Report No.: 69116-BF Burkina Faso Determinants of Cereal Production Stochastic Frontier Approach for Panel Data A Policy Note 3b June 12, 2013 Poverty Reduction and Economic Management 4 Country Department AFCF2 Africa Region Document of the World Bank CURRENCY EQUIVALENTS (Exchange Rate Effective May 8, 2013) Currency Unit = CFA franc (CFAF) 1 US$ = CFAF 500 FISCAL YEAR January 1 – December 31 ABBREVIATIONS AND ACRONYMS CFA Communauté Financière d'Afrique CWIQ Core Welfare Indicator Questionnaire DES Dietary Energy Supply EBCVM Enquête Base sur la Condition des Vie des Ménages EICVM Enquête Intégrale sur les Condition de Vie des Ménages EP Enquête Prioritaire EPA Enquête Permanent Agricole FCFA Franc CFA FAO Food and Agriculture Organization GDP Gross Domestic Product HCPI Harmonized Consumer Price Index HDI Human Development Index HDRO Human Development Report Office INSD Institut National de la Statistique et de la Demographie LDC Less Developed Countries MDG Millennium Development Goals OLS Ordinary Least Squares SCADD Stratégie pour une Croissance Accélérée et une Développement Durable SSA Sub-Saharan Africa UNDP United Nations Development Programme Vice President: Makhtar Diop Country Director: Madani M. Tall Country Manager : Mercy M. Tembon Sector Director /Sector Manager : Marcelo Giugale Task Team Leader: Andrew Dabalen ii Table of Contents 1. DETERMINANTS OF CEREAL PRODUCTION IN BURKINA FASO STOCHASTIC FRONTIER APPROACH FOR PANEL DATA...................................................................................... 5 A. CONTEXT ................................................................................................................................................... 5 B. METHODOLOGY ........................................................................................................................................ 6 C. DATA DESCRIPTION ................................................................................................................................... 7 D. ESTIMATION RESULTS ............................................................................................................................. 14 E. CONCLUDING REMARKS .......................................................................................................................... 19 List of Tables: Table 1.1: Composition of surveyed households .......................................................................................... 7 Table 1.2: Share of cultivated land for main cereals (%) .............................................................................. 8 Table 1.3: Households and parcel characteristics ......................................................................................... 8 Table 1.4: Cereals revenue as share of total household revenue (%).......................................................... 13 Table 1.5: Likelihood-ratio test................................................................................................................... 14 Table 1.6a: Frontier results for panel (2002-2007) ..................................................................................... 15 Table 1.7b: Frontier results with labor (2007) ............................................................................................ 16 Table 1.8: Cereals land ............................................................................................................................... 17 Table 1.9: TFP decomposition .................................................................................................................... 18 List of Figures: Figure 1.1: Cereals yield by gender .............................................................................................................. 9 Figure 1.2: Cereals yield by age groups ...................................................................................................... 10 Figure 1.3: Cereals yield and parcel slope .................................................................................................. 10 Figure 1.4: Cereals yield by cropping systems ........................................................................................... 11 Figure 1.5a: Composition of labor force by cereals farming activities (%) ................................................ 11 Figure 1.6b: Composition of labor force by cotton farming activities (%) ................................................. 12 Figure 1.7: Cereals yield and share of hired labor ...................................................................................... 12 Figure 1.8: Share of males’ labor in cereals activities (%) ......................................................................... 13 Figure 1.9: Effects of provinces and time fixed effects .............................................................................. 18 Figure 1.10: Distribution of efficiency ....................................................................................................... 18 iii ACKNOWLEDGMENTS This policy note was prepared by a core team consisting of Andrew Dabalen (World Bank) and John Ulimwengu (IFPRI). Substantial inputs were also provided by Yele Batana (World Bank). Judite Fernandes provided excellent assistance with document preparation and finalization. The report was prepared under the guidance of Marcelo Giugale (Sector Director) and Miria Pigato (Sector Manager, AFTP4). We are grateful to the following for very helpful suggestions: Josefina Posadas, Nishta Sinha, Koffi Nouve, Ali Zafar, Mariam Diop, and the two peer Reviewers: Lire Ersado and Nobuo Yoshida. The Team would also like to thank the Agricultural Panel Survey unit in the Ministry of Agriculture for generously sharing the data sets used for the analysis in this note. iv 1. Determinants of Cereal Production in Burkina Faso Stochastic Frontier Approach for Panel Data A. CONTEXT 1.1 In the last ten years, Burkina Faso economy experienced an average annual growth rate of about 5 percent, with a peak of 7.4 percent in 2005. This growth bust is mainly driven by the performance in the agricultural sector, where growth averaged 5.2 percent over the past ten years. However, the dependence of agricultural sector to weather variability cannot guarantee a sustained growth trend. 1.2 Agriculture in Burkina Faso is almost exclusively extensive, and is practiced mainly on family farms (about 800,000 farms), dominated by small farms. The arable land is estimated at 9 million ha (approximately 30 percent of the total land area). Only 3.5 to 4 million hectares (one third) are actually cultivated annually. The main crops include cereals (sorghum, millet, maize, rice and fonio) which occupy nearly 88 percent of cultivated land every year; legumes and tubers which cover a very small area (about 2 percent); cash crops (cotton, sesame, groundnut, soybean and sugar cane) account for 9 percent of cultivated land; vegetables (mainly tomato, onion and green beans) and fruit (mainly mango). 1.3 Only 24,000 ha of land are irrigated for an irrigable potential of 233,500 ha. Irrigated crops include rice, sugarcane and vegetables. Market gardening is also practiced in irrigated areas and also in small individual gardens especially in suburban areas. In recent years, the country has been promoting irrigated maize (dry season) through small-scale irrigation. 1.4 During the past decade, grain production grew at an average rate of 4.6 percent per year led by maize which experienced a steady growth of 10.7 percent on average annually and millet which grew by 5.9 percent. Average cereal production was about 2.7 million tons during that period. Still, cotton remains the main cash and export crop of the country; it accounts for 2/3 of export earnings of the country. 1.5 Sorghum, millet, and maize are the main cereals produced and consumed throughout the country. The cereals share in total dietary energy consumption is around 73 percent (FAO, 2010). Sorghum and millet accounted for 26 percent and 22 percent, respectively, of the total dietary energy supply (DES) in 2005-07 (FAO-GIEWS database). On average in 2004-08 per capita consumption of sorghum was 83 kg/year, and that of millet was 62 kg/year. The self-sufficiency ratio of sorghum is 105 percent and 103 percent for millet. Bulk of rice is imported and mostly consumed in urban areas. Rice accounted for 6 percent of the total dietary energy supply (DES) in 2003-05. On average in 2004-08 per capita consumption (as food) of rice was 21 kg/yr. The self-sufficiency ratio of rice is about 24 percent (FAO-GIEWS database). 1.6 The National Strategy for Food Security by 2015 aims to create favorable conditions for sustainable food security and inequality and poverty reduction in Burkina Faso. Endorsing the guidelines of the World Summit on Food in Rome in 1996, the Government has set a target to reduce by 50 percent the number of people suffering from hunger and malnutrition 5 by 2015. The specific objectives of the strategy include i) increase the level of domestic food production; ii) strengthen the capacity of the market to improve people's access to food; iii) improve the economic conditions of the poor and vulnerable groups; iv) improve the prevention and management of cyclical food crises; and v) strengthen the capacities of stakeholders and promote good governance for food security. 1.7 The main purpose if this paper is to analyze factors driving cereals production in Burkina. More specifically, this study intends to generate research-based evidence to guide the design and implementation of strategies to improve the productivity of cereals production for poverty reduction and food security. To do so, we use farmers level panel data collected from 1994 to 2007. 1.8 Our findings confirm that land is indeed the dominant driver of cereals production in Burkina. Like overall agricultural production system in Burkina, cereals production is also rain-dependent. This should be a concern as the country's economic growth depends on agricultural sector, which itself remains highly dependent on climate variability. Finally, our results suggest that cereals production system in Burkina is characterized by significant provinces and time heterogeneity. B. METHODOLOGY 1.9 We follow the methodology developed by Battese and Coelli (1995) and Kumbhakar and Lovell (2000) as described below. The production function is given as follows: (1) 𝑞𝑖𝑡 = 𝑓(𝑥𝑖𝑡 , 𝑡; 𝛽 )𝜀𝑖𝑡 𝑒𝑥�(�𝑖𝑡 ) where i = 1,..., N , represents agricultural households, qit is a production vector (n × 1) at time period t, xit is the input vector (1 × k ) , β is the (k × 1) vector of parameters to be evaluated, 𝑡 represents time trend and ε it represents the efficiency of household i, with 0 < ε it ≤ 1 . If ε it = 1, then household i is achieving optimal production with respect to technology f ( xit , β ) . However, if 0 < ε it < 1 , then household i is failing to gain the maximum possible benefit from input (xi), which means the production level is sub-optimal. Furthermore, the production of household i is 2) also affected by random shocks, �𝑖𝑡 ~� + (0, 𝜎� . In logarithmic form, equation (6) can be formulated as follows: (2) 𝑙𝑛𝑞𝑖𝑡 = 𝛽0 + ∑𝑘−1 𝑗=1 𝛽𝑗 𝑙𝑛𝑥𝑖𝑗𝑡 + 𝑙𝑛𝜀𝑖𝑡 + �𝑖𝑡 . Let 𝑢𝑖𝑡 = −𝑙𝑛𝜀𝑖𝑡 , it follows that, (3) 𝑙𝑛𝑞𝑖𝑡 = 𝛽0 + ∑𝑘−1 𝑗=1 𝛽𝑗 𝑙𝑛𝑥𝑖𝑗𝑡 − 𝑢𝑖𝑡 + �𝑖𝑡 . The growth rate of output into contribution from the growth of inputs versus productivity change is given by 6 (4) 𝑞̇ 𝑖𝑡 = ∑𝑘−1 ̇ 𝑗=1 𝛽𝑗 𝑥̇ 𝑖𝑗𝑡 + 𝑇𝐹𝑃𝚤𝑡 , where TFP is the Total Factor Productivity which TFP growth can be decomposed into technical change (TC) and technical efficiency (TE). Technical change is defined as the marginal change in output with respect to time trend 𝜕𝑓(𝑥 ,𝑡;𝛽 ) (5) 𝑇𝐶 = . 𝜕𝑡 Technical efficiency is defined as the ratio of observed output for the i-th farmer relative to its potential output and given by 𝑞𝑖𝑡 (6) 𝑇𝐸𝑖𝑡 = exp[𝑓(𝑥 = exp(−𝑢𝑖𝑡 ), 𝑖𝑡 ,𝑡;𝛽 )+𝑣𝑖𝑡 ] 2 where 𝑢𝑖 ~� + �𝜇, 𝜎𝜇 �. C. DATA DESCRIPTION 1.10 The agricultural panel survey Enquête Permanent Agricole (EPA) is conducted by the Ministry of Agriculture. The survey is administered to farm households since 1994, constituting two panels at the household level: from 1994/95 to 2000/01 (30 provinces) and from 2001/02 to 2007/08 (45 provinces). The variation in the number household across the years reflects attrition as well as the new households incorporated in the data set as shown in Table 1.1. Attrition rate varies between zero percent in 2005 and 100 percent in 2001. The roster is not available for 1994/95 and 1995/96 have been wrongly merged (Himelein, 2009). Therefore, we use only the 2001/02-2007/08 panel to estimate equation (3). Table 1.1: Composition of surveyed households Year Total of HHs HHs re- HHs re- Attrition New HHs interviewed the interviewed 1 following year year or more after 1996 4,535 4,535 1997 4,799 3,640 0 895 1,159 1998 5,057 4,305 102 596 854 1999 4,855 4,411 273 919 717 2000 4,801 4,199 477 1,133 1,079 2001 3,762 7 1 4,795 3,756 2002 3,807 3,325 4 441 486 2003 3,381 3,258 50 599 173 2004 3,970 2,458 332 1,255 1,844 2005 3,971 3,970 0 0 1 2006 3,916 3,431 182 722 667 2007 4,264 3,791 371 496 844 Source: Nistha and Josefina, World Bank, mimeo. 7 1.11 The survey uses village level sampling in 2-stratus. Villages are selected first according to their production potential (small land holdings high output vs. large land holding lower output) and some households’ characteristics. In the second stage, farm HHs are selected based on a list of 17 variables such as land area for cereals, land area for other cash crops, number of HH members, gender composition, number of HH members farming, etc., and also dividing HHs in two groups (defined from those 17 variables). 1.12 In this paper, we focus on main cereals (millet, maize, rice, fonio, and sorghum); as shown in Table 1.2, millet is the most important followed by maize and sorghum. Millet and maize alone account for at least 80 percent of land allocated to cereals production. Table 1.2: Share of cultivated land for main cereals (%) 2002 2003 2004 2005 2006 2007 Millet 58.0 57.4 54.7 54.8 54.0 54.5 Maize 24.7 26.1 29.8 28.7 29.8 29.3 Rice 2.0 1.8 1.8 1.8 1.9 1.6 Fonio 0.3 0.3 0.5 0.5 0.3 0.7 Sorghum 15.0 14.4 13.2 14.2 13.9 14.0 All 100.0 100.0 100.0 100.0 100.0 100.0 Source: Authors’ computation. 1.13 More than 90 percent of surveyed cereals growers were headed by males (Table 1.3). This is in sharp contrast with FAO reports indicating that women produce over 50 percent of all food grown worldwide (FAO, 2008a) and grow 80-90 percent of the food in Sub-Saharan Africa (FAO, 2008b). However, according to González et al. (2011), in Burkina Faso women cultivate cereals, but do not make any decisions. Our results show that average age of cereals growers is around 49 years old varying from 44 to 51 years old. Table 1.3: Households and parcel characteristics Households Parcel characteristics characteristics Male No anti-erosion Only Age (years) Plain/Plato (%) scheme cereals 2002 94.8 50.9 96.3 98.0 94.6 2003 95.5 43.7 97.0 97.2 95.6 2004 95.4 43.6 96.9 97.5 75.3 2005 94.0 50.9 96.6 98.1 79.1 2006 93.0 51.0 97.1 97.9 NA 2007 93.4 51.2 97.1 97.5 77.6 Source: Authors’ computation. 1.14 On average, 97 percent of cereals growers are farming in the plains/Plato. More than 97 percent of cereals farmers report not using any anti-erosion techniques. Anti- erosion techniques include the construction of small dikes of stones (rock bunds) to slow down run-off water flows, improve water infiltration in the soil and to reduce loss of top soil through wind and water erosion (Rochette, 1989). Vitale et al. (2011) report that the use of marginal 8 lands has increased environmental concerns since the newly introduced lands are typically located in environmentally sensitive areas where exposed soils are prone to erosion and degradation. In 2002 and 2003, the vast majority of farmers practiced mono-cropping; however, since 2004 at least 20% of growers are practicing intercropping. 1.15 The trend of cereals yield in Burkina has been quite erratic over the 2002-2007 period; from 660.4 kg/ha in 2002, the yield increased to 703.9 kg/ha in 2003 before dropping to 526.4 kg/ha in 2007. On average, males headed households were more productive than female headed households; males produced 615.8 kg/ha against 541.4 kg/ha for females. It is only in 2005 that females farmers achieved higher yield than their males counterparts, 787.8 kg/ha against 757.1 kg/ha (Figure 1.2). The observed low yield among females may be explained by the fact that in general women work on land of poor quality, often land that has been left fallow and produces a lower yield (González et al., 2011). In addition, women do not have the physical strength needed to use soil-conservation techniques which means that their land suffers more damage when there are floods or heavy rain. 1.16 On average, cereals growers over 50 years perform better than their counterparts of age between 15 and 50 years; the first group produced 630.3 kg/ha compared to 591.3 kg/ha for the second group. However, it is worth noting that the yield gap between the two groups is significant only in 2005 (Figure 1.3). Figure 1.1: Cereals yield by gender 900.0 Females Males 800.0 700.0 600.0 Yield (kg/ha) 500.0 400.0 300.0 200.0 100.0 0.0 2002 2003 2004 2005 2006 2007 Period Source: Authors’ computation. 9 Figure 1.2: Cereals yield by age groups 900.0 15-50 years 50+ years 800.0 700.0 600.0 Yield (kg/ha) 500.0 400.0 300.0 200.0 100.0 0.0 2002 2003 2004 2005 2006 2007 Period Source: Authors’ computation 1.17 From 2002 to 2004, cereals yield produced in plains is higher than the one in other areas (Figure 1.4); however, from 2005 to 2007 the trend shifted in favor of cereals cultivated in areas other than plains and plato. With respect to cropping system, there is a significant difference in cereals yield between intercropping (mixed) and mono-culture systems (Figure 1.5); on average, farmers produced 857.5 kg/ha when they mix cereals with other crops compared to 588.2 kg/ha when cotton is planted alone. Osman et al. (2011) argue that intercropping of pearl millet with cowpea can improve crop production without increasing input levels. Figure 1.3: Cereals yield and parcel slope 900.0 Other Plain/Plato 800.0 700.0 600.0 500.0 Yield (kg/ha) 400.0 300.0 200.0 100.0 0.0 2002 2003 2004 2005 2006 2007 Period Source: Authors’ computation. 10 Figure 1.4: Cereals yield by cropping systems 1400.0 Mixed Alone 1200.0 Yield (kg/ha) 1000.0 800.0 600.0 400.0 200.0 0.0 2002 2003 2004 2005 2007 Period Source: Authors’ computation. 1.18 Only the 2007/2008 survey rounds collected information on labor. As expected, like cotton (see Table 1.4b) cereals labor force is also dominated by family labor which accounts for 77.0 percent of total compared to 17.0 percent for loaned labor and 6.0 percent for hired labor (Tables 4a). Unlike cotton, the dominance of family labor is observed in all activities, reaching 91.8 percent during the sowing season but reduced to 69.4 percent (crop management) and 73.5 percent (harvesting). We also found that cereals yield increases with the shares of hired and loaned labor (Figure 1.6) but decreases with the share of family labor. The share of hired labor in cereals production is much lower than in cotton production; cotton being an export crop requires probably more qualified labor from the market. As expected (Figure 1.7), males' participation in cereals production is more important in plowing season (80.9 percent) and crop management season (59.6 percent) but less so during the sowing season (39.9 percent) and harvesting season (46.6 percent). Figure 1.5a: Composition of labor force by cereals farming activities (%) Plowing Sowing Crop management Harvesting All Family 77.6 91.8 69.4 73.5 77.0 Loan 17.6 6.8 22.4 18.7 17.0 Hired 4.8 1.4 8.2 7.8 6.0 Total 100.0 100.0 100.0 100.0 100.0 Source: Authors’ computation. 11 Figure 1.6b: Composition of labor force by cotton farming activities (%) Plowing Sowing Crop management Harvesting All Family 81.4 86.5 57.0 32.6 51.4 Loan 13.8 9.0 23.1 30.7 24.0 Hired 4.8 4.5 19.9 36.7 24.6 Total 100.0 100.0 100.0 100.0 100.0 Source: Authors’ computation. Figure 1.7: Cereals yield and share of hired labor 1400 1300 Hired labor 1200 Yield (kg/ha) Family labor Loaned labor 1100 1000 900 0 .2 .4 .6 .8 1 Share of labor Source: Author’s computation 12 Figure 1.8: Share of males’ labor in cereals activities (%) 90.0 80.0 Male labor as share of total (%) 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 Plowing Sowing Crop Harvesting All management Farming activities Source: Author’s computation 1.19 In this study, we use the distribution of crop revenue for 2006 and 2007 to estimate the dependency rate as the ratio of cotton revenue and overall crop revenue (Table 1.5). Our results suggest that the average dependency on cereals revenue is around 20 percent compared to about 60 percent for cotton. The dependence on cereals revenue varies across regions; the most dependent regions include Centre, Sahel and Sud Ouest. Table 1.4: Cereals revenue as share of total household revenue (%) 2006 2007 Centre 44.2 44.4 Nord 10.1 8.7 Centre Sud 20.4 28.8 Centre Ouest 25.7 25.8 Mouhoun 14.8 18.3 Est 34.0 26.8 Centre Est 29.4 36.5 Sahel 40.8 51.5 Centre Nord 19.9 24.3 Cascades 8.0 8.5 Hauts Bassins 16.9 13.2 Sud Ouest 43.8 24.7 Plateau central 34.8 34.1 Average 20.8 18.9 Source: Authors’ computation 13 D. ESTIMATION RESULTS 1.20 In Table 1.7 we present three different specifications: i) without fixed effects (I); ii) with provinces fixed effects (II); and iii) with provinces and time fixed effects (III). The results suggest significant provinces and time heterogeneity as shown by the Likelihood-ratio test reported in Table 1.6. For example, estimates for fertilizer and rainfall change dramatically from the model with no fixed effects to the one with both provinces and time fixed effects. As a result, the discussion below will focus on the specification with provinces and time fixed effects. Table 1.5: Likelihood-ratio test LR Chi2 Degree of freedom p-value (I) against (II) 713.8 42 0.00 (II) against (III) 987.8 84 0.00 (I) against (III) 273.9 42 0.00 Source: Authors’ computation. 1.21 As reported in Table 1.6a, land appears as the dominant driver of cereals production in Burkina; elasticity of cereals production with respect to land is estimated at 1.61. However land elasticity is reduced to 1.19 with provinces fixed effects and 1.20 without fixed effects when accounting for labor (see Table 1.6b). Impact of land on cereals production is mainly the result of increase in total land allocated to cereals production. Over the 2002-2007 period, total cultivated land has indeed increased steadily from 8526.8 ha in 2003 to 11283.4 ha in 2007 (Table 1.8). However, neither the average land holding nor the share of cereals land over total land has significantly changed; average land holding remains around 2 ha and the share of cereals land around 43 percent of total cultivated land. 14 Table 1.6a: Frontier results 1 for panel (2002-2007) With provinces fixed With provinces and No fixed effects effects time fixed effects Dependent variable=cereals production (kg) Coefficient S.E. Coefficient S.E. Coefficient S.E. a a a Land 1.5489 0.0131 1.6123 0.0135 1.6123 0.0134 Fertilizer (kg) 0.1054a 0.0117 0.0626a 0.0122 0.0511a 0.0125 a a a Fertilizer squared -0.0258 0.0033 -0.0172 0.0034 -0.0127 0.0034 a c Rainfall (mm) 0.6231 0.0528 0.0719 0.0839 0.1736 0.0966 Age of household head (years) 0.0042 0.0053 0.0067 0.0052 0.0064 0.0052 Age squared 0.0000 0.0001 -0.0001 0.0000 -0.0001 0.0000 Gender of household head (1 if female, 0 if male) -0.0532 0.0641 -0.0404 0.0632 -0.0494 0.0630 a c Slope (1 if plain/Plato, 0 otherwise) 0.2492 0.0752 0.1137 0.0744 0.1292 0.0746 a a a Time -0.1226 0.0093 -0.1063 0.0093 -0.1553 0.0551 Intercept 0.9962b 0.5020 5.9775a 0.7480 5.2098a 0.8504 2 𝜎 50.89 a 16.60 71.37 b 32.54 483.57 576.44 𝛾 0.95 a 0.02 0.97 a 0.02 0.99 a 0.01 2 𝜎𝑢 48.39 a 16.60 68.89 b 32.54 481.14 576.44 2 𝜎𝑣 2.50 a 0.03 2.48 a 0.03 2.43 a 0.03 Observations 17527 17527 17527 Log likelihood -34190.0 -33833.0 -33696.1 Wald chi2(51) (51) 16091.1 (51) 18191.4 (93) 18628.4 Notes: a,b and c means significant at 1%, 5% and 10% respectively; S.E.=standard error 1 Production and inputs variables are all in log form 15 Table 1.7b: Frontier results with labor (2007) No fixed effects With provinces fixed effects Dependent variable=cereals production (kg) Coefficient S.E. Coefficient S.E. a a Land 1.1998 0.0176 1.1850 0.0179 Family labor (female) -0.0058 0.0084 -0.0178b 0.0079 a Family labor (male) 0.0113 0.0109 0.0390 0.0096 a Loaned labor (female) 0.0290 0.0073 0.0087 0.0071 Loaned labor (male) -0.0049 0.0066 -0.0012 0.0063 Hired labor (female) 0.0220b 0.0107 0.0268a 0.0102 c Hired labor (male) 0.0150 0.0078 0.0039 0.0072 c Fertilizer (kg) 0.0044 0.0077 0.0139 0.0076 c Fertilizer squared 0.0027 0.0027 -0.0042 0.0026 Age of household head (years) -0.0018 0.0050 -0.0011 0.0045 Age squared 0.0000 0.0000 0.0000 0.0000 b Gender of household head (1 if female, 0 if male) -0.1527 0.0598 0.0108 0.0575 b Slope (1 if plain/Plato, 0 otherwise) 0.1009 0.0506 0.0075 0.0485 Anti-erosion scheme (1 if not, 0 otherwise) -0.1191b 0.0548 -0.0645 0.0509 a a Intercept 7.6330 0.1503 7.1524 0.1584 a a sigma_v 0.2559 0.0108 0.2067 0.0101 a a sigma_u 1.7819 0.0232 1.7132 0.0219 sigma2 3.2407a 0.0818 2.9778a 0.0742 a a lambda 6.9635 0.0274 8.2889 0.0257 Observations 3809 3809 Log likelihood -5337 -5116 Wald chi2(DF) 7489 (14) 8928 (58) 16 Table 1.8: Cereals land Average land Share of land holding allocated to cereals Total land 2002 2.5 43.2 9755.9 2003 2.5 43.4 8526.8 2004 2.2 41.1 8616.6 2005 2.1 42.0 9242.2 2006 2.4 44.5 10694.5 2007 2.6 42.5 11283.4 All 2.4 42.8 58119.3 Source: Authors’ computation. 1.22 The marginal impact of fertilizer use on cereals production is significant but non- linear; this reflects the fact that a minimum level of fertilizer use is required before observing increase in cereals production as a result of increased fertilizer use. Fertilizer use for cereals over 2002-2007 period was erratic and rather negligible; 0.03 kg/ha in 2002, 98.2 kg/ha in 2005, and 27.3 kg/ha in 2007. In Brazil, Lopes et al. (2003) report average fertilizer use of 83 kg/ha for rice and 119 kg/ha for maize. Many factors can explain limited fertilizer use in cereals production in Burkina, including high price, road infrastructure and weather instability. 1.23 Like overall agricultural production system in Burkina, cereals production is also rain-dependent. Indeed, a ten percent increase in rainfall is expected to increase cereals production by 17.4 percent. Stroosnijder and Rheenen (2001) contend that crop performance in the Sahel is limited by both water and nutrient availability. The country's economic growth depends on agricultural sector, which itself remains highly dependent on climate variability. Since 1970, the country suffered from five major drought episodes; as a result, poor rainy seasons led to a slowdown in economic growth, whereas good rainfall caused acceleration in economic growth. 1.24 The estimation results for 2007 2 (he only period for which labor data are available) confirm that different types of labor have different impacts on production. Indeed, using estimates with provinces fixed effects, only hire d female and family male labor are positive and significant (Table 1.6b). The marginal productivity of family female labor is significant but negative. 1.25 As mentioned earlier, cereals production system in Burkina is characterized by significant provinces and time heterogeneity. Results reported in Figure 1.8 suggest that accounting for provinces and time fixed effects almost double the estimates of efficiency. Therefore, only results from specification with provinces and time fixed effects. As reported in Table 1.9, cereals production in Burkina experienced negative TFP growth driven by negative technical change. On the other hand, efficiency estimates increased on average from 0.651 in 2002 to 0.661 in 2007; the complete distribution of efficiency estimates is reported in Figure 1.9. On average, efficiency estimates range from 0.594 in the province of Yagha to 0.733 in Bazega. 2 The only year for which labor data are available 17 Figure 1.9: Effects of provinces and time fixed effects 0.6700 No fixed effects Provinces fixed effects Provinces and time fixed effects 0.6600 0.6500 Efficiency estimates 0.6400 0.6300 0.6200 0.6100 0.6000 0.5900 2002 2003 2004 Period 2005 2006 2007 Source: Authors’ computation. Table 1.9: TFP decomposition TFP TC TE 2003 -15.4 -15.5 0.1 2004 -14.6 -15.5 0.9 2005 -15.2 -15.5 0.3 2006 -15.3 -15.5 0.2 2007 -15.5 -15.5 0.0 Source: Authors’ computation. Figure 1.10: Distribution of efficiency 6 2007 2005 4 Kernel density 2002 2 0 .2 .4 .6 .8 1 Agricultural efficiency Source: Author’s computation. - 18 - E. CONCLUDING REMARKS 1.26 Cereals production in Burkina mirrors the characteristics of the overall agricultural sector; it has been growing through land expansion and is highly rain-dependent. As a result, yield growth has been erratic following weather instability. Our results also suggest that productivity as measured by yield is negatively correlated with the share of family labor. This should be a concern as the labor force involved in cereals production is dominated by family labor which accounts for 77.0 percent of total compared to 17.0 percent for loaned labor and 6.0 percent for hired labor. Cereals sector is affected by negative TFP growth as a result of negative technical change and insignificant efficiency change. Differences in production function across provinces are significant suggesting a geographical targeting approach in addressing the sector’s weaknesses. 1.27 To strengthen cereals production system and probably the whole agricultural sector in Burkina Faso, the government needs to: i) reverse fertility loss and resource degradation, ii) improve management of water resources while expanding access to both small- and large-scale irrigation; and iii) improve agricultural research with clear strategy on the dissemination and adoption of drought-resistant technology. - 19 - References Battese, GE & Coelli, TJ, 1995. A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empirical Economics 20(2), 325–32. Food and Agriculture Organization of the United Nations (FAO). 2008a. “Climate Change, Biofuels, and Land.� FAO. ftp://ftp.fao.org/nr/HLCinfo/Land-Infosheet-En.pdf (acessed May 20th, 2012). Food and Agriculture Organization of the United Nations (FAO). 2008b. “Climate Change, Biofuels, and Land.� FAO. ftp://ftp.fao.org/nr/HLCinfo/Land-Infosheet-En.pdf (acessed May 20th, 2012). González, A. R., A. Belemvire and S. Saulière 2011. Climate change and women farmers in Burkina Faso: Impact and adaptation policies and practices. Oxfam Research Reports. Himelein, K. 2009. Burkina Faso: Review of EPA and Recommendations for PAPSA M&E. Back to Office Report . Kumbhakar, SC & Lovell, CAK, 2000. Stochastic Frontier Analysis. Cambridge University Press, Cambridge. Lopes, A.S., Guilherme, L.R.G. & Silva, C.A.P. 2003. da. Vocação da terra., ANDA, 2nd ed. São Paulo, Brazil. 23 pp. Osman, A. N., A. Ræbild, J. L. Christiansen and J. Bayala. 2011. Performance of cowpea (Vigna unguiculata) and pearl millet (Pennisetum glaucum) intercropped under Parkia biglobosa in an agroforestry system in Burkina Faso. African Journal of Agricultural Research 6(4): 882-891. Rochette, R.M. (ed.). 1989.Le Sahel en lutte contre la desertification: leçons d'experiences. Eschborn, Allemagne. Vitale, J.D, M. Ouattarra and G. Vognan. 2011. Enhancing Sustainability of Cotton Production Systems in West Africa: A Summary of Empirical Evidence from Burkina Faso. Sustainability 3, 1136-1169. - 20 -