Applied Econometrics February 23 2005 JinTan Liu Department

Applied Econometrics February 23, 2005 Jin-Tan Liu Department of Economics, National Taiwan University and NBER Email：liujt@ntu. edu. tw Telephone： 02 -23519641 ext. 520 1

Introduction Kennedy, Peter (2003), “Chapter 21: Applied Econometrics, ” A Guide to Econometrics, Blackwell Publishing. 2

Introduction 1. Economic Theory, Interesting Topics: n Wooldridge, Jeffrey M. (2003), Introductory Econometrics: A Modern Approach, South-Western. Chapter 19, “Carrying out an Empirical Project”. n Policy Evaluation: Natural Experiment n NBER web site: www. nber. org n Taiwan Study: Ø 1968 Taiwan 9 -years education extension program Ø 1995 Taiwan National Health Insurance Happiness Research in Economics 3

Topics n Plug and Vijberberg (2003), “Schooling, Family Background, and Adoption: Is It Nature or Is it Nurture? , ” JPE, 111(3), 611641. n Plug (2004), “Estimating the Effect of Mother’s Schooling on Children’s Schooling Using a Sample of Adoptees, ” AER, 94(1), 358 -368. n Persico, Postiewaite, and Silverman (2004), “The Effect of Adolescent Experience on Labor Market Outcomes: The Case of Height, ” JPE, 112(5), 1019 -1053. 4

Introduction 2. Data: n “Econometrics is much easier without data. ” (Verbeek, 2000) n “At least 80 percent of the material in most of the existing textbooks in econometrics focuses purely on econometric techniques. By contrast, practicing econometricians typical spend 20 percent or less of their time and effort on econometric techniques per se; the remainder is spent on other aspects of the study, particularly on the construction of a relevant econometric model and the development of appropriate data before estimation and the interpretation of results after estimation. ” (Intriligator, Bodkin, and Hsiao, 1996). 5

Introduction 2. Data: n Griliches, Zvi (1986), “Economic Data Issues, ” Handbook of Econometrics, Volume III, Elsevier Science Publishers. n Deaton, Angus (1995), “Data and Econometric Tools for Development Analysis, ” Handbook of Development Economics, Volume III, Elsevier Science Publishers. n Deaton, Angus (1997), “Econometric Issues for Survey Data, ” The Analysis of Household Surveys: A Microeconometric Approach to Development Policy, Johns Hopkins University Press. 6

Introduction 2. Data: n Hamermesh, Daniel S. (1999), “LEEping into the future of Labor Economics: the Research Potential of Linking Employer and Employee Data, ” Labor Economics, 6, 25 -41. Ø Why bother with linking employer-employee data? Ø New directions for research and policy. n n Hamermesh, Daniel S. (2002), “International Labor Economics, ” NBER working paper 8757. NBER, Journal of Economic Data. (A forthcoming Journal) 7

Introduction 3. Econometric Method: n “the current disconnect between economics and econometrics” … “in the past two decades, the gap between econometric theory and empirical practice has grown, ”… “command of statistical methods is only a part and sometimes a very small part of what is required to do firstclass empirical work, ” (Heckman, 2001). n “ it is not what you know about something which is important but rather how you use it. ” (Pagan, 1999, p. 374). 8

Introduction 3. Econometric Method: n Hamermesh, Daniel S. (1999), “The Art of Labormetrics, ” NBER working paper 6927. n Jones, Andrew M. (2000), “Health Econometrics, ” Handbook of Health Economics, Volume 1, Elsevier Science Publishers. 9

The Ten Commandments of Applied econometrics Rule 1: Use common sense and economic theory n The role of theory extends beyond the development of the specification; it is crucial to the interpretation of the results and to identification of predictions from the empirical results that should be test. 10

The Ten Commandments of Applied econometrics Rule 2: Avoid type III errors n A type III error occurs when a researcher produces the right answer to the wrong question. 11

The Ten Commandments of Applied econometrics Rule 3: Know the context n Do not try to model without understanding the nonstatistical aspects of the real-life system you are trying to subject to statistical analysis. (Belsley and Welch, 1988). n History, institutions, operating constraints, measurement peculiarities, cultural customs. n How were the data gathered? 12

The Ten Commandments of Applied econometrics Rule 4: Inspect the data n Economists are accused of never looking at their data. n Economists are unique among social scientists in that they are trained only to analyze, not to collect data. n Data generation is a dirty, time-consuming, expensive and non-glorious job. n Environmental Economics: Benefit Estimation, Contingent Valuation Method. n Summary Statistics, Graphs, and Data Cleaning (any observations impossible, unrealistic or suspicious? ) 13

The Ten Commandments of Applied econometrics Rule 5: Keep it sensibly simple n Econometricians employ the latest, most sophisticated econometric techniques, often because such techniques are novel and available, not because they are appropriate. n Think first why you are doing before attacking the problem with all the technical arsenal you have and churning out a paper that may be mathematically imposing but of limited practical use. (Maddala, 1999). 14

The Ten Commandments of Applied econometrics Rule 6: Test the estimation n to check that the results make sense. n The signs of coefficients as expected? Important variables statistically significant? Are coefficient magnitudes reasonable? Are the results consistent with theory? 15

The Ten Commandments of Applied econometrics Rule 7: Be prepared to compromise n Econometric theory courses students are taught standard solutions to standard problem, but in practice there are no standard problems, only standard solutions. 16

The Ten Commandments of Applied econometrics Rule 9: Do not confuse statistical significance with meaningful magnitude n Very large samples can give rise to estimated coefficients with very small standard errors. n Look at the magnitude of coefficient estimates as well as their significance. 17

The Ten Commandments of Applied econometrics Rule 10: Report a sensitivity analysis n Are the results sensitive to the sample period, the functional form, the set of explanatory variables, or measurement of proxies for the variables? n Are robust estimation results markedly different? 18