It is mainly about where you work!: Labor demand in the Colombian manufacturing sector

Arango Thomas, Luis Eduardo; Castellani, Francesca; Obando, Nataly Working Paper It is mainly about where you work!: Labor demand in the Colombian manufacturing sector IDB Working Paper Series, No. IDB-WP-712 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Arango Thomas, Luis Eduardo; Castellani, Francesca; Obando, Nataly (2016) : It is mainly about where you work!: Labor demand in the Colombian manufacturing sector, IDB Working Paper Series, No. IDB-WP-712, Inter-American Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0000503 This Version is available at: https://hdl.handle.net/10419/173813 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. http://creativecommons.org/licenses/by-nc-nd/3.0/igo/legalcode It is mainly about where you work! Labor demand in the Colombian manufacturing sector Luis Eduardo Arango Francesca Castellani Nataly Obando IDB WORKING PAPER SERIES Nº IDB-WP-712 August 2016 Country Department Andean Group Inter-American Development Bank August 2016 It is mainly about where you work! Labor demand in the Colombian manufacturing sector Luis Eduardo Arango Francesca Castellani Nataly Obando Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library Arango, Luis E. It is mainly about where you work!: labor demand in the Colombian manufacturing sector / Luis E. Arango, Francesca Castellani, Nataly Obando. p. cm. — (IDB Working Paper Series ; 712) Includes bibliographic references. 1. Labor demand-Colombia. 2. Manufacturing industries-Colombia. I. Castellani, Francesca. II. Obando, Nataly. III. Inter-American Development Bank. Country Department Andean Group. IV. Title. V. Series. IDB-WP-712 Copyright © Inter-American Development Bank. This work is licensed under a Creative Commons IGO 3.0 Attribution- NonCommercial-NoDerivatives (CC-IGO BY-NC-ND 3.0 IGO) license (http://creativecommons.org/licenses/by-nc-nd/3.0/igo/ legalcode) and may be reproduced with attribution to the IDB and for any non-commercial purpose, as provided below. No derivative work is allowed. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the UNCITRAL rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this CC-IGO license. Following a peer review process, and with previous written consent by the Inter-American Development Bank (IDB), a revised version of this work may also be reproduced in any academic journal, including those indexed by the American Economic Association's EconLit, provided that the IDB is credited and that the author(s) receive no income from the publication. Therefore, the restriction to receive income from such publication shall only extend to the publication's author(s). With regard to such restriction, in case of any inconsistency between the Creative Commons IGO 3.0 Attribution-NonCommercial-NoDerivatives license and these statements, the latter shall prevail. Note that link provided above includes additional terms and conditions of the license. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. http://www.iadb.org 2016 It is mainly about where you work! Labor demand in the Colombian manufacturing sector Luis E. Arango* Francesca Castellani Nataly Obando Banco de la República Inter-American Development Bank Inter-American Development Bank Abstract Using the Colombian Annual Manufacturing Survey (EAM) between 2000 and 2013, this paper investigates the existence of heterogeneity in the labor demand within the industrial sector. Long run own-price, output and TFP elasticities vary across a variety of dimensions such as regions, sectors and plant sizes depending on workers’ skills and contract modalities (open-ended or temporary). Hence, it matters where one works. Such disparities should be taken into account in the design of policies that promote labor market performance, as outcomes, beyond intentions, are unlikely to be homogenous. Key words: labor demand, skilled workers, unskilled workers, regional disparities, sector disparities. JEL classification: J23, R32. * The opinions expressed here are not necessarily those of neither the Banco de la República nor its Board of Directors nor IADB. We appreciate comments, corrections and suggestions by Giulia Lotti. Research assistance by María Paula Medina, Álvaro José Pinzón and Sergio Rivera is greatly acknowledged. Usual disclaimers apply. Corresponding author: Francesca Castellani [email protected]. 1 1. Introduction The Colombian labor market is fraught with regional differences. While these might not be particular to Colombia, their intensity and persistence over time is quite significant (Arango, 2013). Labor market outcomes such as participation, occupation and unemployment rates show geographic disparities above 10 percentage points across cities. National unemployment rates1 (11.9%) are unevenly spread, from 7.6% (Bucaramanga) to 16% (Quibdó). Participation rates range from 59% (Quibdó) to 71.4% (Bogota) and occupation rates from 49% (Quibdó) to 64% (Bogota). Employment patterns also vary considerably depending on gender. Recent research (Arango, 2013 and Cárdenas, Hernández and Torres, 2014) identifies the importance of the functioning of local labor markets in explaining these differences. For policy makers, private sector, workers (trade unions) and job-seekers, understanding such heterogeneity and its features is crucial. In these contexts, labor market policies might produce diverse outcomes, even unintended ones, if not correctly factored in. Moreover, a clear grasp of these elements is likely to affect employment location and skill development strategies. Exploring these disparities requires insights on dimensions such as local productive structures, firm size and location. Labor demand is a natural point to start with. Investigating this at the industry level, given its employment potential, is the next one. While several efforts have been made to estimate the labor demand in the manufacturing sector at national level, few propose a regional focus. This paper sheds some light on these aspects. Using the information of the Annual Manufacturing Survey (EAM from its Spanish acronym) between 2000 and 2013, it addresses the heterogeneity of labor demand across regions, industrial subsectors and plant sizes. Going beyond the estimation of aggregate labor demand elasticities to own-price, output and total factor productivity (TFP), as in previous literature, the paper innovates by discriminating across skills and contract modalities. By doing so, a clearer picture emerges as to the implications of shocks on skilled and unskilled, as well as permanent and temporary, labor force. Results confirm that disparities do exist and labor markets are heterogeneous. Complementing existing studies, the original contribution of this paper is to illustrate the nature of the regional, sector and size differences in the Colombian labor market and draw attention to the implication of this heterogeneity for labor policy design and outcomes. The paper develops in five sections beyond this introduction. The second section presents a short literature review. The third offers stylized facts of the manufacturing sector. The fourth section describes the theoretical approach and the fifth presents the empirical model and the results. The sixth section concludes. 2. Literature review Existing research on the determinants of labor demand in Colombia does not reach unanimous conclusions. Roberts y Skoufias (1997), using EAM data between 1981 and 1987, find that unskilled workers have a higher long run wage elasticity (-0.65) than skilled ones (-0.42). That is, the former are likely to be more effected by increasing labor costs. Output elasticities are higher for skilled labor (0.89) 1 Based on January 2016 data from the Colombian Statistical Office (DANE). 2 than for unskilled (0.76), suggesting that output and skilled labor demand move almost proportionally. Regional dummies were included among the control variables. Using data from the EAM and household surveys between 1980 and 1996, Vivas, Farné and Urbano (1998) find long run real wage and output elasticities of -0.71 and 1.10, respectively. Exploring the effects of trade liberalization on manufacturing labor demand between 1997 and 1999, Arango and Rojas (2004) obtain elasticities of -0.78 and 0.76 for real wage and output respectively. For a more recent subsample, the estimates turn to -0.92 and 0.67. For Bernal and Cárdenas (2003), based on the information of 2570 EAM establishments between 1978 and 1991, the long-run real wage elasticity is 2.27 and the output elasticity 0.24. With a panel of 91 manufacturing sectors, they obtain a short-run real wage elasticity of –0.6 and a long run one of –1.43. Eslava, Haltiwanger, Kugler, and Kugler (2010), focusing on joint factor demand and the incidence of adjustment costs, find that the latter are significant and substitutabilityrather than complementarityamong factors emerge during the adjustment process. Medina, Posso, Tamayo and Monsalve (2013) study the unconditional labor demand for the period 1993-2009, concluding that industrial employment is highly persistent. Production shocks generate a larger response by skilled and unskilled employment than shocks on capital and wages. Less skilled workers face higher adjustment costs. Resting on this literature and the findings of Arango (2013), this paper explores the industrial labor demand taking into account the potential regional, subsector and size disparities in Colombia, in the line of Adam and Moutos (2014) for the case of European countries.2 3. Some facts about the Colombian manufacturing sector Colombian economic structure has considerably changed over the past four decades. In 1960s the manufacturing sector represented over 17% of GDP while in 2013 its share was 11.5%. The natural decline of agriculture and manufacturing sectors in GDP participation matched the rise of construction, trade and services. Between 2000 and 2013, the industrial sector maintained a 3.1% annual growth rate, severely interrupted by the 2008 crisis (Figure 1). In the aftermath, growth rate resumed at 1.8% per year, well below the rest of the economy (4.4%). The lackluster performance since then has been attributed to the real exchange rate appreciation and the consequent loss of competitiveness; traderelated demand disruptions, increasing labor costs, and poor infrastructure supply.3 Still, manufacture remains as one of the sectors that employs most labor in the economy, being the fourth contributor to employment with 12.1%, according to the Great Integrated Household Survey (GEIH for its the Spanish acronym), after trade (27.4%), services (19.4%) and agriculture (17%). In 2013, the industrial sector accounted for 2.5 million jobs at national level. Hereafter, the paper uses the Colombian Manufacturing Survey (EAM) data, which includes 9,158 establishments (681,452 jobs) located in the 9 major cities (Barranquilla, Bogotá, Bucaramanga, Cali, 2 Another strand of literature on labor market disparities relates to this paper. According to Dumais, Ellison and Glaeser (2002) geographic concentration in the U.S manufacturing industries has been declining over time. Dao, Furceri and Loungani (2014) observe lower migration rate among US states during the past 40 years due to higher labor market flexibility. In Europe, the participation rate response to labor demand shocks has been affected by migration. Beaudry, Green and Sand (2014) explore differences in labor demand introducing a special set of instruments and the local population size as a determinant of employment demand (see also Bartik, 2014). 3 See Griffin (2015) for a discussion. 3 Cartagena, Cúcuta, Manizales, Medellin, and Pereira) and “others”.4 The stylized facts from EAM endorse the heterogeneity identified by Arango (2013).5 Around 80% of manufacturing employment concentrates in 4 cities (Bogotá, Medellin, Cali and Barranquilla), and Bogotá accounts for 40%. With the exceptions of Bogotá and Bucaramanga, all cities experienced a decline in their employment share since 2000 (Figure 2). Figure 1. GDP and manufacturing GDP annual growth rates Source: DANE-EAM; authors’ calculations. Figure 2. Contribution to employment of manufacturing sector by city (% of total) Source: Dane-EAM; authors’ calculations. Over the period, industrial employment expanded annually at 2.1%; however, geographically, patterns are uneven. Employment in Bogotá grew at 3.1% per year - with an average annual growth rate of 4.6% before the crisis of 2008-2009, 2.1% during that episode6 and 1.1% since 2010. In Cartagena employment grew at 2.1% but remained constant during 2008 and 2009. Manizales showed an average 4 12,683 jobs are excluded as they are generated in small plants (less than 10 employees). 5 The GEIH and the EAM differ as the first includes employment generated by firms with up to 5 employees (formal and informal). 6 In 2009, the annual growth rate was -0.9%. -6% -4% -2% 0% 2% 4% 6% 8% 10% 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 GDP Manufacturing GDP 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% Bogotá Medellín Cali Barranquilla Pereira Bucaramanga Manizales Cartagena Cúcuta Others 2000 2013 4 growth rate of 1.4% during the period 2000 and 2007 and 6.3% decline during the crisis. In Pereira volatility was significant along the sample period. Barranquilla provides an interesting case: with an unappealing evolution before 2009, employment increased at 3% annually since 2010. Cúcuta and Manizales displayed dismal growth rates of 0.4% and 0.2%, respectively. Six sectors - food and beverages, chemicals, apparel, rubber and plastic, textiles, and mineralsaccount for 66% of total employment.7 While remaining the largest employment generating sectors, their participation to employment varies overtime in an unsteady fashion. Chemicals, rubber and plastic, mineral, metals record sizable increments overtime while food and textile decline (Figure 3). Figure 3. Employment generation by sector and city 2000 – 2013 Source: Dane-EAM; authors’ calculations. Employment generation varies across cities. Barranquilla generates only a few jobs in apparel while Cali does so in textiles. Food is the largest job generator with 44% of total employment in Bucaramanga, 35% in Cartagena and 30% in Manizales (Figure 4). Apparel has been a very important employer sector in Pereira, Medellin and Cali. Bogotá, Medellin, Cali and Barranquilla have plants devoted to the production of all six subsectors. Colombian business landscape is populated by small and medium sized establishments and manufacturing is not an exception. On average, medium size manufacturing plants account for 48% of the total sample, generating 15% of the employment.8 Large and very large ones represent 32% of establishments and 83% of the total employment; small plants - 21% of total - contribute to employment for 1.8% (Figure 5). 7 If we add metal and furniture, the total employment corresponds to 76.7% of the sample in 2013. 8 In this study, small sized establishments are those that employ 10 people or less, medium sized between 11 and 50, large between 51 and 500 and very large more than 500. This size definition allows sufficient degrees of freedom for estimation purposes. Kanbur and Venables (2005) use similar categories. 0% 5% 10% 15% 20% 25% Food & beverages Chemicals Apparel Rubber & plastic Textiles Minerals Metal Furniture Machinery &… Printing Leather Electrical… Motor vehicles Paper Basic metals Other transports Wood Refined petroleum Medical… Tobacco Radio equipment Office equipment 2000 2013 11 Δ𝑙𝑙𝑡𝑡+𝑗𝑗 =𝐸𝐸𝑡𝑡�∑𝐹𝐹𝑙𝑙�𝑙𝑙𝑡𝑡+𝑗𝑗,𝐴𝐴𝑡𝑡+𝑗𝑗�−𝑤𝑤𝑡𝑡+𝑗𝑗 𝑎𝑎(1+𝑟𝑟)𝑗𝑗 ∞ 𝑗𝑗=0 � Thus, if the present value of the marginal product of labor is greater than the present value of wages, the number of employees will increase. Assuming rational expectations, the model (see Sargent, 1978; Nickel, 1984; Hamermersh, 1993; and, Cahuc and Zylberberg, 2004) can be written as: 𝑙𝑙𝑡𝑡=λ𝑙𝑙𝑡𝑡−1 +∑ ∑ 𝜇𝜇𝑚𝑚𝑗𝑗𝑋𝑋𝑚𝑚,𝑡𝑡−𝑗𝑗 𝐽𝐽𝑚𝑚 𝑗𝑗=0 𝐾𝐾 𝑚𝑚=1 By using the notation of Hamermesh (1993, chapter 7), the empirical model can be written as: 𝑙𝑙𝑖𝑖,𝑡𝑡=λ𝑙𝑙𝑖𝑖,𝑡𝑡−1 +∑ ∑ 𝜇𝜇𝑚𝑚,𝑗𝑗𝑋𝑋𝑖𝑖,𝑚𝑚,𝑡𝑡−𝑗𝑗 𝐽𝐽𝑚𝑚 𝑗𝑗=0 +𝜔𝜔𝑖𝑖,𝑡𝑡 𝐾𝐾 𝑚𝑚=1 where index i accounts for plant i, 𝑋𝑋𝑖𝑖,𝑚𝑚,𝑡𝑡−𝑗𝑗 is the vector of observable variables including the average real wage paid by the firm, added value15 of the firm, TFP indicator, etc.; 𝜇𝜇𝑚𝑚,𝑗𝑗 is the vector of coefficients to be estimated, and 𝜔𝜔𝑖𝑖,𝑡𝑡 is the residual term. With the aim of having information about the adjustment costs, the empirical model holds just one autoregressive element (Hamermersh, 1993) which can be separated from the expectations of remaining variables based on past realizations. Potentially, the residual term is the sum of different elements: 𝜔𝜔𝑖𝑖,𝑡𝑡=𝜂𝜂𝑖𝑖+𝜐𝜐𝑖𝑖,𝑡𝑡+𝜀𝜀𝑖𝑖,𝑡𝑡. The first component, 𝜂𝜂𝑖𝑖, accounts for unobserved heterogeneity directly linked to plant characteristics other than the price of inputs and usual determinants of labor demand (see Roberts and Skoufias, 1997). The list of plausible candidates includes especial properties of output, the relative advantages to combine inputs in the production process, etc. The second element, 𝜐𝜐𝑖𝑖,𝑡𝑡, can be related with shocks to the demand product or to the supply of labor specific to the production of the plant, etc. These first two elements of the unobserved heterogeneity could eventually be correlated to the output of the plant as well as the wage �𝐸𝐸(𝜂𝜂𝑖𝑖|𝑋𝑋𝑖𝑖,𝑚𝑚,𝑡𝑡)≠0� and �𝐸𝐸(𝜐𝜐𝑖𝑖,𝑡𝑡|𝑋𝑋𝑖𝑖,𝑚𝑚,𝑡𝑡)≠0� ; as a result the respective coefficients could be biased upwards. The third element, 𝜀𝜀𝑖𝑖,𝑡𝑡, is a well behaved random component. To address the potential endogeneity (reverse causality) of real wage and the establishment output, instrumental variables as well as some moment conditions are used. Arellano and Bover (1995) and Blundell and Bond (1998) indicate that lagged levels are often rather poor instruments for first differenced variables, especially if the variables are close to a random walk. Thus, the paper adopts the Blundell and Bond (1998) GMM estimator where both lagged levels and lagged first differences can be combined as instruments to improve efficiency of the estimator. As the two-step system GMM estimator presents downwards bias, the Windmeijer (2005) correction is also used. The paper combines skilled and unskilled, permanent and temporary employment, and explores the regional, sector and size dimensions of heterogeneity. The resulting estimations provide elasticities for (i) the 9 major cities (Barranquilla, Bogotá, Bucaramanga, Cali, Cartagena, Cúcuta, Manizales, Medellin, and Pereira) and “others”; (ii) the largest employment generating sectors (food and beverages, chemicals, apparel, rubber and plastic, textiles, mineral, metals and furniture) and plant sizes (medium sized, large and very large). For each of these dimensions the elasticities are estimated for skilled and unskilled workers with permanent and temporary contracts. To take advantage of the panel structure of the Annual Manufacturing Survey the paper uses information at establishment level between 2000 and 2013. The smallest number of plants of our 15 In the EAM, added value is calculated as the difference between gross production and intermediate consumption. 12 unbalanced panel is 6801 in 2003 while the highest is 9867 in 2010. The number of plants was 9158 in 2013, the last year of the sample. The empirical specification sets demand for labor in terms of its own lag, contemporary and lagged values of real wages, output, total factor productivity, the real minimum wage, energy prices and the real interest rate. The log of number of employees is the dependent variable. The variables on the righthand side are: the lag of the dependent variable, the log of the current and lagged value of: average real wage, added value of the firm, real minimum wage, real interest rate, TFP and energy price. The specification also includes the current value of the depreciation rate16. Average real wage, minimum wage and added value starting at lag two are used as instruments. Most of the results below correspond to long-run elasticities which are computed as: 𝜖𝜖𝑗𝑗= (𝜇𝜇𝑚𝑚,0+𝜇𝜇𝑚𝑚,1) (1 − ⁄λ), m is real wage, output and TFP, respectively. As the EAM does not have information about individual real wages, these were computed by dividing, respectively, the payroll in real terms among the number of skilled and unskilled workers. Output corresponds to the total added value in each plant, as well as the energy prices and the depreciation rate. The total factor productivity was computed by using the Levinshon and Petrin (2003) algorithm. Finally, the real interest rate corresponds to ordinary and preferential interest rates,17 since the real rental price of each plant is not available. 5. Results18 Table 1 shows own real price, output, and TFP long-run elasticities under conditional and unconditional specifications of labor demand for skilled and unskilled workers. Wage elasticities of skilled (-0.591) and unskilled workers (-0.415) are both significant. Contrary to existing literature (see Litcher et al., 2014 and, for the case of Colombia, Roberts and Skoufias, 1997), skilled employment shows a larger sensitivity to its own price. When estimations take into account contract modalities, the long-run real wage elasticity of open-ended contracts workers is larger than temporary, irrespective of their skills. Increased flexibility in the labor market institutions and growth in labor intermediation over the sample period might explain this larger response. Unskilled workers are generally more sensitive than skilled ones, in line with previous research. Moreover, under the conditional specification, the temporary unskilled workers show an elasticity of -0.438 and the skilled of -0.256. Similarly, the elasticity to the real wage of demand for unskilled workers with open-ended contracts is -1.109 and - 0.682 for the skilled permanent workers.19 16 According to the EAM, the depreciation corresponds the annual value set as replacement for the deterioration, use or obsolescence of the firm fixed assets during their lifetime. 17 The difference between these two interest rates is the size of the plant. Preferential is the interest rate of loans granted to large and very large establishments (about 32% of the plants) while ordinary is the interest rate of credits requested by medium size establishments. 18 Estimations do not include small size establishments (< 10 employees) neither as residual city category nor as “others”, regardless that both are included in the descriptive statistics. 19 In general, according to the Hansen test, the set of instruments is accurate and the Ar2 test shows that the errors do not exhibit autocorrelation of order two. 13 Table 1. Long-run labor demand elasticities (2000-2013). Specification Conditional Unconditional Skilled workers Own-price -0.591*** -0.316*** Output 1.122*** TFP -1.016*** 0.326* No. observations (plants) 14.579 14.579 Sargan test (p-value) 7.68e-10 0.000381 Hansen test (p-value) 0.102 0.808 Ar2 (p-value) 0.540 0.969 Skilled permanent workers Own-price -0.682*** -0.517*** Output 0.949*** TFP -0.773*** 0.753*** No. observations (plants) 13.062 13.062 Sargan test (p-value) 0.000466 1.73e-05 Hansen test (p-value) 0.213 0.514 Ar2 (p-value) 0.700 0.398 Skilled temporary workers Own-price -0.256*** -0.251*** Output 0.710*** TFP -0.621*** 0.590* No. observations (plants) 4.477 4.477 Sargan test (p-value) 3.11e-07 0.000946 Hansen test (p-value) 0.467 0.761 Ar2 (p-value) 0.443 0.464 Unskilled workers Own-price -0.415*** 0.074 Output 1.051*** TFP -0.951*** 0.261 No. observations (plants) 20.470 20.470 Sargan test (p-value) 0.0353 0.730 Hansen test (p-value) 1.21e-05 0.000113 Ar2 (p-value) 0.618 0.346 Unskilled permanent workers Own-price -1.109*** -0.865*** Output 1.198*** TFP -1.043*** 0.429 No. observations (plants) 17.502 17.502 Sargan test (p-value) 6.32e-08 1.06e-07 Hansen test (p-value) 0.637 0.589 Ar2 (p-value) 0.339 0.513 Unskilled temporary workers Own-price -0.438*** -0.107 Output 0.760*** TFP -0.730*** 0.223 No. observations (plants) 12.579 12.579 Sargan test (p-value) 0 0 Hansen test (p-value) 0.000153 0.00401 Ar2 (p-value) 0.702 0.00624 Source: Dane-EAM; authors’ calculations. The output long-run elasticity is higher than the real wage elasticity as in earlier studies. Permanent workers (0.949-skilled and 1.198-unskilled) show larger responsiveness than temporary ones (0.710- skilled and 0.760-unskilled). However, skilled (1.122) are more sensitive than unskilled workers (1.051). The theory predicts a negative sign of the response to TFP in the case of the conditional labor demand and a positive one for the unconditional demand. These predictions match the estimates for both types of workers, with a larger response for permanent employees. The respective elasticities, corresponding to the conditional demand specification, for skilled workers are: -0.773 and -0.621 while for unskilled 14 workers the elasticities are -1.043 and -0.73. Demand for skilled workers, regardless of the type of contract, display lower TFP elasticity and more resilience to changes in technology. Under the unconditional specification, the own-price elasticity is significant for skilled workers and for unskilled permanent. At the same time, TFP elasticity is significant only for skilled workers and has in the three cases the right sign. This labor demand specification does not seem appropriate for unskilled employees. The empirical specification allows estimating the speed of adjustment of labor demand (halfway or half-life of the adjustment) to shocks in output or factor prices. The median length is computed by solving for 𝑡𝑡∗ the expression 𝜆𝜆𝑡𝑡∗= 0.5. Under the conditional specification, the halfway ranks between 0.61 and 1.27 years (Figure 14), corresponding to skilled temporary workers and unskilled permanent ones, pointing at a shorter adjustment period for higher level of skills and more flexible contracts. These results are consistent with existing estimates using annual data which show halfways of 5.5 quarters (Hamermesh, 1993, page, 253). Figure 14. Halfways of labor demand adjustments (2000-2013). Source: Dane-EAM; authors’ calculations. 5.1 Regional heterogeneity In this and the next two sections, regional, sector and size heterogeneity are analyzed based on the conditional labor demand, which corresponds to the constant product specification. The reason for this selection is the consistency of results with the theory. As mentioned before, labor market outcomes present significant disparities across metropolitan areas in Colombia. Elasticities are estimated across 9 major cities to verify labor market regional heterogeneity (Figure 15). Results indicate that besides Cartagena and Cúcuta, whose elasticities are not statistically significant, real wage long-run elasticities of permanent skilled workers are negative and significant and range between -1.3 (Manizales) and -0.632 (Medellin).20 In the case of skilled workers with temporary contracts, the highest own price elasticity corresponds to Manizales (-1.061). For permanent unskilled workers, the greatest value is found in Pereira (-1.843) 20 These results are shown in the Appendix with the corresponding statistics. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Skilled Skilled permanent Skilled temporary Unskilled Unskilled permanent Unskilled temporary Condicional Uncondicional Years 15 and the smallest in Barranquilla (-0.983). In the case of unskilled temporary workers, the elasticity varies between -0.27 (Bogotá) and -0.58 (Barranquilla) and only these two cities, together with Medellin and Pereira, hold significant long-run elasticities. Consistently with Figure 15, the geographic breakdown confirms that demand for temporary workers, both skilled and unskilled, is less responsive to its own price than permanent ones. The length of adjustment to shocks is greater for unskilled and permanent workers. Once there is a change in the real wage, permanent workers are more likely to be displaced and possibly replaced by temporary ones. Bogotá shows consistently faster adjustment speed than the rest of the cities across skills and contract modalities (Figure 16). Figure 15. Own price elasticity of conditional labor demand (2000-2013). Permanent Temporary Skilled Unskilled Note: 95% confidence intervals. Source: Dane-EAM; authors’ calculations. -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta 16 Figure 16. Halfway of labor demands adjustments of a real wage movement by city. Conditional demand for labor Skilled workers Unskilled workers Source: Dane-EAM; authors’ calculations. Output elasticities, with the exception of Cúcuta, all are positive (Figure 17). However, significant long-run reactions to output for permanent skilled labor are obtained for plants in Bogotá, Cali, Cartagena, Medellin, and Pereira. In the case of temporary skilled workers, only Bogotá and Medellin are significant. Demand for permanent unskilled employees is significant in Cali (1.702), “others” (1.241), Bogotá (0.987), Medellin (0.833), Barranquilla (0.814) and Bucaramanga (0.723). The lowest estimate for temporary unskilled workers is found in Medellin (0.619) and the highest in Manizales (4.047). Elasticities to TFP are negative in most cases as predicted by the theory under the conditional specification (Figure 18). The heterogeneity of the determinants of conditional labor demand seems a prevalent characteristic of the manufacturing sector in Colombia. The hypothesis that the difference between each pair of coefficients is equal to zero can be rejected in most cases (Tables 2 and 3). Based on the tests of meandifferences carried out under the assumption that distributions are independent, the null hypothesis that the elasticities of each type of demand for labor are equal for most of combinations of cities can be rejected. In Table 2, for the case of permanent contracts, the lower part of the matrix (below the diagonal), the labels “O” indicates that the null hypothesis that the long-run wage elasticities of skilled workers are equal between each pair of cities cannot be rejected at 5% level of significance. Cali and Barranquilla are the only metropolitan areas where long-run wage elasticities are statistically equal. With respect to output elasticity of demand for skilled workers, the null hypothesis that the coefficients are the same between each pair of cities cannot be rejected at 5% level of significance in the following cases: Bogotá-Barranquilla, Medellin-Barranquilla, Bucaramanga-Cartagena, and Manizales-Pereira. For TFP elasticities, the pairs of cities in which the coefficients are statistically similar are: Bucaramanga- Cartagena, Manizales-Bucaramanga, Cali-Manizales and Cartagena-Manizales. 012345 Bogotá Barranquilla Medellín Cali Cartagena Manizales Bucaramanga Pereira Skilled permanent Skilled temporary Years 012345 Bogotá Barranquilla Medellín Cali Cartagena Manizales Bucaramanga Pereira Unskilled permanent Unskilled temporary Years 17 Figure 17. Output elasticity of conditional labor demand (2000-2013). Permanent Temporary Skilled Unskilled Note: 95% confidence intervals. Source: Dane-EAM; authors’ calculations. -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta 18 Figure 18. TFP elasticity of conditional labor demand (2000-2013). Permanent Temporary Skilled Unskilled Note: 95% confidence intervals. Source: Dane-EAM; authors’ calculations. In the upper-right section of Table 2 (above the diagonal), where the equality of unskilled elasticities is tested, the labels “X” suggest that the null hypothesis cannot be rejected at 5% of significance. As a result, in the case of real wage elasticity, Bogotá and Cúcuta have the same coefficient as well as Medellin and “others” but most of the coefficients are not statistically equal to each other. For output, similar elasticities are found in the following pairs: Medellin-Barranquilla, Pereira-Cartagena, and Cúcuta-Pereira. TFP elasticities are similar in: Cúcuta-Bogotá, Cúcuta-Barranquilla, and Cúcuta- “others”. -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Bogotá B/quilla Cali Medellín B/manga C/gena M/zales Pereira Others Cúcuta 19 Table 2. Similarity of elasticities tests on conditional labor demand for permanent workers Unskilled → Bogotá Barranquilla Bucaramanga Cali Cartagena Manizales Medellin Pereira Others Cúcuta Skilled ↓ w y tfp w y tfp w Y tfp w y tfp w y tfp w Y tfp w y tfp w y tfp w y tfp w y tfp Bogotá X X Barranquilla O X X Bucaramanga X Cali O Cartagena O O X Manizales O O O Medellin O X Pereira O X Others X Cúcuta Source: Dane-EAM; authors’ calculations. The tests in Table 3 -temporary employeessuggest that the null hypothesis cannot be rejected for 8 out of 45 possible combinations of cities for real wage elasticity in the case of skilled workers; these pairs are: Bogotá-Pereira, Cali-Barranquilla, Bucaramanga-Barranquilla, Cali-Bucaramanga, Pereira- Barranquilla, Pereira-Bucaramanga, Pereira-Cali, and Pereira-Medellin. Similar output elasticities are found for Bogotá-Bucaramanga, Bogotá-Medellin, Barranquilla-Bucaramanga, Barranquilla-Medellin, Bucaramanga-Medellin, Cali-Manizales, Cali-Pereira, and Manizales-Pereira. TFP elasticity for temporary skilled workers is similar in Barranquilla-Bucaramanga and Cartagena-Cúcuta. Thus, the employment in the industrial sector shows a high geographic heterogeneity. Nevertheless, Bucaramanga and Barranquilla are similar in some dimensions of industrial temporary employment. Recall, that according to Figure 8 above, lower panels, in these two cities an important reduction of permanent workers took place between years 2000 and 2013. Table 3. Similarity of elasticities tests on conditional labor demand for temporary workers Unskilled → Bogotá Barranquilla Bucaramanga Cali Cartagena Manizales Medellin Pereira Others Cúcuta Skilled ↓ w y tfp w y tfp w Y tfp w y tfp w y tfp w y Tfp w y tfp w y tfp w y tfp w Y tfp Bogotá X Barranquilla Bucaramanga O O O O Cali O O Cartagena Manizales O Medellin O O O X Pereira O O O O O O O Others Cúcuta O Source: Dane-EAM; authors’ calculations. 20 5.2 Sector heterogeneity Disparities of labor demand also emerge across industrial subsectors. To simplify the analysis, estimates are restricted to the largest employment generating subsectors (Figure 3)21 under the conditional specification. The demand for permanent skilled workers displays larger responses to real wages than temporary employees (Figure 19). At the same time, permanent unskilled workers are more sensitive than skilled ones. Food products, chemicals, and rubber and plastic show significant elasticities. The lowest value is found for temporary skilled workers in chemical plants (-0.284) and the highest for unskilled workers in textile plants (-2.157). For skilled workers, the median adjustment length by subsector varies between 0.5 (temporary) and 2.26 years (permanent). The shortest halfway corresponds to rubber and plastic subsector and the longest to chemicals (Figure 20). For unskilled workers, the halfways range between 0.97 (mineral products) and 6.2 (textiles) years. A partial conclusion is that adjustment costs for unskilled workers are lower. For skilled permanent labor force, the highest value of output elasticity corresponds to chemicals (1.7) and the smallest to furniture (0.98). This estimate is greater for permanent skilled workers than for temporary ones in each industrial subsector. Mineral products for unskilled temporary workers (1.87) and metal products (0.82) represent the extreme values of output elasticities. In the case of unskilled permanent workers, besides rubber and plastic whose coefficients are not statistically significant, output elasticities fluctuate between 0.85 and 1.42. This hints at a myriad of technologies and production functions across subsectors (Figure 21). The TPF elasticity of labor demand is negative in general; when slightly positive it becomes not statistically significant. This is the case of metal (temporary skilled) and rubber and plastic (permanent unskilled). The TFP elasticity of permanent skilled workers fluctuates between -1.5 and -1.0 for almost all subsectors (Figure 22). Tables 4 and 5 show the prevalence of heterogeneity across industrial subsectors. In the case of unskilled permanent workers (Table 4, upper-side), the TFP elasticity of labor demand in rubber and plastics is similar to that of food products, textiles, apparel and chemicals subsectors. By the same fashion, the null hypothesis that chemical subsector has the same long run elasticity to TFP as mineral, metal and furniture cannot be rejected in the case of unskilled permanent workers. As for temporary workers, the demand for unskilled labor in the textile subsector shows similarities with others (Table 5, upper-side). 21 Recent evidence for Colombian establishments shows that the ones that exit the market are the least productive ones (Casas, Carranza, and González, 2014). 27 References Addison, J.T., and Teixeira, P., 2003, The Economics of Employment Protection, Journal of Labor Research, Vol. 24, pp. 85-129, Transaction Publishers. Adam, A., and Moutos, T., 2014, Industry-level labour demand elasticities across the Eurozone: will there be any gain after the pain of internal devaluation?, Bank of Greece Working Papers. Alaimo, V., Bosch, M., Kaplan, D.S., Pagés, C., and Ripani, L., 2015, Empleos para crecer, InterAmerican Development Bank. Arango, L.E., 2103, Mercado de trabajo en Colombia: suma de partes heterogéneas, Chapter 4, in Mercado de trabajo en Colombia: hechos, tendencias e instituciones, Arango L.E and F. Hamann (eds.), Banco de la República, co-editor, 2013. Available at: http://www.banrep.gov.co/es/libro-mercado-colombia. Arango, L.E., Castellani, F., and Lora, E., 2016, Desempleo femenino en Colombia, Banco de la República – Banco Interamericano de Desarrollo, forthcoming. Arango, C.A., and Rojas, A.M., 2003, Demanda laboral en el Sector Manufacturero Colombiano: 1977- 1999, Borradores de Economía, No. 247, Banco de la República. Arellano, M., and Bover, O., 1995, Another look at the instrumental variable estimation of errorcomponents models, Journal of Econometrics, Vol. 68, pp. 29-51. Bartik, T.J., 2014, How Effects of Local Demand Shocks Vary with Local Labor Market Conditions, Upjohn Institute Working Paper, No. 14-202. Beaudry, P., Green, D.A., and Sand, B.M., 2014, In search of labor demand, NBER Working Papers WP 20568. Bernal, R., and Cardenas, M., 2003, Determinants of labor demand in Colombia: 1976-1996, NBER Working Papers WP 10077. Blundell, R., and Bond, S., 1998, Initial conditions and moment restrictions in dynamic panel data models, Journal of Econometrics, Vol. 87, pp. 115-143. Cahuc, P., Carcillo, S., and Zylberberg A., 2004, Labor economics, MIT Press, Cambridge, Massachusetts. Cárdenas, C., Hernández, A., and Torres J.E., 2014, An Exploratory Analysis of Heterogeneity on Regional Labour Markets and Unemployment Rates in Colombia: An MFACT approach, Borradores de Economía, No. 802, Banco de la República. Casas, C., Carranza, J.E., González, A., 2014, Relación entre productividad y supervivencia de las firmas industriales colombianas entre 2005 y 2013, mimeo, Banco de la República, BR-CEEII Cali. Dao, M., Furceri, D., and Loungani, P., 2014, Regional Labor Market Adjustments in the United States and Europe, IMF Working Paper, No. 14/26, IMF Washington DC. Dumais, G., Ellison, G., and Glaeser, E.L., 2002, Geographic concentration as a dynamic process, The Review of Economics and Statistics, Vol 84, No. 2, pp. 193-204. 28 Eslava, M., Haltiwanger, J., Kugler, A., and Kugler, M., 2010, Factor adjustments after deregulations: panel evidence from Colombian plants, The Review of Economic Studies, Vol. 51, No. 4, pp. 529-557. Fajnzylber, P., and Maloney, W., 2001, How comparable are labor demand elasticities across countries?, Police Research Working Papers 658, World Bank. Griffin, N., 2015, Determinants of Firm Profitability in Colombia’s Manufacturing Sector: Exchange Rate or Structural?, IMF Working Paper, No. 15/97, IMF Washington DC. Hamermesh, D., 1993, Labor demand, New Jersey, Princeton Academic Press, pp. 444. Kanbur, R and A. Venables, 2005, Spatial Inequality and Development, OUP Oxford Lichter, A., Peichl, A., and Siegloch, S., 2014, The Own-Wage Elasticity of Labor Demand: A Meta- Regression Analysis, IZA Discussion Paper Series, No. 7958. Levinsohn, J., and Petrin, A., 2003, Estimating Production Functions Using Inputs to Control for Unobservables, The Review of Economic Studies, No. 70, pp. 317-341. Medina, C., Posso, C., Tamayo, J., and Monsalve, E., 2103, Dinámica de la demanda laboral en la industria manufacturera colombiana 1993-2009: una estimación panel VAR, Chapter 7, in Mercado de trabajo en Colombia: hechos, tendencias e instituciones, Arango L.E and F. Hamann (eds.), Banco de la República, co-editor, 2013. Available at: http://www.banrep.gov.co/es/libro-mercado-colombia. Nickel, S., 1984, An Investigation of the Determinants of Manufacturing Employment in the United Kingdom, The Review of Economic Studies, Vol. 51, No. 4, pp. 529-557. Pierre, G., and Scarpetta, S., 2013, Do Firms Make Greater Use of Training and Temporary Employment When Labor Adjustment Costs are High?, IZA Journal of Labor Policy, Vol. 2:15 (Springer Open Journal). Roberts, M., and Skoufias, E., 1997, The long-run demand for skilled and less skilled labor in the Colombian manufacturing plants, The Review of Economics and Statistics, Vol. 79, No. 2, pp. 330-334. Sargent, T., 1978, Estimation of Dynamic Labor Demand Schedules under Rational Expectations, The Journal of Political Economy, Vol. 86, No. 6, pp. 1009-1044. Vivas, A., Farne, S., and Urbano, D., 1998, Estimaciones de funciones de demanda de trabajo dinámicas para la economía colombiana, 1980-1996, Archivos de Macroeconomía, No. 092, Departamento Nacional de Planeación. Wickens, M., 2011, Macroeconomic Theory: A Dynamic General Equilibrium Approach, Princeton University Press. Windmeijer, F., 2005, A finite sample correction for the variance of linear efficient two-step GMM estimators, Journal of Econometrics, Vol. 126, pp. 25–51. 29 Appendix Table A1. Estimates properties labor demand elasticities (2000-2013). City Long-run elasticities Properties Own price Output TFP Sample size Sargan test (p-value) Hansen test (p-value) Ar2 (p-value) Permanent skilled Bogotá -0.638*** 0.000 1.092*** 0.000 -1.037*** 0.000 5318 0 0.413 0.454 Barranquilla -0.673** 0.024 1.067* 0.063 -0.254 0.716 731 0 1 0.834 Cali -0.694*** 0.000 1.340*** 0.000 -1.292*** 0.000 1485 0 0.353 0.690 Medellín -0.632*** 0.000 1.052*** 0.000 -0.739*** 0.006 2625 0 0.338 0.253 Bucaramanga -1.026** 0.015 1.493 0.119 -1.599 0.135 343 0 1 0.334 Cartagena -0.249 0.416 1.543** 0.034 -1.528* 0.077 351 0 1 0.569 Manizales -1.300** 0.022 2.426 0.482 -1.632 0.675 355 0.210 1 0.271 Pereira -0.911*** 0.000 2.517*** 0.001 -2.126** 0.037 235 0.0119 1 0.278 Cúcuta 0.398 0.785 -4.128 0.708 5.154 0.627 124 0.0381 1 0.752 Temporary skilled Bogotá -0.105 0.200 0.629*** 0.000 -0.496** 0.015 1589 0 0.311 0.364 Barranquilla -0.349*** 0.001 0.569 0.260 -0.392 0.525 287 0 1 0.443 Cali -0.331* 0.090 0.386 0.130 0.163 0.667 539 0 1 0.859 Medellín -0.168 0.120 0.619*** 0.004 -0.665 0.005 766 0 0.583 0.812 Bucaramanga -0.291 0.715 0.614 0.351 -0.352 0.675 184 0 1 0.388 Cartagena 0.102 0.858 2.525 0.480 -2.785 0.539 147 0.00542 1 0.543 Manizales -1.061** 0.016 0.447 0.430 1.000 0.255 183 0.141 1 0.778 Pereira -0.316 0.871 0.238 0.909 -1.004 0.611 160 0.102 1 0.280 Cúcuta -1.676 0.420 4.048 0.383 -3.929 0.392 55 0.00945 1 0.944 Permanent unskilled Bogotá -1.094*** 0.000 0.987*** 0.000 -0.787*** 0.000 6857 0 0.859 0.921 Barranquilla -0.983*** 0.004 0.814** 0.038 -0.828 0.164 927 0 1 0.297 Cali -1.666*** 0.000 1.702*** 0.000 -1.790*** 0.000 1866 0 0.586 0.659 Medellín -1.163*** 0.000 0.833*** 0.000 -0.661*** 0.000 3736 0 0.395 0.0835 Bucaramanga -1.225*** 0.000 0.723*** 0.008 -0.266 0.368 574 0 1 0.161 Cartagena -1.230** 0.038 1.960 0.123 -1.844 0.214 440 0 1 0.525 Manizales -1.378*** 0.000 3.118 0.427 -3.752 0.464 433 0.00207 1 0.0901 Pereira -1.843*** 0.002 1.924 0.201 -1.401 0.508 349 0.00292 1 0.610 Cúcuta -1.073*** 0.000 1.393 0.217 -0.958 0.560 196 0.00342 1 0.791 Temporary unskilled Bogotá -0.270*** 0.008 0.750*** 0.000 -0.714*** 0.000 4252 0 0.0782 0.0355 Barranquilla -0.580** 0.010 1.214*** 0.000 -1.125*** 0.000 818 0.0204 1 0.385 Cali -0.042 0.847 0.956*** 0.000 -0.874*** 0.008 1549 0 0.479 0.957 Medellín -0.456*** 0.006 0.646*** 0.000 -0.822*** 0.001 2571 0 0.452 0.0586 Bucaramanga -0.854* 0.058 0.415 0.143 -0.076 0.874 543 0 1 0.0169 Cartagena -0.715 0.292 2.108*** 0.008 -2.276** 0.031 317 0 1 0.422 Manizales 0.096 0.803 4.047*** 0.000 -4.032*** 0.000 416 0 1 0.680 Pereira -0.440*** 0.008 2.263*** 0.000 -2.405*** 0.000 375 0 1 0.649 Cúcuta -2.545* 0.054 1.575* 0.086 -1.876 0.123 167 0.000369 1 0.990 30 Table A2. Estimates properties labor demand elasticities by subsector and size (2000-2013). Subsector/ size Long-run elasticities Properties Own price Output TFP Sample size Sargan test (p-value) Hansen test (p-value) Ar2(p-value) By subsector Permanent skilled Food products -0.736*** 0.000 1.182*** 0.000 -1.049*** 0.000 3146 0 0.269 0.305 Textiles -0.475*** 0.002 1.271*** 0.000 -1.074*** 0.000 709 0 1 0.632 Apparel -0.722*** 0.001 1.173*** 0.000 -0.958*** 0.000 640 0 1 0.557 Chemicals -0.693*** 0.000 1.708*** 0.000 -1.572*** 0.000 1856 0 0.204 0.754 Rubber and plastic -0.831*** 0.000 1.214*** 0.000 -1.282*** 0.000 1321 0 0.703 0.998 Mineral products -0.635*** 0.000 1.264*** 0.001 -1.022* 0.073 915 0 0.993 0.588 Metal products -0.711*** 0.000 1.150*** 0.000 -1.175*** 0.000 731 0 1 0.590 Furniture -0.413** 0.022 0.978*** 0.000 -1.048** 0.014 585 0 1 0.920 Temporary skilled Food products -0.365*** 0.004 0.635*** 0.007 -0.615** 0.039 1303 0 0.448 0.0821 Textiles -0.004 0.998 1.202 0.247 -0.493 0.674 276 0.0253 1 0.381 Apparel -0.189 0.440 0.703*** 0.008 -0.316 0.448 307 0.140 1 0.769 Chemicals -0.284* 0.066 1.063** 0.016 -0.991** 0.041 469 0 1 0.819 Rubber and plastic -0.440*** 0.007 0.845*** 0.000 -0.867*** 0.001 414 0 1 0.876 Mineral products -0.044 0.772 1.227* 0.054 -1.208 0.121 235 0.000137 1 0.992 Metal products -0.011 0.972 0.403 0.133 0.069 0.826 222 0 1 0.657 Furniture -0.290 0.383 1.242 0.113 -1.888 0.178 228 0 1 0.655 Permanent unskilled Food products -1.118*** 0.000 1.209*** 0.000 -1.160*** 0.000 4214 0 0.0535 0.326 Textiles -2.157*** 0.000 1.269** 0.032 -0.762 0.189 984 0 0.999 0.535 Apparel -1.347*** 0.000 1.079*** 0.000 -0.873*** 0.000 1059 0 0.933 0.554 Chemicals -0.915*** 0.000 1.180*** 0.000 -1.118*** 0.002 2003 0 0.201 0.438 Rubber and plastic -1.518*** 0.000 0.482* 0.060 0.074 0.829 1762 0 0.251 0.464 Mineral products -1.036*** 0.000 1.417*** 0.000 -1.101*** 0.005 1362 0 0.750 0.878 Metal products -1.164*** 0.000 0.849*** 0.000 -0.791*** 0.000 952 0 1 0.301 Furniture -1.115*** 0.000 1.019*** 0.000 -0.725** 0.024 876 0 1 0.160 Temporary unskilled Food products -0.204 0.192 1.317*** 0.000 -1.379*** 0.000 3109 0 0.736 0.653 Textiles -0.074 0.984 1.207 0.697 -1.364 0.622 824 0 0.231 0.672 Apparel -1.493 0.379 1.268 0.315 -1.335 0.355 1073 0 0.829 0.624 Chemicals -0.436* 0.055 1.247*** 0.001 -1.148*** 0.010 1274 0 0.366 0.345 Rubber and plastic -0.165 0.418 0.829*** 0.000 -1.068*** 0.000 1205 0 0.604 0.418 Mineral products 0.070 0.828 1.866*** 0.003 -2.063*** 0.010 904 0 0.991 0.972 Metal products -0.289 0.373 0.816*** 0.000 -0.165 0.427 628 0 1 0.0484 Furniture -0.275 0.539 1.137*** 0.000 -1.322*** 0.000 611 0 1 0.751 Note: the accompanying value of estimated coefficient corresponds to the p-value. Source: Dane-EAM; authors’ calculations. 31 Table A3. Estimates properties labor demand elasticities by subsector and size (2000-2013). Subsector/ size Long-run elasticities Properties Own price Output TFP Sample size Sargan test (p-value) Hansen test (p-value) Ar2(p-value) Permanent skilled Medium -0.514*** 0.000 1.081*** 0.000 -0.983*** 0.000 4937 0 0.595 0.816 Large -0.722*** 0.000 1.295*** 0.000 -1.105*** 0.000 8435 0 0.247 0.954 Very large -1.106*** 0.000 1.215*** 0.005 -1.113** 0.015 984 0 1 0.794 Temporary skilled Medium -0.382*** 0.000 0.769*** 0.004 -0.601* 0.073 1090 0 0.480 0.652 Large -0.170** 0.014 1.087*** 0.001 -0.914** 0.014 3433 0 0.591 0.806 Very large -0.591*** 0.000 -0.055 0.947 -0.029 0.973 529 0 1 0.426 Permanent unskilled Medium -1.008*** 0.000 0.436*** 0.007 -0.353* 0.086 8320 0 0.249 0.281 Large -1.150*** 0.000 0.986*** 0.000 -0.863*** 0.000 9850 0 0.370 0.536 Very large -1.018*** 0.000 1.503** 0.014 -1.722** 0.014 1023 0 1 0.158 Temporary unskilled Medium -0.351*** 0.002 0.750** 0.011 -0.906*** 0.006 4533 0 0.356 0.986 Large -0.231*** 0.002 0.356*** 0.007 -0.212 0.229 8365 0 0.130 0.0886 Very large -0.387 0.333 0.333 0.658 0.469 0.457 1007 0 1 0.590 Note: the accompanying value of estimated coefficient corresponds to the p-value. Source: Dane-EAM; authors’ calculations.