1 / 17

Econometric Modeling

Econometric Modeling. How do we determine if econometric analysis is credible? Leamer (1983) focused on robustness: What is robustness? The sensitivity of the results to the functional specification and other key assumptions

cleave
Télécharger la présentation

Econometric Modeling

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Econometric Modeling

  2. How do we determine if econometric analysis is credible? • Leamer (1983) focused on robustness: • What is robustness? • The sensitivity of the results to the functional specification and other key assumptions • Concluded that there was little robustness and little empirical work was credible • Angrist and Pischke focus on the design base • Emphasize the identification of causal effects • Need real or natural experiments to clearly separate groups, allowing the analysis of causality • They argue that IO and macro have ignored this perspective

  3. Empirical Microeconomics has experienced a “credibility revolution” • Much work is now considered empirically robust with significant policy relevance • Partly due to greater attention to robustness • But empirical research design has improved as well What is good research design? • Goal is to improve the data available for analysis • Gold standard is Randomized Controlled Trials (RCTs) • Another useful design is “quasi” or “natural” experiments • Use of instrumental variables • Selection correction (Heckman)

  4. Main ideas behind RCTs • RCTs try to bring the controls of hard science research to social science analysis • Some treatment is envisioned • Participants are assigned randomly to the treatment and control group • Because getting the treatment is random, difference in the outcome, after controlling for covariates, is attributable to the treatment • Removes selection effects. What are selection effects • Covariates help control for differences in the way the treatment impacts differ across groups • A problem with RCTs is that there is selection into the experiment – people who agree to participate may be different than those not willing to participate

  5. What is the idea behind natural experiments? • Basically the same as an RCT, but with less control in assignment to group • Looking for something natural that randomly assigns people into separate categories for getting treatment or not. • More rare than people like to think • Need to meet a high standard; many seeming exogenous differences are endogenous • Looking for something unrelated to the treatment that separates groups • Best are natural disasters, etc. Often different political outcomes are used, but that suffers from the “Tiebot” effect • Does eliminate the selection into RCTs problem

  6. Gain in empirical microeconomics (Angrist and Pischke) • (Most fields) Applied economists don’t pin causal interpretation to the results based on econometric methodology alone • Focus is now on design, institutional or data-driven case for causality • But a caution about natural experiments and the Tiebot problem • Solon (1985) estimated effects of unemployment insurance on duration of unemployment spells • Compared states that recently changed standards • Ignores that the changed standards could be endogenous. Long spell states might have purposely tightened standards

  7. Why things are better today • Better and more data • Fewer distractions on functional forms and methods • Understanding that regression and 2SLS approaches are good for finding average effects (and supported by econometric theory) • Differences in differences analysis • Better corrections for statistical anomalies like heteroscedasticity and serial correlation • Nonparametric analysis • Other advances like quantile regressions • Highlighting of specific forms of differences through better design

  8. So what goes on in Macroeconomics? • Sims (1980, Macroeconomics and Reality) argued that structural macro models use assumptions about exogeneity to achieve identification • The “Lucas critique” and Kydland and Prescott, argue that can’t learn anything from econometric analysis of past policy changes • Would need to tease out the structural parameters underlying individual behavior • Policy changes change the constraints of that optimization problem • Inspired the work of Sargent and his group • Helped the move towards calibration models

  9. More about Macro • Lucas claimed there is no way to do experiments in macro – the cost is too high • Angrist and Pischke claim Calibration models don’t do the job • Numbers are chosen so the theoretical model tracks real data • Produces no evidence of the magnitude or existence of causal effects, they are still assumed • So when a parameter is changed, the causal affects are still assumed • “Harmless, but still theory” • Romer and Romer (1989) and Friedman and Schwartz (1963) have looked for “natural experiments” in Fed behavior

  10. Issues about experimental design • External validity: do the impacts that are observed carryover if the magnitude change of the variable used to define the experiment is very different? • Internal validity (the design) makes experiments narrow and idiosyncratic • But empirical evidence is always local to the data • The underlying variation never is completely representative, so extrapolation is always speculative • Calls for repeated experiments, with a range of values • Accumulate more evidence

  11. Issues about experimental design (continued) • Relevance: experimental paradigm leads researchers to seek good natural experiments rather than to focus on important questions • Analyze game shows instead of poverty • But some supposedly “trivial” work has important implications • Applications of behavioral economics issues like presented-oriented bias in health club memberships reveals information about far reaching policy implications about obesity • Accumulate evidence across settings and designs, and we gain more general understanding

  12. Sims Response • Economics is not an experimental science, and cannot be • Natural experiments and quasi-experiments are no more experiments than CGE model changes • Experiments must be replicable • Thoughts about Macro are nonsense • The problem is inference, not modeling • Same data are subject to different interpretations • Hence, use data to narrow subjective disagreements • He is very much in the spirit of Pearson; the best we can do is come up with models that mimic the data. We never have a model of reality.

  13. Kennedy: Ten Rules for Applied Econometrics 1. Use common sense and economic theory • Use good statistical practices • Match like measured variables • Select functional forms appropriate for your dependent variable (beta function for a variable with values constrained between 0 and 1) • Don’t add trends for trendless variables • Don’t use a formula for your empirical work; think about what you are doing

  14. 2. Avoid Type III errors (producing the right answer to the wrong question) • Corollary, an approximate answer to the right question is worth more than a precise answer to the wrong question 3. Know the context, which means get the facts • How was the data collected and imputed? • How were observations selected? • These are parts of my Know your data rule • But also, understand the system you are trying to model

  15. 4. Inspect the data (I need say nothing more on this) • But put together graphs of the data to see patterns and anomalies 5. Keep it sensibly simple • Begin with simple models, then make them more complicated • This is the empirical analog to what I said about theoretical modeling • Conflict between complexity (general) and simplicity (specific) • Use the simplest method appropriate for your analysis

  16. 6. Use the interocular trauma test (what is this?) • Look at the results until the answer hits you between the eyes. • Then check that the results make sense • Signs, magnitudes, significance 7. Understand the costs and benefits of data mining • Goal is not a high R2 • Significance level is contextual • Specification depends on what data you have, and if it is relevant

  17. 8. Be prepared to compromise • Understand the gap between the statistical theory underlying your analysis, and the actual application you are doing • For example, there are few populations that are truly infinite 9. Do not confuses statistical significance with meaningful magnitude 10. Report a sensitivity analysis • Pay attention to robustness • Confess your errors and shortcomings (know the limitations of what you did, and admit to them)

More Related