I’m reading political posts instead of working, and I found something linked somewhere that I found interesting:
Capital gains top tax rates plotted against contemporary economic growth in the postwar period.
I apologize to right-wingers who were sure there was a negative correlation between growth and capital gains taxes. I can see now where people got that idea. It’s a plausible if weak interpretation of the data up to the mid-1990’s. I guess Steve Forbes didn’t pull it out of thin air. But data from the last fifteen years shows there’s not really a correlation after all. We dropped capital gains taxes and the economy failed to grow. That’s long enough a period of time to say that the (weak) correlation was a statistical illusion.
I’m willing to admit that the argument had data apparently on its side for a while. But it’s fallen apart now.
I’d like to see this chart with a moving-average trendline plotted (say, the average of a three- to four-year lookback period for each year). Eyeballing it, it looks like a sinusoidal business cycle fluctuating about a baseline that does correlate with the cap gains rate, with the obvious exception of c. 2009.
I’m not convinced you couldn’t get that from pure randomness. Unless someone can do a statistical analysis of significance, it doesn’t mean anything. Even then, it’s the old saw of causation and correlation.
THe argument about taxes reducing growth has always been sound theory, but in a complex economy it can be hard to get that from the actual data. Regulations also matter, as does human capital, the efficiency of the tax system as opposed to just its rates, federal spending, state and local taxes and spending.
I definitely agree with you on general principle, but simply stating what the top rate is tells you nothing about the quantitative impact. How many people were subject to it in any given year. What was the dollar amount as a percentage of GDP or in some other relevant context.
I mean, maybe the guys who actually study this stuff can say that the other factors in general are a wash so it is in fact meaningful. IDK. I’m just skeptical.
You can’t really look at it like that. For one thing there are a host of other factors that could obscure this one. Also, we need an actual test of significance to test if there is a statistically significant relationship.
But ignoring all that, the figure itself actually does say there is a correlation of 0.12, so there is a correlation according to it.
Not that it really matter, people always have their beliefs and interpret data according to it.
I don’t remember any of this stuff so I had to look it up. If I’m reading it right, a correlation of -1 means 2 variables are perfectly uncorrelated, 0 means they are independent and +1 means perfectly correlated.
From that, 0.12 wouldn’t seem to inspire a lot of confidence - yes, no?
-1 is perfectly negatively correlated (as the independent variable increases/decreases, the dependent variable commensurately decreases/increases)
Between -1 and 0, changes in the independent variable occasion a generally contrary movement in the dependent variable, but there is some variation not accounted for by the independent variable (the size of this unaccounted for variation decreases as one gets further away from 0 and closer to +/- 1).
Zero means that the independent variable explains none of the variation in the dependent variable.
Between 0 and 1, Between -1 and 0, changes in the independent variable occasion a generally same-direction movement in the dependent variable, but there is some variation not accounted for by the independent variable (the size of this unaccounted for variation decreases as one gets further away from 0 and closer to +/- 1).
+1 is perfectly negatively correlated (as the independent variable increases/decreases, the dependent variable commensurately increases/decreases)
It depends on the sample size and what you expect. With enough independent data it could mean that about 12% of the variability of the GDP is due to the changes in the tax rate.
At the risk of seeming pedantic, I think a black swan is a low probability, high impact event. Something that would be lumped in under the rubric of ‘tail risk’ but aside from being even further out on the tail of the bell curve, has the potential to be catastrophically disruptive.
