Disclaimer: Despite the provocative thread title, I’m not really interested in politics. I’m asking this more as a general question, but had to place it here where all the political stuff goes.
In a forthcoming paper to be published in The Journal of Finance, Pedro Santa-Clara and Rossen Valkanov found the following:
A nine percent difference in returns is substantial, and over such a long duration of time would be statistically significant. How do you explain it? Do investors irrationally place an enormous risk premium on stocks whenever a Democratic is elected president?
The authors claim that the difference is statistically significant. I’m not convinced that it is, although it surely is substantial. The trouble is, the movement of the stock market and the Presidency aren’t statistical processes. So, “statistical significance” really means the results *would be * statistically significant if politics and economics were statistical parameters.
I spent a career as a actuary applying statistics to economic and social variables related to the field of property/casualty insurance. I am quite dubious of the authors’ approach of formally applying various statistical formulas. They’d gain a lot more insight by looking at charts of actual results year by year and administration be administration, and comparing the major movements to real-world events that caused them, in my not-so-humble opinion.
Statistical significance can be determined by formulae which are independent of political variables. Basically it just means that the difference is greater than that which can be attributed to chance. It does not make a judgement as to why.
Well, there are several reasons. However, since some of them will start a huge debate as to which economic policy is better for the economy (“Tax & spend” vs “borrow & spend”) I’ll just give my usual one- the stock market symbol of a “bull” isn’t too far off, as in some ways the stock market is like a cattle herd- given to being “spooked” & stampeding witw little or no apparent reason. In the present case, one of the reasons seems to be that the investors EXPECT better performance from a Dem, and invest accordingly.
This last stock market fall for instance. Now- we’ll never know for sure how much of the huge increase was due to Clinton’s policies. But, everyone knew there had to come a “correction”. Such a correction had been due for the last year of the Clinton admin- but dudes held on, hoping it wouldn’t happen. Well, for the NASDAQ it started to in the spring of Clintons last year, but the Dow held on. So then in the fall, it is clear to all that the Dems are out, and GWB is in. So dudes bail. I bailed- I wish I had done so more (I had my 401k invested mostly in stocks- I changed it to mostly govt bonds, and I wish I had changed it 100% to bonds! I lost $10K)
No matter what the argument is over which party’s economic policy seems to benefit the Dow & economy more- there simply seems to be no doubt as to which party holds the higher “investor confidence”- and that is the Dems. I predict that the Dow will beging to rise back if a Dem wins next presidential election, and continue to wobble as long as GWB is in power.
More arguably, it could be that investors- although they personally LIKE the “cut taxes & increase the deficit” GOP policy- they realize that is actually is “voodoo economics” and won’t help, and will perhaps harm the economy. I beleive so- I think that the tax cut plan is economic foolishness.
I don’t really know, and if I did know, I would be in scandinavia next year, accepting the Noble prize for economics. But whatever the reason- investor confidence is much higher during a Dem admin.
I’ve always thought of it as when the POTUS is a Democrat there seems to be more of a happy-go-lucky, party type atmosphere. Anything goes, devil-may-care, etc. which tends to give investors a warm and fuzzy feeling.
Conversely, when a Republican it’s more like the cat is no longer away and the mice don’t play, watch your P’s and Q’s. Just a more stifling atmosphere for free wheeling fun.
This is not based on anything other a gut feeling, much like what I think the OP was looking for.
Of course- and this is just a wild & crazy idea- perhaps the Democratic policies are simply better for the economy than the GOP policies. This explanations fits all the known data- like it or not.
Except it doesn’t, DrDeth. It would certainly be a good point if you could show the stock market always dumps out under Republican Presidents and always does well under Democrats. The problem with this is:
The different economic policies of the two parties have NOT been consistent over time, and in any case are just as affected by who makes up Congress as who’s sitting in the White House. So you can’t even clearly establish what it is Presidents do to affect the stock market in a consistent manner.
The data cited in the report is from 1927 on. It just so happens that the 1929 stock market crash happened on a Republican watch, which is interesting but can’t really be blamed on Herbert Hoover. The dividing line at 1927 also cuts out much of Calvin Coolidge’s administration, when the stock market did extremely well under a Republican, so the arbitrary 1927 divider effectively means that you’re blaming the Republicans for the losses of 1929 but not giving them credit for all those gains that led up to the losses. Why? It’s never really explained why 1927 is the cutoff point, but it’s the perfect year to start counting if you want to make Republican presidents look bad.
Then you have a Democrat President sitting while the rest of the world destroys itself in 1939-1945 and the U.S. gobbles up all the business of blowing stuff up and rebuilding - none of which can be reasonably credited to Roosevelt or Truman. The huge cycle from 1929 to circa 1950 - well, that’s an enormous chunk of the difference.
Of course, it may also be that the supply of Presidents is just too limited a sample to say anything about their parties. since 1927:
Coolidge
Hoover
FD Roosevelt
Truman
Eisenhower
Kennedy
Johnson
Nixon
Ford
Carter
Reagan
Bush
Clinton
Bush 2.0
You don’t have that many Presidents, relatively speaking; that’s eight Republican Presidents, one of whom was just a placeholder for a couple of years, and six Democrats, which includes one extended shot of 20 years under the Depression-To-Half-The-World’s-Economy. Maybe the Republicans just got unlucky this past 80 years and happened to pump out a few turkeys; Bush 2.0 is certainly not doing a very good job with the economy, but that appears (at least to me) to be his own incompetence and greed rather than any ideological issue.
I haven’t waded through that 26-page paper. But the timeframe covered is 1927-1998, and there were exactly 13 Presidents during that time: 7 GOP (Coolidge, Hoover, Ike, Nixon, Ford, Reagan, Bush I) and 6 Dem (FDR, Truman, JFK, LBJ, Carter, Clinton). My first instinct is to say, “Small sample.” And by the time you have a statistically robust sample of Presidencies, the party definitions change.
The paper’s authors deal with this by using monthly observations as their sample, giving them a sample size of 864 observations, which is statistically quite robust. The problem is, is it intellectually valid to treat the movement of the economy during Reagan’s presidency as 96 data points rather than one? I haven’t dug deep enough into the paper to see if/how the authors deal with this question.
Even though I’m pretty partisan on economic issues, I hardly think this sort of thing is evidence that the Dems have the answers, economically speaking, or that the GOP doesn’t.
If you ask the Dems, then it’s because Democratic Party policies are better for the economy, GOP policies are worse.
Of course, if you ask the GOP, then it’s because the Dems screwed things up, it just took until the time they were out of office for the fallout to catch up, and by the time the GOP president had turned things around, he was voted out of office and the Democrats took the credit once again…
The Constitution gives most of the power to Congress. Reality gives it back to the President.
A 535-headed monster can’t decide where to have lunch, let alone make policy. As a result, the President almost always sets the agenda, and Congress reacts to whatever the President sends over, by passing all/some/none of it.
On the flip side, the President has the veto. This can be overriden by a 2/3 majority of both houses of Congress, but Presidential veto overrides are pretty rare. So the President can block almost anything Congress sends him.
In all my decades of watching politics, I can remember only a few brief instances when Congress wrested the legislative initiative away from the White House. The last time was in early 1995, and once Clinton figured out that vetoing what Gingrich & Co. sent him wouldn’t have any repurcussions, the inter-branch balance returned to normal.
Boy, are we paranoid today! I assume by “success” they mean getting a statistically significant correlation.
They do say that it is robust in subsamples too though and is not explained by business-cycle variations (whatever they mean by that), so it seems that it can’t all be due to this.
I will go on record like RTFirefly though as another person whose political convictions would want him believe it could be true but will reserve judgement without further study. After all, I have seen too many people on the other side abuse data to attempt to show that Reaganomics was successful and other such silliness. I don’t want to repeat their mistakes!
Still, it ought to at least give pause to all those “true believers” who think Republican economic policies are so pro-growth and all that BS!
Yeah, that kinda sounds like “getting the results they wanted” to me. Sounds like they formed a hypothesis that the stock market is directly tied to Republican vs Democratic prevalence in politics, and tried to keep looking at data from new angles until they were proved correct. Hmm.
Jeff
One can demonstrate that the primary driver of the market is corporate earnings. The primary driver of corporate earnings is the economy.
Here it gets tough. Policy can have all kinds of effects. For example, the crash of '87 is blamed by some (I make no warranty,) on the closing of the passive loss law that benefitted limited partnerships and subjected their tax benefits to recapture. This rather precipitous change which effected both investment markets and real estate took several years before it’s effects were felt.
Similarly the Fed’s poilcy take 12-18 months before they are felt in the economy. These are the things that move relatively fast. The business cycle on the other hand moves pretty much in a boom to bust cycle of it’s own accord. While it’s severity may be ameliorated (an arguable proposition,) the cycle itself seems immune to policy. Then too, the hypotheses which best explain market action are “Random walk” which suggests that market cycles are to a large degree random, and the weak version of the efficient market hypothesis. This latter implies that the market is something of a good chess player anticipating action several moves ahead, and discounting policy and macro policy effects for the long-term. What this means is that basically not only is the news built into the market, but the consequences and actions several orders down the road also tend to be built it in.
With few exceptions there is not an awful lot that a sitting President can do to make the market go up over the course of a four year term. Nor can he do a lot to make it go down. The changes that he institutes will most likely be felt by his successor, or even more likely his successor’s succesor.
And, it sounds to me that you are just digging for any reason to dismiss these results. (Not that I am ready, as I noted, to go to the mat to defend them. But it is sort of amusing to see the usual suspects who are quite willing, for example, to give Reagan credit for things that happened more than 10 years after he was out of office react to a piece of research that doesn’t fit in to their preconceived notions!)
I actually read through the introduction to this paper and it seems to be a quite careful research effort. I am still skeptical of a result based on correlations mined from necessarily-limited data though.
Doesn’t look like that to me at all, at least from the first few pages. It looked more like an honest attempt to test a piece of “conventional wisdom.” They even admit the possibility of “data mining” as suggested by John Mace.
I’ll take a longer look at the study when I get a chance (after all, someone actually pays me to care about stuff like this), but here are a couple of possibilities come to mind that I’ll be examining as I read it.
One, markets are anticipatory. So if Republican policies are in fact better for big business profits, one would expect the markets to start moving upward at the (believed) end of Democratic administrations. For example, the Dow made a 25% move in the last year of the Carter Administration, perhaps as market participants saw the possibility of his not being reelected (as it happens, the market gave that back in the early years of the Reagan Administration before beginning the great bull market of the '80s).
Another is that war is considered by many to be good for big business, and Democrats have been president during more war years.
Another may be measuring the “wrong thing.” I haven’t got to the methodology section yet, but it looks like they may be measuring “real returns” against short-term interest rates – a common and legitimate measurement tool which nonetheless tell you little about how much money you’re actually making, particularly if there is a large difference in some periods between short rates and the subsequent inflation that the rates are supposed to be anticipating.
Related somewhat to that last and the first is that the authors talk a lot in the intro about the difference between “expected” and “unexpected” returns – “expected” being those “returns using macro variables known to forecast the stock market as controls for business cycle fluctuations.” As a guy who got the economy balls-on correct last year but nonetheless generated only moderate excess return over my asset-class peer group, I’ll be quite interested to see what precisely that means!
Also, the authors say that the data holds (though to a lesser confidence, as expected given the small number of elections) for sub-periods. It’ll be interesting to see how much of the variation is explained by a relatively small number of identifiable events, such as the 60’s go-go market which died with Bretton-Woods, the speculation bubble which occurred during the Clinton administration, etc.