The rationale for public spending to provide public goods and services is well known (Samuelson 1954; Varian 1992). In this context of agricultural development, it hinges on market failures, including imperfect markets and information asymmetries for agricultural technology advancement as well as for promoting the adoption and use by farmers of those technologies and other productive investments. Public spending is also justified on social grounds for income distribution and poverty reduction. Thus, it seems logical that governments or public institutions would favor public goods and services, or target locations or people, where the impacts of their expenditures and interventions are likely to be greatest. This means spending more in areas with higher productive potential (efficiency motive) or spending more in poorer areas or where public goods and services are low (compensatory or equity motives). Whatever the motive for spending is, it underlies the notion of program placement effects (Maddala 1983). That is, the spending decision is likely to be influenced by the performance or outcome associated with the amount spent. In other words, the public spending decision is endogenous-a debatable (for example, is an increase in public spending an outcome or a cause of agricultural growth or poverty reduction?) and an empirical issue (see Ansari et al. 1997; Zhang and Fan 2004). Another issue is that the effects of public investment commonly materialize with a lag rather than contemporaneously, and so public spending decisions at
any time will depend on previous spending decisions and outcomes. The length of the lag is also an empirical issue that depends on the type of spending as well as the performance indicator of interest. For example, the agricultural productivity effect of public spending on, say, agricultural research is expected to start materializing after several years (e.g. 15 years or more) and persist long afterwards, compared to the effect of public spending on, say, agricultural input subsidies, which materializes immediately and lasts only a few years. Spending on a particular type of public good or service or in a particular geographic location can also affect the spending decisions on other types of public goods or in other geographic locations, respectively. The former reflects complementary or substitution effects among different types of public spending. Regarding the latter, spatial interaction effects (Oates and Schwab 1988) reflect mobility of public goods users, and information asymmetries among public goods providers, such as local government officials and politicians (Case et al. 1993; Figlio et al. 1999). From the competitive standpoint, and depending on the type of public spending, local governments are concerned about how their expenditures or taxes compare with those of their neighbors, and tend to adopt positions such that their public goods choices are seen as more or equally attractive relative to those of their neighbors. For example, if some local governments expand their spending on services (e.g. water, sanitation, and education) that tend to attract businesses and residents, there is a strong incentive for their neighbors to do the same in order to stay competitive. Spending on science and research, on the other hand, has strong externalities to the extent that the science and research outputs are accessible beyond the boundaries of the jurisdictions in which they are produced. Thus, jurisdictions would tend to decrease their spending in response to increased spending by their neighboring jurisdictions. Other factors determining public spending decisions include budget constraints and various sociocultural, political, and institutional factors in the context in which the spending decision is made. For example, how much of its resources a government spends on a specific investment program will depend on the total resources available to the government, and also on political economy, institutional, and governance factors. For example, governments in response to political lobbies by the affluent may spend more in wealthier areas or in areas where public goods and services are already high. Similarly, governments may spend more in areas with greatest political support to secure future electoral wins (Cox and McCubbins 1986) or more in areas with lower political support in order to swing future votes in their favor (Lindbeck and Weibull 1993). Governance is one factor that has attracted particular attention during the past decade, and is seen as increasingly important as a factor influencing the efficacy of public spending (see also the discussion in Chapter 1).