing past and future dynamics in farming and forestry practices, and landholding attributes. These dependent variables were empirically associated with other attributes linked to landowner profile, family unit, and forest economics, as well as to planting, silvicultural and harvesting management practices observed in the Mariña Oriental region. Thus, we first tested the strength and significance of the linear correlation among the variables, and subsequently contrasted the significant differences among them at a 95% confidence limit and a minimum level of statistical significance of 0.05. We used Somers' Dcoefficient and its critical significance level, to measure the statistical relationship among nominal, ordinal and/or binary variables (by use of contingency tables). We subsequently computed and detected significant differences in the frequency distribution across nominal, ordinal and/or binary variables of the cross-tabulation using Pearson's chi-square statistic χ 2 and the two-tailed asymptotic significance. Spearman's rho coefficient ρwas used to estimate the statistical association between the continuous variables and these measures against nominal, ordinal, and/or binary attributes at 0.01 and 0.05 significance levels. We also analysed the mean distribution across variables using the Kruskal–Wallis' Htest, a non-parametric test of variance homogeneity equivalent to one-way ANOVA. Confirming this premise, we conducted pairwise comparisons using the Dunnett's T3 test to determine which categories (levels) of nominal/ordinal variables showed behaviour (means) significantly different from that of the continuous variables. Homogeneous subgroups of similar statistical behaviour were finally defined by Tukey's HSD procedure, after the significant differences were tested. Finally, and to complete the investigation included in the present study, we modelled the relation between the variables MWOOD, CSPECIE and IFOREST as binary responses (1 = past change/intended changes); 0 = no past change/no intended changes) with respect to a combination of explanatory variables that enabled statistical characterization of these three management practices on the part of NIPF owners in the Mariña Oriental region. The variable FMEADOW was not used in the present model as it did not fulfil the balance between responses required to develop a statistically significant model. On the basis of the binary nature of the three dependent variables, in the present study we opted to use logistic regression by backward step-wise selection procedure, a method based on an accumulative probability function whose main objective is to model how the presence or otherwise of diverse factors, and the value or levels of these, affect the probability of an occurrence (Ryan, 1997), i.e.: Pi=EY=1xi ðÞ=1 1+e−β0+βixi ðÞ ð1Þ where, x i , independent variables P(Y)=P, the probability that the forest owner carries out the forestry activity, which takes a value of one when the NIPF owner has made the conversion from marginal meadow into woodland, attempted to change the current productive forest species or to enlarge his/her productive forestland in the future, and a value of zero when the NIPF owner has not developed or did not attempt to develop any of these practices β 0 , independent term in the logistic regression model β i ,coefficients of the logistic regression model, significantly different from zero. The estimates of the different regression coefficients were obtained by Maximum Likelihood Estimation (MLE). Testing of the statistical significance of each of the regression coefficients in the model was carried outby Wald's method, at asignificance level of 0.05(Hosmer and Lemeshow, 2000), i.e.: WALD¼ ˆ β1  2 EE ˆ β1  2ð2Þ The previous statistic follows a chi-squared distribution with k degrees of freedom χ 2 k , where: k=1, if the independent variable is quantitative k=number of categories−1, if the independent variable is qualitative, whether nominal or ordinal. The empirical model of participation of the NIPF owners in the three activities analysed was based on a function that includes regressor variables –related to the profile of the owner, the family unit, the forest land-holding, the forest economy, and the three forest management activities considered (plantation, silviculture and harvesting) –significant at a level of 0.05. The definition and descriptive statistics of these independent regressor variables are shown in Table 2. 3. Results and discussion 3.1. Past trends and future allocations in land-use 3.1.1. Past conversion from forestland to meadow None of the three forest management practices analysed in Mariña Oriental appeared to be significantly affected by the past change from forestland to meadow. Nevertheless, descriptive statistics revealed that the F M0 group appears to include the most active forest managers in the region. The F M0 group annually planted and carried out forest improvement treatments seven times more often and in twice the area of forestland, respectively, thanF M1 owners. In addition, F M0 owners also actively harvested woodland, at an annual rate five times higher than the F M1 group. The landowner status as an active farmer and the production requirements and priorities for land management clarify the reason why the F M1 owners transformed forestland into agricultural land (meadows), as well as explaining the lower involvement of these owners in forestry, although not significantly different from that of the F M0 owners. Thus, Beach et al. (2005) affirmed that land allocation between forestry and other uses is dependent on market factors such as the expected rates of return to alternative type of land use, among other 479V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 such bodies. Moreover, those NIPF owners who contracted more than 5 professional work days for forestry activities were 10% more likely to have this intention than those owners who contracted less than 5 professional work days annually. Finally, those owners who received more than 195.1 €per year and unit of forest area from the sale of timber were almost 30% more likely to increase the area of productive forest than those owners who received less than this amount or who did not sell any timber. 3.2. Territorial structure of the ownership 3.2.1. Degree of parcellation of productive forestland As Kendra (2003) stated, forestland is becoming increasingly parcelled, in contrast to the larger-scale management recommended by policymakers and land planners. Land parcellation is directly reflected in owners' decisions and practices regarding land, which generally reduces the likelihood and intensity of management (Conway et al., 2003; Potter-Witter, 2005). Indeed, the annual rates of planting and silviculture decreased slightly when the degree of parcellation of productive forestland increased (ρ=−0.293 and −0.312, respectively, at Pb0.01). A similar trend was also observed in the timber harvesting practice; the amount of woodland harvested per year decreased slightly when the number of plots per unit of productive forestland increased (ρ=−0.216 at Pb0.05). Post-hoc analyses confirmed this result, and showed that these three practices were also significantly different per PLOT group (Table 6). The P L1 owners were the most active planters in the area, and planted significantly more forestland per year than the P L4 owners (H=10.027). In addition, the P L1 group also included the most active silviculturalists in the study area, who were significantly more active than the remaining owner groups (H=9.763). Finally, the annual rate of timber harvesting by the P L1 group was significantly higher than the observed in the P L2 and P L4 groups, on average five times higher (H=13.057). Some aspects of the forest owner's profile again identified the owners in terms of the degree of parcellation of the forest property and enabled a more detailed understanding of the type of forest management that the owners carried out. Thus, we also observed that the number of plots per unit of productive forestland increased slightly in accordance with the landowner's age (ρ=0.257 at Pb0.05), and furthermore, decreased slightly with the level of formal education and primary occupation (ρ=−0.236 and −0.265, respectively, at Pb0.05). On average, the P L1 and P L2 groups were more than eight years younger than P L3 and P L4 owners (H=6.697). All owners younger than 40 years old were included in the P L2 group, whereas 61.5% of P L3 and P L4 owners were older than 65 years old; 66.7% of P L1 owners were 40–65 years old. On the other hand, 16.7% of P L1 owners had not received any formal education,as compared with 50.2% of P L2 and P L4 owners; almost 67% of P L3 landowners had received primary education (although the differences were not significant). Finally, there were almost three non agricultural professionals in the P L1 and P L2 groups to one in each of the remaining owner groups (although again the differences were not significant). The availability of agroforestry machinery also differed significantly among PLOTgroups (χ 2 =9.381). Over 31.3% of P L1 owners had suitable agricultural and forestry machinery for land management, whereas 70.8% of the remaining groups did not have such resources available. Taking into account the agrarianprofile of a large proportionof the owners included in P L3 and P L4 , this appears to confirm that agricultural productivity may be an importantfactor in land parcellation in the study area. The number of plots per unit of productive forestland also decreased slightly in relation to the increase in the annual earning per household (ρ=−0.250 at Pb0.05). As Table 6 shows, the family income annually earned by P L4 owners was significantly lower than that earned by the remaining owners (H=8.788). No P L4 landowner received more than 18,000 €in annual family income, in comparison with 44.8% of the remaining groups, who received more than this amount. Taking into account the more active forest management on the part of P L1 landowners, these results may support, although indirectly, those obtained by other authors who have reported that higher investments in forest are made by those owners with higher income per household (Hardie and Parks, 1996; Gunter et al., 2001; Mahapatra and Mitchell, 2001; Arano et al., 2004; Ross-Davis et al., 2005). However, the statistical analysis revealed that the annual income of the family units in the study was not a significant factor in the activities related to planting, silviculture and harvesting, or in the money annually invested in holding improvement or silviculture. Only the annual unitary expenditure on forest plantation increased slightly in relation to annual household income (ρ=0.220 at Pb0.05), Table 5 Parameter estimates of the logistic regression model that examines the factors affecting NIPF landowners' future intention to enlargethe area of woodland within the holding in the short/mid-term. Variable Coefficient Wald P-value Standard error ASSOC 2.368 2.899 0.039 0.510 PROFESS 1.492 6.666 0.010 0.155 TINCOME 1.681 2.8230 0.041 0.309 Constant 0.121 13.421 0.000 0.576 −2 Log likelihood 118.056 Model chi-square 3.874 a Nagelkerke R 2 0.238 Obs. With IFOREST=1 44.0 Obs. with IFOREST=0 56.0 Overall % correct 73.3 a P≤0.01. Table 6 Homogeneous PLOT subgroups with regard to annual rates of planting, silviculture and harvesting (%), family income (€/year), expenditures on planting and silviculture (€/ha per year), and the size of productive forest holding (ha). PLOT P L1 P L2 P L3 P L4 P-value % of interviewed landowners 14.0 39.5 27.9 18.6 PLANT Small planter 1.89 1.02 0.50 0.363 Large planter 3.21 1.89 1.02 0.054 TREAT Small silviculturalist 0.88 0.38 0.38 0.947 Large silviculturalist 3.54 1.000 HARV Small harvester 2.32 4.86 0.85 0.160 Large harvester 7.88 4.86 0.394 HOUSEHOLD Farmer traditionalist 18,781.63 17,786.20 10,245.38 0.051 New professional 20,997.86 18,781.63 17,786.20 0.759 PEXP Non-planter 240.22 156.09 120.65 0.226 Planter 299.42 240.22 156.09 0.105 TEXP Non-silviculturalist 97.88 67.52 51.98 0.698 Silviculturalist 168.19 97.88 67.52 0.088 SIZE Small landowner 5.25 3.44 3.39 0.404 Large landowner 7.87 5.25 0.130 486 V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 although the mean values were not statistically significant; landowners whose income exceeded 9000 €per year invested almost 100 €/ha per year more in planting their forestland than those groups with household earnings below this figure. The annual unitary investment in holding improvement increased slightly as the degree of parcellation of the productive forestland decreased (ρ=−0.216 at Pb0.05). More than 85% of P L3 and P L4 owners did not invest in their holding for annual development of new forest buildings or infrastructures, in comparison with 62% of P L1 and P L2 owners. On average, the P L1 and P L2 groups annually invested twice as much in their holdings per unit of forest area as the remaining owners (H=7.840). In the same way, annual unitary expenditure on planting and developing silviculture decreased slightly in those holdings with a larger number of plots per unit of productive forestland (ρ=−0.238 and −0.266, respectively, at Pb0.05). Indeed, more parcelled holdings cannot be managed for forestry as efficiently as larger ones, and there are fewer investment and management options available. In this case, P L1 owners were the most active in terms of planting and carrying out silvicultural practices in forestland, and were significantly more active than the remaining owners (H=7.560 and 8.471, respectively). The P L1 and P L2 owners allocated twice as much money annually per unit of forest area to these practices as the P L3 and P L4 groups (Table 6). On the basis of these findings and the characterization of the landowners, retired farmers and also non-agricultural professionals appear more likely to invest in land in order to increase its value and its returns. Active forest managers appeared to resort to professional assistance for carrying out land management practices. Thus, the annual number of days that forestry workers were hired moderately and slightly increased with the annual rates of forest planting and timber harvesting, respectively (ρ=0.408 and 0.302, respectively, at Pb0.01), as well as slightly in relation to the annual rate of silviculture (ρ=0.251 at Pb0.05). These relationships explain why the annual number of days that forestry workers were hired was higher in the less parcelled productive forest holdings (ρ=−0.246 at Pb0.05). Some 37% of P L1 and P L2 owners hiredworkers for more than 50 working days per year, while 80% of the P L3 and P L4 groups hired professional labour for less than 10 days per year. Takinginto account that 71.4% of the nonagrarian professionals were included in the P L1 and P L2 groups, it can again be deduced that forestry was generallyless feasible for managers with occupations outside the property, thereby forcing landowners to make use of professional assistance in order to develop forestry. The relationship between the degree of land parcellation and the intensity of land management may explain the significant differences in the productive possibilities from studied landholdings according to the number of plots per unit of productive forestland. Even though there was no statistical evidence regarding the annual unitary income from timber sales, the unitary stumpage price fixed from previous harvests differed significantly among PLOT groups (H=6.275). A multiple comparison analysis showed that the P L1 and P L3 owners sold timber at significantly higher stumpage prices per unit (twice as high) than the P L2 and P L4 owners. According to Cubbage (2003), technical advice provided by professional foresters may explain this result, as this type of assistance in timber harvesting and marketing may increase the landowner's net revenues from timber sales. Furthermore, Boyd (1984) and Hyberg and Holthausen (1989) reported that provision of technical assistance tends to increase the likelihood that NIPF owners will harvest their timber. Thus in the study region, the annual rate of timber harvesting increased slightly with respect to the landowner's knowledge of timber market information (ρ=0.274 at Pb0.05), and moreover, the stumpage price per unit from previous harvests increased slightly in accordance with this type of information (ρ=0.282 at Pb0.05). On average, the annual rate of timber harvesting by owners who had access to information about the timber market was twice that observed for the remaining owners (H=6.391), fixing the unitary price of the timber at more than 2 €/t (H=6.442). Finally, larger areas of productive forestland were closely linked to a lower degree of land parcellation. Thus, the number of plots per unit of productive forestland increased slightly on those holdings consisting of smaller areas of productive forestland (ρ=−0.316 at Pb0.01). The post-hoc analyses showed that the P L1 group was characterized as owning the largest productive forest holdings in the region, which were significantly larger than those owned by the P L3 and P L4 groups (H=9.553). Some 64.3% of P L1 and P L2 owners managed productive forest holdings larger than 3.5 ha, and more specifically, 41.2% of them owned more than 7 ha of productive forestland. On the contrary, 45.8% and 62.5% of P L3 and P L4 owners, respectively, owned productive forest holdings smaller than 3.5 ha, and a third of them managed less than 1.7 ha of productive forestland. These significant differences classified the owner population as shown in Table 6. 3.2.2. Land-base of productive forestland As Royer (1980) stated, ‘if there has been one factor most often associated with differences in landowner responses in surveys, it has been the size of landholdings’. In Mariña Oriental, the size of productive forest holding was found to be a key factor in making forest management viable, as found in other studies of NIPF owner land decisions (Binkley, 1981; Boyd, 1984; Hodges and Cubbage, 1990; Löyland et al., 1995; Hardie and Parks,1996; Zhang and Pearse,1997). Thus, the annual rates of planting and silviculture in forestland increased greatly and slightly, respectively, according to the size of productive forest holding (ρ=0.802 and 0.355, respectively, at Pb0.01). In addition, the annual rate of harvesting woodland increased slightly in relation to this landholding attribute (ρ=0.255 at Pb0.05). The S I4 landowners were the most active planters in the region, and differed significantly from the remaining owners (H=56.330). The rate of carrying out silvicultural practices was significantly higher in the S I4 group than in the S I2 group (H=11.787).These variables distinguished two significantsubgroupsof planters and silviculturalists in the area (Table 7). Finally, the S I2 group harvested significantly more woodland per year (N2% more) than the remaining groups (H=7.058). The size of the productive forest holding increased slightly in those landowner groups who had received a higher level of formal education (ρ=0.294atPb0.01). For each S I3 and S I4 owner who had not received any formal education, there were three in the S I1 and S I2 groups; the fraction of owners educated to tertiary-level in the S I4 group was twice as high as in the remaining groups (χ 2 =16.198). The fact that the level of education was positively correlated with the owner's primary occupation completed the previous results (D=0.406). All owners educated to tertiary level and half of the owners educated to secondary level worked in areas outside of agriculture, whereas 75% and 44% of the owners who had not received any formal education or only primary education, respectively, were retired farmers (χ 2 =39.148). Thus, 52.4% of the professionals not related to agriculture managed productive forest holdings larger than 3.5 ha, whereas a similar fraction of the active and retired farmers (57%) managed smaller productive forest holdings, probably because they were more likely to be involved in managing agrarian land. A better level of education would generally improve prospects and success within the labour market, which should be reflected in the annual earnings of the owners. As indicated in Table 7, the annual income of S I4 owners was significantly higher than that of the S I3 owners, a mean difference higher than 8300 €/year (H=11.272). Some 33.3% of S I4 owners earned more than 30,000 €/year in family income, in contrast to 11.3% of the remaining owner groups. Therefore, Nagubadi et al.'s (1996) findings were corroborated in the study region: landowners with a higher income were more likely to manage larger forest holdings. As already mentioned, involvement in agriculture may also explain why annual unitary reinvestment in forest activities increased slightly in those holdings of smaller size of productive forestland (ρ=−0.226 at Pb0.05). As seen in Table 7, the S I1 owners annually took advantage of unitary forest reinvestments 487V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 three times as often as the S I4 group (H=6.811). In fact, landowner status as an active farmer could be the main factor that explains the significant differences in the likelihood that the owner will convert forestland into meadow according to the SIZE group (χ 2 =7.735). None of the S I2 ,S I3 and S I4 owners converted forestland into meadow, in comparison with 11% of S I1 owners. Large holdings required larger annual inputs of money to improve them, as expected due to the relationship between land size and intensity of management (Hyberg and Holthausen,1989; Kuuluvainen and Salo, 1991; Prestemon and Wear, 2000; Zhang and Mehmood, 2001; Arano et al., 2004; Potter-Witter, 2005; Arano and Munn, 2006). The annual unitary investment in holding improvement increased slightly with the size of the productive forest holding (ρ=0.216 at Pb0.05). The S I4 owners spent per unit of forest area twice as much per year on new forest buildings and infrastructures as the remaining owners (H=7.638). Furthermore, annual unitary expenditures on planting and forest improvement treatments moderately and slightly increased as holding size increased, respectively (ρ=0.524 and 0.369, respectively, at Pb0.01). Larger holdings would therefore be more likely to be managed, and more specifically, to be planted and improved, as found by Boyd (1984),Löyland et al. (1995),Arano et al. (2004) and Potter-Witter (2005). The S I4 owners invested per unit of forest area significantly more per year in planting and silviculture than the remaining owners (H=23.997 and 10.217, respectively). The significant differences found in annual unitary expenditure on planting and silviculture enabled classification of landowners as shown in Table 7. The landowners appeared to resort to public subventions in forestry to implement these activities and to pay off their corresponding expenditures, as previously mentioned. We found that the likelihood of a landowner applying for public subventions significantly differed according to the size of the productive forest holding (χ 2 =12.851). None of the S I2 ownersand 11.1%oftheS I1 groupapplied forsubventions, compared with 30.4% of the remaining owner population. As regards the annual unitary amount granted per subsidy, the S I3 and S I4 owners annually received twice as much as the S I1 group, although the difference was not significant. Owners of large landholdings also appeared to beactively supported in technical guidance in forestry. The number of days that forestry workers were hired for per year increased slightly in accordance with the size of productive forestland (ρ=0.277 at Pb0.01). Over 45% of S I1 and S I2 owners hired forestry workers for fewer than 5 days a year; a similar percentage of S I4 owners annually hired workers for more than 50 days, and 22.2% of S I3 owners hired workers for between 11 and 100 days a year (χ 2 =23.103). Therefore, large landowners were more likely to rely on professional assistance for forest management, in accordance with the findings of Nagubadi et al. (1996), Zhang and Mehmood (2001) and Van Gossum et al. (2005). As mentioned throughout this study, the main reason for contacting forest professionals was linked to the landowner's primary occupation. Larger productive forest holdings were characterized by higher timber income at a higher unitary stumpage price, probably because of the greaterlikelihood of being harvested and obtaining a higher timber volume to negotiate a better price at a lower harvesting cost. The positive significance of the size of the landowners' property on timber harvesting has been analysed in several studies (Binkley, 1981; Boyd, 1984; Kuuluvainen and Salo,1991; Löyland et al.,1995; Prestemon and Wear, 2000; Conway et al., 2003; Potter-Witter, 2005; Bolkesjø et al., 2007; Størdal et al., 2008). Thus, the annual unitary timber income and timber selling price per unit increased slightly in relation to the size of the productive forest holding (ρ=0.372 and 0.285, respectively, at Pb0.01). As Table 7 shows, the S I4 owners annually obtained a significantly larger unitary income from timber sales at significantly higher unitary stumpage prices than the other owners (H=13.802 and 9.958, respectively). Taking into account the guidelines for forest management observed in the area, attractive timber prices, in addition to professional support in forest management may motivate landowners to become involved in forestry, as also pointed out by Munn and Rucker (1994) and Löyland et al. (1995). Finally, larger land-bases of productive forestland were generally characterized by a lower degree of parcellation, as already stated. Thus, the number of plots per unit of productive forestland decreased slightly in those holdings comprising a large area of productive forestland (ρ= −0.316 at Pb0.01). As expected, the S I4 group owned productive forest holdings that were significantly less parcelled than those owned by the S I1 group (H=13.027). Some 66.6% of S I1 and S I2 owners managed holdings comprising more than 3.4 productive forest plots per hectare, compared with 73% of the owners in S I3 and S I4 groups whose holdings comprised less than 1.52 plots per hectare of productive forestland. On average, the S I1 group owned holdings with two plots per unitof productive forestland more than S I4 owners, and one plot more per unit of productive forestland than S I2 and S I3 groups (Table 7). This finding, together with the previous characterization of the landowners, appears confirm that owners involved in farming were certainly more likely to manage small and parcelled holdings to intensify their agrarian production. Table 7 Homogeneous SIZE subgroups with regard to annual rates of planting, silviculture and harvesting (%), family income (€/year) and forest reinvestment for household consumption (€/ ha per year), expenditures on plantation and silviculture (€/ha per year), income (€/ha per year) and stumpage price (€/t) from timber sales, and the degree of parcellation of the productive forest holding (no. plot/ha). SIZE S I1 S I2 S I3 S I4 P-value % of interviewed landowners 20.9 23.3 31.4 24.4 PLANT Small planter 0.17 0.33 1.09 0.463 Large planter 4.58 1.000 TREAT Small silviculturalist 0.46 0.32 0.65 2.65 0.065 Large silviculturalist 0.46 0.65 0.982 HOUSEHOLD Farmer traditionalist 19,891.83 14,319.11 13,784.32 0.193 New professional 19,891.83 14,319.11 22,130.,12 0.057 REINVEST Non-forest consumer 94.73 58.20 33.03 0.055 Forest consumer 108.19 94.73 58.20 0.164 PEXP Non-planter 95.25 148.80 183.50 0.339 Planter 371.06 1.000 TEXP Non-silviculturalist 40.67 69.52 67.09 0.658 Silviculturalist 69.52 67.09 108.23 0.360 TINCOME Non-wood seller 86.98 132.04 231.10 0.152 Wood seller 132.04 231.10 301.66 0.066 TPRICE Non-timber industrialist 2.25 3.41 4.35 0.336 Timber industrialist 3.41 4.35 5.92 0.192 PLOT Small landowner 3.20 3.71 2.43 0.101 Large landowner 4.30 3.20 3.71 0.197 488 V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 4. Conclusions and implications As a prior step in developing and executing successful forest measures in land management and planning policy, the main objective of the present study was to examine structural attributes of the forest holding, as well as past changes and future intentions for land use, in relation to individual forest management in the Mariña Oriental region, in Northeast Galicia (Northern Spain) for the 1999–2003 period. The data used was obtained, in 2004, by personal interviews with 103 NIPF owners, each responsible for more than 1 ha of productive forestland in the study area. On the basis of the results obtained, three main types of NIPF owners were identified: retired farmers who maintained forestry activity as their only link with the agrarian sector, and who were actively involved in carrying out the work themselves; active full-time farmers who viewed their forestry activities as an extra source of income, and finally, owners occupied in other professional sectors (outside of agrarian activities) who own forestland, but which is largely managed by contracted personnel. Secondly, the correlation and dependency analyses revealed that the greatest and most efficient forest activity was associated with three key parameters: larger areas of productive forestland in the property, less-divided forestland and greater availability of time to dedicate to forestry activities, whether by the owners themselves or by contracted workers. As regards past changes and future intentions for land use, conversion of forestland into meadow responded to the demand for increasing agrarian land-base, and this past land-use change was significantly related to the landowner's occupation as an active farmer. In the region, owners who were farmers generally managed smaller productive forest land-bases, because of their full-time commitment to agriculture, and more parcelled productive forest holdings, probably because of their aim to improve and increase the overall agrarian productivity, in contrast to the remaining landowner population. Furthermore, both past transformation of marginal meadow to woodland and the future intention to increase the area of productive forestland in the short-medium term clearly depends on past experience. Thus, those NIPF owners who had profited by selling timber from previous harvests (high annual unitary income from timber sales at better unitary stumpage prices), had either increased the area of woodland on their properties or intended to do so in the near future. Retired farmers and non-agricultural professionals were the owners most likely to carry out this type of land-use conversion through use of public subsidies. Finally, the intention to change the productive forest species in the next productive cycle was also linked to the forest profitability. Thus, those NIPF owners who devoted more time to forestry work, either themselves or through contracted workers, and who invested more money in the activity were more inclined to improve theprofitability oftheirland. The profile thatincludedthis type of owner mainly corresponds to active farmers, who because they do not have more land to reforest because the land is required for livestock purposes, aim to improve the profitability of the forestland by the type of change mentioned. In summary, the main conclusions of the present study are that the NIPF owners in the Mariña Oriental appear to respond clearly to timber markets, as other authors have concluded in other regions, with the particular characteristic of “a moral responsibility”of looking after the land or using it for productive purposes —in contrast with the alternative of the decapitalization of abandoned farm land. The forest area and its configuration (number of plots per unit of land) are also important factors in motivating NIPF owners to initiate or continue forestry activities, with the so-called economies of size being particularly important, as also found in other studies. Forest professionals, researchers and policymakers should take into account the existence of different profiles of private forest owners and managers, and therefore, the existence of different types of land holdings and land management practices, in making decisions and developing programmes involving sustainable land planning and forest resources. The present study demonstrates that it is essential to reinforce the role of the associations and of the technicians and/or professional forestry workers in forest planning and management, given the importance of this in initiating and continuing forest activity, by providing information and training, generating and professionalizing the work, and finally producing timber and other forest products and services. In addition, the land-base of productive forestland must necessarily be improved or restructured, either by direct land planning measures (agroforestry land consolidation) or by indirect forest cooperation measures (landowner involvement in management-related associations). Consequently, public programmes aimed at ensuring the continuity of private forest management and that may improve holding profitability and viability should be designed and executed. Such public programmes should involve the use of economic incentives, training and practical tools, and professional assistance specifically aimed at NIPF owners. To maximize and make forestry competitive within rural development, further and continuous research is required in order to discover and test plausible means of encouraging NIPF owners to invest appropriately in a land-base in keeping with their personal and family circumstances or needs. References Arano, K.G., Munn, I.A., 2006. Evaluating forest management intensity: a comparison among major forestland owner types. Forest Policy and Economics 9, 237–248. Arano, K.G., Munn, I.A., Gunter, J.E., Bullard, S.H., Doolitle, M.L., 2004. Comparison between regenerators and non-regenerators in Mississippi: a discriminant analysis. Southern Journal of Applied Forestry 22 (2), 132–138. Beach, R.H., Pattanayak, S.K., Yang, J., Murray, B.C., Abt, R.C., 2005. Econometric studies of non-industrial private forest management a review and synthesis. Forest Policy and Economics 7, 261–281. Beiras-Torrado, X.M., 1975. A emigración: o seu papel na dinámica da formación social. In: García-Sabell, D. (Ed.), A Galicia rural na encrucillada. Galaxia, Vigo, Spain, pp. 39–73. Biĉík, I., Jeleĉek, L., Ŝtěpánek, V., 2001. Land-use changes and their social driving forces in Czechia in the 19th and 20th centuries. Land Use Policy 18, 65–73. Binkley, C.S., 1981. Timber supply from private nonindustrial forests. Bulletin No. 92. Yale University School of Forestry and Environmental Studies, New Haven CT. Bolkesjø, T.F., Baardsen, S., 2002. Roundwood supply in Norway: micro-level analysis of self-employed forest owners. Forest Policy and Economics 4, 55–64. Bolkesjø, T.F., Solberg, B., Wangen, K.R., 2007. Heterogeneity in nonindustrial private roundwood supply: lessons from a large panel of forest owners. Journal of Forest Economics 13, 7–28. Boyd, R., 1984. Government support of non-industrial production: the case of private forests. Southern Economic Journal 51 (1), 89–107. Butler, B.J., Swenson, J.J., Alig, R.J., 2004. Forest fragmentation in the Pacific Northwest: quantification and correlations. Forest Ecology and Management 189, 363–373. Cao-Abad, R., 2002. Análisis multivariante. Curso de postgrado en estadística aplicada. University of A Coruña, A Coruña, Spain. Cihlar, J., Jansen, L.J.M., 2001. From land cover to land-use: a methodology for efficient land-use mapping over large areas. The Professional Geographer 53 (2), 275–289. Conway, M.C., Amacher, G.S., Sullivan, B.J., 2003. Decisions nonindustrial forest landowners make: an empirical examination. Journal of Forest Economics 9,181–203. Cubbage, F.W., 2003. The value of foresters. Forest Landowner 62 (l), 16–19. Dewees, P.A., 1992. Social and economic incentives for smallholder tree growing: a case study from Muranga District, Kenya. Community Forestry Case Study Series 5FAO, Rome, Italy. Etxezarreta, M., 1979. La evolución del campesinado. La agricultura en el desarrollo capitalista. Servicio de Publicaciones del MAPA, Madrid, Spain. Finley, A.O., 2002. Assessing private forest landowners' attitudes towards, and ideas for, cross-boundary cooperation in western Massachusetts. Ph. Thesis. University of Massachusetts, Amherst, 28 pp. Gunter, J.E., Bullard, S.H., Doolitle, M.L., Arano, K.G., 2001. Reforestation of Harvested Timberlands in Mississippi: Behaviour and Attitudes of Non-industrial Private Forest Landowners. Mississippi State University, Mississippi. 25 pp. Hardie, I.W., Parks, P.J., 1996. Program enrollment and acreage response to reforestation cost-sharing programs. Land Economics 72, 248–260. Healy, R.G., 1985. Competition for Land in the American South. The Conservation Foundation, Washington, DC. 333 pp. Herbohn, J., 2001. Prospects for small-scale forestry in Australia. In: Niskanen, A., Väyrynen, J. (Eds.), Economic Sustainability of Small-scale Forestry. EFI Proceedings no 36. European Forest Institute, Finland, pp. 9–20. Hodges, D.G., Cubbage, F.W., 1990. Adoption behaviour of technical assistance foresters in the Southern Pine region. Forest Science 36 (3), 516–530. Hosmer, D.W., Lemeshow, S., 2000. Applied Logistic Regression. John Wiley & Sons, Inc., New York, The United States. 2 a ed. Hyberg, B., Holthausen, D., 1989. The behavior of non-industrial private forest landowners. Canadian Journal of Forest Research 19, 1014–1023. 489V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 INE, 2004. 2004. [Internet site]. Consumer Price Index. Instituto Nacional de Estadística, Madrid, Spain. Available at: http://www.ine.es/daco/ipc.htm. [Cited 6 December 2004]. Jansen, L.J.M., Di Gregorio, A., 2003. Land-use data collection using the “land cover classification system”: results from a case study in Kenya. Land Use Policy 20, 131–148. Karppinen, H., 1998. Values and objectives of non-industrial private forest owners in Finland. Silva Fennica 32 (1), 43–59. Kendra, A., 2003. Newlandowners in Virginia's forest: a study of motivations, management activities, and perceived obstacles. Ph. Thesis. Faculty of the Virginia Polytechnic Institute and State University, 163 pp. Kittredge, D.B., 2005. The cooperation of private forest owners on scales larger than one individual property: international examples and potential application in the United States. Forest Policy and Economics 7, 671–688. Kline, J.D., Alig, R.J., 2001. A spatial model of land use change for western Oregon and western Washington. Research paper PNW-RP-528. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, 32 pp. Kline, J.D., Butler, B.J., Alig, R.J., 2002. Tree planting in the south: what does the future hold? Southern Journal of Applied Forestry 26 (2), 99–107. Koontz, T.M., 2001. Money talks—but to whom? Financial versus nonmonetary motivations in land use decisions. Society & Natural Resources 14 (1), 51–65. Kurttila, M., Hamalainen, K., Kajanus, M., Pesonen, M., 2001. Non-industrial private forest owners' attitudes towards the operational environment of forestry —a multinomial logit model analysis. Forest Policy and Economics 2 (1), 13–28. Kuuluvainen, J., Salo, J., 1991. Timber supply and life cycle harvest of non-industrial private forest owners: an empirical analysis of the Finnish case. Forest Science 37, 1011–1029. Kuuluvainen, J., Karppinen, H., Ovaskainen, V., 1996. Landowner objectives and nonindustrial private timber supply. Forest Science 42, 300–308. Löyland, K., Kringstad, V., Öy, H.,1995. Determinants of forest activities —a studyof private non-industrial forestry in Norway. Journal of Forest Economics 1 (2), 219–237. Madsen, L.M., Adriansen, H.K., 2004. Understanding the use of rural space: the need for multi-methods. Journal of Rural Studies 20, 485–497. Mahapatra, A.K., Mitchell, C.P., 2001. Classifying tree planters and non planters in a subsistence farming system using a discriminant analytical approach. Agroforestry Systems 52, 41–52. Marey-Pérez, M.F., 2003. Tenencia de la tierra en Galicia: modelo para la caracterización de los propietarios forestales. Ph. Thesis. University of Santiago de Compostela, Santiago de Compostela, Spain, 633 pp. Marey-Pérez, M.F., Rodríguez-Vicente, V., 2008. Forest transition in Northern Spain: local responses on large-scale programmes of field-afforestation. Land Use Policy 26 (1), 139–156. Marey-Pérez, M.F., Rodríguez-Vicente, V., Crecente-Maseda, R., 2004. El monte en Galicia en el siglo XXI: balance evolutivo y consideraciones para el futuro. In: MayaFrades, A. (Ed.), ¿Qué futuro para los espacios rurales? University of León, León, Spain, pp. 117–125. Marey-Pérez, M.F., Rodríguez-Vicente, V., Crecente-Maseda, R., 2006. Using GIS to measure changes in the temporal and spatial dynamics of forestland: experiences from north-west Spain. Forestry 79 (4), 409–423. MMA, 1998. III Inventario Forestal Nacional. Madrid. Dirección General de Conservación de la Naturaleza. Ministerio de Medio Ambiente. Munn, I.A., Rucker, R., 1994. The value of information in a market for factors of production with multiple attributes: the role of consultants in private timber sales. Forest Science 40, 474–486. Newman, D.H., Wear, D.N., 1993. Production economics of private forestry: a comparison of industrial and non-industrial forest owners. American Journal of Agricultural Economics 75 (3), 674–684. Nagubadi, V., Mcnamara, K.T., Hoover Jr., W.L., Mills, W.L., 1996. Program participation behaviour of nonindustrial forest landowners: a probit analysis. Journal of Agricultural and Applied Economics 28 (2), 323–336. Pattanayak, S.K., Murray, B.C., Abt, R., 2002. How joint in joint forest production: an econometric analysis of timber supply conditional on endogenous amenity values. Forest Science 48 (3), 479–491. Potter-Witter, K., 2005. A cross-sectional analysis of Michigan nonindustrial private forest landowners. Southern Journal of Applied Forestry 22 (2), 132–138. Prada, A., Vázquez, M.X., Soliño, M., 2005. Beneficios y costes sociales en la conservación de la Red Natura 2000. Fundación Caixa Galicia, A Coruña, Spain. Prestemon, J., Wear, D., 2000. Linking harvest choices to timber supply. Forest Science 46 (3), 377–389. Ross-Davis, A.L., Broussard, S.R., Jacobs, D.F., Davis, A.S., 2005. Afforestation motivations of private landowners: an examination of hardwood tree plantings in Indiana. Northern Journal of Applied Forestry 22 (3), 149–153. Royer, J.P., 1980. Surveying nonindustrial private forests and their owners. Nonindustrial privateforests: data and information needs. Center for Resource and Environmental Policy Research, School of Forestry and Environmental Studies. Duke University, Durham, NC. Ryan, T.P., 1997. Modern Regression Methods. John Wiley, New York. The United States. Størdal, S., Lien, G., Baardsen, S., 2008. Analyzing determinants of forest owners' decision-making using a sample selection framework. Journal of Forest Economics 14, 159–176. Siry, J.P., Cubbage, F.W., Ahmed, M.R., 2005. Sustainable forest management: global trends andS opportunities. Forest Policy and Economics 5, 551–561. Sukhatme, P.U., 1953. Sampling Theory of Surveys. FAO, Rome, Italy. Van Gossum, P., Luyssaert, S., Serbruyns, I., Mortier, F., 2005. Forest groups as support to private forest owners in developing close-to-nature management. Forest Policy and Economics 7, 589–601. Wear, D., Parks, P., 1994. The economics of timber supply: an analytical synthesis of modelling approaches. Natural Resource Modelling 8 (3), 199–223. Wiersum, K.F., Elands, B.H.M., O'Leary, T.N., 2002. Landowners' perspectives on the future of rural Europe: consequences for farm forestry. In: von Teuffel, K. (Ed.), Proceedings of the International Symposium on Contributions of Family–Farm Enterprises to Sustainable Rural Development. Gengenbach, Germany, p. 13. Zhang, D., Flick, W., 2001. Sticks, carrots and reforestation investment. Land Economics 77 (3), 443–456. Zhang, D., Mehmood, S.R., 2001. Predicting non-industrial private forest landowners' choice of a forester for harvesting and tree planting assistance in Alabama. Southern Journal of Applied Forestry 25 (3), 101–107. Zhang, D., Pearse, P., 1997. Differences in silvicultural investment under various types of forest tenure in British Columbia. Forest Science 42 (4), 442–449. 490 V. Rodríguez-Vicente, M.F. Marey-Pérez / Forest Policy and Economics 11 (2009) 475–490 MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research I MODELO DE GESTIÓN PARA EXPLOTACIONES FORESTALES EN GALICIA: NUEVO ENFOQUE PARA LA INVESTIGACIÓN FORESTAL Modelling non-industrial private forest management in Galicia: a new approach for forest research V Journal of Forest Economics 16 (2010) 269–295 Contents lists available at ScienceDirect Journal of Forest Economics journal homepage: www.elsevier.de/jfe Analysis of individual private forestry in northern Spain according to economic factors related to management Verónica Rodríguez-Vicentea,∗, Manuel F. Marey-Pérezb,1 aGalician Sectorial Forestry Association (ASEFOGA), Doutor Maceira 13-baixo, 15706 Santiago de Compostela, Spain bDepartment of Agroforestry Engineering, University of Santiago de Compostela, Campus Universitario s/n, 27002 Lugo, Spain article info Article history: Received 13 October 2008 Accepted 8 June 2010 JEL classification: Q23 Keywords: Forest expenditure Forest income Holding investment Non-industrial private forest (NIPF) owner Stumpage price Subsidy abstract In addition to being motivated by profit, the management decisions taken by non-industrial private forest (NIPF) owners involve other considerations beyond timber, such as non-timber goods and services, as well as factors that affect the level of timber output from the land. Ensuring and improving forest profitability to make NIPF management viable is one of the main challenges faced by this type of landowner. This study empirically explores and assesses management by NIPF owners, through analysing attributes of forest economics (investment in holdings, expenditure on planting and silviculture, public subsidies, along with timber and non-timber incomes). With the aim of predicting outcomes, a multiple regression model was also constructed to investigate and quantify the relationship between socioeconomic and holding factors, and the planting activities carried out by NIPF owners. For this, 103 resident forest landowners in a forest region in northern Spain were interviewed in person, during March 2004, about their commitment to and involvement in land management during the period 1999–2003. The results mainly revealed that attractive forest returns and favourable market conditions for timber production are significant factors for investment in and development of forestry, with personal and family conditions also being important factors in explaining the type of land management carried out. In particular, the multiple linear regression model for forest planting activity correctly explained 84.5% of the variability observed in the study population, indicating that both the investments in and ∗Corresponding author. Tel.: +34 981 530 500; fax: +34 981 531 724. E-mail addresses: [email protected] (V. Rodríguez-Vicente), [email protected] (M.F. Marey-Pérez). 1Tel.: +34 982 252 303; fax: +34 982 285 926. 1104-6899/$ – see front matter © 2010 Department of Forest Economics, SLU Umeå, Sweden. Published by Elsevier GmbH. All rights reserved. doi:10.1016/j.jfe.2010.06.001 270 V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 the incomes from forestry play an important role in the activity, as does the size of the holding. The findings may be of interest in promoting public measures related to timber markets and economic incentives for forest management, which will allow landowners to develop economically viable practices, as well as enabling fulfilment of social and environmental demands for sustainable forestry and rural development. © 2010 Department of Forest Economics, SLU Umeå, Sweden. Published by Elsevier GmbH. All rights reserved. Introduction The need to consider environmental issues as well as incorporating social and economic criteria in forest decision-making and management has been recognized and accepted as a paradigm for rural development in recent times, a salient theme in forestry throughout the world today. Sustainable forest planning guidelines are particularly complex for non-industrial private forest (NIPF) owners because their objectives are much more diverse than those of other types of private landowner, given that they are heterogeneous by nature (Arano and Munn, 2006). As pointed out by Alig et al. (1990), NIPF owners and their holdings are diverse both within and across regions, where intentions vary widely and often change over time. Furthermore, many of them do not cite timber production as one of their primary aims. The situation is particularly complex within regions where this type of private forest ownership predominates and contributes significantly to rural development. In many rural areas, NIPF land management generates numerous benefits that complement the economy, contributes to social welfare and improves the natural environment (Marey-Pérez and Rodríguez-Vicente, 2008). Thus, in addition to being motivated by profit, the decisions made by NIPF owners with regard to production are affected by other considerations beyond timber, such as non-timber goods and services, as well as factors that affect the level of timber outputs from the land (Newman and Wear, 1993). Therefore, the balance among forest productivity, management and monitoring, and profitability is more complex to model and forecast for NIPF ownership than for other types of land tenure. Two basic theoretical models have been used to analyse and model the types of NIPF land management within the extensive literature concerning this type of private individual ownership: utility maximization and profit maximization. In the utility maximization approach, NIPF owners select from among timber and non-timber options that forests offer to maximize perceived utility for themselves (financial and non-financial benefits from the land), whereas the profit maximization assumption views the landowner as a firm or commercial entity and the forest as a unit of production, usually of timber products (Alig et al., 1990). Studies of NIPF owners commonly profile and model them as utility-maximizers of forests, given that non-timber products may be of equal or greater importance to NIPF owners than timber products (Binkley, 1981; Boyd, 1984; Pattanayak et al., 2002; Conway et al., 2003; Potter-Witter, 2005). In a more globalized economy, the current and future competitiveness of the forest management practices carried out by many NIPF owners are nevertheless threatened, as forest practices that are socially acceptable and environmentally respectful may not be economically profitable. Moreover, high investment in silvicultural treatments is required throughout the productive cycle and there is a long delay between planting and timber harvesting in the rotation of forest species. This means that landowners cannot generate a constant economic cash-flow, which would encourage and ensure continuous management and monitoring, as in other agrarian practices. Factors such as the long-term nature of any profits, lack of professionalism, the use of forestry practices that are based on family requirements, as well as the increasing proportion of landowners (who do not earn their living from agriculture as they have more profitable primary occupations), and market competition based on low prices but high costs (Bolkesjø and Baardsen, 2002; Marey-Pérez et al., 2004) all contribute to destabilizing the economic sustainability of forest management, and hence, social and environmental sustainability in rural areas. V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 271 The present study attempts to mitigate the general concern in scientific and political fields about landmanagement practices by NIPF owners,through empirical analysis and discussionof the following key issues: 1. Forestryasan ‘economicactivity’ (profit model) thatsupports andmatches theNIPF economy,that is, forest decision-making and management based on the economic balance of total forest production, taking into account the benefits from production for own consumption, intermediate expenditure, depreciation and taxes, plus subsidies, repaid with added interest and at certain risk, and 2. Forestry as a ‘moral norm’ (utility model) in which the land is managed and maintained within the NIPF heritage, that is, forest decision-making and management based on a close personal links between the landowner and his/her property, as well as family assets that will be passed on to future generations. The aim of the study was therefore to explore management by NIPF owners by empirical analysis of variables concerning the balance sheet,and the profit and loss account in forestry, represented as investment in holding, expenditure on planting and silviculture, public subsidies, along with timberand non-timber-related incomes, by targeting surveyed resident NIPF owners in an area in northern Spain. However, according to Karppinen (1998), forest management, as a voluntary action, is primarily driven by the motivations of the landowner, i.e. their values and goals. Thus, analysis of the economy of NIPF forest management practices would not be complete without analysis of the agroforestry system and the landowners’ personal goals and circumstances. We therefore explored possible statistical relationships or distinctions between these economic variables and other factors related to the landowner profile, family unit, forest property, and land-use changes. In order to complete the results, we moreover included the three practices traditionally used to predict forest management behaviour of NIPF owners in the relevant literature, i.e. planting and silviculture on forestland and timber harvesting on woodland (Löyland et al., 1995; Hardie and Parks, 1996; Kuuluvainen et al., 1996; Prestemon and Wear, 2000; Zhang and Flick, 2001; Kline et al., 2002; Conway et al., 2003; Arano et al., 2004; PotterWitter, 2005; Ross-Davis et al., 2005; Størdal et al., 2008), and examined these empirically in relation to economic factors of relevance in forestry activities. In summary, we used sociodemographic and territorial data to evaluate and therefore address the role of the forest economy in NIPF management behaviour, whilst considering the forest practices carried out by NIPF owners by modelling the role and weight of economic forest attributes, which are always and inseparably determined by the preferences and circumstances of the manager, the family unit, and the territorial system. Ensuring and improving forest profitability to make NIPF management viable is one of the main concerns in decision-making and practices carried out by this type of private owner. Characterizing land decision-making and management by NIPF owners in this way may allow policymakers to design suitable measures or tools to implement profitable forest practices under the current criteria of sustainability and within the framework of rural development. As Lillandt (2001) stated, individual landowners are not motivated to participate in forest practices without economic profitability. After outlining the background information about NIPF land management behaviour and centering the research objective, the article is structured as follows. The second section describes the study area and explains the empirical data, variables, and methodological framework employed. The third section presents and discusses the study results. Finally, the conclusions and implications are drawn in the last section. Materials and methods Study area and data collection Following the research initiated by Marey-Pérez (2003), concerning individual private land ownership and its forest management in the Autonomous Community of Galicia (northern Spain), data for this survey were collected by interviews in person with randomly selected NIPF owners within the Mari˜ na Oriental area of northeast Galicia (Fig. 1). This area was chosen for the study as it is a forest 278 V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 Table 4 Significant differences among study variables (Pearson’s 2and Kruskal–Wallis’ Htests). INVEST (O) PEXP (O) TEXP (O) REQUEST (B) SUB (O) TINCOME (O) TPRICE (O) NTINCOME (O) Landowner land management practices PLANT (C) H16.647* 44.154* 16.787* 5.165* 5.270* 18.843* 15.897* 1.414 TREAT (C) H14.471* 22.972* 29.402* 2.747* 5.883 3.192 4.708 1.857 HARV (C) H2.847 7.112* 2.397* 0.398 2.718 67.637* 66.328* 21.944* Landowner profile AGE (C) H2.107 2.769 0.084 4.490* 5.162* 2.845 2.258 1.616 EDUC (O) 2/H4.463 13.030 2.550 2.621 5.461 6.987 8.165 2.981 OCCUP (N) 2/H4.410 30.986* 19.207 7.134 7.212 23.363* 20.178 10.379 FARM (B) 2/H0.136 5.160 4.950 2.184 2.711 1.071 5.783 3.041 ASSOC (B) 2/H1.866 2.293 0.908 0.360 6.209 5.889 1.070 4.589 TRAINING (B) 2/H7.642* 1.824 2.156 0.203 5.240 1.263 4.226 0.893 IMARKET (B) 2/H0.871 3.706 1.171 0.002 1.734 7.984* 6.635* 3.462 TECHNIC (N) 2/H3.980 18.982* 6.063 1.060 4.837 11.345* 17.207* 0.394 Family unit INHERIT (N) 2/H9.757* 13.780* 2.888 6.853* 6.921 8.547 4.650 1.926 BEQUEST (N) 2/H3.883 7.419 2.711 0.808 0.544 5.560 4.024 3.522 HOUSEHOLD (C) H0.071 9.038* 5.110* 0.599 1.765* 3.217 2.791 0.380 REINVEST (C) H0.054 4.123* 4.032* 1.992 4.994* 0.252 0.293 18.888* PERSONAL (O) 2/H14.427 15.855 7.908 3.997 23.563 19.653 38.867* 6.854 FAMILY (O) 2/H8.751 37.273* 15.872 2.548 15.230 27.182* 27.779* 12.410 MACHINERY (B) 2/H3.445 0.597 0.980 0.696 6.974 2.254 1.943 3.043 PROFESS (O) 2/H30.016* 42.573* 27.407* 8.849 17.798 18.461 18.332 9.420 Forest property and land-use changes FMEADOW (B) 2/H4.807* 0.623 2.749 0.542 0.210 2.533 0.353 2.842 MWOOD (B) 2/H3.029 4.036 3.190 3.095* 7.238 1.454 1.139 1.558 CSPECIE (B) 2/H2.125 3.267 3.869 0.000 2.948 2.453 4.632 0.195 IFOREST (B) 2/H6.289* 9.550* 7.929* 2.644 4.176 5.744 4.008 2.043 PLOT (C) H4.544 8.333* 5.437 0.011 5.205* 4.119 8.624* 2.194 SIZE (C) H6.844* 19.627* 6.366* 3.672* 3.592 13.720* 14.683* 0.378 Forest economics INVEST (C) H21.154* 8.130* 8.188* 13.304* 4.607 2.603 0.848 PEXP (C) H21.000* 37.598* 4.463* 5.347 12.384* 10.061* 4.825* TEXP (C) H10.567* 50.134* 1.923 2.712 2.026 1.401 5.108* REQUEST (B) 2/H10.808* 8.368* 0.920 2.415 7.484 0.417 SUB (C) H8.112* 2.127 1.631 2.564 2.833 1.649 TINCOME (C) H4.536 10.694* 5.552 0.265 0.715 13.039* TPRICE (C) H4.774* 7.701 1.257 0.177 0.830 9.538* NTINCOME (C) H0.716 11.737* 9.849* 0.105 1.975 4.636* 1.356* Note: Variable subscripts indicate the type of variable used in the statistical analysis (C, continuous; N, nominal; O, ordinal; B, binary). *Statistically significant coefficient: *P< 0.05. V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 279 significance of the F-statistic. At each step, the regression algorithm selects the independent variable which yields the largest reduction in the unexplained variance of the dependent variable, i.e. the independent variable most highly correlated with the dependent variable. If adding an explanatory variable contributes to the regression model (the significance level associated with F-statistic is lower than 0.05, entry probability), then it is retained, but all other explanatory variables in the equation are then re-tested to check if they are still contributing to the success of the regression model; if they no longer contribute significantly they are removed (the significance level associated with F-statistics is higher than 0.10, exit probability). The regression algorithm finishes when no variables meet the entry or exit criteria of probability. The global importance of the set of independent variables and the relative importance of each of them in the regression model were measured by means of the global F-test and the individual t-test, respectively, at P< 0.05. In order to detect high correlations between explanatory variables that could cause problems with regard to the success of the regression equation, the collinearity analysis was controlled with the tolerance criterion (a measure of the proportion of variance of a variable that does not depend on the remaining variables included in the model). In this way, an independent variable formed part of the regression model if the level of tolerance was higher than 0.0001 (the closer to zero the tolerance value is for an explanatory variable, the stronger the relationship between this and the other explanatory variables). The selected reduced model was computed and subjected to residual analysis to ensure a reasonable degree of independency (Durbin–Watson’s DW-test), homoscedasticity and normality (through graphics of standardized residuals plotted against predicted values, and a histogram of standardized residuals, respectively), and linearity (Adjusted-R2of the regression model) of residuals. Thus, the robustness of the regression equation was ensured. Results and discussion Investment in holding improvement In the study region, management decisions in terms of planting and carrying out silvicultural practices in forestland were both influenced by the annual amount invested in improving the holding by means of new forest buildings and infrastructures. Thus, the investment annually devoted to holding improvement increased moderately with the annual rates of planting and carrying out forest stand improvements (=0.415 and 0.413, respectively, at P< 0.01). The post hoc analyses revealed that the two forest practices differed significantly according to the INVEST group (H=16.647 and 14.471, respectively). In our analysis, IV1 landowners were the most active planters in the area, with a significantly higher annual rate of planting than the IV0 landowners. On the other hand, the IV2 owners were the most active in terms of silvicultural treatments, differing significantly from the IV0 group. The homogeneous subgroups of landowners according to these three variables are illustrated in Table 5. The annual investment in holding improvement was moderately and positively correlated with the annual expenditure on planting (= 0.474 at P< 0.01), whilst being weakly and positively correlated with the annual expenditure on silviculture (=0.341 at P< 0.01). In this case, IV0 owners spent significantly less per year on forest planting than the remaining population (H=21.000). As regards the annual expenditure on silviculture, the mean expenditure on forest stand improvements differed significantly between the IV0 and IV2 groups (H=10.567). The classification of the study population according to annual investment in holding improvement and annual expenditures on planting and silviculture which is summarized in Table 5. The active involvement in forest management by IV1 and IV2 landowners may be explained by considering the landowner’s primary occupation, production requirements and goals for land management. The IV2 group included a large proportion of retirees (60%), whereas professionals not involved in agriculture were mainly included in the IV1 group (41.7%); farming was the most common professional category among IV0 owners (50.8%). According to the management behaviour and expenditure on forestry observed in Mari˜ na Oriental, we suggest that retirees, older landowners, and non-agrarian professionals who worked part-time on their properties invested in marginal land in order to maintain their productivity, as reported by Karppinen (1998),Gunter et al. (2001) and Marey-Pérez et al. (2004). 280 V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 Table 5 Mean annual rates of planting and silviculture (%), expenditure on planting and silviculture (D/ha per year), amounts awarded through public subsidies (D/ha per year), and size of productive forest holding (ha) per homogeneous INVEST subgroup. INVEST IV0 IV1 IV2 P-value No. of interviewed landowners 76 15 12 PLANT Non-investor 0.91 2.31 0.245 Planter investor 4.51 1.000 TREAT Non-investor 0.47 1.84 0.374 Silviculturalist investor 1.84 3.54 0.223 PEXP Non-investor 148.37 1.000 Planter investor 399.86 304.38 0.294 TEXP Non-investor 62.55 128.61 0.184 Silviculturalist investor 128.61 151.83 0.807 SUB Non-investor 0.66 9.95 0.361 Investor 26.21 1.000 SIZE Small investor 4.20 4.99 0.823 Large investor 7.50 4.99 0.148 In fact, the likelihood of converting former meadows into woodland was negatively correlated with the condition of the landowner as an active farmer (D=−0.233); more than 62% and 29% of the landowners who considered this type of land-use change within their holding were, respectively, retired farmers and non-agricultural professionals (2=4.667). Thus, significant differences in the likelihood of converting forestlands into meadows according to the INVEST group were observed (2= 4.807). As expected, none of the IV2 owners and 5.6% of IV1 owners carried out this type of conversion on their holding, in contrast to the 50% of the IV0 group who opted for this procedure. Because of their close association with the land, farmers tend to manage their property themselves and they generally have more time to dedicate to forestry (Zhang and Mehmood, 2001). In addition, spending more time working on the property may result in better forestry training, which would qualify landowners to manage the land and take a more active role in forest-related activities. In the Mari˜ na Oriental area, the annual number of personal working days devoted to forestry is positively correlated with the landowner’s training in forestry (D=0.209). All trained owners worked for more than 50 days per year on forestry in the holdings; in contrast, 58.3% of owners who did not have any knowledge of forestry spent less than 50 working days per year on the holding (2= 10.113). This pattern of personal forest management may explain the observed significant differences in the landowners’ forest training according to the INVEST group (2=7.642). Some 30.0% of IV2 landowners were trained in forestry, in comparison with 6.3% of IV0 landowners and none of the IV1 group. Annual investments in holding improvement increased slightly with the hired labour employed on the holding (= 0.308 at P< 0.01). Whereas 77.8% of IV0 landowners hired forestry workers for less than 10 professional days per year, 70% and 41.7% of IV1 and IV2 owners, respectively, hired forestry workers for between 11 and 100 days per year (2=30.016). According to the previous landowner profile, the landowner’s occupation outside the property may mean having less time available for working the land (Löyland et al., 1995; p. 226), and therefore, he/she is less likely to carry out forest practices him/herself. In fact, we observed significant differences in the landowner’s main occupation according to the annual amount of hired help obtained (2=36.601). Some 67% of retired and active farmers hired forestryworkers forless than5 days peryear, incomparison with 44.5%of non-agriculturallandowners who hired forestry workers for more than 50 days per year. Professional assistance in forestry proved to be an important factor in making forestry viable in the area, as suggested in other studies concerning V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 281 NIPF management (Löyland et al., 1995; Hardie and Parks, 1996; Zhang and Flick, 2001; Zhang and Mehmood, 2001). The annual investment in holding improvement and the land acquisition pattern were positively and weakly correlated (= 0.246 at P<0.05). Whereas 72.5% of IV1 and IV2 owners managed inherited and purchased holdings, inheritance was the sole pattern of land acquisition for 60.3% of IV0 owners (2= 9.757). This is probably associated with the landowner’ condition as a retired farmer. In Mari˜ na Oriental, retired farmers represented a specific group, with holdings that were mainly acquired throughpurchase and inheritance. Briefly, retired farmers are morelikely to carry out land transactions in order to improve and increase their agricultural productivity, and hence display a higher degree of land mobility (Marey-Pérez et al., 2004). Thus, the main occupation of owners varied significantly on the basis of the land acquisition pattern (2=22.191). While 59.1% of the retired farmers managed inherited and purchased land, 25.8% of the remaining occupational groups managed inherited and purchased land. The key role of public funding in forestry was supported considering that the annual investment in new forest buildings and infrastructures in the region increased slightly with the likelihood of applying for public subsidies and with the annual amount granted (=0.334 and 0.297, respectively, at P< 0.01). There were four IV1 and IV2 owners for every IV0 owner who applied for this type of subsidy (2= 10.808). With regard to the amounts granted per subsidy, the mean amount received by the IV2 group was significantly higher than that received by the IV0 group (H=8.112), as seen in Table 5. The main reason for the land management behaviour in the region may be attributed to previous harvestsand timber sales, i.e. to the owners’ interest in timber production.Thus, the annual investment in holding improvement increased slightly in relation to the unitary stumpage price from previous harvests (= 0.232 at P< 0.05). There were significant differences between the IV2 group and the IV0 and IV1 groups (H=4.774); the price at which the IV1 and IV2 owners sold the timber was almost twice the price obtained by the IV0 owners. The interest in timber production may also explain why the owner’s intention to extend the area of woodland within the holding increased slightly with the annual investment in holding improvement (= 0.234 at P<0.05). Over 50% and 80% of IV1 and IV2 owners, respectively, aimed to extend the area of woodland, compared with 38.1% of the IV0 group (2= 6.289). According to this finding, large amounts of money invested in forestry do not appear to reduce the intensity of management in the region, but favourable market conditions for timber production, particularly attractive timber prices, appeared to motivate involvement in land management by NIPF owners, as explained in the following sections. As expected, due to the relationship between the area of land and the intensity of land management (Kuuluvainen and Salo, 1991; Prestemon and Wear, 2000; Zhang and Mehmood, 2001; Arano et al., 2004; Potter-Witter, 2005; Arano and Munn, 2006), the annual investment in holding improvement decreased slightly with the degree of parcellation of the productive forestland (=−0.216 at P< 0.05), and increased slightly with the size of the productive forest holding (= 0.216 at P< 0.05). The number of plots per unit of productive forestland was greater than 3.4 for 55.5% of IV0 owners, and for 16.7% and 27.3% of IV1 and IV2 owners, respectively. This is probably associated with the agricultural occupation of the owner. As reported by Butler et al. (2004),Marey-Pérez et al. (2006) and Marey-Pérez and Rodríguez-Vicente (2008), the desire to improve and increase the current or former agricultural productivity may be one of the main predictors and reasons for dividing the land. With regard to the area of productive forest, the IV1 group also owned significantly larger landholdings than the IV0 landowners (H=6.844), as illustrated in Table 5. Expenditure on planting The annual expenditure on planting appeared to affect all three forest management practices analysed in the region. Annual rates of planting and silviculture increased greatly and moderately, respectively, with annual expenditure on planting (=0.835 and 0.571, respectively, at P< 0.01). In addition, the annual rate of harvesting woodland increased slightly in relation to annual expenditure on planting (= 0.354 at P< 0.01). Pairwise comparisons indicated that EP4 owners were significantly more active planters and silviculturalists than the other landowners in the area (H= 44.154 and 22.972, respectively). As regards timber harvesting, the EP3 owners were the most active harvesters in the 282 V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 Table 6 Mean annual rates of planting, silviculture and harvesting (%), expenditure on silviculture (D/ha per year), income from land sales (D/ha per year), and size of productive forest holding (ha) per homogeneous PEXP subgroup. PEXP EP0 EP1 EP2 EP3 EP4 P-value No. of interviewed landowners 5 61 25 7 5 PLANT Non-planter 0.00 0.53 1.74 0.150 Farm planter 0.53 2.25 1.74 0.159 New planter 10.42 1.000 TREAT Farm silviculturalist 0.04 0.41 1.03 0.80 0.935 New silviculturalist 9.23 1.000 HARV Non-harvester 1.53 2.34 4.17 0.914 Harvester 4.17 11.73 2.98 0.117 TEXP Retiree 0.00 32.40 0.977 Farmer 32.40 154.71 0.184 Absentee 154.71 259.47 265.26 0.273 NTINCOME Non-land seller 50.60 140.59 290.13 74.92 0.996 Land seller 404.22 1.000 SIZE Small landowner 2.29 3.67 6.60 3.79 0.108 Large landowner 6.60 11.11 0.083 region;there were significant differences in the mean rates of timber harvesting between the EP3 group and both the EP1 and EP2 groups (H=7.112). The homogeneous subgroups of landowners according to the annual expenditure on planting and annual rates of the three forest management practices are shown in Table 6. The amount designated annually to forest planting increased moderately in relation to the amount spent annually on holding improvements (= 0.474 at P< 0.01). The EP2 and EP4 owners invested significantly more money in new forest buildings and infrastructure than the remaining owners, on average three times more a year than the amount invested by groups EP0,E P1 and EP3 (H= 21.154). Furthermore, the annual expenditure on planting increased greatly according to the annual expenditure on silviculture (= 0.799 at P< 0.01). As shown in Table 6, annual expenditure on silviculture by EP1 owners differed significantly from the remaining groups, excluding group EP0; significant differences were also observed between EP0 owners and both EP3 and EP4 owners in terms of this variable (H= 50.134). As well as investments in holding improvements, high expenditure on forestry (planting and silviculture) did not appear to reduce the intensity of management in the region. We found that the owner’s primary occupation and the annual amount spent on planting forestlands were weakly and positively correlated (= 0.235 at P< 0.05). All active farmers in the study population were included in groups EP0,E P1 and EP2, because they were more likely to manage agrarian land, whereas all EP3 and EP4 owners were retired farmers and non-agricultural professionals (2= 30.986). This finding supports the idea that both retired farmers and professionals who did not earn their living from agriculture were more likely to invest in forestry in order to maintain the land as a capital asset. Moreover, the annual expenditure on the forest plantation increased slightly in relation to the annual household income (= 0.220 at P< 0.05), and there were significant differences between EP0 and EP2 owners in this respect (H= 9.038). The EP2 and EP4 groups received almost 8500 D/year more in household income than the remaining population; the lowest income per family unit corresponded to the EP0 group, with a mean income of 9700 D/year. These results indirectly support those obtained by other authors who have pointed out that high forest investments are made by those owners with a larger income per household (Hardie and Parks, 1996; Gunter et al., 2001; Mahapatra and Mitchell, 2001; Arano et al., 2004; Ross-Davis et al., 2005). V. Rodríguez-Vicente, M.F. Marey-Pérez / Journal of Forest Economics 16 (2010) 269–295 283 Involvement in agriculture may be the main reason why the annual fraction of forest products for own consumption differed significantly according to annual expenditure on planting forestlands. The annualamount of forest products for household consumption by EP4 owners differed significantly from that observed for EP1 and EP2 groups (H=4.123). The EP0,E P1 and EP2 owners actively benefitted from forests for their own use, a mean benefit of 76.2 D/ha per year, which was more than two and eight times higher than the mean value obtained for EP3 and EP4 landowners, respectively. As described by Dewees (1992),Kurttila et al. (2001) and Marey-Pérez et al. (2004), although some owners are less dependent on forestry because of a higher proportion of other incomes, farmers may become more dependent on forestry as a source of revenue because of lower benefits obtained from agricultural activities. In Mari˜ na Oriental, the annual amount of forest products for own consumption was slightly higher for landowners actively involved agriculture than for other owners (= 0.304 at P< 0.01). Active farmers annually took advantage of forest products twice as often as the remaining occupational groups (H= 7.840). The annual income from land sales and the annual expenditure on planting were weakly and positively correlated (=0.266 at P< 0.05). According to the land management behaviour observed in the area, the landowners could sometimes sell forestland that they were unable to manage and could investin forestry as the remaining part ofthe holding. As seen in Table6,the annual non-timber income of EP3 owners was significantly higher than the corresponding income of the remaining landowner population (H=11.737). The likelihood of marketing forestland may explain the small positive relationship between the pattern of land acquisition and the annual expenditure on planting (= 0.224 at P< 0.05). Whereas 79.2% of EP3 and EP4 owners managed holdings acquired through inheritance and purchase, 43.3% of the remaining landowner groups had holdings acquired through inheritance alone (2= 13.780). Considering the owner profile of the EP3 group, improving and increasing agricultural productivity increased land mobility among retired farmers, as already explained. The owner’s primary occupation may also explain why the annual number of working days dedicated to forestry by both the owners themselves and by hired labour increased slightly and moderately, respectively, with annual expenditure on planting (= 0.225 at P< 0.05 and 0.402, at P< 0.01, respectively). Forestry is generally less feasible for managers with occupations outside the property, which obliges them to hire forestry workers, whereas owners involved in agriculture are more likely to devote time to forestry, probably because of a greater attachment to the land. All EP0 owners spent less than 5 working days on forestry annually, while 44.5% of EP1 and EP2 owners and 66.7% of EP3 owners devoted between 11 and 100 working days annually; in contrast, 75% of EP4 owners dedicated less than 10 working days to forestry per year. With regard to hired labour, none of the EP0 owners hired forestry workers for more than 2 days per year, in comparison with all owners in the EP4 group who hired forestry workers for more than 50 days a year; 82% of EP1 owners and 48.8% of EP2 and EP3 owners hired labour for less than 10 days a year (2= 42.573). Family aid was also important with regard to the way in which the forestland was managed. Thus, the annual number of working days that the family spent on forestry differed significantly according to the annual expenditure on planting (2= 37.273). All EP0 owners and 65.4% of EP1 and EP2 owners benefitted from a family labour-force working in forestry for less than 10 days per year, while 83.4% of EP3 owners annually benefitted from their families working on the holding for between 11 and 100 days; half of the EP4 owners had family assistance for less than 5 working days per year and the other half part-benefitted from a family labour-force working for more than 100 days per year. This may be explained by taking into account that the input of family labour on the holding depended to a certain extent on the availability of equipment, and moreover, that the annual inputs of personal and family labour were interrelated. Thus, for each owner who benefited from a family labour force for less than 5 days a year and had machinery on the holding, there were two owners who had these resources as support for forest management in the groups who benefitted from more than 50 annual days input from family labour-force (x2=11.158). In relation to the second item, the relatives of owners who devoted between 11 and 100 days per year to working on the holding, worked for 11–50 labour-days on forestry per year, while owners who spent less than 5 personal days on forestry per year benefitted from fewer than 11 family labour-days per year. Owners who worked on the holding for between 6 and 10 labour-days or more than 100 days per year, benefited from family input that was slightly less than the input of labour by the owner (2= 101.468). [Document text truncated for crawler view.]