This is a difficult subject. You need to understand the meaning of causality in econometrics.

Your results are conditional on your economic model. If another economist has an alternative model that implies alternative causal links to the policy effects and this model is also data consistent, his results may be different from yours. His results are conditional on his model. The causal links are determined by the economic model. Econometrics provides a means of quantifying the effects conditional on the model.

Pearl and Mackenzie (2018), The Question of Why, Hachette, is a readable introduction to the problem of causality.

Lee (2005), Micro-econometrics for policy, Program, and treatment effects, Oxford, deals with the use of micro-econometrics for policy effects.

Banerjee and Hendry (1996), The econometrics of public policy, Blackwell highlights a lot of the problems that arise in macro-econometric policy evaluation.

The book recommended by Yudhi Dharma Nauly is also a good reference

I find this book useful indeed. I am sure you will too.

https://www.worldbank.org/en/programs/sief-trust-fund/publication/impact-evaluation-in-practice

With research limitation, you can use panel data type for your research analysis.

Your research finding is not only determined by the type of data that you use. However, your choice of econometric models may influence significantly your research result. Hence,for identifying such a cause-effect relation, using appropriate model is essential. Since household opinions is not exogenously determined, multi-step regression models are recommended. Meanwhile, Propensity score matching (PSM) is advisable for impact.

I have the same question: how to measure the impact when having panel data available? I think the key method here is establishing granger causality. But in case of panel data the test gets quite counter-intuitive..

All econometrics methods can deliver the policy implications, if you ask the right question.

Panel data or not - this is a technical issue, more or less (by my opinion). Evaluation of policy effects is not an easy task - but, anyone interested could start from somewhere - e.g. https://economics.mit.edu/files/13924

And, there is always the question - which is "causing" what, ceteris paribus, and by what mechanism - and modeling such interrelations is "tricky"...

Wishing you success!

This is a difficult subject. You need to understand the meaning of causality in econometrics.

Your results are conditional on your economic model. If another economist has an alternative model that implies alternative causal links to the policy effects and this model is also data consistent, his results may be different from yours. His results are conditional on his model. The causal links are determined by the economic model. Econometrics provides a means of quantifying the effects conditional on the model.

Pearl and Mackenzie (2018), The Question of Why, Hachette, is a readable introduction to the problem of causality.

Lee (2005), Micro-econometrics for policy, Program, and treatment effects, Oxford, deals with the use of micro-econometrics for policy effects.

Banerjee and Hendry (1996), The econometrics of public policy, Blackwell highlights a lot of the problems that arise in macro-econometric policy evaluation.

The book recommended by Yudhi Dharma Nauly is also a good reference

Hi,

Please you have to also bear in mind that the method to be adopted depends on the type of data you are using for your analysis. so kindly check that as well.

Thank you and stay safe

: مجموعة كتب قيمة في الاقتصاد القياسي

الاقتصاد القياسي - جوجارات

الجزء الأول

https://www.mediafire.com/download/v61lb51pjeplo2n

الجزء الثاني

https://www.mediafire.com/download/7siz7lce8okdv5f

الاقتصاد القياسي - حسين بخيت وسحر فتح الله

https://www.mediafire.com/download/vke1azqnbz9c9st

طرق الاقتصاد القياسي - شيخي محمد

https://www.mediafire.com/download/pxip2p8aa13dbap

الحديث في الاقتصاد القياسي - عبد القادر عطية

https://www.mediafire.com/download/26nbd3xfts8n0gr

إيفيوز والقياس الاقتصادي - خالد السواعي

https://www.mediafire.com/download/whw2jw99k9ujcy2

مقدمة في الاقتصاد القياسي - كلجيان وأوتس

https://www.mediafire.com/download/yua6v17fyy8lh8t

مدخل لنظرية القياس الاقتصادي - تومي صالح

https://www.mediafire.com/download/a993gesijvilsbl

مبادئ الاقتصاد القياسي والرياضي

https://www.mediafire.com/download/7vxsn4gm8bkvd3v

مادة EViews في الاقتصاد القياسي

https://www.mediafire.com/download/jgaptprms2pf5s4

محاضرات في الاقتصاد القياسي

https://www.mediafire.com/download/8slsxe5j23jy9bt

Dear Hakam,

The policy impact analysis is too broad. Hence, are you looking macro, micro or panel or cross-sectional or any other sort of data. Even the richness of the data matters a lot. Hence, different factors tend to play as what type of econometrics model to choice.

It is wrong to begin a project with the question, which method to use.

First, one should be clear what one wants to analyse (Sule: Ask the right question?) and to develope a theory or search for a good theory. From that, you should develope a (rough) model (1. of Frain). One can and should do that, even there are no data. If you have data, you can look, if they fit into your model. Econometric estimations, if possible, may be the last step of the inquiry.

Choosing the right model really depends on the composition of social assistance.

If there is no discrimination in the age of those who must receive social assistance, then the Diff in Diff model would suffice.

On the other hand, if there is an age of eligibility or an income threshold for receiving social assistance, then the Regression Discontinuity Design model is appropriate.

I suggest to you these articles, I think they can help you.

Lemieux and Milligan (2008)

Article WORKING PAPER NO. 70 Incentive effects of social assistance:...

Fortin and al. (2004)

Article Welfare Benefits and the Duration of Welfare Spells: Evidenc...

Always start with the Economic problem you are trying to solve ? Then ask your question so you can shed light on the problem to improve decision making ?

حسب البيانات والمعلومات المتوفرة لدى الباحث

I agree with Sule Akkoyunlu

I agree with Akkoyunlu. However for general equilibrium framework, DSGe models would be appropriate

DSGE models are a dead end of economics. See my article "Nonsense Economics" on RG.

It’s generally best if you can apply multiple methods with adequate design and obtain consistent results. This will make your results more convincing. If I had to choose one methodology for your study (given the very limited information that I have), my guess is that a DiD or related event study methodology will be the best.

You’re trying to claim a causal effect of a certain program on crime rate; thus, evidence of a pronounced change in crime after program implementation should be relatively convincing (particularly if you have data on a number of similar programs implemented in different areas).

If you have 'with and without treatment' observation, better you use PSM. Meanwhile, if you want to estimate the 'change' and time effect, better for you to use DiD