Working Papers

Constructing structural models of educational choice allows to explore design features
for educational programs and to predict how the program would perform in alternative
contexts, for instance when accompanied by new complementary programs. We use the
experience of Progresa, Mexico’s ambitious conditional cash transfer program for
education in poor rural communities, to construct such a model. The impact of transfers
on decisions to enroll in secondary school and to repeat a grade in case of failure is
accurately measured due to randomized treatment in a subset of communities. While
impact measurements of Progresa on educational attainment are available from reduced
form estimates, the structural model allows to decompose the channels of influence in
decision making and to measure their relative importance on observed outcomes. We
measure the gains from a design where future transfers can be credibly committed in
spite of political cycles, and from complementary supply-side programs providing
improved off-school support to students and access to better information about job
opportunities outside the community offered by education.

There has been a lively debate about the potential impact of global climate change
on U.S. agriculture. Most of the early agro-economic studies predict large damages (see,
for example, Richard M. Adams, 1989; Harry M. Kaiser et al., 1993; and Adams et al.,
1995). In an innovative paper Robert Mendelsohn, William D. Nordhaus and Daigee Shaw
(1994) - hereafter MNS - propose a new approach: using the variation in temperature and
precipitation across U.S. counties to estimate a reduced form hedonic equation with the
value of farmland as the dependent variable. A change in temperature and/or precipitation
is then associated with a change in farmland value which can be interpreted as the impact of
climate change. Adams et al. (1998) characterize the hedonic approach as a spatial analogue
approach, and acknowledge that "the strength of the spatial analogue approach is that
structural changes and farm responses are implicit in the analysis, freeing the analyst from
the burden of estimating the e®ects of climate change on particular region-speci¯c crops and
farmer responses." On the other hand, one of the potential disadvantages of the hedonic
approach is that it is a partial equilibrium analysis, i.e., agricultural prices are assumed to
remain constant.1 While year-to-year °uctuations in annual weather conditions certainly
have the potential to impact current commodity prices, especially for crops produced only
in a relatively localized area, (such as citrus fruits which are grown mainly in California
and Florida), changes in long-run weather patterns (i.e., changes in climate) might have a
smaller e®ect on commodity prices because of the greater potential for economic adaptation,
particularly shifts in growing regions.2 The hedonic approach as implemented by MNS
predicts that existing agricultural land on average might be more productive and hence result
in bene¯ts for U.S. farmers.3 The hedonic approach has received considerable attention in our
judgment in part because the conclusions are at variance with those of some other studies that suggest warming will lead to damages and in part because of the new methodology.

Although the approach is appealing, it is at the same time vulnerable to problems related
to misspeci¯cation. Several authors have questioned the particular implementation in MNS (William R. Cline,
1996; Robert K. Kaufmann, 1998; Darwin (1999b); and John Quiggin and John K. Horowitz,
1999). Speci¯cally, they suggest that (i) the hedonic approach cannot be used to estimate
dynamic adjustment costs; (ii) the results are not robust across di®erent weighting schemes;
and (iii) the inadequate treatment of irrigation in the analysis might bias the results. The ¯rst
criticism alludes to the fact that some farmers might not ¯nd it pro¯table to switch to new
cropping patterns given their existing crop-speci¯c ¯xed capital. However, climate change
will occur only gradually and most costs can thus be seen as variable. In this paper we focus
on the latter two points, especially the role of irrigation. Previous comments have raised
theoretical concerns about potential sources of misspeci¯cation related to irrigation. We
provide an empirical test. Once irrigation is accounted for, we show that results also become
robust across weighting schemes or models. Elsewhere we extend the analysis in various
directions: construction and use of climate variables tied more closely to agronomic ¯ndings;
development of more accurate measures of both climate and soil conditions; adjustment
for spatial correlation of the error terms in a hedonic regression; and use of recent climate
scenarios that go beyond the traditional assumption of uniform impacts across regions of
a doubling of greenhouse gas concentrations in the atmosphere (Wolfram Schlenker et al.,
2004). We note here that none of the implied changes in the analysis a®ects the arguments
concerning irrigation discussed in this paper.

We begin the paper by critically reviewing the unitary model (Vermeulen, 2002), which (1) assumes that a single preference represents all agents in the group, or equivalently, that all agents have identical preferences; (2) imposes the choice of a benevolent dictator on the group; (3) assumes the group’s budget to be a single pooled value; (4) does not allow bargaining or negotiation, (5) nor permits knowledge or experience differences between members to lead to the use of different decision-making strategies. The impact of policies affecting group members differentially, for example, cannot be correctly assessed in the unitary model because its use will lead to erroneous welfare inferences (Vermeulen, 2002). Clearly, the need for alternative models is pressing.

In the remainder of this paper we will synthesize the literature across disciplines with a view towards characterizing the shortcomings of current approaches to modeling group decisions. Subsequently we broach a number of issue areas that we believe future research must address to improve our understanding of group decisions and to enhance our ability to model and predict outcomes from such decisions. We then present a conceptual model of group decisionmaking that arose from our discussions and, we feel, synthesizes the multidisciplinary views represented in the workshop. We conclude by proposing a number of specific research questions 3 that we believe should be addressed in the short term.

Among others who point to environmental variability and managerial uncertainty as causes of ¯shery collapse, Roughgarden and Smith (1996) argue that three sources of uncertainty are important for ¯sheries management: variability in ¯sh dynamics, inaccurate stock size estimates, and inaccurate implementation of harvest quotas. We develop a bioeconomic model with these three sources of uncertainty, and solve for optimal escapement based on measurements of ¯sh stock in a discrete-time model. Among other results we ¯nd: (1) when uncertainties are high, we generally reject the constant-escapement rule advocated in much of the existing literature, (2) inaccurate stock estimation a®ects policy in a fundamentally di®erent way than the other sources of uncertainty, and (3) the optimal policy leads to signi¯cantly higher commercial pro¯ts and lower extinction risk than the optimal constant-escapement policy (by 42% and 56%, respectively).

Economists sometimes find themselves in the position of having to extend the
neoclassical model of consumer demand to settings where, in addition to the conventional budget
constraint, there are one or more additional linear constraints that restrict the consumer’s utility
maximization problem. Examples include point rationing [Tobin-Houthakker (1950-51), Tobin
(1952)]; models of time allocation where the time constraint cannot be collapsed into the budget
constraint [de Serpa (1971); de Donnea (1972); McConnell (1975); Lyon (1978) Larson and
Shaikh (2001)]; and multi-period portfolio allocation problems [Diamond and Yaari (1972)].
Without exception, the existing literature has focused on differential properties of the resulting
demand functions-i.e. issues such as the effect of rationing on demand elasticities, the Le
Chatelier--Samuelson Theorem, the generalization of the Hick-Slutsky decomposition, and other
comparative static results [see Kusumoto (1976), Chichilnisky and Kalman (1978), Hatta (1980),
Wan (1981) and the references cited above]. By employing some “tricks with utility functions”
in the spirit of Gorman (1976), I am able to obtain a global characterization of these demand
functions. Specifically, I develop an algorithm for deriving the demand functions that apply
when there are M linear constraints from those that apply when these is only a single constraint.
The algorithm permits one to derive all of the existing comparative static results in a simple and
compact manner. It also has some value for empirical demand analysis, because it shows how to compute the demand functions associated with maximization problems involving multiple linear constraint based on direct or indirect utility functions associated with known conventional demand functions.

The paper is organized as follows. Section 2 presents some preliminary results which are
needed for the main analysis, but are also of interest as "tricks” in their own right. Section 3
considers the utility maximization problem with two linear constraints, summarizes the existing
comparative static results, develops the new Global Representation Theorem, and shows how
this can be used to derive and sharpen the existing comparative static results. Section 4 considers
a utility maximization problem with three linear constraints and develops the analogous Global
Representation Theorem for the solution to this problem in terms of known demand functions
associated with a conventional single-constraint problem; the results developed here provide the
basis for extension to problems involving more than three linear constraints. Section 5 offers
some concluding observations.

We define the irreversibility effect and demonstrate its importance in problems involving investment
decisions under uncertainty. We establish several analytical and numerical results that suggest both
that the effect holds more widely than generally recognized, and that an existing result (Epstein’s
Theorem) giving a sufficient condition for determining whether the effect holds can be applied more
widely than previously indicated, in particular to problems involving intertemporally nonseparable
benefit functions. We further show that a low elasticity of intertemporal substitution will however
result in failure of the effect, but that the effect will hold if the value of information increases in
the degree of flexibility.

The persistence of rural poverty, concentration in rural areas of the most extreme forms of
poverty, and rising inequality in the distribution of rural incomes remain vexing aspects of rural
development in Latin America, in spite of expensive programs intended at reducing poverty and inequality.
Mexico is no exception to this observation. This widespread failure calls upon exploring alternative
approaches to rural development that may have greater chances of success. Taking an approach that
distinguishes between marginal and favorable areas, and that seeks to integrate rural and urban activities in
a territorial dimension centered around regional economic projects and the economic incorporation of the
poor is one such option that deserves further consideration. It has been introduced in Mexico through the
Microregions Strategy. While it is too early to evaluate this program, we derive lessons from international
experiences that provide guidelines to assess the Mexican strategy.

We do this by first characterizing the recent evolution of rural poverty and inequality in Latin
America. We then proceed to explore a set of qualitative changes in rural poverty that need to be taken into
account in a new approach. This is complemented by analyzing a set of new opportunities for rural poverty
reduction that should also be factored into a new approach. On the basis of international experiences with
territorial development, we derive a set of principles for success of the approach. We use these principles
to discuss the methodology followed in Mexico for the Microregions Strategy.