Friday, September 12, 2008

Bad economics makes good editorials

Like in today's WSJ where we are told "Divided Government is Best for the Market" and treated to a lovely exercise in torturing the data. It turns out that, since 1948, the stock market has done better under Democratic presidents (15.6%) than Republican ones (11.1%). Now this number really means nothing. First, we'd like to know real, inflation adjusted returns, not nominal returns. Second, and most important, we need to control for other relevant factors. Despite the best efforts of the MSM, we do not live in a unicausal world; politics is not the unique cause of all outcomes.

However, the editorial writer has a unique approach to attacking the problem:

"But it's not so simple... First, not all Democrats act like Democrats, and not all Republicans act like Republicans. John F. Kennedy, for example, was an enthusiastic supply-side tax cutter, and George H.W. Bush raised taxes. Bill Clinton promoted free trade, and Richard Nixon imposed wage and price controls.

If you assign those four presidents to the opposite party based on that -- make the two Democrats into Republicans and the two Republicans into Democrats -- the numbers completely reverse. Now stocks average 14.7% under Republicans and only 10.5% under Democrats.

In fact, it turns out that if you do just one single switch -- if you make Richard Nixon into a Democrat -- it's enough to reverse the numbers."

Awesome!! When all else fails, massage the data!

People, no matter how you slice them, these numbers can't at all be taken as evidence about what market performance would be in the future under different Presidents or about how the market would have faired if George McGovern would have beaten Nixon because they don't take into account any of the other massively important factors that affect market performance. Give a couple of examples, would there have been no oil shocks in the 70s if Nixon were not President? Would the tech boom not have happened if Clinton were not president?

Of course our beloved editorial writer is a sophisticated guy and does bring in another factor later in the article: Congress!!!

"If the electorate were really smart, it would elect a Democratic president and a Republican Congress. Under that deal, stocks have averaged a 20.2% total return"

Again, awesome. That combo in this sample is the last 6 years of Clinton. So the tech boom happened because of a Democratic prez and a Republican congress. Elect that combo again and presto, another boom. It's just that simple eh?

Why oh why oh why oh why does the WSJ publish trash like this? And why oh why oh why would anyone put a single penny into the hands of Trend Macrolytics LLC?


12 comments:

Anonymous said...

Massaging the data is one thing, but this guy's pretty much just flipping the labels. Are you sure it's not a spoof?

Anonymous said...

From Luskin's CV:

Education
1973 to 1974
Yale University
New Haven CT
Completed freshman year; dropped out to pursue career


http://www.trendmacro.com/a/about/luskincv.asp

Anonymous said...

To be fair, he has done very well for himself. I just thought that line was funny.

Norman said...

Why should the electorate want stock prices to rise so much? Since when does increasing stock prices mean increased economic well-being (unless of course your income comes from day trading stocks)? Stock prices? Really?

Anonymous said...

Well think of how Krugman talks about all those jobs Clinton "created".

Anonymous said...

Angus, I think you must drive poor graduate students nuts. How else can someone get their PhD in poli sci without a few "innocent" massages (of data).

Michael Stack said...

Oh Lord, was that written by Donald Luskin? If you want a good belly laugh or two, search Delong's site for Donald Luskin. Long story short: Delong doesn't have much use for Donald Luskin, and IIRC, Delong's criticisms were right on point.

Michael Stack said...

RE: Delong & Luskin, here is one link to get you started:
Delong on Luskin

Anonymous said...

FYI, I just had sex with your cat.

-Ibod Catooga

datacharmer said...

Looks like the WSJ guys are reading my blog!

http://bluematter.blogspot.com/2007/06/what-women-do-to-president.html

Here's what I wrote last June, and it's very interesting to see that the WSJ adopts these high standards of analysis.

And of course note the date: it's June 2007, when I hadn't even heard of Barack Obama and didn't know that Hillary would be the runner up. My theory, in common with the WSJ's, is really solid.

Anonymous said...

Flipping the "Democrat" and "Republican labels, but not entirely without reason. Democrats are associated with higher taxes, and more economic regulation. Republicans with lower taxes and less attempts at economic regulation.

The problem is that he shouldn't just flip the labels, he should put in place a different label that describes the breakdown of the different presidencies. Something like "Free trade supporters vs opponents", or "Tax cutters vs. those that increase taxes", or "heavy regulators, vs, moderate regulators". Maybe none of my suggestions get to his point exactly, he's trying to combine different things, and several of these presidents are on the more "Republican" side on one thing and the more "Democratic" side on the other (for example Clinton who raised taxes, but also was a strong supporter of free trade, and signed entitlement reform when a Republican congress voted for it)

- Tim

Anonymous said...

Great article you got here. It would be great to read more about that theme.
By the way look at the design I've made myself Companionship in London