Central authority prefers that all be "with the program." This puts more pressure to capture detail. In the case of Morocco, we had two sets of models 11 . One at the national level and several at the regional and farm category models, all of which were consistent, in terms of variable de±nition, with the national model. That way we could (a) increase/decrease gov. subsidies to farmers of a certain type in a region of the country using the farm level models. (b) assume the policy was made available to all farm- ers using the national model. Then, (c) return to (a) with new prices from general equilibrium and solve (a) again to assess the total e/ects (direct and indirect that are caused by the clearing of national markets). This map- ping provided the ground work for directing dialogue between national and regional authorities. There are countless microeconomic puzzles that have macroeconomic im- plications. For example, Deaton (2010, JEP, page 10) notes that total 11 Roe, T. A. Dinar, Y. Tsur, and X. Diao (2005) Feedback links between economy-wide and farm-level policies: With application to irrigation water management in Morocco, J ournal of Policy Modeling , 23: Nov.:905-928. 22
household consumption rises with household size while per capita household expenditure falls with household size, i.e., economies of scale. The implica- tion is that if we compare larger with small households who have the same per capita income, the larger households are better o/. Economic growth typically causes a decrease in fertility, and an increase in human capital per capita. What is the nature of the "interaction" among household level forces that result in the macro level outcomes? 7.4 The "technology" of applied general equilibrium- growth models The modern approach takes advantage of "scale economies" obtained from theory-data-solution. General equilibrium invariably entails a larger number of endogenous variables than partial equilibrium models. Static general equi- librium models of relatively small scale (for example, a three sector economy employing capital, labor and land), are relatively tractable analyticity, and we can apply the insights from this framework to larger dimension models. To assess the magnitude of changes in endogenous variables, requires that we ±t them to data. Dynamic models have equilibria along a transition path to long-run equilibrium. So, imagine a static model cast into an endogenous savings (hence dynamic) model. This framework typically is not analytically tractable. Thus, we must ±t it to data, doing so such that in period t = 0 ; the model reproduces the observed data exactly, and then solve the model to obtain transition equilibria t = 0 ; 1 ; ² ² ² ; T: This means one needs to 1. Know the theory 2. Know the data and how to organize it in a "general equilibrium" man- ner 3. Know how to obtain parameter values from the data in a way consistent with the theory 4. Know computer code 5. Know the algorithm(s) to solve di/erential equations, and 6. Develop a strategy for proceeding from 1. to 5. in a way that is e¢ cient, meaning, in a way that allows you to minimize the number of errors and maximizes your ability to ±nd errors.