Our educational and research institutions tend to mirror this shortcoming,8 with the result that the larger system contexts of research questions are infrequently investigated and poorly understood. Difficulties in apprehending and resolving problems whose constituents are grounded in several interrelated systems are compounded by the international community’s disparate, competitive political and economic systems. Nations act to promote their own priorities but affect, often negatively, globally shared resources and globally interdependent societies. Although nations and other sociopolitical groups generate impacts beyond their borders, they are generally incapable or unwilling to assess and react equitably to the results of their actions. Pierre Crosson and Norman Rosenberg 18 note the inadequacy of information feedback about significant environmental problems in modern societies, an inadequacy which characterizes feedback about social problems as well. Accounting for the system-wide implications of local actions should be a primary objective for sustainable agricultural systems. The definition of sustainability offered here places a priority on broad-based equity considerations. We believe it is inadequate to exclude social justice as a priority and that there is an ethical requirement for greater equity in the agricultural system. Some have combined concern for how we treat the environment with how we treat our fellow human beings.19, 20, 21,22 For those focusing on the latter, it is essential to look beyond sustaining our environmental and economic ability to produce agricultural goods. It is equally important to ensure that those goods are produced and distributed in an equitable manner. A concern with this human values aspect of agriculture involves a sweeping rather than localized concept of who constitutes “us.” Typically, resource conservation is dis- cussed in terms of its implications for farmers’ profit- ability or our descendants’ food-producing capabilities. The sustainability definition offered in this paper does not limit equity considerations to these groups. A concern with equitable social relations in agriculture requires defining “us” in terms of all fellow humans – not only farmers and future generations, but also farm workers, consumers, non-farm rural residents, Third World urban poor, and others.
Sustainability in this sense is framed in terms of both intergenerational and intragenerational equity. Thus,greenhouse vertical farming issues such as farm worker rights and inner-city hunger are as central as issues of soil erosion and groundwater contamination to the goals of agricultural sustainability. One of the most profound challenges facing agriculture is creating a decision-making process which will fairly resolve equity issues. Such a process must assess competing interests; evaluate agriculture’s costs and benefits, and the recipients of each; decide fairly what the compromises must be; recognize and encourage shared goals and common ground. In most discussions of sustainability either environmental quality or social justice issues are emphasized, but neither can be sup- ported wholly at the expense of the other. Nourishing humans, ensuring social justice, and providing a reasonable quality of life cannot be accomplished if agriculture’s resource base and environmental constraints are neglected. Likewise, few would argue that environmental considerations should be pursued at the expense of satisfying basic human needs. An equitable agricultural system must foster a decision-making process which is truly democratic, one which identifies not only what the costs and benefits are but how to distribute them fairly among all sectors of society.Many sustainability definitions, particularly those which guide applied sustainable agriculture programs, are based on the primacy of farm production and short-term profitability. As sustainable agriculture programs have increasingly been incorporated into long-established agricultural institutions they have manifested the largely unquestioned intellectual assumptions and infrastructural constraints which characterize their parent institutions. This is problematic because conventional agricultural institutions have fostered many technologies and policies counter to sustainable agriculture goals.Such institutions have, for example, contributed to concentration within agriculture; have not generally benefited agricultural labor; and have systematically failed to examine their impact on the environment, the structure of rural households and communities, and the consequences of rural resident displacement.
To situate new pro- grams designed to address these problems within the framework which produced them is of questionable value unless steps are taken to change the nature of that framework, for it determines the way its re- searchers see the world, pose questions, and define problems. When agriculture is viewed in a whole-systems context and sustainability is defined comprehensively, it is clear why the current popular focus on farm production practices is insufficient for achieving agricultural sustainability. Developing non-chemical pest management methods, for example, will effectively reduce pesticide use only if economic structures and policies encourage their adoption by farmers. More importantly, one cannot conclude that improved production practices will transform the agricultural system into one that meets all environmental, economic, and social sustainability goals. Social goals must be addressed explicitly. This is why production techniques such as organic farming, while a likely component of a sustainable food and agricultural system, cannot be thought of as synonymous with sustainable agriculture. Given the conventional institutional context of most state and federal sustainable agriculture programs it is not surprising that they tend to focus research on conventional priorities such as production practices and efficiency and have not, for the most part, aggressively addressed social and economic issues. Sustainability priorities – and the definitions which embody them – must be expanded to encompass the many factors affecting production and distribution as well as the larger environmental, economic, and social systems within which agriculture functions. This has been the focus of the Agroecology Program since its inception in 1982. Through conferences and publications* we have worked to expand the discussion and practice of integrating these aspects of sustainability. Recently, the University of California Sustainable Agriculture Research and Education Program has broadened its agronomic focus to include social, economic, and policy issues. SAREP defines sustain- able agriculture as integrating “…three main goals – environmental health, economic profitability, and social and economic equity.”Their grant program, which encourages research and education on social, economic, and public policy issues affecting food and agriculture, could become a model for other sustain- able agriculture programs such as LISA. We believe that it is important to continue exploring the meaning of agricultural sustainability. Before an improved agricultural system can be developed the biases and structures that have led to agricultural problems must be closely examined and concrete goals articulated, based upon a broadened concept of agricultural sustainability. The concept of sustainability offered in this paper emphasizes that social goals are as important as environmental and economic goals, and widens the opportunity to move beyond the narrow agricultural priorities expressed in the past.
It is based upon the whole-systems, interactive nature of all aspects of the agricultural system – that problems and their resolutions must be conceived not only in terms of their immediate time frames and local impacts, but just as importantly, in terms of their future time frames and their global impacts. It encourages emphasis on optimum production over maximum production, the long term along with the short term, the public’s best interest over special interests, and the contextualization of disciplinary work within interdisciplinary frameworks. Our hope is that this definition helps advance the discussion on developing a food and agriculture system that is sustainable for everyone. While aggregate growth in an economy may improve the welfare of both wealthy and poor households,vertical agriculture the latter are most usually rural, and rural households have employment and incomes that depend disproportionately on agriculture. It is natural to wonder if growth in aggregate agricultural income has a different effect on the welfare of poorer households than does growth elsewhere in the economy. The question is an important one for many policy issues. Faced with continuing extensive poverty, many development agencies and scholars have suggested the need to refocus growth on agriculture , arguing that the alternatives of redistributing income generated outside of agriculture or migration out of agriculture to urban areas are difficult to achieve and create other problems. Of course, we are not the first to wonder whether growth in agriculture may be more effective than growth in the rest of the economy in reducing poverty; an extensive theoretical and empirical literature already exists on the subject which we discuss in Section 2. The theoretical literature focuses on the different transmission mechanisms of an exogenous gain in agricultural productivity on poverty, while the empirical literature analyzes the reduced form relationship, and generally documents a stronger association between poverty reduction and growth originating in agriculture compared to growth originating in non-agriculture, with the exception of Latin American countries. In this paper we tackle this question by comparing changes in the level and distribution of household expenditures due to growth in both aggregate agricultural and aggregate non-agricultural income. We use growth in household expenditures as the outcome of interest because we believe expenditures to be the best available indicator of material well-being; also, these are the data generally used for poverty calculations for most low-income countries. However, our analysis differs from most other studies in several aspects. First, we consider growth in expenditures across the entire distribution rather than the simple poverty headcount ratio, giving a richer picture of the effect of sectoral growth on welfare. Second, we use the deciles as defined within each country, rather than a common international benchmark of expenditures.
To correct the underlying assumption that deciles of very different countries have similar relationship with agriculture, we then pursue some heterogeneity analysis. Finally, we tackle the issue of simultaneity between sectoral income and expenditures using an instrumental variable approach, allowing us to take a stand on the causality of sectoral growth on welfare. The simple regression we would like to estimate relates expenditure growth for differently positioned households to growth in sectoral income, the latter weighted by its share in total aggregate income; this is described in Section 4. The question of whether the poor benefit more from agricultural income growth than growth in other sectors could then be answered simply by examining the relative size of the coefficients on aggregate income growth from agriculture and from other sectors. In practice, there is a series of challenges we must face before estimating such a regression. First, we do not have household level data that would allow us to make comparisons across countries. Instead, we use data from the World Bank’s Povcal Net project and consider estimates of household expenditures from different expenditure deciles; in effect we construct a panel of ten representative ‘households’ for each country, each representing an expenditure decile.1 We discuss these data in Section 3.1. Second, the resulting ‘panel’ is extremely unbalanced, since the underlying expenditure surveys are conducted at irregular intervals. This creates some important accounting issues when we turn to estimation, treated in Section 4.1. Third, some countries, some years, and perhaps some deciles can naturally be expected to have different expenditure growth rates for reasons unrelated to sectoral income growth. A global financial shock may cause expenditure growth to slow for everyone; households’ risk attitudes or time preferences may imply different rates of expenditure growth across deciles ; the endowments of a particular country or some aspect of the structure of its economy may imply systematically different rates of expenditure or income growth even over long periods; and variation in the global price of agricultural commodities will change the composition of income across sectors for many countries. We attempt to deal with these kinds of alternative sources of variation in expenditure and income growth in a fairly agnostic manner, by using fixed effects and related methods for dealing with what Wooldridge calls “unobserved effects.” So: we account for aggregate cross-country shocks using a collection of time effects; and for systematically different rates of expenditure growth across the distribution we use a set of decile fixed effects. We would be inclined to also use a complete set of country fixed effects to deal with differences in endowments, but with these we reach the limits of our dataset; instead we employ a set of continent fixed effects, which in practice seems to be effective. Fourth, the stochastic process governing country-level agricultural income exhibits more time-series variance than does income from other sectors.