ترکیب سهام کالا در یک مدل تعادل عمومی اقتصاد جهانی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|28611||2005||19 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Economic Modelling, Volume 22, Issue 4, July 2005, Pages 646–664
Applied General Equilibrium (AGE) models are increasingly used for short-run commodity market analysis, although they are generally not designed for such purposes. This study remedies a key shortcoming of such work by incorporating a commodity stockholding model into a short-run global AGE model. An approach to calibrating and validating the model is demonstrated for the staple grains sector. Its historical weather-induced supply shocks form a natural vehicle for reconciling model outcomes with observed behavior. The stockholding model, in conjunction with a Gaussian Quadrature approach to characterizing supply-side uncertainty, will prove useful to researchers seeking to conduct short-run analysis of commodity market policy in an AGE context.
Applied General Equilibrium (AGE) models are increasingly used for short-run policy analysis of international commodity markets (OECD (Organization for Economic Cooperation and , 2001, Burfisher and Jones, 1998, Martin and Winters, 1997 and Francois and Shiells, 1994). However, most AGE models are not designed with such a purpose in mind. Comparative-static models, for example, are typically oriented towards a “medium run” scenario of 2 to 3 years, depending on the factor market assumptions employed. Dynamic AGE models focus on long-run growth effects, not short-run dynamics. In turn, stockholding is typically missing from AGE models, yet this is a principal means of accommodating short-run volatility in international commodity markets (Williams and Wright, 1991). In developing an AGE model, changes in commodity stocks are typically eliminated in the initial data, or subsumed into aggregate investment. For these reasons, standard AGE models are of limited usefulness for short-run commodity market analysis. However, 1 year or less is usually the time frame many policy analysts have in mind. The purpose of this paper is to bridge the divide between short-run trade policy analysis, and AGE models, by incorporating stockholding behavior into a widely used AGE model of the global economy. The first step is to develop a model of commodity stockholding from first principles, and incorporate this into the Global Trade Analysis Project (GTAP) model. GTAP is a relatively conventional, multi-region AGE model that plays a role in commodity market analysis in dozens of institutions worldwide, including: national ministries of trade and agriculture, as well as international public agencies concerned with trade and development and academic research centers (Hertel, 1997). In addition to incorporating stockholding, the model's closure (i.e., classification of variables into exogenous and endogenous) is adapted to short-run analysis. The paper's next steps are to demonstrate how the stockholding function can be parameterized. This is demonstrated for “staple grains,” which is an amalgam of wheat, rice, maize, and other grains (Appendix Table A3). Staple grains are particularly useful for illustrative purposes because it has a volatile supply, is widely traded, and is stored over time. This is also a composite commodity that receives considerable attention from the UN Food and Agriculture Organization (FAO), as well as the World Bank. One approach to parameterizing the stockholding model is econometric estimation, and this is carried out for a country–commodity combination that offers reliable and extensive time-series data (U.S. wheat). There are substantial challenges in this work, however, including a lack of time-series data on non-price determinants of stockholding behavior, and how to correct for the simultaneous determination of price and stock changes. Another concern is that estimation of the stockholding model as a distinct entity may not be appropriate for use in the rather different environment represented by the AGE model (Browning et al., 1999). Ideally, one would simultaneously estimate all elasticities and share parameters using time-series data in conjunction with the full general equilibrium model. However, AGE studies that attempt this are often plagued by lack of data, and face other significant conceptual and computational difficulties (Arndt et al., 2002). Therefore, the sub-systems of the AGE model typically have to be estimated separately, instead of incorporating the full set of cross-equation equilibrium restrictions in the procedure (Jorgenson, 1984 and McKitrick, 1998). As an alternative to the econometric approach, we exploit the fact that staple grains production is characterized by weather-induced supply volatility, and use this feature to calibrate the stockholding function as it is embedded in the AGE model. This approach draws on techniques used in the real business cycle literature, wherein models are calibrated in a “heuristic” fashion to mimic the properties of business cycles exhibited in time-series data, including volatility and correlations among consumption, output, investment, and labor. Our approach begins by developing historical distributions of grains production for individual regions. Using methods from the literature on systematic sensitivity analysis (DeVuyst and Preckel, 1997), the model is simulated repeatedly under multiple draws from these distributions. The base parameterization of each region's stockholding function corresponds to the initial econometric estimate for U.S. wheat. The moments of the resulting region-specific distributions of stock changes can then be compared to the moments of actual distributions based on historical data. The stockholding parameters are then adjusted in a heuristic manner away from the base setting, until simulated moments are brought into congruence with the corresponding observed moments. The principal calibration criterion is the standard deviation of proportional changes in commodity stocks relative to production. Similar criteria are adopted for a subsequent validation exercise. As such, we demonstrate a procedure for orientating an AGE model towards short-run commodity market analysis. Once fully developed, such a model has many potential applications. It can address questions of trade policy and price volatility in the presence of production uncertainty (e.g., Claessens and Duncan, 1993). It can examine issues related to the vulnerability of low income households due to international price volatility for staple grains (e.g., Berck and Bigman, 1993). The model can also be used without uncertain production for the purpose of quantifying short-run policy impacts. Finally, this work can serve as a template for conducting short-run analysis of other non-perishable commodity markets where stockholding is an important component of year-to-year adjustment. The remainder of the paper is organized as follows. In the following section, a simple two-period model of stockholding is developed from micro-foundations. A functional form consistent with the results of this model is then proposed, and estimated for the case of U.S. wheat, using data from the FAO. The subsequent section describes the procedures used in calibrating the stockholding model within the AGE model. The final sections present results, discussion, and conclusions.
نتیجه گیری انگلیسی
AGE models are increasingly used for short-run policy analysis of international commodity markets. The time frame for such analyses is often 1 year (or less), and for many commodities, stockholding is a significant determinant of price volatility. Yet most AGE models are specified with a medium-run (two-to-three year) time frame in mind, and stockholding is typically ignored. In developing an AGE database, changes in stocks are also typically eliminated from the initial data, or subsumed into aggregate investment. As a step towards resolving this problem, this paper introduces a stockholding relationship into one of the most widely used AGE models of global trade—the GTAP (Global Trade Analysis Project) model (Hertel, 1997). Although we focus on the GTAP framework, our approach is adaptable to other such models. We develop an exponential stockholding function inspired by Dixon and Rimmer's (2002) general-equilibrium modeling of sector-specific changes in capital stocks. In turn, supply-side volatility is modeled with a Gaussian Quadrature procedure (DeVuyst and Preckel, 1997). Together, these features form an appealing, flexible vehicle for evaluating alternative short-run policy scenarios in a way not possible with most existing AGE models. The framework is quite general, and easily adapted to a variety of non-perishable commodity market settings where stockholding is an important component of year-to-year adjustment. In this paper, the framework is oriented to short-run market analysis of staple grains (an amalgam of rice, maize, wheat, and other grains). While this application is meant to be illustrative, it forms a starting point for a researcher wishing to explore alternative price stabilization schemes intended to benefit the rural poor in developing countries14, to name just one example. An important challenge in this work is to calibrate and validate the stockholding model, and two complementary strategies are demonstrated within the paper. First, we show that the stockholding relationship can be econometrically estimated using time-series data on stock changes and prices. Ideally, however, the model is parameterized within the AGE model as a whole, that is, within the framework that it ultimately operates. In this vein, we demonstrate an approach to calibration that exploits the natural experiments offered by year-to-year, weather-induced production variability in staple grains. Stockholding parameters are adjusted in a heuristic manner until the model's simulated stock change variation (given by standard deviation) is reconciled with observed, historical stock change variation. Upon meeting this objective, a subsequent validation exercise compares simulated moments of staple grains price changes with a sample of observed, historical outcomes. Results for the staple grains example suggest that model-obtained price volatility is too low compared with available historical observations. While this may be partly related to aggregation problems, it suggests that too much weight is put on the price parameter in attempting to replicate stocks volatility. An important area for future research is to incorporate alternative determinants of stockholding behavior into the model. The paper also presents an additional approach to calibration in which observed price volatility is no longer held in reserve as a validation criterion, but is used as a calibration criterion alongside stocks volatility. This facilitates a reasonable compromise between the two criteria, and the resulting framework forms a useful foundation for AGE analysis of international grain market issues in which the stochastic nature of production and stockholding plays an important role.