Importantly, rather than resting on an inverse farm size – productivity relationship, policy that seeks to impact both equity and efficiency should focus on ensuring that smallholders have access to the productivity gains experienced by their larger counterparts. Thus, policies that help build human capital, facilitate adoption of new technologies, and enhance access to markets via a reduction in transactions costs will continue to be indispensable for reducing rural poverty in developing countries.The regularity with which an inverse relationship between farm size and land productivity is observed led to many theoretical explanations for the phenomenon. Early explanations centered around multiple market failures , asymmetric information , and risk aversion among farmers . A second set of explanations emphasized empirical issues such as systematic measurement error in farm size and/or output and omitted variables . Empirical studies have typically found that existing theory fails to fully explain the observed inverse relationship, generating a body of mixed and at times contradictory evidence. Chapter 1 illustrates how the choice of productivity measure can alter the relationship observed and how it can obscure a changing relationship between farm size and total factor productivity, the more relevant productivity measure. A dynamic relationship was found between farm size and total factor productivity in the rapidly modernizing agricultural regions of Brazil, contributing to an emerging literature that documents changing farm size – productivity relationships as agricultural sectors modernize and develop . This is consistent with Helfand et al. , hydroponic bucket whose findings suggest that both the larger commercial farms and smaller family farms in Brazil have advantages in harnessing technical change and achieving rapid gains in productivity.

In this paper the hypothesis of a dynamic farm size – productivity relationship is extended to the context of Mexico, identifying the relationship in a panel of family farms from the Mexican Family Life Survey and testing for changes over the sample period of 2002-2009. Mexico is an interesting case for assessing changes in the farm size – productivity relationship because of its long history of land reform and the recent agricultural policy reform associated with the North American Free Trade Agreement in the 1990s. These policies are a prime example of the Washington Consensus, liberalizing markets for land, agricultural inputs, and agricultural output in Mexico with the objective of spurning the modernization, competitiveness, and productivity of the agricultural sector and the broader economy. An environment with such market reforms, if successful, is expected to diminish the multiple market failure explanation of the inverse relationship between farm size and productivity, and any observed inverse relationship might weaken accordingly. We test for changes in the farm size – productivity relationship and, contrary to expectations, find that an inverse relationship exists and has remained strong in the wake of Mexico’s market reforms. We explore the relationship further by estimating a stochastic production frontier, an approach often applied in developed economy agriculture but infrequently applied in developing economy contexts. While frontier productivity growth has increased rapidly for larger farms, eliminating the inverse relationship at the frontier, the average relationship has remained unchanged due to more rapidly increasing technical inefficiency amongst the larger farms in the sample. This finding highlights the need for, and echoes calls for, policies that support family farms’ transitions towards modern agriculture and adaptation to market liberalization in Mexico.

The proceeding section discusses agricultural policy in Mexico, providing context for the empirical analysis. This is followed by an introduction of the empirical methodology, a description of the data, and the presentation of empirical results. Policy recommendations for Mexican agriculture and research implications conclude.The institutional structure of Mexican agriculture continues to reflect agricultural policies implemented after the Mexican Revolution of the early 20th century. Land policy of the 1934 Agrarian Code established the ejidos – tracts of communally held land with individual plots farmed by designated households – as a principle tool for redistributing land and property rights to peasants. Agrarian communities, a distinct form of land tenure located predominantly in the South where farmers had pre-existing claims to agricultural land, were similarly formed although to a lesser extent. A dual system of agricultural tenure emerged, with ejido farmers on the one hand and private landowners on the other. Within both the ejido and private farm sectors there exists both the larger, commercially oriented farms and the smaller predominantly subsistence farms. It is in this context that Berry and Cline first studied the farm size – productivity relationship in Mexico. Drawing from the Mexican Agricultural Census of 1940 and of 1960, they compared land productivity of small and large private farms. They found land productivity of small farms to be 6.5 times larger than that of larger farms in 1940, but just 3.5 times as large by 1960. More importantly, when output per unit of land and capital was measured, a more comprehensive measure of productivity, small farms were 1.7 times more productive than large farms in 1940 but just 0.8 times as productive in 1960. This early evidence illustrates that an inverse relationship between farm size and land productivity is neither necessary nor sufficient for an inverse relationship between farm size and more comprehensive productivity measures, similar to the findings of chapter 1 in the context of Brazil.

Berry and Cline hypothesized that the changing productivity ratios reflected a shift from livestock to crops on large farms, facilitated by government investment in infrastructure, provision of credit, and other supportive policies. As the birthplace of the Green Revolution, Mexican agriculture experienced productivity growth throughout this period, notably becoming net exporters of important staples such as wheat and maize. A weakening of the IR between farm size and land productivity accompanied this period of agricultural modernization and development, as did a reversal of the IR between farm size and output per unit of capital and labor. More recent research using farm-level panel data from the Mexico National Rural Household Survey , a household survey statistically representative of 80% of rural Mexico, showed evidence of an inverse relationship between farm size and productivity in 2003 and 2008 . By estimating an average production function and a stochastic production frontier, they find an inverse relationship between farm size and land productivity, farm size and average TFP, and farm size and TFP along the production frontier. They conclude that the observed farm size – TFP relationship was driven, in part, by larger farms being further from the frontier . Mexican agriculture in the early 20th century is an interesting setting for studying the farm size – productivity relationship because of the policy changes and market reforms associated with The North American Free Trade Agreement going into effect in 1994. As part of an economy-wide reduction in tariffs, agricultural tariffs were gradually eliminated over a 14-year span ending in 2008. The liberalization of agricultural trade exposed the Mexican agricultural sector to increased competition and imports from Northern neighbors. As a result, a flood of cheap imports has led to a decline in the price of staple crops for many Mexican farmers . For Mexican agriculture, NAFTA was part of a broader program of reform and market liberalization. One important change was the Program for the Certification of Ejido Rights and Titling of Urban Plots , which included reform of the ejido system of land tenure. Following a constitutional amendment, Procede facilitated the new option for ejidos to privatize individual parcels that could then be mortgaged, rented, or sold. Further, agricultural rights to ejido parcels ceased being contingent upon actual agricultural production, strengthening property rights for the ejido sector. Importantly for the private sector, stackable planters the practice of expropriating large private holding for the formation of ejidos was ended. By securing property rights and integrating ejidos into the market, these changes were expected to increase opportunities throughout the rural farm sector. A World Bank evaluation of the ejido reforms found that, while Procede had been widely successful in securing property rights, often in the form of certificates of agricultural rights, the program had not led to widespread land transfers and ejido farms remained credit constrained at the turn of the century. A second set of policy changes affected the manner in which government supported agricultural input and output markets. Policy shifted away from heavily subsidizing inputs and providing price supports for output towards a system of direct transfers for those impacted by increased international competition. In general, producers of staple products have suffered due to increased competition with relatively cheap imports whereas exports of high-valued horticultural products have benefited . The Program for Direct Assistance in Agriculture , primarily an income support program, offered per hectare payments to any farms with a history of producing any of nine key staples prior to 1993 that were actively farming one of those crops. An important change to the program in 1995 allowed participation of any farm producing any legal crop that had previously qualified for the program.

Further changes to the program in 2001 included higher per-hectare payments for farms under 5 hectares and a shift of the timing of payments to the start of the planting season. Alongside Procampo is Alianza para el Campo, a suite of programs designed to increase agricultural productivity primarily through investment in infrastructure and extension assistance. As government programs withdrew, farms became increasingly reliant upon markets for access to key agricultural inputs such as fertilizer, pesticides, and seed and for access to credit. Although government credit programs have scaled back, well functioning credit markets have not appeared in rural Mexico and access to credit markets is not widespread, inhibiting access to key inputs and modern technology. As in other developing country contexts, market concentration in both input markets and post harvest processing and marketing has hurt the profitability of family farms and generated economies of scale in transacting with the agricultural supply chain. We hypothesize that the farm size – TFP relationship is likely to be changing over time in the wake of Mexico’s NAFTA-era reforms, much as it appears to have done in Mexico during the Green Revolution and in Brazil’s modernizing agricultural regions . This hypothesis rests upon the assumptions that market imperfections contribute to any pre-existing IR in Mexican agriculture and Mexico’s NAFTA-era market liberalization has improved the efficiency of agricultural input and output markets. Beyond the farm size – productivity relationship, agricultural productivity is important to Mexico for both rural economic development and poverty reduction. According to data from the World Bank,2 agricultural output made up 3.6% of Mexico’s GDP but employed 13-14% of the workforce in 2015. Further, approximately62% of Mexico’s rural population is impoverished when using the national poverty line, suggesting that agricultural productivity has a potentially important role in Mexico’s rural economic development and efforts to reduce poverty. There are similar implications for trends in migration, as increasing agricultural productivity on family farms facilitates the ability to generate adequate livelihoods and effectively support families, reducing an important push factor in migration decisions.As discussed in chapter 1, land productivity is a partial measure of productivity and does not account for the use of inputs other than land. Where other inputs are used in production, failing to account for the use of those resources potentially introduces bias into estimates of the relationship between farm size and productivity if the intensity of input use varies with farm size. Controlling for all inputs in agricultural production can be accomplished with estimation of a production function, uncovering TFP, the comprehensive and preferable measure of productivity. We use two complementary approaches to explore the relationship between farm size and TFP with a panel of Mexican farms. First, we use an average production function to estimate average TFP and its relationship with farm size. Second, we use a stochastic production frontier to estimate both TFP along the technological frontier and technical inefficiency, identified as deviations from the frontier. The frontier analysis identifies any relationship between farm size and frontier TFP and any relationship between farm size and technical inefficiency. As is standard in the literature , we view TFP change as a combination of changes in the technological frontier and changes in the deviations from the frontier.