Hi,
I am developping an economic growth model following Robert Barro's Model in his "Economic Growth" textbook and orientated on his growth determinants.
The majority of them fits into my model pretty well, but I was wondering, what could otherwise influence GDP growth. At the moment, I have variables including consumption, government expenditure, political variables, investment, inflation, capital stock, population, and trade.
It's just a brainstorming question, that I do not forget "important" variables, that would reduce the OVB, improve the fit of the model, and change the interpretation of my variables.
Thank you so far!
Don't forget the importance of low interest rates. Low rates encourages borrowing for investment, new business, plus home and automotive purchases...
Thank you both for your answers and attached files. I will take a look at it.
Interest rates would be an interesting topic to look at, when I also look at saving rates of countries. The problem that I have is, that for some constraints of my model, the constant variable of my regression is high and sometimes statistically significant and therefore at least hinting, that I ommitted some variables. I will check these and also the Barro-articles, I have not found before, and check if they further improve the fit of my models.
I don't know about the fit of your regression analysis as judged by the size of the equation constant, but if the R-squared is close to one (ex.: 0.9997) then you have a good fit. If you think there may be other significant factors that you may have left out, then gather data on all possible factors and do a correlation analysis of those factors against GDP over a reasonable time period.
Dear Matthias,
The topic is of course inspirational. But could you share a bit more about your regression model you are thinking about? What data do you use, which time periods/lags in the model?
Not sure but it seems that with relying on consumption (growth), government expenditure (growth) and such we are dangerously moving towards an accounting identity rather than an explanatory model.
@Brian: Well, the ideal model would be close to 1 for the R-squared. Or it would be theoretically, but normally regressions have a R-Squared closer to 0 than to 1. My model luckily has a higher R-Squared than the model I refer to in the beginning.
@Mate: I will give an update to my project very soon, where I will further explain the aim of the project and its regression model. But the data is mainly from the World Bank because it's preferable to regress data from the World Bank on GDP data from the World Bank. Something that the model I was inspired did not do. And yes, the accounting identity is similar, but only in a way to be accountable for the fundamental drivers of growth.
Again, thank you for all of your comments and I will post an update on the project very soon.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude