These workshops are open to any interested faculty and graduate students at the University of Kentucky. Their purpose is to introduce workshop participants to econometric and statistical techniques being used increasingly across the social and behavioral sciences. As the focus of the workshops is on applications of these techniques, the instructors will provide a variety of applied examples along with several hands-on exercises using STATA.
Below are descriptions of the two workshops being offered in the summer of 2010. If you have questions about the content of the workshop please feel free to contact the workshop instructors. If you have questions about the workshop series please contact William Hoyt at whoyt@uky.edu or 257-2518.
Last summer, the Martin School offered a successful workshop on hazard models. This summer, the Martin School is joined by Economics and the Research Office to offer two workshops in the first and third weeks of June. Learn a new technique and meet people in other departments working on similar problems from slightly different perspectives.
Workshop on Spatial Econometrics
Chris Bollinger and Jihai Yu, Department of Economics
Date and Time: June 1 – 4, 2:00 – 4:00 PM
Location: Gatton 305
We are offering a short course in introduction to spatial modeling and estimation. The course is designed to introduce students to concepts of spatial econometrics, spatial models, testing for spatial dependence, and estimation of spatial models. Spatial models are heavily used in social sciences such as economics, political science and sociology. Data which derive from geographic regions, such as census tracts, counties or states do not typically fit the independent sampling assumptions made for most typical estimation strategies.
This class will begin by exploring concepts of spatial dependence, standard models of spatial dependence and their impact on estimation of model parameters. We will discuss and interpret the standard model of spatial dependence and spatial autocorrelation. Implications of choosing particular spatial weights matrices and models are discussed as well. We will examine estimation including both Maximum Likelihood and Instrumental Variables. The final day will be spent in the computer lab, applying the techniques to some example data using STATA.
Questions? Contact Chris Bollinger at chris.bollinger@uky.edu or 257-9524.
Workshop on Frontier Functions
J.S. Butler, Martin School of Public Policy and Administration and Department of Economics
Date and Time: June 16 – 18, 2:00 – 4:00 PM
Location: Gatton 313
Downloads
Frontier Functions (Powerpoint Presentation)
Frontier Functions (PDF Document)
Frontier Data, Excel (.zip)
Frontier Data, STATA (.zip)
Suppose that a researcher wishes to estimate a production function or cost function. However, rather than estimating average production or cost suppose the researcher wishes to estimate the maximum possible production given a set of inputs or the minimum possible cost of a set of outputs. OLS estimates the average production or average cost. The purpose of frontier functions is to estimate maxima or minima of a dependent variable given explanatory variables, usually for the purpose of estimating production or cost functions.
Stochastic frontier analysis (SFA) and data envelopment analysis (DEA) will both be covered, with accompanying statistical, econometric, and theoretical background.
An Outline of the Workshop: Frontier functions in published research
0. Introduction: the purpose of frontier functions
1. Review of OLS regression
2. Review of MLE
3. Review of probability distributions from 0 to infinity
4. Disturbances in frontier functions:
5. One sided disturbances, e.g. gamma distribution
6. Composite disturbances
7. Fixed and random effects in frontier functions
8. Corrected OLS and fixed effects as measures of efficiency
9. Stata: frontier and xtfrontier
10. Data envelopment analysis (linear programming)
11. Data envelopment analysis with statistical properties
12. What Stata allows and what Stata does not allow
13. When estimation of frontier functions fails