Research

On a Set of Methods to Address Problems with Two Stage Data Envelopment Analysis (Job Market Paper)

Abstract:

The usefulness of two-stage DEA or SF has been a source of some debate in the literature. If a formal statistical model is posited then the statistical integrity of any two-stage procedure requires that the environmental variables used in the second stage should be independent of the input variables used to estimate technical efficiency in the first stage. This is easily done in a regression setting but less easily implemented when using linear programming approaches. We introduce a propensity scoring approach to control for environmental variables when constructing the frontier so that the influence of environmental variables can be naturally estimated in a second-stage regression, dealing of course with limited nature of the dependent variable (Simar and Wilson 2007). We examine our new estimating procedure in terms of its statistical properties, a set of Monte Carlo experiments, and revisit several studies that have utilized traditional and newer approaches to addressing this problem.

Texas Electricity Generation Market after the Reform of 1995

Abstract:

This paper is concerned with the effects of electricity market reform of 1995 on electric power generators in Texas. Electricity market of Texas is on its own with a negligible overlap of other markets and is a good example of self-contained environment. With the introduction of the wholesale and retail electricity markets and an opportunity for consumers to buy electricity from whichever company they prefer, regulators anticipated an increase in efficiency of production and transmission of electricity due to the increase of private participation in these activities. It is transmission of electricity that is considered to have economies of scale not generation, but our expectation is that with more private firms in the market the competition will induce generators to operate as best as they can to get the contracts for their electricity. We find that after the introduction of the reform the overall efficiency of the market didn’t change (estimated with both DEA and SFA). Malmquist Index decomposition shows that TFP and efficiency deteriorated by about 4%, while technical change was positive 1%.

Semiparametric Estimation of Production Functions with Shape Constraints

Abstract:

Fidelity to economic theorems and flexibility of functional forms are both desired in empirical modeling of production functions. In this paper we present a flexible semiparametric estimator for production functions that reserves two important traits, monotonicity and concavity. In particular, our estimator extends Ramsay’s (1998) series-based smooth monotone estimator to incorporate concavity and further generalizes it allow for multiple inputs. We suggest a simple algorithm to solve the proposed nonlinear estimator, and present its large sample properties. Numerical examples demonstrate good performance of proposed estimator. Lastly, a simple empirical application to production function estimation is provided.

Pavlo Demchuk & Valentin Zelenyuk, 2009. “Testing differences in efficiency of regions within a country: the case of Ukraine,” Journal of Productivity Analysis, Springer, vol. 32(2), pages 81-102, October

Abstract:

In this paper we synthesize and adopt the recently developed methods in efficiency analysis to the case of comparison of regions within a country. We take Ukrainian regions as a subject of investigation, yet the same toolkit can be applied to test disputable differences in productivity for many other countries where such questions can be of national concern (e.g., Belgium, Great Britain, Spain, etc.). Contrary to common perception of economists focusing on Ukraine, we find no significant differences in distributions and aggregate efficiencies between the agricultural and industrial regions, neither between western (mostly Ukrainian speaking) and eastern (mostly Russian speaking) regions of Ukraine. However, we find strong support for a rapidly increasing gap between the capital (Kyiv) and all the regions since 2001. Using truncated regression analysis with bootstrap we also find robust evidence that the inefficiency of regions is positively related to alcohol and tobacco consumption, the amount of foreign direct investment and inversely related to criminality in the region. On the other hand, we also find strong evidence that amount of capital in the region and its wealth is positively associated with efficiency level of this region.