Effects of policy at the
household and village levels

George Dyer and J. Edward Taylor

Introduction

Economic activity in developing rural areas is largely based on household-farms and firms. As a result, households’ responses to policy largely determine i) effects on the well being of the household; ii) spill over effects on other members of the rural population, and iii) consequences on aggregate regional and national economic variables, such as market supply and international trade (Singh et al 1986). Since household-farms consume part of their production and provide part of their own production inputs, their behaviour is relatively complex. Moreover, household decisions on what to buy or sell depend on the market context in which they are embedded. Agricultural household models (Singh et al 1986) and village-wide models (Taylor and Adelman 1996) reflect efforts to capture these complexities.

Household responses

An idealized household that produces based entirely on its own means and consumes only what it produces - an autarchic household-simultaneously decides on production and consumption levels: one dictates the other. Household decisions reflect only endowments, technology and preferences; the marginal valuation of goods and factors is therefore subjective and household - specific. Autarchic households are rare because households gain from trade among themselves as long as their endowments, technologies or preferences show some heterogeneity and the cost of trade transactions is low. Even if markets are not available, pairs of households can benefit from individualized exchanges of goods and factors. The production and consumption decisions of bargaining households are determined simultaneously in this process. The terms of trade reflect households’ preferences, technologies, endowments and bargaining abilities: each household’s marginal values of goods and factors adjust in the process, but they do not necessarily equate those of other households; ‘prices’ may vary across space, creating inefficiency and welfare losses. In a situation where markets are complete and households are price takers, households plan production optimally - according to prices and available technology - and trade to satisfy demand, improving their well being in the process. In real life, an intermediate situation where some actors have some market power is probably the norm.

In an economy where no markets are missing and prices are fixed, household production and consumption decisions are related through income alone and only in one direction: production, based on prices, determines farm profits - a component of household income - which in turn influence consumption and supply-supply decisions. As a result of this recursive decision-making process, an increase in the price of staples unleashes to two opposing forces: the traditional substitution and income effects of consumer theory, and the opposing effect of profits. The ultimate effect on marketed surplus is a matter of empirical investigation (Singh et al 1986).

When one or more markets are missing, market prices for goods and factors may not reflect their value - or that of production activities - to households. Risks associated to price fluctuations, for instance, alter the value of household production in ways that reflect household heterogeneity (Finkelstein and Chalfant 1991). Idiosyncratic shadow prices become a two-way link between each household’s production and consumption decisions (i.e. they become simultaneous). More generally, markets are not entirely absent but transaction costs prevent households from participating in them. Transaction costs specific to certain commodities can affect all households equally, or many markets can fail selectively for particular households - i.e. imperfections can be household specific but not commodity specific (de Janvry et al 1991). In both cases, the scope of transaction costs effectively creates price bands within which the household or households respond to their own shadow prices. While fixed and variable but otherwise exogenous transaction costs create well-defined price bands (Key et al 2000), other market failures can create endogenous price bands or price thresholds (Dyer 2002). All things considered, the nature of household-farms and the realities of market conditions in developing rural areas make household responses to market signals an empirical question.

Spill over effects

The interaction of economic agents through markets in a partially closed economy guarantees that policies and other exogenous shocks have ‘spill over’ effects in addition to any direct effects. The study of spill over effects requires consideration of the entire economy in question, which sets it beyond the scope of a single-household model. Two types of spill over effects have been considered in the literature: multiplier and general equilibrium (or price) effects. Although multiplier analysis and computable general equilibrium (CGE) models - both based on the social accounting matrix or SAM - originally focused on national economies, they were soon applied to smaller, regional economies.^[10] The application of multi-sectoral models to developing rural areas allowed a level of detail and disaggregation not possible with nation-wide models, revealing to what extent these economies are isolated and self-contained, on one side, and inward integrated, on the other (Adelman et al 1988; Taylor and Adelman 1996).

Analysis with both SAM and CGE models addresses precisely the structural integration of an economy and the nature of its markets. Multiplier models represent either supply-constrained or demand-driven markets with fixed-prices and Leontief technologies. Regional CGE models relax the assumption of fixed prices and incorporate microeconomic behaviour absent from nation-wide models (Taylor and Adelman 1996). Practical application of multiplier and CGE models in the literature on developing rural areas reflects their differences and limitations. The assumption of fixed prices has restricted the use of multiplier models to analysis of the implications of economic structure on development, as regards the role of migrant remittances (Adelman et al 1988), agricultural investment, rural industrialization and tourism. Although CGE models have also been used to this end (e.g. to study the indirect implications of land erosion and tourism (Taylor et al 2003)), price flexibility has extended their range of application. The stylized model presented by Taylor and Adelman has been applied with only minor changes to study the implications of structural change and trade liberalization policies (i.e. removal of subsidies and price supports, changes in trade tariffs and exchange rates), and those of compensatory or social polices (e.g. cash transfers and public works), particularly in Mexico and Central America, but also in Zambia.

Modelling policy

Globalization has meant that even seemingly isolated households in remote rural areas might be affected by macroeconomic events in distant places. The level of economic integration accomplished with globalization makes a compelling case for modelling economic processes at different levels of aggregation. Household, micro-regional and macroeconomic models have both advantages and disadvantages in explaining those processes. Ideally, these three types of models can be integrated to produce a more complete picture of developing rural economies. Micro-regional models have served this purpose well, but much remains to be done.

Household-farm models are ideal to explore specific motivations influencing the decisions of agents in rural areas, but in the process, they often ignore other household motivations and abstract from the complexity of their interactions with other agents. On the other side, there are clear practical limits to the complexity of household behaviour that can be incorporated to understand macroeconomic processes. Micro-regional models were conceived as an alternative in-between household-farm models and aggregate CGE models for policy analysis, blending microeconomic behaviour with economy-wide modelling. These models are particularly suitable to predict policy outcomes because they represent agents in the full context of the economy.

Since micro-regional models must include all relevant economic activities, simplicity is necessary to insure that simulation results remain tractable. In practice, this has implied that instead of including full-blown household models within a micro-regional model, specific results from household models are incorporated as stylized facts. This is the case of the profit effect described earlier: micro-regional models have often assumed a linear expenditure system that is compatible with the profit effect. This simplification is not absolutely necessary (especially if this is focus of the model), but it allows inclusion of the profit effect while focusing on other economic issues, such as migration (Taylor and Adelman 1996).^[11] Research on the impact of migrant remittances is yet another instance where household and regional models have been a perfect fit. Household models have demonstrated that migrants play the role of financial intermediaries, enabling rural households to overcome credit and risk constraints, albeit reducing the availability of family labour (Rozelle et al 1999). These observations have been incorporated into regional models in the form of investment flows linked directly to the household’s labour supply and indirectly - through the wage rate - to other local households (Taylor et al 1999a). Ensuing changes in capital stocks have been modelled in an ad hoc manner, that is, exogenously. This is an area where refinements would be advantageous.

Other times, micro-regional and macroeconomic models have been linked via policies: price forecasts derived from macroeconomic models are incorporated as the exogenous shocks of micro-regional models (Taylor et al 1999a,b,c.). Household and regional models have also been presented as alternative or competing descriptions of the same process. Alternative descriptions of the type of transaction costs faced by Mexican staple growers have been put forward using household (de Janvry et al 1995), village (Taylor et al 1999a) and regional models (Taylor et al 1999c). Differences arise from conflicting interpretations of the facts. A village model is justified by the fact that government-purchasing points are either present or absent, creating the same transaction costs for all households in a given village or region. Household models are justified by the assumption that individual households have differing capabilities to transport their product to distant purchasing points. It is entirely possible that one or the other model applies to different rural areas in Mexico.

Holden et al (1998) describe a ‘typology’ of market environments to guide choice of model, based on i) the type of transaction costs faced by households and ii) household heterogeneity. Their typology ranges from a situation where households are relatively homogeneous and there is no trade among them, to one where there are complete and perfectly functioning markets and heterogeneous farms. They find that the situation in a Zambian village, where homogeneous household-farms sell a cash crop in distant markets (i.e. there are no local markets), can be addressed by summing up individual household models. They construct a village-wide model to show that, since there are no local markets, results are the same as those from individual household models. The absence of local markets in their study may be an artefact of their village-wide model, since aggregation of households into types obscures trade within same-type households.

For the most part, distinction of household types is an advantage of micro-regional models. Different households are subject to different exogenous shocks and have different responses to these shocks. Choice of household types reflects specific research focuses, e.g. migration, agricultural policy, poverty alleviation. However, most micro-regional models abstract from household heterogeneity, which restricts the way household types adjust their production and consumption decisions to exogenous shocks. This limitation was recently addressed by integrating simple individual household models into a micro-regional model (Dyer and Taylor 2002). Incorporation of individual models requires that activities within each household be modelled separately and later aggregated in a market-equilibrium equation. Production, consumption and marketed surplus of individual households respond to idiosyncratic shadow values - in the case of subsistence producers - or market prices - in the case of commercial growers. This extension of previous micro-regional models also permits one to analyse the full distribution of the impacts of policies across a sample of the local population.

Another direction in which micro-regional models have evolved is geographical. Originally applied to villages, micro-regional models have been adapted to counties (Dyer et al 2001) and archipelagos (Taylor et al 2002; Taylor et al 2003), reflecting efforts to analyse policies at different levels of government, but also the geographical scope of the various markets in which any given households participates. In a way, micro-regional and macroeconomic models, but also household-farm models can be construed as models of partially open economies involved in trade with other economies (Taylor and Adelman, 2003). They are all general equilibrium models to the extent that household shadow-prices, local prices or domestic prices reflect supply and demand within the economy.

A case study

Macroeconomic models of structural adjustment and market liberalization of the Mexican maize sector helped guide policy during the early 1990s. These models predicted that maize output would decline by up to 20 percent with the implementation of the North American Free Trade Agreement (NAFTA) (Levy and van Wijnbergen 1994). Commercial maize growers in rain-fed areas would be hurt slightly, since lower wages would partially offset lower output prices. Subsistence growers and landless workers would suffer job losses and lower wages, migrating to urban areas and the United States in response. De Janvry et al (1995) criticized these models for neglecting the microeconomics of the household-farm, its diversified economy and its market-participation decisions. Based on their own household model, they suggested that the effect of price declines would be felt differentially across household types: surplus growers (net sellers) would be seriously and adversely affected, but market failures (i.e. transaction costs) would buffer subsistence producers (‘about half of all maize producers’) from the decline in sales price. Adding up different groups’ responses obtained from microeconomic forecasts, de Janvry et al (1995) expected rain-fed maize agriculture to contract substantially less than macroeconomic models predicted.

Government purchase points implementing support prices in certain regions were closed as part of structural adjustment. This motivated Taylor et al (1999b,c.) to suggest that whole villages might experience an increase in transaction costs. The implications of their village-wide models differ from those of household-models. The adequacy of one or the other model evidently depends on the nature of transaction costs experienced in reality. Estimating a household model, Key et al (2002) found econometric evidence that Mexican maize growers are subject to household-specific transaction costs, but there is little factual evidence to back this suggestion.^[12] Dyer (2002) suggested that high shadow prices might arise from maize’s non-market benefits to markets and not from transaction costs. Dyer and Taylor (2002) incorporate price and income elasticities derived by Dyer (2002) as stylized facts in a micro-regional model that aggregates nearly fifty individual household models.^[13] They suggest that the decline of commercial maize production has opened the way - through the land rental market - to subsistence maize production. Household heterogeneity and commercial growers’ market power are determinant in influencing wages and land rental rates to yield this result. Clearly, the variety of models used to analyse the liberalization of Mexican maize markets has widely differing implications. It is clear that the forecasts of macroeconomic models were well off the mark, but it is unclear what this implies for the future. The forecasts arising from current interpretations of household motivations are contradictory.

Analyses of social policies implemented to counteract the undesirable effects of agricultural policies in Mexico have also been based on different types of models and, not surprisingly, reach different conclusions. A household analysis by Sadoulet et al (2001) suggests that Procampo’s cash-transfer program to growers has had an important within-household multiplier effect through investment in off-farm activities. Dyer and Taylor (2002) do not incorporate this finding into their micro-regional model. Instead, they suggest that Procampo payments have a limited regional multiplier effect since Procampo recipients tend to spend a large share of their income on consumption goods ‘imported’ into the region. They also compare the merits of alternative compensatory policies. Surprisingly, an across-the-board lump-sum transfer does a better job compensating lost household income after market liberalization than Procampo, which compensates growers alone for the loss in crop value.

Conclusions

Global economic integration makes a compelling case for modelling economic processes at different levels of aggregation. Ideally, models at different levels can be integrated to produce a more complete picture of developing rural economies. Micro-regional models are well suited for this purpose, but their potential has not been fully realized. Existing applications of micro-regional models do not reflect well the diversity of issues affecting household-farms, such as risk, access to information, and learning. Recent integration of individual household models into a micro-regional model - allowing for household heterogeneity and interaction among households - promises to extend the range of applications.

References

Adelman, I., Taylor, J.E. & Vogel, S. 1988. Life in a Mexican village: A SAM perspective. Journal of Development Studies 25:5-24.

De Janvry, A., Fafchamps, M. & Sadoulet, E. 1991. Peasant Household Behaviour with Missing Markets: Some paradoxes explained. The Economic Journal 101:1400-1417.

De Janvry, A., Sadoulet, E. and Gordillo de Anda, G. 1995. NAFTA and Mexico’s maize producers. World Development 23(8):1349-1362.

Dyer, G. 2002. The cost of in situ conservation of maize landraces in the Sierra Norte de Puebla, Mexico. Ph D. thesis, Univ. of California, Davis.

Dyer, G. & Taylor J.E. 2002. Rethinking the Supply Response to Market Reforms in Agriculture: Household Heterogeneity in Village General Equilibrium Analysis from Mexico. Dept. Ag. and Res. Econ. Working Paper, Univ. of California, Davis.

Dyer, G., Yúnez, A. & Taylor, J.E. 2001. Land degradation in a diversified rural economy: a village-town CGE analysis from Mexico. In Heerink, N., H. van Keulen, M. Kuiper (eds.) Economic policy reforms and sustainable land use in LDCs: recent advances in quantitative analysis, Physika-Verlag, Heidelberg.

Finkelshtein, I., and Chalfant, J.A. 1991. Marketed surplus under risk: Do peasants agree with Sandmo? American Journal of Agricultural Economics 73(3):558-67.

Holden, S., Taylor, J.E. & Hampton S. 1998. Structural adjustment and market imperfections: A stylized village economy-wide model with non-separable farm households. Environmental and Development Economics 4:69-87.

Key, N., Sadoulet E. & de Janvry A. 2000. Transaction Costs and Agricultural Household Supply Response. American Journal of Agricultural Economics 82(2):245-259.

Levy, S. & Van Wijnbergen, S. 1994. Labour markets, migration and welfare: Agriculture in the North-American Free Trade Agreement. Journal of Development Economics 43:263-278.

Rozelle, S., Taylor, J.E. & de Brauw, A. 1999. Migration, remittances, and productivity in China. American Economic Review 89(2), 287-91.

Sadoulet, E., de Janvry, A. and Davis, B. 2001. Cash transfer programs with income multipliers: PROCAMPO in Mexico. World Development 29(6):1043-1056.

Singh, I., L. Squire, & J. Strauss (eds.). 1986. Agricultural household models: Extensions, applications and policy,.The Johns Hopkins University Press, Baltimore.

Taylor, J.E. & Adelman, I. 1996. Village economies: The design, estimation and use of villagewide economic models. Cambridge University Press, Cambridge.

Taylor, J.E. & Adelman, I. 2003. Agricultural household models: Genesis, evolution and extensions. Review of Economics of the Household 1(1):33-58.

Taylor, J.E., Yúnez, A. & Dyer G. 1999a. ‘Agricultural price policy, employment, and migration in a diversified rural economy: a village-town CGE analysis from Mexico. American Journal of Agricultural Economics 81:653-662.

Taylor, J.E., Yúnez, A. & Hampton, S. 1999b. Agricultural policy reforms and village economies: A Computable General-Equilibrium analysis from Mexico. Journal of Policy Modelling 21(4):453-480.

Taylor, J.E., Zabin, C. & Eckhoff, K. 1999c. Migration and rural development in El Salvador: A micro economywide perspective. North American Journal of Economics and Finance, 10:91-114.

Taylor, J.E., Dyer G. & Stewart M. 2002. Economic study of the Bay Islands: a preliminary report for the Inter-American Development Bank. IDB, Washington, D.C.

Taylor, J.E., Yúnez, A., Dyer, G., Ardila, S. & Stewart, M. 2003. Eco-tourism: A Galapagos Island economy-wide perspective. Economic Development and Cultural Change 51(4):977-998