Or did it all regress to the mean over time?

The research looks at stock prices in the short run. This is very different than long-run business results.

You can't from this research assert that a republican administration is good for "republican" companies. All you can say is that, in the short run, the institutional investors and money managers who do most of the volume stock trades BELIEVED that it would and were willing to pay marginally more for the stock in the timeframe of a few days.

This leaves unanswered two more important questions:

1) Did the profit gains which would justify a higher stock price actually materialize?

2) Did the stock changes for "republican" companies persist, or did they

Until you look at that, you can't honestly say that partisan companies reap real benefit. All you can say is that partisan companies reap share price gains in the short run, which says less about the company and more about investor perception.

Agreed, wholeheartedly. Stock price is not the same as "profit". Stock price is the expectation of profit. In the short run, where speculators and momentum traders have an inordinate impact on price, people believe certain connected companies will profit. Fine, but that is definitely not the same as it actually happened.

We simply cannot use short term stock price movements as a proxy for profit. I know there is wisdom in markets, but there is also stupidity, otherwise there wouldn't be any day traders. I don't doubt that political connections do have real value, but looking at short term price movement doesn't really prove that.

That was what first popped into my head upon reading the article, too. Of course, to study what you two are really interested in would probably prove tricky... presumably, we'd need to find similar companies in the same industries, with some headed by dems and some headed by republicans for a real apples to apples comparison. And of course you couldn't just compare two companies on one side to two on another; you'd need a large sample size to discern anything from the data. Not saying it couldn't be done, but grad students could certainly pick easier theses!

For sure, actually testing the hypothesis would be extremely time consuming. I wouldn't want to do it either.