pageok
pageok
pageok
The Economy under the GOP vs Dems:

On today's NY Times op-ed page, Tommy McCall has a very interesting graphic on the performance of the stock market under Democratic and Republican presidents. Each of the two parties has been in control of the White House for 40 of the last 80 years; on average, during Democratic administrations, the S&P index has increased 8.9% annually. During Republican administrations, the corresponding figure is 4.7% — and that's if you exclude the Hoover years (during which time the market declined an astonishing 77%). If you include Hoover, the Republican "annualized return on investment" is an even more dismal 0.4%. Put another way, if you put $10,000 in the market and left it in only during Democratic administrations, you'd have $300,000, compared to around $12,000 (or $51,000, again excluding Hoover from the calculations) if you adopted the same strategy for Republican administrations.

This follows a number of articles this summer addressing the similarly-pointing argument advanced by Larry Bartels' in his book "Unequal Democracy," in which Bartels' data show average economic growth of 2.78% under Democrats, 1.64% under Republicans (roughly a 67% difference). (I don't recall discussion of Bartels' book here on the VC, but I might have just missed it)

Most people, I think it fair to say, would have predicted just the opposite. Republicans would then say "You see?? Our policies really are better for the economy," and the Democrats would say "Well, that's just Wall Street — what about the poor and the marginalized, and the social safety net . . ." etc. etc.

But with the actual results, it's not so clear what to say — in particular, what Republicans say. On its face, it's a very damning indictment of Republican economic policies — not an indictment in theory, but an indictment in fact. The economy grows more slowly, and the stock market increases in value more slowly, when a Republican is in the White House than when a Democrat is in the White House.

Has there been, or is there, a consensus response on the Republican side? I'm curious to know how those of you who are staunch Republicans respond. I can think of some responses: one can concede the point, but argue that Republicans are preferable on other grounds; one can point to flaws in the data (the sample is small (only 80 years); the effects from one administration "spillover" into later administrations, blurring the significance of any one year's data; the comparison ignores differences that may be due to control of Congress by the other party, etc.). But it does seem worthy of thought and discussion, even if we weren't simultaneously in the midst of an economic crisis and a presidential election.

UPDATE: A number of readers pointed to responses published here, and here, both of which give some good arguments for looking at the results of these studies at least with some skepticism. I don't find it too persuasive to argue, as Luskin does, that if you treat Nixon and GHW Bush as 'Democrats,' and JFK and Clinton as 'Republicans,' the results turn around -- if you call elephants 'mice,' and mice 'elephants,' then 'elephants' weigh less than 'mice.' But the more cogent criticisms seem to fall into two categories: first, that the Nixon presidency, like the Hoover years, had a disproportionate effect on the Republicans' poor performance and (b) that at least as far as the GDP numbers are concerned, one needs to take into account a President's actions and the effects of his policies, and that with a 2-year lag the data show almost equal GDP growth under Republicans and Democrats.

FantasiaWHT:
Oh I'll take a couple pretty pathetic stabs

Wartime spending?

The market actually reacts to impending change to democrats by failing, and then to the impending change to republican by surging?

Or the reverse, a cross-presidency delay in stock market, such that the raises during democrat years are the result of earlier republican policies?

Have democrats had higher rates of inflation such that the actual gains are less?

I dunno.
10.14.2008 3:51pm
Ilya Somin:
I'm no staunch Republican. But there are several obvious responses:

1. We only have a handful of Republican Administrations in the data set and some of them (e.g. - Nixon) pursued very liberal economic policies. Some of the Democrats (e.g. - Clinton and JFK) pursued relatively more market-oriented policies.

2. The effects of policies enacted under one president tend to have spillover effects. The study naively asssumes that each president is "responsible" for anything that happens during his time in office, and for nothing that happens afterwards. Most economists believe that presidents have little or no influence over short term economic trends, but that's precisely what the study attributes to presidents (mostly wrongly in my view).
10.14.2008 3:52pm
A Law Dawg:
The better numbers to compare are GDP, inflation, and unemployment.

They may be quite favorable to Dems, I dunno. But I think the Dow is not the most important indicator.
10.14.2008 3:54pm
CJS (mail):
Aside from the silliness of trying to assign causation for the stock market to the president, who has to coordinate with Congress on any budgets and spending and who is subject to Fed policy, based on the correlation between the dates: Where is the lag for the effect of implementation of any particular economic policies? I.e., the President comes in to office in January of a given year, and he automatically gets credited with the stock market performance from day 1?
10.14.2008 3:54pm
commontheme (mail):
No, no, no - democrats are big spending socialists who destroy the economy and republicans are miserly guardians of public funds who grow the pie for everyone.

And I am just going to continue repeating this again and again with my fingers in my ears so I can't hear any of the nonsense that other people are saying...

Oh, and Obama is a Muslim, too!
10.14.2008 3:54pm
RebelRenegade:
What about Repub vs Dem House and/or Senate? I'd figure they would have more of an effect than the POTUS. But I could be wrong.

What if you excluded wartime?

FWIW I'm not a "staunch Republican" but this seems interesting to discuss.
10.14.2008 3:55pm
Joe1616 (mail):
The most obvious flaw in this analysis is that Democratic and Republican policies from 80 years ago have little or no resemblance to their policies today. The Dems of the past were the party of segregation. Are you saying that going back to segregation would be beneficial for the economy?

Also, try reading "Fooled by Randomness." You have no logical explanation for these results. It is probably random fluctuations. I doubt that the results are even statistically significant.

My third response, which you touched on, would be to examine control of Congress. I seem to remember the stock market doing pretty well between 1994 and 2000, just to take one example. Specifically, it would be interesting to see a breakdown of all combinations of executive and legislative branch control. Dem-Dem, Dem-Rep, Rep-Dem, and Rep-Rep. Also, what happens when one party controls 2 out of three or all three branches.
10.14.2008 3:56pm
Dan Schmutter:
I'm not a Republican (you requested Republican responses), but my immediate thought is that the "spillover," as you call it, has an extremely important effect on these data.

For example, a good case can be made that the "prosperity" of the Clinton years, during which the stock market rose greatly, represented little more than the creation of the very housing bubble that has resulted in the mess we now face.

This probably happens quite a bit, so measuring and comparing the performance of the stock market during Democratic or Republican administrations probably tells us very little about the true impact of their policies. Rather, it tells us literally only what happend to the stock market during their adminstrations which may not correspond to the real (especially long term) impacts of their policies.

Dan Schmutter
10.14.2008 3:56pm
finman:
This WSJ column addresses that very issue.
10.14.2008 3:58pm
wooga:
In addition to the points raised by others, there is also a positive link between wartime spending and stock market performance. Historically, Democrats tend to wage more wars than Republicans. That skews things significantly.
10.14.2008 4:00pm
wooga:
finman's link FTW.
10.14.2008 4:01pm
Randomness:
My first thought is that people seem to like the Dems in one of two situations: economy is doing great or the economy is in a total rut. When its doing great, any nominal threat of tax increase is offset by the benefit of general "do gooding." When its in a total rut, tax increases are less meaningful than social help.

In both situations, it's likely that Dems will reap the benefit of any "increases." In the latter case, they'll see the most increases since they started from something lower.
10.14.2008 4:03pm
Joe345:
I look at those Republicans presidents that has been positive, and say, when Republican leaders behave like conservatives they have better results, over 10%. When they behave like Democrats the results are much worse.

However the fact that these graphs ignore the make up of the Congress, really tells you all that you need to know. It's not an honest comparison. It would be much more interesting to see how all the combination of president, house and senate interact. I suspect that the real answer is gridlock is good.
10.14.2008 4:04pm
JB:
I agree with Ilya. In addition, under FDR we climbed out of the Great Depression (whatever his policies may have done to prolong it). What would the data look like starting in 1944 or 1960?
10.14.2008 4:05pm
Assistant Village Idiot (mail) (www):
Rubbish. No president influences the economy much on the day he is inaugurated, so any measure "how he is doing" must necessarily be delayed until his policies begin to take effect. The economy still "belongs" to President A more than to President B for months afterward. There is no sharp break point, and there are of course some actions that have their effect sooner than others. But I mentally insert a 12-18 month delay before the new guy gets full credit or blame. The stock market is a somewhat fairer measure, as it is an earlier indicator than unemployment or some other lagging indicator. But it's still not instantaneous.

You will notice that this puts us up to the next congressional election, which is why we often punish or reward the wrong people in elections.

Here in NH we regularly have primary candidates pointing out what was happening in the economy when the incumbent was last running in NH - as if what happened in the year before they even took office should be held to their account.

Insert that 12-18 months, BTW, and the country's economic history since Gerald Ford has a whole different spin.
10.14.2008 4:05pm
MarkField (mail):

1. We only have a handful of Republican Administrations in the data set


Since 1928 (the last 80 years) Republicans have controlled the Presidency for 40 years. It's 50-50.


The better numbers to compare are GDP, inflation, and unemployment.

They may be quite favorable to Dems, I dunno. But I think the Dow is not the most important indicator.


I believe Dems win all those too. IIRC, Republicans only won the bond market.
10.14.2008 4:07pm
M. Gross (mail):
I'd merely argue spurious correlation.
10.14.2008 4:08pm
cirby (mail):
It's funny how they kept it to "Presidency only" when making their chart.

On the other hand, the current economic troubles started really taking shape early last year - right after the Democrats took over Congress.

Nixon (the other recent President with a negative index) had a Democratic Congress all through his terms.

The wonderful Clinton years? Most of that time, Republicans were in control of Congress.

I also wonder about some other things, like the positive economic results in the Carter years. I was in high school and college back then, and "good" was not one of the words anyone could use for the Carter years, economically (inflation was certainly not part of the charts in the article, and neither was unemployment). And, once again, Democratic Congress...
10.14.2008 4:10pm
Houston Lawyer:
You could easily assume that Republican presidents have spend the majority of their terms cleaning up the messes left by their predecessors. Bill Clinton gets all of the credit for the good side of the tech bubble and Bush gets the bad side.

Also, the whole ad hoc proctor hoc nature of the argument makes you wonder whether the writer was required to take logic classes in college.

The last president to implement a radically different economic agenda from his predecessors was Reagan. Bush I, Clinton and Bush II just tinkered around the edges. Reagan's agenda of lower taxation was so successful that it has been copied by almost every industrialized nation in the world. Everywhere I go I hear people pining for the good old days of Jimmy Carter :)
10.14.2008 4:10pm
MarkField (mail):

What would the data look like starting in 1944 or 1960?


The linked article provides data from 1948-2005. The GDP performance under Dems was much better.


No president influences the economy much on the day he is inaugurated, so any measure "how he is doing" must necessarily be delayed until his policies begin to take effect.


IIRC, Bartels accounted for this.
10.14.2008 4:10pm
Calderon:
Not a staunch Republican, but your last paragraph ticks off most of the responses I'm aware of. For example, this link responds to the arguably related question of the effects of the president on economic growth. Most of them apply here as well:

Likewise, the President isn't a dictator, and the Congress plays a huge role in what economic policies actually are passed. For example, from 1994 to 2000, I think it's fairer to say congressional Republicans were setting the economic agenda more than Clinton. And if you take out Clinton from the Democrat column, the increase for Democrats drops sharply.

On the Republican side, Nixon is about as far away from free market economics as you're going to get from any president on that list besides FDR. The New York Times had another chart today showing Nixon increased regulatory spending more per term than any other president, and of course he implemented wage and price controls. If you take Nixon out, then the Republicans go up sharply. As for Bush Jr., there's two months to go and with volatile stock swings there's no telling where they'll end up.

So in short, because of the lag in effects for most economic policies, that the President and both houses of Congress are involved in settling economic policy, and small sample sizes allowing a single president to cause large swings in the numbers, I don't think it means much.
10.14.2008 4:14pm
big bob (mail):
As usual the NYT is a day late and a dollar short. This topic was hashed to death in the blogs over two weeks ago. I'm not going to re-hash all the stats here, as it shouldn't take long to dig up last week's fashionable cocktail talk. Obviously there is no reason why any letter after someone's name, even when that someone is the President, should have any real economic effect on the stock market. This theory is just a hair's breadth away from sunspots, hem lines, and Super Bowl winners. Ilya is on the right track. Switch JFK, Clinton, GHW Bush, and Nixon and you get exceedingly strong returns for tax cutters versus tax raisers. Having a divided government incapable of meddling is another way to see large gains. The best combo is a tax cutting President with an opposition congress.
10.14.2008 4:18pm
Elliot123 (mail):
I think general ignorance of the economy is why we see these people who both credit and blame a president for economic performance during his administration. If one wants to say a president caused something to happen in the economy, then tell us what he did, how it related to the economy, how and why the economy reacted, and what the resolution was. Also tell us the situation when he came into office, and the lead time necessary for any action by him or his predecessor to take effect.

I realize there is a human tendency to loook to a leader for guidance, and it's far easier to attribute events to a single person rather than acknowledge a complex web of interrelationships which we might not understand. I suggest we have lots of intelligent people who really don't understand, but pretend they do by focusing on the president.

But, maybe we can extend the game. Who wants to attribute to the current US Supreme Court the results of decisions made by prior courts? Wouldn't that make life easier for people? It makes just as much sense as attributing good or bad economic conditions to the sitting president without spcific reason.
10.14.2008 4:19pm
The Unbeliever:
The linked article provides data from 1948-2005. The GDP performance under Dems was much better.
In case people didn't bother going to finman's link:
How many times have you heard some Democrat pull out some "study" (they always call it a study, it sounds so scientific) by some professor or some "nonpartisan" think tank that purports to show that since 1948 (it's always 1948 for some reason) stock performance or economic growth has been better under Democratic presidents than Republican ones?

...But it's not so simple when you study that "study." 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. Then stocks average 14% under Republicans and only 12.1% under Democrats. This fact discredits this whole study more than it does Republicans, or even Richard Nixon himself. Any analysis that can be undone by omitting or changing a single data point isn't very robust.
10.14.2008 4:21pm
Sebastian H (mail):
I'm on my way to not being a Republican, but there are some serious flaws.

First, it equates Democrats of the past with Democrats of the present and the same with Republicans. Equating Clinton with Carter and Truman for example doesn't make any sense if you start actually looking at the underlying economic policies. The economic things Clinton was most famous for on a policy level are almost precisely the opposite of what you normally think of as "things inidicative of being a Democrat" over that time period: NAFTA and welfare reform being the biggest.

Second, the data set is much smaller than you suggest. It isn't 80 years, it is 12 Presidents (the limiting factor).

Third, the findings rely almost entirely on an outlier data point (Clinton). Statistically speaking if your whole finding relies on a single outlier data point you have to be suspect. And we know empirically that in some respects he got lucky he got the upswing from Bush's austerity, he got the downtic in military spending, he and Greenspan pushed the aftermath of the tech bubble into Bush II. Which isn't to say that Clinton did a horrible job economically. He didn't make lots of stupid mistakes. But he also benefitted from the times. And without him the data set falls into statistical insignificance.

Fourth, there is a causation ques
10.14.2008 4:32pm
MarkField (mail):
Here's a link which collects the data in charts.

Some of the responses here are delusional. First, anyone who relies on Donald Luskin is, to use the word of the week, deranged. Second, Luskin's argument is itself absurd. Think about it: if only you convert Ds into Rs, and Rs into Ds, the result changes. Duh.
10.14.2008 4:34pm
SP:
Professor Post, couldn't you have found a more credible study, like the correlation between the Dow and whether the team that wins the World Series is from the AL or the NL?
10.14.2008 4:35pm
Greg Q (mail) (www):
Let's highlight that list bit from Luskin

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. Then stocks average 14% under Republicans and only 12.1% under Democrats. This fact discredits this whole study more than it does Republicans, or even Richard Nixon himself. Any analysis that can be undone by omitting or changing a single data point isn't very robust.


Given that Nixon (wage and price controls) government much more like a Democrat than a Republican, it's even a fair switch.

IOW, like so much else posted by Prof Post, this "study" is crap.
10.14.2008 4:38pm
roy (mail):
I'll confess to being surprised by these numbers, but before taking them very seriously, I'd want to see the relationship to the makeup of Congress. If I still don't like the results, I'll demand statistics on governors, mayors, and so on down to notary publics until I see a correlation that fits my previously held beliefs.

If I can take a cheap shot at liberals, these numbers may just mean that Democrats tend to create bubbles. Or should we count the Dot-Com bubble under Clinton as a sign of Democratic economic prowess?

And I second the recommendation to read "Fooled by Randomness".

(My bias: I'm a free marketeer, but not a Republican)
10.14.2008 4:43pm
Ben P:

Given that Nixon (wage and price controls) government much more like a Democrat than a Republican, it's even a fair switch.


*eye roll*

so if we include democrats that were good for the economy and exclude republicans that were bad for the economy, and find that because of the exclusions Republicans were actually better for the economy?

and the argument is still "you can't vote for a democrat because it's bad for the economy?"
10.14.2008 4:43pm
Jamey:
One major reason for the data looking bad is that the end point for Bush is not his last day in office, but a day chosen when the stock market is very low. Roll back one month and Bush is basically 0% growth and the Republicans clock in at just under 8% average growth (leaving aside Hoover), or not significantly different from the Democrats.
10.14.2008 4:47pm
JRL:
There's no doubt it's the spillover. Our economy is not a little scooter that can turn on a dime. It's a huge ocean liner that takes time to move and for its effects to be felt.
10.14.2008 4:47pm
A.S.:
To the extent there is a causal relationship between the party of the President and differences in growth, I think the difference results from a greater effort by Republican presidents (as compared to Democrats) in trying to constrain inflation.

Tyler Cowen makes the same observation here, in response to Bartels.
10.14.2008 4:48pm
PubliusFL:
Other posters have addressed some of these, but here are some of the key failings of McCall's graphics:

1. Failure to account for inflation. Truman and Carter had the worst inflation records of all of the presidents during the covered time period.

2. Failure to account for control of Congress.

3. Failure to account for lags.

4. Change in the policies of the parties. Hoover was a protectionist, which was the root of much of his economic problems. The Republicans since then have been free-traders.

5. Statistical artifacts of date selection. The market is so volatile right now, George W. Bush's percentage is highly sensitive to the exact end date selected. McCall's chart happens to end on 10/10/08, which happens to be the end of several days of free-fall in the indexes. If it had ended on the next market day, 10/12/08, when the Dow and S&P 500 rose 11% in a single day, Bush's numbers would still be negative but would look a lot better.
10.14.2008 4:49pm
Sasha Volokh (mail) (www):
Yup, a lot of people have already noted this ("finman's link"), but I'll say it again. The Luskin article is what you should read to interpret these numbers. Let me say it again. Read the Luskin article. In other words, read finman's link. Then come back and we can talk.
10.14.2008 4:56pm
PubliusFL:
One more:

6. Failure to account for bubbles which do not represent real meaningful wealth. Clinton gets all the upside of the upside of the dot-com bubble, but none of the downside since most of its bursting occurred during Bush's term. There's nothing Bush could have done to preserve the illusory market capitalization created by the dot-com speculation. Bush had his own bubble but was unfortunate enough to have it burst a couple of months too early, otherwise his successor would have had the blame for it under McCall's methodology.
10.14.2008 4:57pm
1 1/2 cheers for Krugman:
the comparison ignores differences that may be due to control of Congress by the other party,

Doesn't this tell you all you need to know about the usefulness of the "study"?
10.14.2008 5:01pm
KevinM:
Hmm, yes. We really should elect a new president every month - then we'd have enough data points for a really good study.

Seriously, I don't think the author was even purporting to present this as a rigorous economic study, but just a set of considerations to give us pause in this silly election season. I think he sought only to rebut a certain simple-minded and partisan mindset: vote R because it's good for the economy. For that purpose, even these crude statistics are probably good enough.
The critics in the comments above, by reassigning party labels retroactively according to how the economy did, really reinforce the point, which is: watch what they do, not what they say, and beware of party labels.
10.14.2008 5:08pm
DrGrishka (mail):
Well, it may be that the effect of the policies is delayed. In other words, good policies show returns only a few years after being enacted, rather than right away. In that case, good performance under Democrats would be an indication of good policies enacted under a previous (GOP) administration.

Of course, this is just wild speculation. But I suppose a theory can be tested. We need to take a look at the periods of long control by one party (i.e., where the "previous" administration was of the same party) and see if the trend holds. Thus, under my theory, the latter part of FDR's and Truman's administration would see low growth, while the latter part of Reagan or Bush Sr. administration see high growth. (In fact, we could probably include early Reagan too, since Carter was in the office for relatively short time and buffered on each end by the GOP administrations).
10.14.2008 5:16pm
Randy R. (mail):
"Historically, Democrats tend to wage more wars than Republicans."

The Republicans are working hard to address this imbalance.

And of course, these wars that we were supposed to have won within one year and now costing us upwards of a trillion dollars. So does that count as helping or hurting the economy?

As to the larger issue, I don't believe the president can do much to make a good economy, but he can do a lot to screw it up. Beyond that, there are forces way beyond his control. Clinton benefited from the technology revolution, for instance. The Fed controls interest rates and affects the economy greatly.

Houston Lawyer: "You could easily assume that Republican presidents have spend the majority of their terms cleaning up the messes left by their predecessors. Bill Clinton gets all of the credit for the good side of the tech bubble and Bush gets the bad side."

If true, then it was a mistake of Bush to cut taxes so much that he and the republican congress turned a surplus into a deficit. And that has hobbled us in dealing with our current problems. (Or you could argue that Bush was perfectly happy to ride the same tech boom, and that's why he cut taxes. Either way, the Republicans are at fault for creating at least some of the deficit, no?)
10.14.2008 5:16pm
Elliot123 (mail):
"And that [tax cut] has hobbled us in dealing with our current problems."

Can you tell us what economic activity over the past seven would have been hobbled by the larger tax?
10.14.2008 5:22pm
kurious:
Repeat as necessary: "Correlation is not causation."
10.14.2008 5:25pm
Sarcastro (www):
Elliot123 taxes help with the GDP, and the stock market, and national well being, teenage pregnancies, crime, can give selected individuals the ability to fly and defeat terrorists!
10.14.2008 5:26pm
Randy R. (mail):
We couldn't spend money on upgrading SCHIP because it would cost too much.

Our infrastructure needs billions of dollars to prevent more bridges from collapsing, but were is the money for it?

Those are just two issues that can't be solved because of the deficit. Now, one can argue that SCHIP (providing health insurance for poor kids) isn't a function of gov't, but infrastructure certainly is. And I would hate to have anyone I know get killed because a bridge fell out from under him.
10.14.2008 5:26pm
The Unbeliever:
Some of the responses here are delusional. First, anyone who relies on Donald Luskin is, to use the word of the week, deranged.
Anyone who spends his time railing against the equally deranged Paul Krugman is fine by me.
Second, Luskin's argument is itself absurd. Think about it: if only you convert Ds into Rs, and Rs into Ds, the result changes. Duh.
That's a critique of the statistical quality of said studies; which is to say, it's crappy. Goes along the lines of suggesting there are insufficient data points. And the critique is valid, but I suppose it's only interesting to those using such studies as actual scientific data points instead of cheap partisan points.

More importantly, it answers David Post's implicit question: why does it seem that the classically stereotypical "Republican economic policies" lose (free market, deregulation, etc), whereas the stereotypical "Democrat economic policies" win (price controls, increased entitlements, socialism lite, etc)? And Luskin's answer is that the results are unexpected because for four of those 12 data points the actual implemented economic policies were opposite of the expected policies--the two Democrats were free market/supply side, and the two Republicans were more centralized command-and-control types.

Throw in all the other objections listed above, most significantly lag time and party control of Congress, and the inevitable conclusion is that these kind of partisan comparisons are a waste of time. Or: when it comes to economics and economic policies, the most simple answers are often the most misleading.
10.14.2008 5:27pm
Randy R. (mail):
" whereas the stereotypical "Democrat economic policies" win (price controls,..."

When has ANY democrat advocated for price controls? The only politician in modern times who actually pushed it through was NIxon.

I believe Nixon was a Republican.

I guess I have a problem with the story line that Republicans are good for the economy and Democrats are bad. Says who? Where is the proof? It's just a political stance, nothing more.
10.14.2008 5:32pm
Splunge:
Christ, I can't believe someone can become a college level teacher without grasping the elementary observation about reality, noted above by kurious, that correlation is not causation.

You might as well ask in a puzzled way (as the New York Times, hilariously, does every so often) why the number of people imprisoned rises even when crime falls.

How about we just run the correlation link the other way, viz.:

Why is it that in times of good economic growth, people generally tend to elect Democrats (the party of expansive government, income redistribution, higher taxes, and pursuing utopian goals of justice, health care, and world peace), whereas when times are tough they generally elect Republicans (the party of government belt-tightening, lower taxes, and being practical instead of utopian)?

Now that question kind of answers itself, without any need for imagining some strange new counter-intuitive theory about how the economic growth occurs.
10.14.2008 5:42pm
wfjag:

My third response, which you touched on, would be to examine control of Congress.

Dear Joe:

I saw an article that examined that -- sorry, didn't save the article or the link. The article reported that since WWII, when Republicans controlled Congress, the average annual increase in the DJ was a little over 14%, while when the Democrats controlled Congress, the average annual increase in the DJ was a little over 7%. I don't recall the average when different parties controlled different Houses of Congress.

The role of Congress would appear to be more important, since Congress enacts the laws, appropriations and budgets.
10.14.2008 5:44pm
PC:
How about national debt? Anyone want to compare that between D vs. R presidents?
10.14.2008 5:47pm
Sarcastro (www):
What about the number of pirates? I'm burning to know which party cuts down on those scurvy dogs and which one surrenders like scallywags!
10.14.2008 5:50pm
Derrick (mail):
How about national debt? Anyone want to compare that between D vs. R presidents?


Oh you really don't want to do that, ie. Reagan and Bush vs. Clinton's surplus.

This is an interesting topic, for I'm not sure we've seen such a large proportion of lame excuses on this thread in awhile. Truthiness Returns!!!
10.14.2008 5:53pm
byomtov (mail):
The Luskin article is what you should read to interpret these numbers.

You can't be serious. The man is a buffoon, as this article among many others demonstrates.

Even overlloking the utter absurdity of "if we count some Democrats as Republicans and vice versa it comes out different," consider this:

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.

JFK is a "Luskin Republican" because he cut taxes. (Saying he was an enthusiastic supply-sider is typical Luskin idiocy. See Herbert Stein's take here.). Clinton is a "Luskin Republican" because he promoted free trade. But wait. Clinton raised taxes, over strenuous Republican objection. So how can he be a Luskin Republican?

Forget it, Sasha. Luskin really is a moron.
10.14.2008 5:55pm
PC:
Derrick, national debt as a % of GDP.

Sarcastro, What about the number of pirates?

I would suggest doing the same comparison with ninjas.

I wonder if Luskin would consider President Bush to be a "Luskin Democrat" since Bush is nationalizing banks?
10.14.2008 6:07pm
PubliusFL:
Randy R." When has ANY democrat advocated for price controls?

We can start with the rent controls in effect mostly in traditionally Democrat-controlled states and cities, proceed to the price controls that have been part of various health care reform plans offered by prominent Democrats (ClintonCare, for example), and finish with any number of proposals to cap energy prices in the past several years.
10.14.2008 6:08pm
Joe Bingham (mail):
I'm no diehard Republican, but I'd think the obvious response would be that if Republicans take economic policy positions that're more advantageous over the longer run, and Dems take policy positions that're more advantageous over the shorter run, this is exactly what you'd expect to see as the parties trade places.

IDK if that fits with the data, but some variation of that seems like the simplest response.
10.14.2008 6:10pm
The Unbeliever:
When has ANY democrat advocated for price controls?
Well, FDR for one, the ur-Democrat that gets trotted out whenever Clinton worship gets stale or they need some foreign policy cred. Or have you forgotten when the most visible surviving form of price controls (rent control) was started?

Maybe you wanted the dumbed down cable news network version of the stereotype. Which goes something like this: Price controls==centrally planned economy==Democrats. And I already addressed the problem with this stereotype.
10.14.2008 6:11pm
Joe Bingham (mail):
Randy R.,

It's not just Democrats that support anti-gouging laws, a form of price controls, but I think support for them is more universal among Dems than Repubs.
10.14.2008 6:12pm
Sarcastro (www):
PC you can't study ninjas! The problem is their awesome stealth.

To whit: for every 3 ninjas you can't see, there are three other ninjas you can't see.
10.14.2008 6:16pm
Randy R. (mail):
"We can start with the rent controls in effect mostly in traditionally Democrat-controlled states and cities, proceed to the price controls that have been part of various health care reform plans offered by prominent Democrats (ClintonCare, for example), and finish with any number of proposals to cap energy prices in the past several years."

Oh, okay.
10.14.2008 6:46pm
roy:
It's even harder to measure than that -- not all pirates act like pirates. Nobeard-San sailed under the Jolly Roger and plundered booty, but he had a katana instead of a scimitar and a throwing star instead of a hook hand. If we count him as a ninja, the pirate-to-ninja ratio ratio actually reverses and shows that Republican(ish) presidents grow the ninja population.
10.14.2008 6:48pm
Jay Myers:

if you call elephants 'mice,' and mice 'elephants,' then 'elephants' weigh less than 'mice.

When making an examination such as this you have to decide if your goal is to score partisan political points or to decide which policies, and thus which candidates, would best serve the republic. If the former then you categorize your data by political party. If the latter then you categorize your data by policy approach. I guess it shouldn't surprise anyone that you prefer to go for partisanship.
10.14.2008 7:11pm
Joe Bingham (mail):
"When has ANY democrat advocated for price controls? The only politician in modern times who actually pushed it through was NIxon."

I think I've read that price controls had bipartisan support during the Nixon era.
10.14.2008 7:12pm
A.S.:
Derrick, national debt as a % of GDP.

This is the wrong statistic - you need to look at debt held by the public. The graph would probably look similar to the linked graph, but I just wanted to point out that the linked graph is misleading.
10.14.2008 7:12pm
David Hecht (mail):
This sort of analysis shows the essential futility of trying to divvy things up by party. As many have pointed out, not all Republicans follow free-market policies, and not all Democrats are statists.

BTW, are these figures adjusted for inflation? Becasue if not, then they aren't a very realistic measure.

That being said, I say you should read the stats differently: four out of six Republicans had double-digit Dow growth during their administrations. Only one Democrat out of six did. So--by that measure--your chances of a big jump are two out of three with the GOP in the White House, and only one in six with the Dems!

How's that for bogus statistical analysis?
10.14.2008 7:16pm
Calderon:
Having read the link with Luskin, I don't quite get some commenters' visceral dismissal of him. Because of the small sample sizes, I don't think he proves anything, but he certainly illustrates a lot of the problems with the Bartel/Blinder claims, such as lagged responses to fiscal stimulation, what happens when you count Congress, etc. As far as Nixon being more like a Democrat and Kennedy or Clinton being more like Republicans, isn't the point that one should look at the actual policies of the presidents rather than whether there's just a D or R by their name?
10.14.2008 7:18pm
A.S.:
Also, to the extent that we are trying to use the graph to determine which party is better for the economy in the future, what we should really be studying is Democrats *who don't implement supply side tax cuts* vs. Republicans *who don't implement wage and price controls*.

After all, in the future, we surely don't expect Democratic Presidents to implement supply side tax cuts, and we don't expect Republican Presidents to implement wage and price controls. So, if somebody would please rerun the numbers with those changes, I'd appreciate it.
10.14.2008 7:19pm
Brian K (mail):
First, not all Democrats act like Democrats, and not all Republicans act like Republicans.

[and to many others with spout the same argument]

if your only attack on this study is a form of the no true scotsman fallacy, then you'll have to forgive me for not believing you.

as you so readily admit, not all democrats act like "democrats" and not all republicans act like "republicans". which is more likely to be wrong, this study or your biased conception of each side?
10.14.2008 7:23pm
Brian K (mail):
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.

so in other words, what this guy is saying is that if you change the data to get the result you want then you'll get the result you want. although true, it is not a very intelligent critique of the study in my opinion.
10.14.2008 7:25pm
Bart (mail):
There are so many problems with the NYT graphic, it is hard to know where to begin:

Parties are not the proper methodology for this measurement. Rather, the methodology should be conservative (restrained government and free markets) and liberal (expansive government and regulated markets). For example, under this methodology, Nixon was an economic liberal far closer to LBJ than to Reagan and Clinton was an economic conservative far closer to Reagan than to previous Dem Presidents.

As noted above, an Administration can only be given credit for the results of its own policies, which take a year to implement and a year to take effect. The periods examined for each administration should begin and end two years later than in the graphic. This gives Mr. Carter credit for the recession he dumped on Reagan and Clinton credit for the recession he dumped on Bush.

Occasionally, the policies of a previous administration have an even greater lag time into the following administration. The 2008 market panic was caused by a mortgage failure which was in turn caused by the government prodded growth of the sub prime market starting in 1994.

There are far too many variables outside the control of the President affecting these measurements. For example, Roosevelt's New Deal turned what historically should have been a 2-3 year recession as with previous panics into a far deeper decade long Depression. What increased the stock market under his watch was the advent of WWII and war production, not the New Deal.

Congress has more than a little say in implementing a conservative or a liberal economic policy. The liberal economic policy of Clinton and the Dem Congress turned into the completion of the Reagan Revolution under Clinton and Gingrich.

I could go on, but the entire NYT graphic is an academically worthless exercise in partisan election year gotchya.
10.14.2008 7:29pm
Lior:
The main issue to be addressed, in my opinion, is showing that this is a case of causation rather than correlation. Perhaps the voters prefer Democrats when they anticipate good economic times, but Republicans when they anticipate bad economic times? I'd love to see a correlation between the party in power and the economy in the year before the election.
10.14.2008 7:31pm
Brian K (mail):
to the argument that there is spillover that accounts for the difference:

it may or may not be true, but it definitely cuts against one argument that many republicans are fond of making. that drilling in ANWR will decrease the price oil now rather than 10-15 years down the road when the oil becomes available. does the market react to expected future changes or does it only react to what is happening right now? you have to pick one and stick with it...you can't choose whichever theory makes your side look the best at the time. well at least not if you want to maintain any integrity.

remember, this is the stock market we're talking about here. which, if the efficient markets theory is true, is supposed to be very good at pricing in future events.
10.14.2008 7:34pm
MarkField (mail):

that at least as far as the GDP numbers are concerned, one needs to take into account a President's actions and the effects of his policies, and that with a 2-year lag the data show almost equal GDP growth under Republicans and Democrats.


This is fine up to a point. I think a lag time is reasonable. The question is, what is the extent of the lag, and why? Until someone has an answer to that, I don't think you can just assume a two year lag. Why not 4? Or 1?


Having read the link with Luskin, I don't quite get some commenters' visceral dismissal of him. Because of the small sample sizes, I don't think he proves anything, but he certainly illustrates a lot of the problems with the Bartel/Blinder claims, such as lagged responses to fiscal stimulation, what happens when you count Congress, etc. As far as Nixon being more like a Democrat and Kennedy or Clinton being more like Republicans, isn't the point that one should look at the actual policies of the presidents rather than whether there's just a D or R by their name?


I'll respond to this as representative of several who defended Luskin.

Luskin himself makes an idiotic argument. What he seems to be grasping for is a way to articulate the argument that you cogently raised in your last sentence, namely that the focus should be on the specific policies followed rather than the partisan identification of the President. This, I suspect, is true, but many people forget it in election season.

There is, though, a problem with the "it's all policy" counterargument. That argument assumes (I think) that the same policy will always be effective. It may very well be that policy A will be successful under condition X, but policy B will be effective under condition Y. IOW, it's a lot more complicated than internet blog comments would have you believe.
10.14.2008 8:01pm
Observer:
This is all because the market anticipates the future. A Democratic administration is more likely than a Republican administration to be followed by a Republican administration (or Republican takeover of Congress); hence, the market performs well. A Republican administration is more likely than a Democratic administration to be followed by a Democratic administration (or Democratic takeover of Congress); hence the market performs poorly. Isn't this precisely what we see now with the market going downhill as the outcome of the November election becomes more certain?
10.14.2008 8:07pm
PC:
Bart, so everything good that happens in the economy is because of Republicans and everything bad that happens is because of Democrats?
10.14.2008 8:09pm
Brian K (mail):
The question is, what is the extent of the lag, and why? Until someone has an answer to that, I don't think you can just assume a two year lag. Why not 4? Or 1?

Bart answers this question for you:
"The periods examined for each administration should begin and end two years later than in the graphic. This gives Mr. Carter credit for the recession he dumped on Reagan and Clinton credit for the recession he dumped on Bush."

the why can reasonably be inferred.
10.14.2008 8:33pm
Elliot123 (mail):
"And I would hate to have anyone I know get killed because a bridge fell out from under him."

The city, county, or state is free to rebuild any bridge they choose.
10.14.2008 8:33pm
Calderon:
MarkField -- your point about lagtime is completely right. There's no particular reason to assume 1, 1.5, 2, or 4 year lag for all policies. One could imagine policies such as improvements in elementary and secondary education that would take many more years to have an impact. Or for more recent events, some have blamed the repeal of Glass-Steagall (10 years ago), passage of the Commodity Futures Modernization Act (12 years ago), regulatory changes to the Community Reinvestment Act (13 years ago) for the current financial problems. But all that proves is the problems with Bartels' study or any other study that measures a president or party's success by what happened during his term.

For Luskin, maybe we're just reading him different ways. To my view, all he's arguing is that there are a lot of problems with a Bartel-type argument, and then he illustrates it with various examples and different assumptions. His second to last sentence seems to establish that he doesn't actually think that one kind of President/Congress organization is better or worse than another. ("That tells us that economic and stock market success isn't really about partisan politics at all.") All of his criticisms seem valid to me; are there ones you or others disagree with?

Also, on the last point, looking at the candidates' policies doesn't assume that the same policies work all the time. The only point is that sometimes a candidate or president's personal policies are inconsistent with what it believed to be the general policies of his party, and so one should look at the presidents actual policy preferences instead of assuming that Nixon will get government out of the economy because he's a Republican or Clinton would regulate more and clamp down on free trade since he's a Democrat. (And of course what a president actually does is a function of Congress, etc.)
10.14.2008 8:40pm
PC:
Bart answers this question for you:

Except for the current crisis which requires you to look back 14 years. I'm sure if there's another period of economic slump Bart can explain how we need to adjust the numbers to lay the blame on Dems.
10.14.2008 8:41pm
Calderon:
One error in my last post: "passage of the Commodity Futures Modernization Act (12 years ago)" should be 8 years ago in 2000.
10.14.2008 8:41pm
Calderon:
One last point that also needs to be brought up, of which I was reminded by some of the other posts. In addition to Congress, the role of the Federal Reserve on the economy is another factor that's independent of the President. The recession during Reagan's first term is widely attributed to Volker's tightening to stop inflation. Some blame Greenspan for raising rates during Clinton's first term causing lower growth. Growth from 2003-06 as well as the credit crunch now have also been pegged to Greenspan's easy credit.
10.14.2008 8:51pm
byomtov (mail):
There is, though, a problem with the "it's all policy" counterargument. That argument assumes (I think) that the same policy will always be effective. It may very well be that policy A will be successful under condition X, but policy B will be effective under condition Y. IOW, it's a lot more complicated than internet blog comments would have you believe.

I agree with this, and would add that defining "policy" in a way that makes it amenable to this sort of analysis is at best very difficult. Governments don't have one economic policy. They have several, and the economy is also affected by things that are not part of "economic policy," like wars, or technological innovation, for example.

You can't just categorize policies the way Luskin wants to do.

My response to those who say he's just criticizing the method is that he's using a worse method to promote his preferred ideas. We can at least tell which party Presidents belonged to. We can't easily sort their economic policies the way he wants.

All that said, I agree that this sort of analysis isn't particularly meaningful, though I do think it suggests that, contrary to many opinions, Republicans have no particular claim to superior economic management.
10.14.2008 9:00pm
Brian K (mail):
PC,

good point.
10.14.2008 9:04pm
CJS (mail):
Greg Mankiw takes on this dubious claim on his site, which for some reason I can't link to right now.

Ultimate quote:

"The bottom line: Trying to isolate the differences between the parties with stock market data is silly at best."
10.14.2008 9:16pm
Sasha Volokh (mail) (www):
David: I don't think your update is a sufficient rebuttal to Luskin's points. Reclassifying presidents as Democrats vs. Republicans is not a matter of calling mice elephants. Rather, the only reason it makes sense to talk about Democrats' or Republicans' effects on the economy is that we believe that Democrats and Republicans have different approaches to policy.

So, properly, we don't want to talk about party labels as such, but the characteristic that these labels are trying to capture (for example, regulatory vs. deregulatory, or high-tax vs. low-tax, or protectionist vs. free-trade, though you could look at other dimensions too). Ideally, we'd want to give every president a score along a continuum, or along a set of continua, and figure out the effect of movement along the continuum on economic outcomes.

But of course the original chart wasn't doing that, opting instead for the political label-based approach. A second-best thing would be to try and redo the exercise using real characteristics, not labels. Therefore, let's try to put the deregulatory or low-tax people together, and the regulatory or high-tax people together. That's why it makes sense to put Nixon (EPA, wage and price controls, etc.) with the more regulatory types, for instance. (The reassignment of Nixon was enough to make the results disappear.)

Another item you didn't mention was also very important -- it's treating the "party in power" as not just who controls the White House but as who controls the White House and Congress. This gives us results on divided government and the like, and makes more sense.
10.14.2008 9:46pm
PC:
Prof. Volokh, so all we need is an n-dimensional matrix to solve this equation? Some hedge funds should be clearing out pretty soon, maybe we can use their cycles for the number crunching.
10.14.2008 10:10pm
Brian K (mail):
Rather, the only reason it makes sense to talk about Democrats' or Republicans' effects on the economy is that we believe that Democrats and Republicans have different approaches to policy.

what this study and your (and others) attempts to redefine the data in a certain way show is that these policy beliefs aren't true (or at least highly influenced by personal bias).

attempts to rationalize away the results doesn't change the fact that the results run contrary to what the majority of posters want to believe. and when discussing which party results in a greater increase in stock prices you can't change which party a president belonged to just because it gives results you dislike. the president was (re-)elected as a R (or D) and governed as an R (or D) and therefore was an R (or D).

that's not to say there aren't legitimate criticisms of the study. this just isn't one of them.
10.14.2008 10:19pm
Sasha Volokh (mail) (www):
Brian K: Yes, if all we knew was that someone is an R or a D, then yes, we have to run the numbers using the label R and D. You're right that, in that case, we can't use stereotyped views of the party's views: R is as R does.

But if we know that someone favors policies X and Y, then it makes sense to, instead, rank different Presidents (or President-Congress combinations) on scales of X-ness and Y-ness. That will be a lot more sensible.

Unless... it turns out what a candidate _says_ about the policies he favors turns out to have nothing to do with what he actually _does_. For instance, who would have guessed what government interventions in the economy Bush ended up favoring? Suppose that a candidate's statements had _nothing_ to do with his actions, and the best predictor of his actions was still the R or D label. Then the label-based approach is still the best way to go.

But... I still suspect that, if we consider the candidate's statements about what he'll do, it's still a better predictor of how he governs than his R or D label alone. This is an empirical guess, but I suspect that it's true. So my approach would be to use the candidate's statements, possibly together with the R or D label, and of course also merging it with who controls Congress. This would give us a better view.

Barring the ability to do that -- we can do it, just not in the blog comments, from our armchairs, in our pajamas -- the best strategy is to point out how meaningless the R/D labels are as such, by doing exactly what Luskin does: Consider the variation among Ds and among Rs, such that some of them are hard to tell apart from each other. Then try reassigning some and see whether your results change.

The point is not to show that the new, rejiggered results are correct. Rather, the point is to show that the old results show less than they seem. (One could say "are meaningless"; perhaps this is too strong, but in any event, relying on the labels-only approach obscures so much variability (that's known beforehand!) that it's not very useful.)
10.14.2008 10:43pm
Tatil:

Isn't this precisely what we see now with the market going downhill as the outcome of the November election becomes more certain?

No. McCain was going neck and neck in the polls until the financial crisis took on a huge size and required a gigantic bail out.
10.14.2008 10:48pm
MarkField (mail):

For Luskin, maybe we're just reading him different ways. To my view, all he's arguing is that there are a lot of problems with a Bartel-type argument, and then he illustrates it with various examples and different assumptions.


I would agree that I'm not giving Luskin the benefit of the doubt. That's because I've seen him make silly arguments on several occasions before. It's possible he was making your argument and just chose a bad example. But it was a lousy example; the posters in this thread have done a much better job and they don't get national column space.

I mostly agree with the rest of your points. This really is an n-dimensional problem, which is what makes it so hard. Trying to boil it down to a simple R v. D is what's silly, and I think that was partly Bartels' point (here I am giving him the benefit of the doubt).
10.14.2008 10:51pm
Tatil:

Christ, I can't believe someone can become a college level teacher without grasping the elementary observation about reality, noted above by kurious, that correlation is not causation.

However, if there is no causation, there is no correlation. Therefore, based on this data, either Dems are good for the economy or this is just a coincidence and it actually makes no difference. Either way, "Republican presidents are good for the economy theory" is shot.
10.14.2008 10:52pm
marty (mail) (www):
There are so many numerous mistaken assumptions floating around here that it's hard to know where to begin, so let's just start with a few things:

1. "we don't want to talk about party labels as such, but the characteristic that these labels are trying to capture"

or-

Can we make mice elephants and vice versa and have this chart be valid?

Those would both be true if we wanted to make a point about what sorts policies are best for economic growth. But we don't. We want to talk specifically about political parties. Talking about anything else is just going off the point.

2. "Correlation is not causation."

Again, that would be true if we wanted to talk about what were the causes of economic prosperity. But we don't. We specifically want to talk about correlation only. That's what the graph is about.

3. We need to factor in Congress, Fed Reserve Chairman, tax policy, etc.

Again, that would be true if we wanted to find out what role Congress, Fed Reserve, tax policy has vis a vis the economy. But we don't. We're just talking about the President.

Folks, the only point of this graph is to establish whether or not there is any kind of correlation between the party of the President and a single gauge of prosperity - the stock market. We are not looking into causes, other factors, etc. Just Presidential Party and correlation, that's the only point that needs to be discussed here.

Why? The reason why is because it is very frequently a *Republican* argument that there will be some kind of *correlation* between electing a Democrat as President versus electing a Republican on the overall performance of the economy. That correlation is often posited to favor Republicans: the graph is simply meant to show that that argument is mistaken, the correlation in fact favors Democrats. (No Republican who makes this point stops to say, oh, well, but "don't consider Clinton a Democrat when you vote, we think you should vote for him because he acts like a Republican." - No each party paints the other with the same broad accusations.)

All of you who argue that there are other factors, lag-time, causation elements, policy differences are just supporting that point, that there is in fact is NO POSITIVE correlation between voting for the Republican for President and the economy doing better (according to you all the reason for the improved economy would be policies, congress, etc., and not the party in power). So you are merely supporting the point of the article in the NY Times: That there is no reason to vote Republican if your reasoning is that would increase the odds of a better economy...since THAT FACTOR ALONE (the presidential candidate being a Republican) doesn't correlate. In fact, instead, you might do as many here suggest...and actually think about where candidates stand on addressing the economic issues, rather than their party. But then, that seems to be the POINT the graph and the article (and in fact Obama himself) is making.
10.14.2008 10:52pm
AK (mail):
Alternate, equally valid conclusion from the data:

Presidents who serve exactly one full term are immediately followed by presidents whose policies pair with the highest rates of DJIA growth.

So this proves that the presidents who follow one-termers are the best for the economy. It's, like, SCIENCE!!
10.14.2008 10:57pm
ed (mail) (www):
Hmmmmm.

It's an idiot study because Presidents have very little power over domestic issues. A study based on control of Congress would actually make sense.
10.14.2008 11:01pm
Bart (mail):
PC:

Bart, so everything good that happens in the economy is because of Republicans and everything bad that happens is because of Democrats?

Not really.

LBJ's implementation of the Kennedy cut in marginal tax rates caused an economic boom.

Nixon was arguably the worst of any of the Presidents in this area.

Clinton was a far better economic President than the GOP gives him credit for. If not for the marginal tax rate increases, Clinton would be near the top.

Apart from the 2003 marginal tax rate reductions, W was far worse than Clinton.

This is why I noted at the outset of my post that the better methodology is conservative (restrained government and free markets) vs. liberal (expansive government and regulated markets).
10.14.2008 11:06pm
AK (mail):
The reason why is because it is very frequently a *Republican* argument that there will be some kind of *correlation* between electing a Democrat as President versus electing a Republican on the overall performance of the economy.

I'm sure you can locate a few people who will say "Republicans good, Democrats bad" and actually believe it. Those people are idiots.

I don't know if any serious person actually believes that Barack Obama will be bad for the economy simply because he's a Democrat. If you hear them say that, it's just shorthand for "he's going to soak the most productive individuals and corporations, which is bad for the economy." You certainly don't hear that sort of argument "frequently."
10.14.2008 11:07pm
Bart (mail):

Brian K (mail):

The question is, what is the extent of the lag, and why? Until someone has an answer to that, I don't think you can just assume a two year lag. Why not 4? Or 1?

Bart answers this question for you: "The periods examined for each administration should begin and end two years later than in the graphic. This gives Mr. Carter credit for the recession he dumped on Reagan and Clinton credit for the recession he dumped on Bush."

PC:

Bart answers this question for you:

Except for the current crisis which requires you to look back 14 years. I'm sure if there's another period of economic slump Bart can explain how we need to adjust the numbers to lay the blame on Dems.

Which part of it takes one year to enact the policies and another year for them to take effect did you willfully ignore?
10.14.2008 11:09pm
Pauldom:
Even if we were checking correlation with Congress (which would be interesting, even if not the point of this graph), it wouldn't be enough to see which party controlled Congress; we'd need to see whether D or R proposals were enacted. IIRC, a Dem congress passed pretty much all of Reagan's 1st term initiatives. So does that mean R gets credit? Or does D get credit for being wise enough to support R?

But overall, I agree with Marty: the utility of the graph is that it shows no correlation between Republican president and good economy.
10.14.2008 11:18pm
Brian K (mail):
Which part of it takes one year to enact the policies and another year for them to take effect did you willfully ignore?

which of your own post do you not understand? i see no evidence that you have a reasoned belief in the 1 year time frame for either of these events other than the fact tha 1+1=2 and 2 makes democrats look bad. the other examples pointed out by PC from your own post further provides evidence. the time lag is always just coincidentally long enough to blame whatever bad happens on a democrat.
10.14.2008 11:32pm
PC:
Bart, the point where you dipped back to 1994 to make a partisan point. There are many things that caused the current financial crisis, but anything that happened in 1994 is low on the list.
10.14.2008 11:33pm
Brian K (mail):
Sasha,

marty in point 3 provides a fairly effective response to your argument so i won't rehash the entire thing. but to briefly summarize, the point of this study appears to be an attempt to debunk the myth that R's are naturally better than D's when it comes to the economy. it fairly effectively destroys the so often used claim that Ds will destroy the economy and Rs will usher an a golden era.

Your critique appears to boil down to the fact that you wish the authors would have done a completely different study altogether. yes, they could have looked at all kinds of other things and done an analysis of underlying policies and other stuff (although i recall posts on volokh.com that classifying anything as liberal or conservative is highly problematic), but that is another study for a different purpose.

For the most part I agree with you, a candidates views/beliefs/past history is the best method to judge a candidate by. (problems result however when their future plans don't match their past actions) but, to beat an already dead and beaten horse, that wasn't this study's purpose. if you wish to do one, i'll be happy to read it though.

The point is not to show that the new, rejiggered results are correct. Rather, the point is to show that the old results show less than they seem.
I disagree, many of the above posters appear to want to show that the results mean the opposite of what they say, not just that they mean less than they seem.
10.14.2008 11:45pm
EricPWJohnson (mail):
Some refer to the 2 year lag as one year is under another budget and the next year (the second year) the economy is still acting and reacting to the previous administrations 4 or 8 years of policies.

BTW Luskin feels Krugman's work, ethics and his politics are not Nobel Prize worthy either. I was reminded in the article of companies who hired him for advice in his brilliance

Ask Enron, GM, hows that working out for them?
10.14.2008 11:48pm
CJS (mail):
marty,

Exactly who makes the argument that electing a Republican president would lead to a correlation with better economic times? I've heard a lot of arguments, by a lot of Republicans, and they go something like this: If X Democrat is elected, he will support Y policy that will be bad for the economy. How would this address such arguments at all? Unless your point is that the whole NYT piece was engaged in arguing against a ridiculous straw man, and didn't add anything to the debate, in which case I agree.
10.15.2008 1:05am
Sasha Volokh (mail) (www):
Brian K: I can't take responsibility for what the other commenters want to read the study for!

But seriously, "all those other factors" are very highly relevant, even for the very limited question of whether D or R is better for the economy. For instance, maybe the answer is "neither D nor R is better as such, but the best is to elect D when Rs control Congress." Seriously, after having observed recent history, maybe I'd affirmatively vote for a Democrat if I knew there was going to be a Republican Congress!

Plus, suppose a candidate says: "I'm a Republican, but under no circumstances will I ever do what Nixon did -- that was just craziness!" Or suppose he doesn't say this explicitly, but that there are strong reasons to believe that this is the case, say because of his voting record (what if he was a key conservative critic of Nixon's economic policy?), his advisors, etc. Then it's appropriate to judge him not by R standards, but by R-except-Nixon standards.

What the study shows effectively is that, if all I knew was that a guy was D or R, it would be irrational for me to vote R on the ground that this would be better for the economy. (Even then there are problems, like lags.) But this is never reality -- we always know more about people. Once you control for that stuff, both what we know about who controls Congress and what we reasonably believe about the candidate's own beliefs, I suspect that a correlation -- supported by a good causal theory -- would emerge. Of course this is speculation on my part, because I haven't done that study, but hey, that's my guess.

So, far from "effectively destroy[ing] the so often used claim that Ds will destroy the economy and Rs will usher an a golden era," it merely shows that "here's one empirical test that doesn't support the claim." By the way, my suspicion is not that a good study would support Republicans generically. As to partisan labels, I think it would support divided government; and as to actual policies, I suspect that it would strongly disfavor high-tax, high-regulation policies. Again, that's just my guess.
10.15.2008 4:21am
David Warner:
"But overall, I agree with Marty: the utility of the graph is that it shows no correlation between Republican president and good economy."

Nor does it show a correlation between a Democratic president and cleaner air and water. Glad we got that canard debunked!

Logic. It's hard but useful.
10.15.2008 4:45am
PC:
Nor does it show a correlation between a Democratic president and cleaner air and water.

Still waiting for ninjas and pirates.
10.15.2008 5:00am
David Warner:
PC,

"Still waiting for ninjas and pirates."

No correlation there either! Is there anything the chart can't do?
10.15.2008 5:29am
Cody:
Another embarrassing post from Post. (It's better than Palin blogging though!)

So let's see. The study used crude political labels to analyse policies instead of the actual policies. It looked at the executive branch when the legislative branch has at least as much influence. And it forgot to account for policy lag. All three errors are glaringly obvious, easy to correct, and any one of which would have reversed the results if corrected.

Ooh, now that is damning! (Well, actually, no. I wouldn't dream of making a sweeping condemnation of Democratic policies based on stock market returns. True, the data does point in that direction, but the sample size is tiny, I'd be shocked if the results are statistically significant, there's massive external factors like oil prices not being accounted for, and the entire idea of using the Dow Jones to evaluate political parties is absurd.)
10.15.2008 8:55am
Elliot123 (mail):
"However, if there is no causation, there is no correlation."

Could you elaborate on that?
10.15.2008 12:01pm
Sasha Volokh (mail) (www):
Cody: As someone who's criticized this in the posts above, let me note that the crude political labels can be useful. If all you knew were the political labels of candidates -- and you knew nothing else -- then this might be the right thing to use. Because you have to know whether to vote R or D. Using policies would be harmful in that case, because it's stereotyping to assume that a R votes for free-market policies or whatever.

Relatedly, it allows one to judge the naked statement "Vote for Rs because Rs are better for the economy" -- as opposed to a policy-based statement like "Vote for people who believe in low taxes because low taxes are better for the economy." (Ignoring all the other valid critiques....) If nothing else, this sort of analysis would allow one to disprove the correlation between being an R and doing stereotypically R-style policies.

Now allow for people who are interested in making more nuanced statements.... You might want to code for actual policies, but there's the danger that the campaign rhetoric may not correlate with actual policies, so it's not useful to make voting decisions before the election. This is why you want to code for _statements_ about policies, in addition to possibly the candidate's own policy record over the course of his career.

Also, the party label may have extremely different effects -- in fact, maybe opposite effects -- depending on who controls Congress. Perhaps a R Congress will primarily fight a D President's big-government instincts, but when an R President is in office, they'll all collaborate in funding their own pet projects, so an all-R government is affirmatively harmful while an R-Congress-with-D-President is affirmatively beneficial? Maybe or maybe not, but it's a possibility.

I suspect that when you put candidate statements together with party record and legislative control, and also put in some appropriate policy lag, then you may well find something useful. As it is, it's only useful for arguing against the crudest possible partisan rhetoric. Admittedly, some people may engage in that sort of stuff, but it's not the sort of thing I'm much interested in.
10.15.2008 12:48pm
astrangerwithcandy (mail):
10.15.2008 2:45pm
Opher Banarie (mail) (www):
The DOW 30 Industrials is a useless indicator of stock market performance. Thirty companies out of thousands. And looking at it's "performance" over 80 years you wash out the changes that the index has undergone (all those years that IBM was not part of it, or that Kodak was part of it) and the fact that the index does not consider things like dividends and merger payouts. Already stated were the inflation and other macro-economic factors that affect the numbers without regard to the party "in power".

Oh, yes, and what does it mean to be "in power"? The NYT conveniently only uses Presidential party affiliation, without noting that Congress may or may not be the same party - or how well the party coordinates with the president. For example, LBJ got much more from a Republican Congress (66-68) than he did with a Democratic Congress (64-66).

Nice chart. Useless conclusions.
10.15.2008 7:44pm
Kathryn Cramer (mail) (www):
Theodore Gray, co-founder of Wolfram Research, decided to build a model in Mathematica to check Tommy McCall's figures. What he found was, unsurprisingly, that reality is more complex than infographics and that different assumptions in the calculations lead to differing results on financial returns by party, but that what works best in the very long haul is to leave one's money in the market no matter who is president.

Wolfram Research has published Theo's piece in two parts: as a blog post
and as a Mathematica Demonstration.
10.16.2008 7:00pm