DSGE model has now become the hallmark of policy antidote. The macroeconomists now feels intimidated that the legacy and the traditional belief no longer hold. Many efforts to persuade the researcher to integrate macroeconomic analysis into the model prove abortive. Lately, earlier theoretical notion and past macroeconomic model are fading away.
Paul: I don't think anything is wrong with the Economic models. Each model has its own merit. Sometimes the government doesn't follow the economic models. That's what is wrong, in my opinion.
The problem with macroeconomic models is that they are based on assumptions that are to a large extent not realistic. It holds certain factors constant without accounting fully for the effect of other factors outside the model. Until Economists are able to account for the dynamic nature of economic phenomenon in reality, macroeconomic models will continue to be flawed.
"Macroeconomic models" seems to mean econometric models of the Klein-Goldberger type, which were used for economic forecasts and advice and had their peak end of the 1970's. Important contents were the (Keynesian) multiplier and the Phillips curve. They came under pressure because of the stagflation, and after the Lucas Critique they were superceded by GE models.
I think that already from the beginning, the expectations concerning macroeconomic models were exagerated. As they are, at best, solutions for functions of predicted exogenous and past endogenous and exogenous variables, they become obsolete, if there are breaks in policy, behaviour or international relations etc. Another problem are the (mostly) considerable confidence intervals of the econometric estimates (the effect of which are, in general, not mentionned, because politiciens and the public are not interested in interval forecasts). And many econometricians are too strongly method-oriented and do not care much about economic theories. They are often naive in this respect: I heard a speach of Larry Klein in 1980 (after he had received the Nobel Price), where he said that one could make perfect econometric models (with exact forecasts), if one had as many data as are measured in astronomy.
Nevertheless, it is unfortunate, that the (New Classical/New Keynesian) DSGE models have won the competition, because for me that is nonsense economics (see my contribution on RG under this title).
Models operating with several variables cannot perfectly describe complex, open and non-linear economic systems. Another important problem of macroeconomic models is measurement. They provide data that does not reflect the real situation. Until we know human nature, we will not be able to create perfect social (macroeconomic) models. And it is so difficult and interesting ...... Or maybe it’s good that we cannot do it.
I believe there isn't anything wrong in macroeconomics models, as mentioned by Professor Paul Stock . These models are made with assumptions so that they are quite general and easy to comprehend. Without these assumptions the models would be too complicated. Trying to relax these assumptions would only enable us to explain specific cases. What went wrong is that decision makers forget all these assumptions in formulating policy.
Old economic models are not fading away; they are still with us because they have no new developments to replace these old and inaccurate models. Current literature still recites them as foundational works despite the fact that the current conditions of today's world are quite different than 50 years ago. We use words like "innovation" and "creativity" in today's discussion, but we are still bound by the thoughts of people in the past. We have technology and data that our predecessors did not have, yet we cannot think of anything new or better. We are using new data to plus in old models and claim this is "empirical" work.
Paul (Stock) and Yudhi think that there is nothing wrong with economic models, and that the fault is only on the side of the politicians (decision makers). I think one should differentiate between models and how they are used. If model builders make forecasts, they do not, in general, also publish the model and assumptions (with the exception of some forecasts of exogenous variables - e.g. international developments). Politicians - like other people - can believe or not believe the forecast. They cannot "foreget all these assumptions", which they do not know. And after the forecasting periods, one could know, if the forecasts were correct or wrong (which is often foregotten). If a model is used for policy simulation, its quality is not dependent on whether politicians follow (or not) recommandations or take measures which show favourable results in model simulations. But in case they follow, the real developments can often be rather different from those calculated by model simulations. These resuts are quite different for different types of models: Austerity policy has (almost) no influence in New Classical, but is negative for GDP in Keynesian multiplier models.
Anton Rainer Interesting observation: I hope respected Professor and Yudhi respond to your arguments soon.
I'm sorry for this late response. State budget is a very common example of a good macroeconomic application where the underlying assumptions are often forgotten. A finance minister will always publish the budget along with the. assumptions as a complete package. Nonetheless, politicians, and even economists, intentionally leave behind the assumptions when criticizing the budget. They realize that by keeping the assumptions out of the debates, the budget would seem illogical and unjustified.
I'm sorry for this late response.
I think the quality of the economic model in forecasting depends on the efficiency of the market and what is the ability of economic variables to explain the real market movement
Dear Colleagues and Friends from RG, Built, complex, multi-faceted macroeconomic models do not always work in practical applications because each subsequent major macroeconomic processes, strong increases and decreases in the economic development of countries are determined at least in part by new factors that are difficult to predict, and even more difficult to forecast. Such a factor that will seriously affect the global economy in 2020 is the current (mid-March 2020) development of the SARS-Cov-2 coronavirus pandemic causing COVID-19. The possibility of practical application of macroeconomic models is directly proportional to their degree of reliability in model reflection of the surrounding economic reality. On the other hand, the likelihood of erroneous application of a specific macroeconomic model in forecasting economic processes and / or improvement of specific economic and financial policies is directly proportional to the issue of the historicity of the data on the basis of which a specific macroeconomic model was built, and in relation to the degree of complexity and multi-aspect of a given model.
Do you agree with me on the above matter?
Please reply.
I invite you to discussion and scientific cooperation.
Thank you very much.
Best wishes.
Dariusz Prokopowicz
- Badges
- Science topic
- Similar topics
- Psychology
- Emotion