Econometric Modeling - Science topic

Nobody (I hope) has said that explanatory variables should be symmetrically distributed. I am curious on how you can interpret percentages of a variable taking negative values. Anyway, don't expect to be published with your model. I will stop here this discussion.

Chuck A Arize Thank you! Do you have some examples of recent empirical papers on segmented market theory (maybe your own research if relevant)?

The suggestion by deleted profile on February 23 is not correct. TraditionalARDL estimation has been used to approximate the behaviour of dynamic systems involving I(0) variables for over 40 years. The ARDL Bounds test involves 1(0) and I(1) variables. It works when there is one cointegration expression because that cointegrated expression is I(0). If there is no cointegration the coefficient on the error correction term is zero and everything in the estimated equation is again I(0).

Can I remind you that I(2) variables an those with higher orders of integration can not be included in an ARDL bounds test. You must also satisfy certain exogeneity conditions. There must be no feedback from the dependent variable to any of the explanatory variables.

Implementing ARDL is very easy in many software packages - perhaps too easy. My advice would be to steer clear of the article.

Sustainable climate risk management at the firm level is a complex and evolving field. It typically involves identifying, assessing, and mitigating climate-related risks while also seeking opportunities for sustainability and resilience. There's no one-size-fits-all model, as each firm's specific circumstances, industries, and locations may require different approaches. However, you can consider the following general framework as a starting point:

Climate Risk Assessment: Begin by conducting a thorough assessment of climate-related risks. This includes physical risks (e.g., extreme weather events, sea-level rise, and resource scarcity) and transition risks (e.g., policy changes, market shifts, and technological advancements).

Collect Climate Data: Gather data related to climate trends, both historical and future projections, that are relevant to your business and industry. Consider using climate models and scenario analysis.

Financial Impact Analysis: Assess how these climate risks might impact your firm's financial performance, assets, and operations.

Develop various climate change scenarios to understand how different climate outcomes could affect your business. This is critical for risk management and resilience planning.

Physical Risk Mitigation: Implement strategies to protect your physical assets from climate-related damages, such as investing in resilient infrastructure or relocating vulnerable assets.

Transition Risk Mitigation: Develop strategies to adapt to policy changes and market shifts. This may include diversifying products and services, transitioning to low-carbon technologies, and staying informed about relevant policy developments.

Consider integrating sustainability measures to reduce your firm's carbon footprint and environmental impact. This can include energy efficiency initiatives, renewable energy adoption, waste reduction, and supply chain sustainability.

Develop resilience plans that ensure business continuity in the face of climate-related disruptions.

Consider voluntary reporting through initiatives like the Task Force on Climate-Related Financial Disclosures (TCFD) to communicate climate risks and strategies to stakeholders.

Engage with industry peers, government entities, and non-governmental organizations to stay informed about the latest climate developments and best practices.

Be prepared to adapt your strategies as climate risks evolve and new information becomes available. Continual learning and flexibility are key.

Regularly monitor and assess the effectiveness of your climate risk management and sustainability measures. Transparency in reporting can help build trust among stakeholders.

Ensure compliance with relevant climate-related regulations and standards. Stay informed about evolving regulatory requirements.

It's important to tailor your firm's climate risk management model to your specific industry, location, and risk profile. Consider seeking advice from climate risk experts, and collaborate with stakeholders to develop a comprehensive and sustainable approach to climate risk management.

here is a resource

It depends on how the shock influenced the treated and control unit.

If you can argue that the shock influenced all units in the same way, then the inclusion of two way fixed effects (particularly time fixed effects in this case) can help you to rule out the effects of such 'common' shock.

However, if there are reasons to think that the structural shock differently affected treated and control units, then this could result in potential biases in the estimation and identification of the average treatment effect.

There is a novel literature dealing with the correct identification of such effects. I would suggest you reading the following paper:

"VISUALIZATION, IDENTIFICATION, AND ESTIMATION IN THE LINEAR PANEL EVENT-STUDY DESIGN" from Simon Freyaldenhoven, Christian Hansen, Jorge Pérez Pérez and Jesse M. Shapiro. You can find the paper here: http://www.nber.org/papers/w29170

I particularly suggest you to read section 3.1. of the paper as long as accompanying the paper there is the Stata command "xtevent" which is particularly useful to deal with such econometric analyses.

Answer from ChatGPT

It is possible to build a GMM with the number of observations (n) being lower than the number of time periods (t). However, in general it is recommended to have a larger number of observations than the number of parameters being estimated. This is known as the "curse of dimensionality" and can lead to overfitting and poor model performance. Additionally, a GMM requires estimation of the mean, covariance, and mixing proportion for each component, which may be difficult to do with a small number of observations. Therefore, it is generally recommended to have a sufficient amount of data when building a GMM model.

I recommend you NOT to convert quarterly to monthly data. You will have more observations but the increase of precision will be illusory. You say to have an I(2) variable but you don't say anything about the other variables. For example, if all of them are I(0), I don't think that cointegration will help. You should say more about your problem if you want appropriate suggestions.

Dear Simon, you confirm my contribution. Statstistics was applied by many sciences long before the term econometrics was invented. Like mathematics (one could, of course, say that statistics is part of mathematics), it is universal in the sense, that it can be applied for rahter different scientific studies. Nevertheless, it can only be applied, if there is a theory of the respective science.

You should read the Global Burden of Disease articles on health care financing. Their appendices fully disclose their variables (& they've spent a lot of time & effort dettermining the best ones) and they provide their (sophisticated) statistical code. Lead author on most of that work is Joseph Dieleman or something like Global Burden of Disease Health Financing Collaborator Network. A full biblio of this work is at https://www.healthdata.org/about/joseph-dieleman if you click on the publications tab. The pubs are all open-access & most are on ResearchGate. :-)

It is not clear what you want to study. There is no econometric model to help you, if you do not know what you want to do.

I asked that question to the person who code that package, and he said C(9) coefficient does not have any meaning here, just ignore. It comes up because the package is written for the old version of Eviews and has not been updated that is why.

I think the similarity of your robustness check to the main method depends on the properties of both estimators. For instance, using augmented ardl as a robust test for bootstrap Ardl could yield similar result. Same might go for LSDV and GMM because they are both dynamic model method and can also correct for endogeneity problem. So the similarity depends on the similarity of the estimators. The coefficient doesn't really matter though but the similarity of the signs of the coefficient is very important..

Regards

Kindly visit..

http://www.laits.utexas.edu/orkelm/kelmpub/VSM2013_Manual.pdf

I think you are guided by your objectives and also following appropriate statistical procedure. there is hardly a statistical tool that incorporate every thing eg. breaks regimes, data in panels, cross sectional, time series, etc.

Hi Sultana Razia

I would recommend you read about the difference between EFA and PCA first. Whether or not you should run an EFA has nothing to do with the number of response options on the indicators, five or otherwise. In general, EFA is preferable to PCA as it is considered to be the 'real' factor analysis. The are many threads on RG on this issue.

Best

Marcel

YES OF COURSE WE CAN THERE IS A FIELD CALLED MICRO-ECONOMETRIC

in this case, you must use a dummy variable that is 1 and 0. Those are adopter is treated as 1 and non-adopter treated as Zero. Than Run multiple linear regression model

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

Dear, conducting any standard model depends on an important factor, namely the number of observations included in the model, for example, if the observations are small, the Phelps-Peron test can be conducted to test the stability, and if the observations are large, the whooping-full test can be conducted, and in light of the stability results, we can determine the model that can be conducted Julius Hogan

Before deciding on the econometric model, you should go through the stationarity test (ADF test). If the data are stationary, OLS Regression with a log-linear model would be fine. But, if not, you may go for VAR or ARDL. You should also check the robustness of the model by going through residual tests such as Autocorrelation LM Test.

It seems a very small sample to apply microeconometric techniques. Having 27 seven observations and 10 covariates, at most, you will have 27 - 1 - 10 = 16 degrees of freedom. This is pretty low. If I had to decide to pursue a project based on that, I would try to avoid it.

It is really closer to multiple times series than panel data. Have a look at this link:

Dear Rand,

I could not open your data set, because I have no unpacking program. But that is not so important, as I have looked on some of the data in the World Bank Statistics.

My first proposal is to choose a higher dimension for the tourism revenues (e.g. billions of units - $?). This would lead to a better readability of the results.

For the first step of the specification of the model/equation to estimate, you need no data. Simply ask yourself and others, how the political stability and other political circumstances in a country would influence the decision to go there as a tourist and how much to spend.

Then I would look on the time series (from which one can get insights on changes over time) for some of the countries to find patterns between the exogenous variables and the dependent variable. The main problems may be: 1. The political variables might have a rather limited influence. 2. They may show too little variances. 3. A change of e.g. political stability will not lead to an instantaneous change of tourism. Improvement will likely have its positive influence only after some years, whereas a change to instability will have a shorter-run effect (asymmetric behaviour). 4. You will likely find high correlations between the political variables, which means that you should omit some of them. 5. The indices are far from an ideal cardinal scale, it is even not clear, whether a higher value really measures an improvement.

If you (hopefully) find-out some relations, you can try simple OLS estimations with one of the political variables lagged for several years and you may try to do that for transformed variables (percentage change of revenues with absolute differences of the indices).

The comparison between countries could, in principle, give you an impression of the political influence on (the level of) tourism, but I fear that other factors (landscape, seaside, history, culture) are by far dominating. If your aim is only to exercise econometric methods, you can try an estimate for the whole data set. For this purpose, I recommend, to calculate indices instead of the absolute revenue values (to neutralize the “non-political” differences between countries. But you should take into account that you have more explaining variables than you used in your first approach – namely all these political indicators with different lags. I think, that for such an approach you lose too many degrees of freedom and, therefore, you would not get reasonable results.

If I had this difficult (not to say: impossible) task, I would never do a pooled estimation, but I would try to find good time series estimates for all 10 countries, and then compare the estimated coefficients to find-out similarities.

R is one of the best tools when it comes to spatial modeling in General, and as I have worked with R multiple times you just need to find the right cran file that serves well your manner, Also check GIS software in literatures, many studies used it for Bayesian spatial modeling. Good luck

How to make ICT index, I mean what weights to assign internet, mobile and fixed phone line?

I notice the following about your specification. (1) Your inclusion of a constant (and its subsequent significance) means you allow for (and find) a trend in the real exchange rate independent of any trends in the other variables. Is that economically reasonable? (2) I assume the CRISIS variable is a zero-one dummy for time periods with a "crisis" of some sort. Apparently it is not in the cointegration vector. Why not? If it were, then I'd expect to find CRISIS differences in the error correction equation. Instead, you have it in levels. Thus you specify that a temporary crisis has a permanent effect of the level of the real exchange rate independent of the other variables. Is that what you intend? (3) You do not include the lagged difference of the real exchange rate in the error correction equation. Why not? Normally it would be there.

Dear Gonçalo Maria ,

Please note that transforming variable(s) does NOT make the series stationary, but rather makes the distribution(s) symmetrical. Application of logarithmic transformation needs to be exercised with extreme caution regarding properties of the series, underlying theory and the implied logical/correct interpretation of the relationships between the dep variable and associated selected regressors.

Reverting to your question, the proposed solution would be to use the Autoregressive Distributed Lag (ARDL) model approach, which is suitable for datasets containing a mixture of variables with different orders of integration. Kindly read the manuscripts attached for your info.

All the best!

R-H bicorrelation test -https://faqs.tips/post/rolling-hinich-bicorrelation-test-can-anyone-help.html

Hello Robert Traistaru ,

It is shown as lnpop, so i think you need to transform the variables into log. Simply follow this steps if it is the case,

quick<generate series<lnpop= log(pop). and log the rest of the variables in the same way.

after that put lnpop in the estimate equation.

Best wishes for you.

Ibrahim Niftiyev , probably the reviewer wants to see not only whether a variable affects or not the other (i.e. the results of the Granger causality tests), but also to which extent (the magnitude and temporality of the dynamic relationship, something you can obtain from the IRFs of a VAR model). If you want to apply a VAR but you have a small sample size/data limitations, you want to consider a Bayesian VAR. Bayesian VARs are very popular and Bayesian methods are valid in small samples.

The Ljung-Box test is aimed at testing the independance of errors using residuals of an ARMA model estimated on the same data. But it makes use of autocorrelations so it is not powerful when the errors are uncorrelated but not independent. When applied to squared residuals, it can reveal ARCH and GARCH effects. Note that the errors of a ARCH-GARCH model are uncorrelated but not independent. Have a look at the excellent book by Francq and Zakoian entitled "GARCH Models: Structure, Statistical Inference and Financial Applications" published by Wiley in 2010.

See these links:

Depends on the research question and the target audience. While policy and social related issues are better handle using qualitative methodology it loses the generality and reliability of the findings to extend to other situations and context.

Generally speaking, MAPE may not be the best choice. Please let me recommend this guide to measuring forecasting accuracy:

Btw, if you have time series with non-negative actuals, you can try data transformations to improve forecasts. If you apply data transformations, I would recommend MAE or the AvgRelMAE.

Maybe, some helpful information is included here.

Are you attempting prediction or are you trying to do some form of “causal” relationship. If you are estimating a ”causal“ model you conclusions are conditional on the model estimated. Strictly speaking, it would be better to use the adjusted degrees of freedom - particularly with your small sample. In this case, a non-significant coefficient does not necessarily imply that the coefficient is truly zero. It is more likely that your sample is too small t establish a significant result. Its p-value must not be very far from your significance level. If the estimate is of the sign and magnitude expected by theory I would accept the value and report the p-value. Esimating 5 coefficients is a big demand from a sample of 23 obs.

If you are simply doing prediction or forecasting and are not attributing explanatory power to your coefficients you might be better with a simpler model which might have better predictive ability.

Dear Ibrahim Niftiyev ,

Econometric models estimated with non-stationary data are profoundly invalid and misleading (Greene, 2002). An example of a simple scenario: - in a regression with one regressor, there are three variables that could be stationary or non-stationary; namely the dependent variable (Y), the regressor (X), and the disturbance term (u). A suitable econometric treatment of such a model depends critically on the pattern of stationarity and non-stationarity of these three variables (incl. the dependant variable). Since quite often variables can be non-stationary at I(0), it is important to understand the forces behind such non-stationarity, which largely include structural breaks, deterministic trends, and stochastic trends. Differencing (including the explained variable as in your case) is a common appropriate in nonstationary models, and this is often correct (Granger & Newbold, 1974; Green, 2002; and Stock, James & Watson, 2011).

Granger, C. W. J., and Paul Newbold. 1974. Spurious Regressions in Econometrics. Journal of Econometrics, 2(2):111-120.

Greene, W. 2002. Time series Models. (pp. 608-662) In Econometric Analysis, 5th edition. Prentice Hall, Upper Saddle Rive, NJ.

Stock, James H., and Mark Watson. 2011. Introduction to Econometrics. 3rd ed. Boston: Pearson Education/Addison Wesley.

Thanks Prof. John for your info!

Check if the bivariate scatterplots are linear or at least monotonically increasing or decreasing.

The tool I normally use is public meetings with focus group discussions as a priority.

Dear sir,

I guess an intrinsic inclusive approach, founded/guarded through regulations would provide a sound participatory model that would help elevate socioeconomic balance. This should be considered a quite good approach for developing and underdeveloped countries.

If you are intended to go for some related publication, I would love to be a part of it.

My best regards,

To devleop these type of models on a single country, you have to start your study by a strong motivation to convince the reader why you do such a work on a single country. This has its limitations.

normally, we these type of questions Gravity models to be applied and Dynamic Panel data models for robustness.

As you are facing a single country, consult time series tools: Cointegration, ARDL, NARDL or VECM

It is not clear, whether your aim is to explain an extraordinarily high or low growth in 2016 (for that you will not need econometrics) or general effects on real and nominal GDP growth for a longer time period. Without an economic theory and a carefully derived model, econometrics will not help. To apply statistics, the main trouble will be to define what measures are reform and to transform these measures (which need not be directly measurable, e.g. agreements or treaties) into numbers. Another problem is that one should not mix up the effect on GDP and GDP growth. A (reform) measure would hopefully lead to a higher GDP level, but hardly to a higher permanent growth. It may take several years until the higher GDP is reached, then you have, of course, a higher growth during the adjustment, but not a higher growth afterwards (The percentage growth could even be lower because of the higher base). The timing of the GDP-effects is likely rather different for different measures.

Farshad Fathian Can you check, please. This might be the same as we discuss

What are the best econometric models for adoption study?

You may want to have a look at these two booksS:

I think you may need to program it by introducing a known link function on your frontier model. If you are using a parametric stochastic frontier model (SF), say, y=x'b+v-u, where v is error term and u is inefficiency random variable, you may need to specify a link function F such that y=F(x'b)+v-u, so that it can satisfy the support of y. Not sure if available packages are there in R. It would be much simplified if you can observe the agriculture income directly.

In terms of control variables, you have to look at " asymmetric information theory", and for DID, it is a good method but you check the robustness with PSM-DID, Two-Step System GMM, even DDD.

A VAR model will be a good alternative. The shocks to the model is well captured in VAR modelling and to see at what stage the model settle down after a shock. It is not for forecasting but to trace the impact of the shock using Impulse-response function and variance decomposition.

Thank you!

Hi Charles, I visited Switzerland twice for 1-month and 6-month internships. Nice country. Yeah, Switzerland's industrialization should be a nice research topic, but I wonder about your period of study. Do previous studies consider 1890-1913 as a period of insustrialization in Switzerland? Do you have quarterly or monthly data for running the VAR model? For how many Swiss industries do you have the gross added value? Of course (since Einstein's job in Switzerland), do you have the time series of Swiss patents for that period? By industry or broken by type of patents? Considering the empirical section of your study, (if you only have annual data) I would strongly recomend to add the 40 or 50 years after the period of study because a plain VAR model is not good for small samples of data (a Vector Error Correction -VEC- model should work better for such data, instead). But there is another reason explaining the cold reaction of the reviewers: a VAR model is not quite useful for discovering the industrialization determinants since it is an "atheoretical" statistical model: you may need a "Structural VAR" (aka SVAR model) or better a Structural VEC (SVEC) model. Then you will be able (abstracting from the data frequency I already mentioned) to propose a theoretical structure (meaning, useful research estimated parameters) to discuss following your hypotheses. Those hypotheses should come from the previous section, the core section, which is the "theoretical" section. Such a section should consider at least the hypothesis you can select from reading many studies about economic history of Switzerland BUT also review some growth models, like the classic ones usually reviewed in macroeconomics' textbooks (Solow model, etc.) and other more recent ones like Romer's model (which consider the idea of Research & Development (please remember my idea about patents). For discovering those models, Romer's book for graduate students is the best. Finally, allow me a personal comment: I do like reading about economic history because it is a great companion for macroeconomic economists. It is such a down-to-Earth topic !

Try to get a complete case analysis data (removing the missing values in all models) and then obtain the listw matrix. Then build a model and estimate the marginal impacts. That should work for you.

Post estimation diagnostic tests are performed to provide proof for validity of estimates. The following diagnostic tests are necessary when GMM technique is applied: 1) first order/ second order serial correlation tests. 2) Sargan test/ Hansen test for over-identification restrictions. 3) Wald chi square /F ratio test for join significance for the model. You have to check the instruments to ensure they are less than the groups employed in the analysis.

If it seems to refer that they are almost equal constructions, which should first consider evaluating the dimensionality of each variable.

There is an interesting inquiry on this issue in the following paper of Christer Gerde's

Christer Gerdes, 2011. "Using “shares” vs. “log of shares” in fixed-effect estimations," Journal of Economics and Econometrics, Economics and Econometrics Society, vol. 54(1), pages 1-7.

This could make absolutely sense...

Partial F-test will be useful hear. After the 1st variable is in, you add other variables ONE at a time. after 2nd bariable is added you have your y-variable as function of 2 variables giving model with 2 d.f. & certain Sum of Squares (SS). from 2-variable SS subtract 1-variable SS. that change in SS will have 1d.f. cost. So extra variable SS divided by '1' is the "Change in regression mean squares (regMS). Further, divide 2-variable residual SS (RSS) by 2-variable residual d.f to get Residual mean Squares (resMS). Now divide Change in regMS by resMS to get partial F-value & look up Tables for probability of partial F-value. If significant keep the 2nd variable in & do the same for any further independent variable you may want to add to your model. Adjusted Rsq is = 100*[1 - {(regSS/regDF)/(totalSS/totalDF)}]

To assume that other variable do not change in order to allow for an evaluation of partial variation in a dependent variable due to variation in the only independent while other variables do not change

Stata, Eviews and R can do time series measurement well. You need to know how to implement event study using these software from the literature review of other similar studies.

Volatility by their very nature is unpredictable. Perhaps a more manageable task is scenario building rather forecasting the path of the volatility.

Dear Prof. Raymond,

My aim is to see what happens to returns of stocks when they join a certain stock exchange index, do they generate abnormal returns?

I’m trying to study that for the last 3 years.

So the event is “joining the index” which happens with 2 dates (1) announcement date (AD), on which stock exchange announces the news of those stocks will join the index, and (2) change date (CD), which is the date for really including those stocks in the index, this CD is is decided and announced by the stock exchange during the AD.

I have attached a similar work for your kind reference.

Thanks for your reply.

Hello Alexa Drk. Generally speaking, Statalist (https://www.statalist.org/forums/forum/general-stata-discussion/general) is a better place to post questions about how to do X using Stata.

Second, I would think that if you are using -xtset-, you would want to use -xtreg- rather than -regress-.

Third, if you use the # or ## operators to include interactions (as in the file Aymen Ammari linked to), you'll be able to use -margins- and -marginsplot- to explore the nature of the interactions. See section 11.4.3 here:

HTH.

You can consider using the country's stock market index as a proxy for market portfolio.

There is no single, universal, multi-faceted model built of many indicators selected from fundamental and technical analysis, including market, economic and financial analyzes, indicator analyzes, due diligence and others that could be used to analyze the situation, economic situation, volatility, etc. for every capital market, including the currency market, stock market, market of raw materials and other production factors, and for every situation, i.e. bear market, bull market, balance, strong changes in trends, different levels of investment risk, various phases in the economic cycle of the economy, and irrespective of the national specifics of the operation of financial markets. Specific analytical models should be built for a specific economic situation, for a specific market. Then it is possible to achieve a high level of decision-making accuracy on the basis of forecasts formulated from this type of multi-faceted, complex indicator models for analyzing the current and prospective situation on a specific stock exchange market or other specific capital market operating in a specific macroeconomic environment, in a specific financial system, in a specific financial system country.

Greetings.

Dariusz Prokopowicz

Am not sure Santiago I follow what you said. Elasticity can be used and Beta weights can be used. If I understand you? I will interpret elasticity as I % increase in the RHS variable changes the regressand by the estimated sign and coefficient of the RHS eg LnY= 2 -0.5Ln(X) -- Here a percent increase in X decreases y by .5 on the avg

I agree with Matteo Mazzarano . Ten years is not enough, You could use moor than two or three independent variables , you need moor than 25 years as a ( time series data) for every country.

"I don't have the theory to include more control variables that may me important for this model." -- so this is a statement. You know the field and the arguments why or why not a variable might have to be considered. You state that, at the best of your knowledge. So you can defend it and clearly point the advantages and limitations of your model. If reviewers (and, later, readers) have different ideas, they are invited to discuss it.