What is the most acceptable method to measure the impact of regulation/policy so far?
I only know the Difference-in-Difference (DID), Propensity Score Matching (PSM), Two-Step System GMM (for dynamic) are common methods. Expecting your opinion for 20 years long panel for firm-level data.
In the empirical literature I have had exposure to, the DID has more extensive application relative to other methods like the PSM, etc.
Randomized control trials (RCTs) is called Gold Standard for measuring the impact of social interventions. IPA has many studies. Pls explore following link:
https://www.poverty-action.org/about/what-we-do
There are two issues I would raise. First, search the literature for what proponents of a policy would view as an exogenous measure. For example, full employment "surplus" might be a reasonable indicator of the position of fiscal policy. Second, consider the literature on event studies in finance, especially look for announcement effects.
Salam. If your data include the economic variables you can use ARDL model for long database
Difference in difference (DID), Matching and Regression discontinuity will be suitable for to measure the policy impact for long panel data.
Suppose an event, a policy, occurs during your time series. What can you know about the impact of that policy . . . no matter how you transform your data? The answer is nothing. Your detractors will correctly argue that when X policy occurred Y also changed, or Z or whatever. That same policy has to recur many times for you to identify its effect. It may recur in different places (countries? states or provinces?). You need to learn about "event time," because these policies are not contemporaneous. You may have to model key variables in these different places to shift the effect of the policy.
These techniques are very good and often used in microeconometrics. They seek to identify effects in a context of little information about each component of the sample. In this context of little information, estimating the mean treatment effect or the mean effect on the treated is already a challenge. When the panel is long you can do better and control the trends of each unit or estimate effects for each unit by changing the DID techniques and/or using techniques more like time series. See for example the text below:
Leonardo Oliveira thank you so much for your contribution, but is there any recent development or limitations of mentioned methods?
recent development
(1) Wooldridge two-way Mundlak regression and fixed effects and dif-in-dif
(2) synthetic control
(3) Cerulli, G. 2015. Econometric Evaluation of Socio-Economic Programs: Theory and Applications.
(4) Pesaran (2015) Time Series and Panel Data Econometrics
Hello, I think the most efficient method is a Panel Ardl if the integration of the variables must be in the 0 or 1 level.
Dear Sir Oluwaseyi
I have one question, if I found that all my variables have an order of intagration I(1) but only one of them is I(0) can I apply ARDL model in such case?
There are several methods that can be used to measure the impact of regulation/policy for long panel data. One such method is the two-way fixed effects estimator, which is often used in panel data analysis.
Another method is difference-in-differences (DID) estimation, which is a popular method for evaluating the impact of policy interventions using panel data.