Identifying economic hurdles to early adoption of preventative practices: The case of trunk diseases in California winegrape vineyards

Kaplan, Jonathan D. et al. Article Identifying economic hurdles to early adoption of preventative practices: The case of trunk diseases in California winegrape vineyards Wine Economics and Policy Provided in Cooperation with: UniCeSV - Centro Universitario di Ricerca per lo Sviluppo Competitivo del Settore Vitivinicolo, University of Florence Suggested Citation: Kaplan, Jonathan D. et al. (2016) : Identifying economic hurdles to early adoption of preventative practices: The case of trunk diseases in California winegrape vineyards, Wine Economics and Policy, ISSN 2212-9774, Elsevier, Amsterdam, Vol. 5, Iss. 2, pp. 127-141, https://doi.org/10.1016/j.wep.2016.11.001 This Version is available at: https://hdl.handle.net/10419/194522 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ HOSTED BY Available online at www.sciencedirect.com Wine Economics and Policy 5 (2016) 127–141 ?? Identifying economic hurdles to early adoption of preventative practices: The case of trunk diseases in California winegrape vineyards $ Jonathan Kaplan a, n , Renaud Travadon b , Monica Cooper c , Vicken Hillis d , Mark Lubell d , Kendra Baumgartner b a Department of Economics, California State University Sacramento, 6000 J Street, Sacramento, CA 95819, USA b United States Department of Agriculture- Agricultural Research Service, Davis, CA 95616, USA c University of California Cooperative Extension, Napa, CA 94559, USA d Department of Environmental Science & Policy, University of California, One Shields Avenue, Davis, CA 95616, USA Received 31 March 2016; received in revised form 11 November 2016; accepted 18 November 2016 Available online 25 November 2016 Abstract Despite the high likelihood of infection and substantial yield losses from trunk diseases, many California practitioners wait to adopt fieldtested, preventative practices (delayed pruning, double pruning, and application of pruning-wound protectants) until after disease symptoms appear in the vineyard at around 10 years old. We evaluate net benefits from adoption of these practices before symptoms appear in young Cabernet Sauvignon vineyards and after they become apparent in mature vineyards to identify economic hurdles to early adoption. We simulate winegrape production in select counties of California and find widespread benefits from early adoption, increasing vineyard profitable lifespans, in some cases, by close to 50%. However, hurdles may result from uncertainty about the cost and returns from adoption, labor constraints, long time lags in benefits from early adoption, growers’perceived probabilities of infection, and their discount rate. Development of extension resources communicating benefits and potential hurdles to growers likely reduces uncertainty, increasing early adoption. Improvements in efficacy of preventative practices, perhaps by detecting when pathogen spores are released into the vineyard, will increase early adoption. Lastly, practice cost reductions will increase early adoption too, especially when the time it takes for adoption to payoff and infection uncertainty are influential in adoption decisions. &2016 UniCeSV, University of Florence. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Keywords: Grapevine trunk diseases; Early adoption; Plant-disease management; Preventative practices 1. Introduction Vineyards suffer from damaging wood diseases, which present serious challenges to grape production in every grape-growing region of the world (Bertsch et al., 2013). These diseases, collectively referred to as “trunk diseases” include, among others, Botryosphaeria dieback, Esca and Petri diseases, Eutypa dieback, and Phomopsis dieback. In California, which accounts for approximately 90% of US winegrape production (USDA, 2015), yield losses in susceptible cultivars can reach over 80% in mature vines, during what should be the peak years of production (Munkvold et al., 1994). Siebert (2001) estimated that California winegrape production would generate 14% greater annual gross producer value in the absence of Eutypa dieback. www.elsevier.com/locate/wep http://dx.doi.org/10.1016/j.wep.2016.11.001 2212-9774/&2016 UniCeSV, University of Florence. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ☆ This research was supported in part by a grant from USDA SCRI. Peer Review under the responsibility of UniCeSV, University of Florence. n Corresponding author. Fax: þ1 916 278 5768. E-mail addresses: [email protected] (J. Kaplan), [email protected] (R. Travadon), [email protected] (M. Cooper), [email protected] (V. Hillis), [email protected] (M. Lubell), [email protected] (K. Baumgartner). For all of these diseases, the causal agents are fungi that establish chronic infections of the wood, which represent mixtures of different trunk pathogens (Bruez et al., 2016); rarely is one trunk disease present in a vineyard. Infection occurs primarily through pruning wounds, which are made every dormant season when vines are pruned, starting in year 3 as part of the normal production practices in the vineyard. To minimize such pruning-wound infections by the fungal spores, preventative practices have been developed and are used by practitioners: (i) delaying pruning until late in the dormant season, when the risk of infection is low (Petzoldt et al., 1981), (ii) double pruning, a modified version of delayed pruning using a mechanical pruning machine to nonselectively trim canes to a uniform height during a first pass in early winter, followed with a second hand-pruning pass in late winter to remove wood infected after the first pass and adjust to traditional 2-bud spurs (Weber et al., 2007), and (iii) applying fungicides to fresh pruning wounds as a protective barrier (Amponsah et al., 2012;Halleen et al., 2010;Pitt et al., 2012; Rolshausen and Gubler, 2005;Rolshausen et al., 2010; Sosnowski et al., 2008,2013). As these practices are preventative in nature, they must be used before vines are infected to ensure optimal efficacy. 1 Pest-control advisers (PCAs) working in grape production systems acknowledge the widespread nature of trunk diseases in California vineyards and their impact on yields (Hillis et al., 2015). Nonetheless, PCAs have a greater tendency to recommend preventative practices in vineyards where vines with symptoms are widespread, which is typically when the vineyard is 10 or more years old (Duthie et al., 1991). 2 However, by definition, the benefits of prevention are minimal when the vines are already infected. This habit of recommending preventative practices in mature, diseased vineyards can be explained in part by the fact that trunk diseases are not typically apparent until a vineyard is 8 years old or older; infections occur when the vineyard is young, but symptoms take several years to appear. By year 10, approximately 20% of vines present symptoms (Duthie et al., 1991) and up to this point, yield losses are relatively minor (Munkvold et al., 1994). Recommendation of preventative practices in diseased vineyards by PCAs may also be explained by a gap in the research. Although preventative practices have been tested by researchers in many short-term experimental trials (Rolshausen et al., 2010;Urbez-Torres and Gubler, 2011;van Niekerk et al., 2011;Weber et al., 2007), their long-term efficacy has been the subject of far fewer studies (Gu et al., 2005). Practitioners may thus be hesitant to adopt preventative practices in younger vineyards because improvements to yields and net returns have not been quantified. Reluctance to adoption early may also stem from how long it takes for the symptoms to appear in the vineyard. The relatively long time it takes for the disease to grow in a vineyard implies annual benefits from early adoption will take many years to accumulate and offset the annual additional cost of the practice, which is incurred immediately. Our work addresses the economic factors that may result in a delay to adopt preventative practices in young vineyards by providing a more transparent description of the costs and benefits. We simulate winegrape production for representative vineyards in five of California's winegrape growing counties, which are aligned with Grape Pricing Districts or ‘crush districts’as follows: Napa (Crush District 4), San Joaquin (Crush District 11), Central Coast (Crush District 8), Lake (Crush District 2), and Sonoma (Crush District 3). 3 Our parameters include disease-control efficacies from published experimental field trials and vineyard practice costs from economic budgets for producing Cabernet Sauvignon, one of the most widely-planted winegrape cultivars in California. Cabernet Sauvignon is not known to be the most susceptible cultivar to any of the trunk diseases (Travadon et al., 2013), but we use it as an example of winegrape production because it has similarly large production acreage in all five counties. Also, it is the cultivar most widely considered in the published University of California Cooperative Extension (UCCE) Cost & Return studies (UCCE, 2004–2014), which form the basis for the economic analysis. We derive annual net returns for a healthy vineyard and an infected vineyard in which preventative practices are adopted in years 3, 5, or 10. These ages were selected to evaluate conditions when vines are fully trained onto the trellis system and winter pruning begins (3 years old), when vines reach maturity (5 years old), and when trunk disease symptoms typically appear in vineyards (10 years old). In this way, we quantify the cumulative yield and revenue gains or losses due to adopting in young vineyards rather than waiting until year 10 to do so. 2. Background The research described in this paper adds to the literature on adoption of disease-prevention practices. Past research on adoption of agricultural technology and innovation has primarily analyzed annual crops (Alston et al., 2010). More recent work on perennial crops, wine grapes in particular, has considered managing Pierce's disease (Alston et al., 2013, 2014;Tumber et al., 2014), powdery mildew (Fuller et al., 2014;Lybbert and Gubler, 2008), and grapevine leafroll disease (Atallah et al., 2014;Fuller et al., 2015;Ricketts et al., 2015). Siebert (2001) provided insight into the economic impact of Eutypa dieback to California's wine grape industry. Sipiora and Cuellar (2014) examined farm-level impacts of 1 Although, in general preventative practices prevent secondary infections in an infected vineyard and re-infection in a vineyard where vine surgery is performed, here it refers to preventing prevent infection of new pruning wounds on vines with infections that started at pruning wounds in the past given vines are pruned every dormant season; each year there is a new set of wounds. 2 Recommendations to use preventative practices after the disease is meant to reduce secondary infections. 3 Winegrape growing areas within California have been delineated in a variety of ways. In addition to counties or crush district, others (such as Alston et al. (2015)) have delineated regions based on variety and value. Table 4 shows these regions as well. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141128 preventative practices against Eutypa dieback in a Napa vineyard on annual yields and net present value. Our economic analysis contributes to this literature by providing the first study, to our knowledge, to evaluate economic hurdles to adopting preventative practices in young versus mature vineyards. 2.1. Economic simulation model We develop simulation scenarios that consider future management costs and benefits (i.e., amelioration of cumulative yield losses by adopting preventative practices) based on past observations (i.e., increasing disease incidence and the associated yield losses over time), similar to other recent research on grapevine diseases (Alston et al., 2013,2014; Fuller et al., 2015,2014;Ricketts et al., 2015), given that field experiments would take decades to complete. Like these past studies, we establish baseline conditions and scenarios to capture the dynamics of trunk disease infections and net returns in the different winegrape districts using information on currently available practices, their costs, and effects on yields and lifespan taken from the UCCE Cost & Return Studies, historical data from the California Department of Food and Agriculture (CDFA) and the United States Department of Agriculture, National Agricultural Statistics Service (USDA– NASS), the scientific literature on plant pathology and the efficacy of pruning practices, and interviews with winegrape growers, farm advisors, and other stakeholders. Our approach to modeling the economics of trunk diseases requires a different framework, however, given that trunk diseases may not have measurable impacts on yield until many years after infection. Our model captures time-varying yield and practice costs through adopting preventative practices (Table 1)in young vineyards, relative to adopting in year 10 when symptomatic vines are visible. We examine changes in returns and costs to the grower over a 25-year vineyard lifespan, holding all other factors constant except practice costs and yield losses from adopting preventative practices in young vineyards relative to returns and costs from adopting when a vineyard is 10 years old. In this way, this model allows us to compare long-run average outcomes without incorporating unknown and unpredictable future events, and alleviate the inherent challenges in modeling current and future expectations. An important factor in studying winegrapes in California is the regional variation in yield and price per ton (see Table 2 for parameter value details). For example, at one extreme in Napa and Sonoma Counties (Crush Districts 4 and 3, respectively), establishment decisions and management practices restrict vineyard yields (approximately 4.5 to 5 t of Cabernet Sauvignon per acre in mature vineyards) with the goal of achieving higher wine quality that sells at a high average price per ton ($2,355 and $5,192 for Sonoma and Napa, respectively). At the other extreme, in San Joaquin county (Crush District 11), fruit prices are much lower ($650 per ton) and vineyards produce higher yields (10 t per acre (CDFA/NASS, 2015). The other counties face prices and yields within these two extremes. 2.2. Disease-control efficacy of preventative practices Our survey of the scientific literature on preventative practices provided a range of disease-control efficacies (DCEs), which were calculated from multiple experimental trials on different trunk diseases (Table 3). DCE is the proportion of pruning wounds which do not become infected as a result of a preventative practice but would otherwise become infected. In the empirical analysis, we use DCEs of 25, 50, and 75%, which reflect the range of natural variation across study years [e.g., DCEs ranging from 29 to 88% for delayed pruning against Phaeoacremonium minimum (Larignon and Dubos, 2000)] or across pathogens [e.g., DCE of 52% for Topsin M against Phaeomoniella chlamydospora versus DCE of 80% against Lasiodiplodia sp. (Rolshausen et al., 2010)]. The high extreme of our range in DCEs is truncated at 75% to reflect that all infections may not arise through pruning wounds. For example, planting material may be infected in the nursery (Gramaje and Armengol, 2011) and, thus, it is unrealistic to assume a practice can prevent 100% of infections. Table 1 Description of preventative practices for management of grapevine trunk diseases. Practices Description Delayed Pruning Prune late in the dormant season (February or later, before budbreak) by hand, when both pathogen inoculum and wound susceptibility are lower, hence minimizing the risk of infection compared to December and January. Double Pruning Prune early in the dormant season (December or January) with a mechanical-pruning machine; partially prune canes to a length of approx. 0.4 m. Prune again late in the dormant season (February or later) by hand to twobud spurs, to remove potentially infected canes. Topsin M Topsin M is a fungicide that provides a protective barrier on pruning wounds against infection by the spores of trunk pathogens. After pruning and before rain, the latter of which induces spore production, liberation and dispersal, apply Topsin M by hand with a paintbrush or sponge to cover pruning wounds. a a Protectants registered for hand application during the dormant season in California are Thiophanate-methyl (Topsin M WSB; United Phosphorus, Inc., King of Prussia, Pennsylvania), Boric acid (Tech-Gro B-Lock; Nutrient Technologies, Inc., Dinuba, California), and VitiSeal (VitiSeal International LLC, San Diego, California). Topsin M is also registered for spray application. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141 129 The experimental trials on preventative practices are fragmented. They were conducted by different labs, on different cultivars, in different regions, and in different years. All trials involved controlled inoculations, which ensured that the pruning wounds were ‘challenged’by individual species of trunk pathogens and, thus, the practice efficacy in preventing infection was tested. Nonetheless, trunk diseases occur in mixed infections in the vineyard, where individual vines are often infected by multiple trunk pathogens, which attack vines through different pruning wounds in different years. Cultivar susceptibility is not consistent across trunk pathogens, based on the few studies that have been done [e.g., Travadon et al., 2013]. 3. Methods: bioeconomic model We develop a representative farm mathematical program to simulate the dynamic economic decision making involved when investing in perennial crops, such as winegrapes. The perennial nature of the crop, its relatively long life-expectancy (on the order of decades), and the multi-year delay between infection and symptom expression suggest a dynamic model is more appropriate than a static model. A dynamic model allows us to capture the effects of decisions made today and in the future on investments in preventative practices in vineyards. Table 2 Parameters used in simulated economic analysis. Additional cost/acre (in 2013 dollars) for preventive practices relative to industry winter pruning standard by county. a Source: UCCE cost and return studies and semi-structured interviews with growers, viticulture advisors, and other stakeholders in California winegrape production. Practice Delayed Pruning Topsin M Double Pruning Napa (4) $0 $71 $478 San Joaquin (11) $0 $90 $243 San Luis Obispo (8) $0 $117 $268 Lake (2) $0 $90 $279 Sonoma (3) $0 $74 $335 Tons/acre of Cabernet Sauvignon by age and county. Source: UCCE Cost and Return Studies. Vineyard Age Year 0 Year 1 Year 2 Year 3 Year 4 Year 5þ Napa (4) 0 0 1 4.5 4.5 4.5 San Joaquin (11) 0 0 5 10 10 10 San Luis Obispo (8) 0 0 2.5 5 7.5 7.5 Lake (2) 0 0 0.75 1.5 3.5 5.75 Sonoma (3) 0 0 1.5 3 5 5 Total cash costs/acre (2013 dollars) by age and county. Source: UCCE Cost and Return Studies Year 0 Year 1 Year 2 Year 3 Year 4 Year 5þ Napa (4) $32,303 $5,264 $5,304 $7,784 $7,784 $7,784 San Joaquin (11) $12,213 $3,370 $3,395 $3,505 $3,505 $3,505 San Luis Obispo (8) $9,998 $2,554 $3,501 $4,625 $4,625 $4,625 Lake (2) $7,301 $6,942 $3,252 $3,404 $4,053 $4,053 Sonoma (3) $26,780 $4,204 $5,186 $6,280 $6,280 $6,280 Five-year weighted average price/ton for Cabernet Sauvignon (2013 dollars) Source: USDA/NASS Annual Crush District Reports (USDA-NASS 2015) Napa (4) $5,192 San Joaquin (11) $650 San Luis Obispo (8) $1,262 Lake (2) $1,623 Sonoma (3) $2,355 Remaining Parameter Values Symbol Value Ref. Carry Capacity K0.92 Duthie et al. (1991) Initial Percentage of Symptomatic Vines Y 0 0.001 Duthie et al. (1991) Constant of Integration B 0 ¼(KY 0 )/Y 0 919 Duthie et al. (1991) Simulation Constants of Integration (B age ) B 3 305.9085 Derived from Eq. (1) B 5 58.7497 Derived from Eq. (1) B 10 7.46919 Derived from Eq. (1) Trunk Disease Growth Rate g 0 0.55 Duthie et al. (1991) Real Discount Rate δ0.03 Fuller et al. (2014) Fuller et al. (2015) a All dollar amounts adjusted to 2013 dollars using the United States Department of Commerce, Bureau of Economic Analysis implicit GDP deflator (USDC-BEA 2015). Table 3 Disease control efficacies (DCEs; % pruning wounds protected) for preventative practices against three trunk diseases and six trunk pathogens. Trunk disease Preventative practices a Trunk pathogen Topsin M Delayed pruning Double pruning DCE (% pruning wounds protected) Botryosphaeria dieback Lasiodiplodia sp. 80% b 59–75% d – Neofusicoccum luteum 60% c –– Neofusicoccum parvum –55–79% d – Esca Phaeoacremonium minimum 57% b 29–88% e – Phaeomoniella chlamydospora 52% b 40–58% e – Eutypa dieback Eutypa lata 100% b 90% f 33–85% g a For Topsin M, values are calculated as a reduction in pathogen recovery from treated-inoculated pruning wounds, relative to that of nontreatedinoculated pruning wounds. For delayed pruning and double pruning, values are calculated as a reduction in pathogen recovery from late-winter pruning wounds, relative to that of early-winter pruning wounds. Ranges reflect data from replicated studies in the same vineyard across two years. b When applied to Chardonnay in 2005 and Zinfandel in 2006; averaged across both cultivars/years (Rolshausen et al., 2010). c When applied to Chardonnay (Amponsah et al., 2012). d When pruning Cabernet Sauvignon and Chardonnay in March versus December, in 2007 and 2008 (Urbez-Torres and Gubler, 2011). e When pruning Cabernet Sauvignon in March versus January, in 1997 and 1998 (Larignon and Dubos, 2000). f When pruning Grenache in March versus December (Petzoldt et al., 1981). g When pruning Chardonnay and Merlot in February versus December, in 2001 and 2002 (Weber et al., 2007). J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141130 Although, productivity is theoretically stable after a vineyard matures, symptoms of trunk diseases are not apparent until vines mature, and they worsen over time because the infections are chronic. With a dynamic model, we can capture the effects of these diseases on time-varying yield per acre and of currently available preventative practices adopted at different vineyard ages. We are then able to compare early adoption scenarios with that of year 10, and measure the changes in costs and returns not just today, but in the future as well. 3.1. Biological model To approximate the spread of trunk diseases and corresponding yield effects throughout the vineyard, we adopt the trunk disease logistic growth model estimated by Duthie et al. (1991) using test plots of Chenin blanc and Barbera varietals grown throughout Merced County, California. The yield loss function comes from Munkvold et al. (1994), who derived yield losses due to the combined effect of Eutypa and Botryosphaeria Diebacks from the same test plots as used in Duthie et al. (1991). Although we do not explicit model the infectious disease in a susceptible, infected, and recovered model [see Atallah et al. (2014), for example], we rely on the plant pathology and infectious disease literature which provides numerous empirical studies using the logistic function to capture the spread of infection. 4 Following the estimates in Duthie et al. (1991), it is assumed that 92% of vines are susceptible when first planted, declining over time as the percentage of infected vines grows. In addition, because trunk diseases go many years before detection, removal of infected vines when pruning slows the spread of the infection but cannot eradicate it. We apply this relationship to Cabernet Sauvignon grown in the sample regions across all trunk diseases, following discussions with growers, managers, and farm advisors on their experiences with trunk diseases. Mathematically, disease incidence grows over time according to Yt¼K ð1þB0eg0tÞ¼0:92 1þ919e0:55t ðÞ ð1Þ where Ytis the percentage of symptomatic vines per acre, Kis the carrying capacity, tis the age of the vineyard, B 0 is the constant of integration and equals (KY 0 )/Y 0 , where Y 0 is the initial percentage of symptomatic vines is set at 0.001. Lastly, g 0 is the growth rate. Fig. 1 shows this growth over the 25-year lifespan evaluated in the empirical analysis. 5 Growth is negligible over the early years with a little over 1.5% of vines presenting symptoms by the time a vineyard is 5 years old. The rate accelerates rapidly shortly thereafter with 7.5% of the vines having symptoms by year 8, nearly 20% by year 10, and 75% by year 15. This growth rate estimated by Duthie et al. (1991) represents the average scenario, in a vineyard where disease incidence increases rapidly due to a variety of factors (e.g., high susceptibility of the grape cultivar, optimal climate conditions for infection, absence of management practices against trunk diseases), the impacts of which have not been quantified. This increase in disease incidence translates into yield reductions based on Munkvold et al. (1994) as follows YieldI t¼ð100:198:81YtÞYieldH tð2Þ where YieldH t and YieldI tare annual tons per acre produced by a healthy and an infected-untreated vineyard, respectively. This function takes into account that vines may compensate for lost fruiting positions, toxins from trunk pathogens may affect apparently healthy shoots, and in more severe cases, symptomatic vines may produce less photosynthate, thereby negatively affecting yield. When preventative practices are adopted, there are fewer symptomatic vines over time, lowering the reduction of yields throughout the 25-year lifespan of a vineyard. Fig. 2 illustrates reduction in yields as disease incidence increases for one of the winegrape growing counties. 6 How preventative practices affect this relationship is discussed below. Yield per acre values for the different counties used in the empirical analysis are contained in Table 2 above. 3.2. Economic model When deciding whether to adopt one practice over another, a grower may weigh the cumulative expected present value of annual net returns over a 25-year vineyard lifespan across the possibilities based on their perceived risk of infection. Annual Fig. 1. Trunk Disease incidence (in % symptomatic vines/acre by the age of the vineyard (Duthie et al. 1991). Note: Duthie et al. (1991) measured Eutypa dieback symptoms and dead spur positions, the latter of which is now known as a general symptom of three trunk diseases (Botyrosphaeria dieback, Eutypa dieback, and Phomopsis dieback). 4 Numerous examples from the literature use or suggest the use of a logistic growth function to capture the disease growth (e.g., Madden et al., 2000; Murphy et al., 2016) 5 This lifespan is consistent with California winegrape production as reported in the UCCE Cost and Return Studies. 6 Figures showing the effect of trunk diseases in other regions are available on request. Note, the percentage change in yield is the same for all counties, however, different counties yield different tons per acre and thus we will see different absolute reductions but not relative reductions. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141 131 net returns per acre (NR) are defined as NRtðA;c;dceÞ¼PricetxYieldtðA;dceÞ Cos ttðA;cÞð3Þ where Adenotes the age when adoption occurs, cthe annual preventative practice cost, dce the DCE, and tthe age of the vineyard. Fig. 3 shows streams of net returns in 2013 dollars over a 25-year vineyard lifespan for the San Joaquin County. A grower with a healthy vineyard versus one with an infectedvineyard that adopts preventative practices in year 10 with 50% DCE can expect to make $33,019 per acre instead of between $336 and $2,004, depending on the different practice costs, respectively, over this time. A grower is likely to replace or abandon the vineyard before the 25th year is reached if annual returns are negative. However, we extend production out to 25 years so we can compare across similar lifespans and evaluate years of lost profits for a given initial investment. The cumulative discounted stream of net returns (PVNR) or simply net benefits (NB) across the scenarios are NBðA;c;dce;δÞ¼ X 25 t¼0 NRtðUÞ ð1þδÞtð4Þ Table 4 shows NB per acre when the real discount rate (δ)is assumed to be 3%, 7 for a healthy vineyard and an infecteduntreated vineyard, across the five counties examined. Clearly taking no action to prevent trunk diseases results in significant economic losses. The greatest potential losses are in Napa, reaching over $160,000 per acre. As noted in Hillis et al. (2015), many growers adopt preventative practices once trunk disease is apparent. PCAs also tend to recommend these preventative practices more often in vineyards with a greater percentage of symptomatic vines (Hillis et al., 2015). As such, $160,000 per acre is an upper bound on potential losses over the 25 years. Other growers, alternatively, replant an infected vineyard or use vine surgery [physically cutting out infected wood and retraining a new cordon or a new vine from the trunk (Sosnowski et al., 2011)] to treat symptomatic vines (and hopefully restore yields) before the 25 years have passed. This latter approach to managing trunk diseases can be prohibitively costly and not guaranteed to restore yields as the replanted vine or retrained sucker may be infected, or the remaining vine may not produce suckers. Given the scope of this analysis is to understand why growers do not adopt preventative practices in young vineyards before symptoms are apparent, we leave the evaluation of vine surgery and replanting for future analysis. The expected net benefits (E½NB) for each scenario are then E½NB A;c;dce;δðÞ¼1πðÞNBHðA;c;dce;δÞ þπNBIðA;c;dce;δÞð5Þ where the superscripts denote healthy or infected vineyards, respectively, and πis the grower's perceived probability of infection. Fig. 4 shows E½NBðÞ for a known preventative practice adopted in a 10-year old vineyard (A 10 ) and a 5 years old vineyard (A 5 ), for different perceived probabilities of Fig. 2. Effect of Trunk Diseases on yield by vineyard age in a Cabernet Sauvignon vineyard in San Joaquin County. Note: Effect of infection on yield from Munkvold et al. (1994). Yield data come from UCCE (2012). Fig. 3. Cumulative undiscounted net returns (Total RevenueTotal Cost) per acre for healthy versus infected-untreated Cabernet Sauvignon vineyards in San Joaquin County (in 2013 dollars). Table 4 Cumulative discounted net benefits (NB) per acre (in 2013 dollars) for healthy and infected-untreated vineyards over a 25-year lifespan, by county (crush district number). County (Crush District number) Region a Healthy vineyard Infected-untreated vineyard NB per acre Napa (4) North Coast $203,982 $42,271 San Joaquin (11) North Central Valley $33,019 $11,957 San Luis Obispo (8) Central Coast $59,372 $6,144 Lake (2) Other California $40,375 $4,601 Sonoma (3) North Coast $49,496 $31,975 a These winegrape-growing regions are based on value and variety as delineated in Alston et al. (2015). 7 Fuller et al. (2014, 2015) also assume a 3% real discount rate. However, the literature provides a range of values for the discount rate. Some as high as 5.75%. In our analysis we considered other discount rates and found no qualitative change in results. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141132 infection (π). This model can provide both prescriptive and predictive information. With this model we can see how a grower will respond to changes in model parameters (A;c;dce;δ) given their knowledge of costs and returns, and perception of disease infection. This model also provides an infection probability threshold (π 0 ) that shows, given a grower's knowledge of their costs and benefits, whether it is better to adopt early or to wait until symptoms appear. In this framework, a grower maximizes his or her wellbeing by selecting the scenario with the greatest ENBðÞ ½ . The intersection of these lines, at π 0 , divides the population of growers with varying perceptions of the probability of infection. In general, if a grower who knows the cost and benefits from adopting these preventative practices, has a perceived probability of infection less than π 0 , they would be expected to wait until year 10 to adopt a practice, those with a perceived probability of infection greater than π 0 would adopt in year 5, and those at π 0 would be indifferent. Over time, grower perceptions of the probability of infection will likely increase as a result of experiential or scientific evidence, extension services, or networking, and thus a greater share of growers would be expected to adopt in the future as well. We derive a general expression for π1that divides adopters and nonadopters by equating the expected net benefits from adopting in year 10 E½NB A10;c;dce;δðÞ¼1πðÞNBHðA10;c;dce;δÞ þπNBIðA10;c;dce;δÞ;ð6Þ with the expected net benefits from adopting a practice in year y, which is earlier than year 10 E½NB Ay;c;dce;δ  ¼ 1πðÞNBHAy;c;dce;δ  þπNBIAy;c;dce;δ  ð7Þ and given the assumption that adoption of a preventative practice does not affect yields in a healthy vineyard we rewrite Eq. (7) as E½NB Ay;c;dce;δ  ¼ 1πðÞNBHNA10;δðÞCA y;c;δ  þπNBIAy;c;dce;δ  ð8Þ where CA;c;δðÞare the cumulative discounted practice costs over the additional years of adoption, which increases with decreases in Aand δ, and increases in c, while NBðÞ increases with increases in dce and decreases in A, c, and δ. 8 Solving for π0produces the general expression π0ðAy;c;dce;δÞ¼ CA y;c;δ  NBIAy;c;dce;δ  NBIA10;c;dce;δðÞþCA y;c;δ  ð9Þ that varies with changes in the age of the vineyard when early adoption occurs (A y ), practice cost (c), disease control efficacy (dce), or the discount rate (δ). Evaluating the comparative statics with respect to these factors shows: 1) when vineyard age at time of adoption changes, the change in π0and the proportion of adopters is ambiguous, suggesting that adopting earliest may not be optimal; 2) when practice costs increase,π0 increases, reducing the share of adopters; 3) when dce increases, the change in π0is ambiguous; and 4) when δ changes, the change in π0is also ambiguous. To see this, we first take the derivative of the equilibrium condition with respect to A, yielding 9 ∂πAy  ∂Ay ¼ ∂CA y ðÞ ∂AyNBIAy  NBIA10 ðÞþCA y  CA y  ∂NBIAy ðÞ ∂Ayþ∂CA y ðÞ ∂Ay hi ½NBIAy  NBIAy10  þCA y  2 ð10Þ The two products in the numerator are both positive, while the term in the denominator is positive. As such, to infer the conditions for the direction of this change, we set the numerator less than zero and solve ∂CA y  ∂Ay NBIAy  NBIA10 ðÞþCA y  CA y  ∂NBIAy  ∂Ay þ ∂CA y  ∂Ay  o0ð11Þ Fig. 4. Expected net benefits for infected-untreated scenario and representative adoption scenario as a function of grower perception of disease risk. The probability π1separates growers into adopters with perceived probabilities of infection greater than π1and non-adopters with perceived probabilities of infection less than π1. 8 Note, C(.) appears in the first term on the right-hand side of Eq. (8) but not the second as it is part of NBIA;c;dce;δðÞ:We do so to provide a simpler depiction of the infection probability threshold and easier interpretation of the comparative static results. 9 Some subscripts are removed to simplify presentation. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141 133 Rearranging terms yields, ∂CA y ðÞ ∂Ay CA y  o ∂NBIAy ðÞ ∂Ay hi NBIAy  NBIðA10Þ  ð12Þ When the percentage increase in the cost of the practice (given it is adopted sooner rather than later) when the vineyard is healthy is less than (greater than) the percentage increase in the net benefits from adopting earlier when it is infected, then the threshold will fall (rise). This shows theoretically that a grower acting in their best interest may not adopt at the earliest possible vineyard age. That is, a practice that has greater overall economic benefits in an infected vineyard when adopted early may not be adopted early by some growers because the expected relative gains in an infected vineyard from adoption are not enough to compensate them for the expected relative cost they face if the vineyard is healthy. Not surprisingly, when we evaluate a change in the practice cost, c, ∂πcðÞ ∂c¼ ∂CA y;c ðÞ ∂cNBIAy;c  NBIA10;cðÞþCA y;c  CA y;c  ∂NBIAy;c ðÞ ∂c hi ½NBIAy;c  NBIA10 ðÞþCA y;c  240 ð13Þ we see the potential share of growers who adopt falls as the threshold moves outward from zero, given the first term in the numerator of Eq. (13) is positive because the change in the overall cost increases with a change in the practice cost and taking action results in greater net benefits, and the second term in the numerator is negative because an increase in the practice cost decreases the net benefits from adoption in an infected vineyard. If DCE were to increase, the change in π 0 is ambiguous as both terms inside the brackets in the numerator of Eq. (14) are positive. ∂πAðÞ ∂dce ¼ CA;cðÞ ∂NBIAy;c;dce ðÞ ∂dce ∂NBIA10;c;dceðÞ ∂dce hi hNBIAy;c  NBIA10 ðÞþCA y;c  i2ð14Þ An increase in DCE results in an increase in π1(and a decrease in earlier adoption) when ∂NBIAy;c;dce  ∂dce o ∂NBIA10;c;dceðÞ ∂dce ð15Þ That is, when the cumulative discounted net benefits for a vineyard that adopts preventative practices in mature vineyards increase more than those in a vineyard that adopts earlier, we would see a movement toward later adoption. This relationship becomes more apparent when comparing adoption in earlier and earlier years since the net benefits from adopting earlier and earlier decline given the slow initial growth in the infection. Further, when dce is high and increases, we might expect the condition in 15 to hold since the nature of the disease growth means adopting earlier is likely to produce fewer additional benefits than when we adopt late. Lastly, we consider a change in δ,reflecting both changes in a grower's time preference as well as differences in growers’ time preferences, as growers are not all likely to have the same intertemporal preferences. The comparative static with respect to δis ∂πðÞ ∂δ¼ ∂CðÞ ∂δNBIAy  NBIA10 ðÞþCðÞ  CðÞ ∂NBIAy ðÞ ∂δ∂NBIA10 ðÞ ∂δþ∂CðÞ ∂δ hi ½NBIAy  NBIA10 ðÞþCAðÞ 2 ð16Þ and given the decrease in ∂NBIAy ðÞ ∂δis greater than the decrease in ∂NBIA10 ðÞ ∂δ, the two products in the numerator are both negative. To determine the sign we set the numerator less than zero and solve ∂CðÞ ∂δNBIAy  NBIA10 ðÞþCðÞ  CðÞ ∂NBIAy  ∂δ∂NBIA10 ðÞ ∂δþ∂CðÞ ∂δ  o0ð17Þ Rearranging terms yields, ∂CðÞ ∂δ CðÞo ∂NBIAy ðÞ ∂δ∂NBIA10 ðÞ ∂δ hi NBIAy  NBIA10 ðÞ  ð18Þ When the percentage decrease in the cost of the practice is less than (greater than) the percentage decrease in the net benefits, then the threshold will fall (rise) with a change in the discount rate. The long term nature of the effect of the disease on yields means the benefits from adoption are not realized until later in a vineyard's lifespan. Further, the costs are uniformly distributed throughout that lifespan. If these future benefits from adoption are large (i.e., the practice is highly effective) and exceed the additional costs, then we would expect that an increase in the discount rate decreases π 0 , and vice versa. Without knowing the distribution for grower perceptions of the probability of infection, or how it might change over time, we cannot determine the number of growers who will adopt now or in the future. We might assume most growers perceive the probability of infection to be relatively close to 1, based on the findings of Hillis et al. (2015), that PCAs report many vineyards have trunk disease. Nonetheless, this may not be the case as the PCAs are not recommending preventative practices until after the disease is apparent and some growers may be relatively new to the industry and thus have not yet seen the effects or prevalence of trunk disease. We see, in the empirical analysis to follow, that π 0 is, in many cases, very close to zero. These low values and lack of early adoption suggests another reason is likely compelling growers to wait to adopt. 3.3. Simulated economic experiment In the simulated economic experiment, annual costs and benefits from winegrape production over a 25-year lifespan are estimated using budgets taken from the University of California Cooperative Extension (UCCE) Cost and Returns Studies J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141134 North, Cabernet Sauvignon 〈http://coststudyfiles.ucdavis.edu/uploads/ cs_public/c7/0b/c70be1f3-aa8c-40c5-b7ac-5fade99a2e2b/grapewi nevn2012.pdf〉(Available from) (accessed March 2016). Urbez-Torres, J.R., Gubler, W.D., 2011. Susceptibility of grapevine pruning wounds to infection by Lasiodiplodia theobromae and Neofusicoccum parvum. Plant Pathol. 60, 261–270. van Niekerk, J.M., Halleen, F., Fourie, P.H., 2011. Temporal susceptibility of grapevine pruning wounds to trunk pathogen infection in South African grapevines. Phytopathol. Medit., 50. Weber, E.A., Trouillas, F.P., Gubler, W.D., 2007. Double pruning of grapevines: a cultural practice to reduce infections by Eutypa lata. Am. J. Enol. Vitic. 58, 61–66. J. Kaplan et al. / Wine Economics and Policy 5 (2016) 127–141 141