Ph.D. Dissertation: The Use of Panel Time-series Data in Modeling Agricultural Markets
Essay 1: Revisiting the Neoclassical Model of Out-farm Migration: Evidence from Nonlinear Panel Time Series Data
The first essay provides a model of inter-sectoral migration from farm to non-farm sector. The neoclassical model of inter-sectoral migration of labor argues that farm workers will move to non-farm jobs if their expected returns outside of agriculture exceed those achieved in the farm sector, net of migration costs. However, previous studies point out that even in the presence of a positive wage differential between the origin and the destination sectors, workers do not always migrate to the sector offering the higher rate of return. I argue that this finding is likely because migration is not completely reversible due to large sunk costs and uncertainty involved in switching to a different occupation or sector. These characteristics of migration suggest that the real options approach may potentially be more plausible to model out-farm migration. In the empirical model, the validity of real options framework would manifest itself with nonlinearities in response to relative wages, where relative non-farm wages need to surpass a significantly large threshold to trigger out-farm migration. The objective of the first essay is modeling state-level determinants of U.S. out-farm labor movements using an occupational migration model that is consistent with large sunk costs and uncertainty involved in changing occupations. For the empirical application, I use a dynamic panel threshold first-difference general method of moments (GMM) estimator that allows for both an endogenous threshold variable and endogenous regressors. I use two different panel time-series data sets from Occupational Employment Statistics (OES) and the Quarterly Census of Employment and Wages (QCEW) on the farm and non-farm employment and wages. I compute the threshold level of sectoral wage gaps that trigger out-farm migration and the elasticities of out-farm migration with respect to wage differentials. Results have significant implications for policy objectives intended to inhibit the flow of labor and other resources out of the farm sector through the use of price supports or direct payments.
Essay 2: Are Fundamentals Driving Agricultural Land Values? Evidence from Panel Data with Cross Section Dependence
The second essay is on modeling farmland values using nonstationary panel time-series data. Empirical investigations of farmland valuation models typically involve testing for cointegration between cash rents and farmland values, or simply testing the stationarity of cash rents-to-value ratios. The literature on farmland values points to a divergence between the present value of future cash flows and the market price of farmland. A possible explanation for the implied absence of empirical support for these models might be that standard econometric tests of net present value hypothesis may not be powerful enough to detect long-run equilibrium when applied to a single time series. A promising approach, in this case, is to combine the sample information from the time-series dimension with that from a cross-section. A major issue that arises in every panel data study with potential implications on inference is the possibility that the cross-section units are interdependent. So, called second-generation panel unit root tests have been proposed to address the problem of cross-section dependence in numerous ways depending upon the underlying structure of the dependence. In this paper, I aim to investigate the suitability of this newer class of panel unit root tests for testing the Net Present Value (NPV) hypothesis for U.S. farmland, and to provide empirical guidance on how to tackle the cross-section dependence issue when using panel time-series data. I focus on i-) tests based on common factor extraction, ii-) tests based on spatial dependence across cross-section units, iii-) Block bootstrapping approach. While some of these are quite flexible in allowing for more general forms of cross-unit correlation structures, others are relatively restrictive. Some models require a large cross-section dimension to be implemented, while others can be applied to smaller panels. I compare the estimates to those from earlier studies and discuss if the farmland values can only be explained by its economic fundamentals as suggested by the NPV hypothesis.
Essay 3: SUR Random Effect Model Estimation for Food Consumption Patterns: Evidence from Unbalanced Pooled OECD Data
The third essay is motivated by food consumption patterns across time and countries with varying income levels. I use unbalanced cross-country panel time-series data to estimate the system of food demand equations. I propose an unbalanced panel data Seemingly Unrelated Regressions (SUR) estimator with random effects. Cross-section dependence across among sampled countries is non-trivial and will be addressed. Utilizing multi-stage budgeting, I develop and use two extended versions of Working’s (1943) model on unbalanced panel-time series data from the Organization for Economic Co-operation and Development (OECD) between 1985 and 2014. A significant contribution of this essay to demand literature is that I do not throw away available information for the sake of achieving a balanced panel; the pooled data are unbalanced with ten phases and diﬀerent number of countries in each phase. I analyze the demand for nine broad categories of goods in the first-stage budgeting and eight detailed sub-categories of food in the second-stage budgeting. Estimated model parameters will be used to compute income and price elasticities, which in turn will allow for direct comparison of food consumption patterns across lower, middle, and higher-income countries. I propose to illustrate the usage of these elasticities to simulate spillover impacts of a country- or group-specific policy. I also explore the sustainability of current food consumption in OECD and policies needed for the more sustainable food system. Precisely, I draw my attention to the issues such as public health, food and nutrition security, sustainable diet and the role of current consumption pattern in the OECD on them.
1) Tayebi, Z., Bakhshoodeh, M., 2015. Factors Affecting Procrastination of Rural Women’s Decision Making to participate in Micro Credit Funds and Self-Financed Groups in Fars, Iranian Journal of Rural Economics, 2(4), pp.1-14. http://ruraleconomics.kiau.ac.ir/article_527728_0.html
2) Tayebi, Z., Najafi, B., 2012. Determination of vulnerability and risk management in microcredit programs: applying risk sharing model and panel data approach, Iranian Journal of Agricultural Economics,5(4), pp.25-49. http://www.iranianjae.ir/article_9485_en.html
- Working Paper
My work focuses on the investigation of determining the contribution of TFP growth to output growth, assessing the potential impact of weather, and water scarcity on agricultural production in Greater Middle East. This is a region where food security and political stability have very much been affected by water availability and where droughts have the potential of inciting revolutions. I also consider economic, as well as political and social factors that might be affecting agricultural performance and incorporate various environmental characteristics of each country to understand efficiency differences across them. The results indicate increasing agricultural productivity during the period with innovations contributing approximately 30% to agricultural output growth. Most frequent extreme drought episodes and irrigation affect agricultural performance substantially in the region.