But there are others who believe that a (D) is short-hand for something else, namely that the politician generally hews a bit more closely to a particular set of policies that have worked a little better than the policies adhered to by people who have an (R) after their name. Its not a guarantee of performance.
I have noticed that many people who take the second approach have been criticizing Obama's approach even before he took office. I for one have posted (over and over and over) on the folly of bailing out the financial institutions that have caused this mess (since when is taking from the poor and middle class and sending to the fabulously wealthy compatible with the policies that generally define Democrats?), and I've had a few posts noting that Christina Romer was a very poor choice for CEA chair (Mankiw's endorsement alone should have been a tip-off even to someone who never read her "narrative economic history" paper).
Sure, exogenous factors do matter. Truman was a Democrat, but the economy was awful for the first few years of his administration; transitioning from WW2 to a peacetime economy ain't easy. And it sucked in 1933 too, which was FDR's first year. This despite the fact that annualized, real GDP per capita grew much faster under FDR than any other President over the time period for which the BEA computed GDP, even leaving out the War years. Which also indicates something else - the economy might still turn out alright despite the many Obama mis-steps. Time will tell.
But there are a few other things - for one, you don't have only one (R) or (D) administration to look at. You have probability and statistics. (And to forestall the point that inevitably gets brought up again and again and again and again, that's what degrees of freedom are for.)
And while no theory is perfect, the better ones require fewer exceptions. If Obama starts acting like a Democrat and the economy still sucks by the end of his term, that means Democrats have to explain away Obama and Truman. Republicans still have to explain why Republicans generally do so much worse when it comes to real GDP per capita growth rates. The BEA has only been calculating real GDP per capita since 1929, and yet the best performing Republican, Reagan, is beat out by four Democrats, and all but one of the entire bottom half of the sample is taken up by Republicans. Tip to Megan McArdle: nobody can tell you for sure who's going to win next year's World Cup in South Africa, but betting on Grenada to pull it off is a bad idea.
And then there's the question that McArdle herself keeps bringing up - the mechanism. The mechanism doesn't apply if you don't have the right inputs. As I noted above, Obama has not yet behaved like a Democrat; his policies on the bail-out have been essentially the same as those of his predecessor who most certainly was no Democrat.
So in the end, here's what I say to Megan McArdle: ask your question again when Obama starts doing things a Democrat would do. Pushing through the "public option" on a healthcare would be a start.
So Obama doesn't count because he's not really a Democrat. But Bill Clinton was. But Richard Nixon--the chap who implemented price controls and massively expanded Social Security and Medicare--was definitely a Republican. Jimmy Carter, who deregulated like mad: definitely a Democrat.
What are these policies that neatly define Democrats to exclude only the ones who happen to have crappy growth? On what metric does Barack Obama register as farther to the right than Bill Clinton? Because from what I remember of the 1990s, I spent most of the decade listening to my genuinely left-wing friends weep that he'd betrayed them. Remember Edelman's resigning in protest of welfare reform?
I submit that there is no intellectually credible way that you can throw Obama out of the sample because he didn't include a public option in the health care plan--something that's actually driven by Congress, anyway--while keeping Clinton, his welfare-reformin' ways, and his lovely, lovely late 1990s growth rate. On actual domestic policy, Bill Clinton was well to the right of Richard Nixon. You could argue that he was more socially liberal, but then Jack Kennedy was less socially progressive than Richard Nixon--time is a more important variable than party identification. And at any rate, I'm aware of no work indicating that abortion policy is an important factor in economic growth.
Now, of course, it may be that the economy starts growing like gangbusters in the next year. In which case I expect that Cactus and his merry band of madmen will continue with their arguments. But if, as most people expect, growth continues to stall for the next few years, it seems I can look forward to more explanations of why Democrats--and only Democrats--can be thrown out of the sample if they have low growth and betray The Faith; and why the economic results of Democratic presidential administrations--and only Democratic presidential administrations--are sensitive to exogenous starting conditions.
The problem with saying that the economy fares better under Democratic President's is that it is an isolated item. Granted, Presidents have agendas and seek to implement them, but it is similar, in my mind, to saying that the economy fares better on a sunny day.
The fact that a Democrat or Republican is president is only part of the picture. There are a host of other factors that could contribute to the fate of the economy - one in particular comes to mind, and that is Congress.
For example, you could argue that the economy does better under Democratic Presidents, but when you look at Clinton, who controlled Congress for the majority of his presidency? Republicans. So you could then argue that the economy does better when Republicans run Congress.
It is simplistic to say that simply having a Democrat in the White House will result in a stronger economy. There is much more to it than that.
I too would like to see an analysis which includes congress. As much as I have to listen to Democrats blame things on Bush, the contention that his final two years when the Democrats totally controlled congress were the best, is quite rare.
Iraq War: We could have been out, no super majority needed, just don't approve the funds.
"Bush Deficits": Again, when you control congress,complaining about fiscal policy doesn't quite work.
Financial Crisis: Perhaps I just missed Barney Frank and Chris Dodd's calls for contracting home lending or increasing rates in the period after their respective chairmanships began and before the bubble burst.
In the minds of the democratic politicians, the strategy is clear - why waste political capital when one can likely gain support for favored social programs by supporting the wars, wiretapping, etc., and still get to blame the Republican president because Americans still think the president rules by fiat.
For the Democratic base, this is quite uncomfortable, as it leads to the recognition their leaders are humans rather than angels, the acceptance of public choice theory, and many other thought processes which they decided against years ago.
Of Cactus and Yglesias, you responded to the far worse argument. Yglesias says that even if Obama collapses, adding Bush's failures to the Bartels chart will not change the overall pattern. He also wonders why you're worried about disingenuous hypothetical liberals instead of the many real ones who criticize Obama now. Both points strongly undercut your argument about the impending demise of the Bartels argument. In response, you ignore both and track down a disingenuous liberal to argue with?
Gotta say, Megan, as much as I love your blog, you have a habit of saving your longest posts for your weakest critics.
All of these "studies" have the same problems: the number of data points is too small, the start points are somewhat arbitrary, and the "D" label is very poorly specified.
And the irony is that when this data first started making the rounds, the right wing response was that Kennedy really was a conservative while Ike and Nixon really were liberals. In other words, Yglesias made the same argument as Grover Norquist, he just cherry picked different presidents.
Much as Angry Bear protests, he does in fact only have a smattering of statistics.
He doesn't seem to realize that most of his tests require IID (Independent Identically Distributed) samples (which he doesn't have).
The IID requirement is almost never satisfied in non-laboratory studies.
Macroeconomics wouldn't exist if you required IID. It's silly to point that out as a flaw. Rather, these studies are always forced to assume that sample correlations are not high enough, and variances do not vary wildly enough, that they invalidate all conclusions. There are a number of post hoc methods you can use to test this assumption.
Right, but he doesn't use appropriate tests. He uses tests designed for independent distributions, and doesn't control for sample size.
Translation: "The problem is hard, so we should be less careful with our methodology."
Sorry, but no. Angry bear rans some basic hypothesis tests on a case when they don't apply. As you point out, this renders his conclusions invalid. Some other studies are invalid for the same reason. So what? That doesn't make Angry Bear's results any more valid.
I also found his "degrees of freedom" comments a bit odd. Using 11 data points and one regressor, okay, fine ... but then he's saying either that presidential party is all that matters or that anything else that matters is independent of or at least orthogonal to presidential party. But he obviously doesn't think this is the case (nor should he). So how many degrees of freedom does he think he has here? Does he think you can get reasonable results from 11 data points and 5 regressors?
Actually, over a period of a few years, I ran all sorts of tests (some assumed iid, some were non-parametric, etc.) using all sorts of tools (OLS, MLE, systems of equations, impulse response functions, etc.) on all sorts of data sets (real gdp per capita, real wealth per capita, etc.) at all different levels of granularity (length of administration, yearly, quarterly) going back a number of different years. Sometimes I looked at D and R alone, sometimes other factors were taken into account. Most of those are presented in one or another blog post.
I know $#%^ well that when you get an interesting and unexpected (and when I started, it was to check on some stuff that I read from such luminaries as Thomas Sowell and the folks at Heritage) you should check it again from multiple angles. Beat it to death so to speak. So that's been my hobby.
As to the fact that Nixon looks like a D or Clinton like an R, I've been trying to point out at least since 2007 that there is one thing that distinguishes every single R from every single D at least going back as far as I had data - the tax burden - that is, the share of people's income they actually paid in taxes. Every single D, whether they cut marginal rates or raised them, increased the tax burden over the length of their administration. Every single R, whether they cut marginal rates or raised them, decreased the tax burden.
Its very easy to construct a story about how tax burden affects economic growth. Republicans and libertarians have been telling such a story for a very long time. Of course, the story they tell doesn't quite fit the facts when you throw in growth rates, and they don't like the story lines that do fit the facts.
But that's not my fault. I am telling it like I see it. If that makes me a mad man, so be it.
Answer the question cactus. What evidence do you have that your regression is valid and statistically significant? What statistical technique are you using that does not require IID samples?
Also, if you believe that tax burden is the explanatory factor, then why not do that regression? And for that matter, do that regression without dividing time into 4 year chunks, just do that regression directly.
I have to agree with Ninja Zombie here (as much as I don't like to agree with someone who chooses Ninja Zombie as their nick).
If you believe that the difference between Rs and Ds comes down to one single aspect of policy, you might as well test for it directly. Maintaining the emphasis in the political color of the administrations doesn't make sense under this scenario, and it has the disadvantage of exacerbating partisanship bias.
"What evidence do you have that your regression is valid and statistically significant? What statistical technique are you using that does not require IID samples?"
That was stated in the comment you're responding to:
"some were non-parametric"
If you're aware of non-parametric tests that require some sort of assumption about the underlying distribution of the sample, I think you're aware of something no statistician has seen before.
"Also, if you believe that tax burden is the explanatory factor, then why not do that regression? "
I have. And I've posted on it. So now I'm writing a book because the same questions come up over and over.
Ok, so can you provide links?
Nonparametric estimators may solve the IID problem, but I'd be really surprised to see nonparametric estimators working effectively with 11 data points (or fewer, for a large number of the particular confounding factors you claim to refute).
As for tax burden explaining economic growth, I'm genuinely interested to see this. It seems quite surprising that such a simple predictor exists, and has gone undiscovered for so many years.
Bush didn't really act like a Republican on domestic policy and spending. The "No True Scotsman" fallacy is great for cherry-picking your inputs. We could compare economic performance to the party of the President and rearrange all the Presidents by what they "really" were and probably get any result we wanted.
A useful comparison would include the makeup of Congress and account for the time lag for many policies to affect the economy. And it would have to account for international factors, like the economic performance of other countries and world trade. And wars. And the fact that the parties have changed over time since 1929. And solar flares.
What are these policies that neatly define Democrats to exclude only the ones who happen to have crappy growth? On what metric does Barack Obama register as farther to the right than Bill Clinton? Because from what I remember of the 1990s, I spent most of the decade listening to my genuinely left-wing friends weep that he'd betrayed them. Remember Edelman's resigning in protest of welfare reform?
Exactly. I've written about Cactus pushing this meme before:
http://togetrichisglorious.blogspot.com/2008/10/presidents-and-growth.html
Clinton pushed welfare reform, telecom reform, banking deregulation, NAFTA, the Uruguay Round conclusion and a capital gains tax cut. Dubya gave us a new Medicare entitlement. Carter pushed for deregulation, vetoed a public works bill and called congressional tax committees "ravenous wolves." George HW Bush raised taxes and signed the clean air act and americans with disability act.
If they can toss out Obama for not being a genuine Democrat then I want to do the same with the Bush family.
"Clinton pushed welfare reform"
That's not quite fair, is it? Clinton vetoed welfare reform twice, and let it through the third time because his advisers said that it would hurt his poll numbers to veto it a third time. The fact that he later touted welfare reform as one of his biggest accomplishments just shows how good he is at spinning.
Cactus has had a couple of problems that he never resolved well. First almost all of his statistics are driven by one data point: Clinton. If you don't have Clinton the difference between the two becomes very small.
So in order to have the worth-talking-about difference, you have to closely examine Clinton. The problem for cactus is that the only plausible causal explanation would be through policies. Clinton's economic policies were not run-of-the-mill Democratic party policies. His most important economic decisions were to push NAFTA (not traditional Democratic at all), push welfare reform (traditional Republican reform), and to put Greenspan at the Fed (not a popular thing to do in retrospect for a huge number of Democrats who want to blame him for current woes). So as far as policy goes, Clinton looks more like a traditional Republican than a traditional Democrat.
Now if you want to appeal magic (that the 'D' on the television screen makes things better) cactus has a point, or at least he did before Obama. But at that point you might as well stick to the Super Bowl effect for predicting the economy, which happens to have a tighter correlation that presidential party.
The problem is that the whole argument is based on a huge fallacy. The President doesn't control the economy. He really doesn't. And in good times (Clinton for example) it is probably best for the president to step way back and let things operate in large part without meddling at all. (See the stupidity that was much of Nixon's economic policy).
SH,
Long time no hear from you.
"Cactus has had a couple of problems that he never resolved well. First almost all of his statistics are driven by one data point: Clinton. If you don't have Clinton the difference between the two becomes very small."
How can that be if economic growth was faster under FDR, JFK, and LBJ than under Clinton?
Why leave out Calvin Coolidge, Warren Harding, Woodrow Wilson, Theodore Roosevelt . . . the postwar economy is the standard measure, not least because FDR got twice as many terms as everyone else to average out the bad growth years. Throw out Clinton, and your sample reads LBJ/Kennedy, Reagan, Carter, Ike, Nixon/Ford, Bush, Bush, Truman.
And the Award for Most Blatant Misinterpretation of Data goes to...
This comment has already accumulated its share of bad arguments, but none of them nearly as bad as saying that Clinton is the one driving the Democratic statistics up. I believe Sebastian H deserves some sort of price for this.
I'm holding the award for a few more hours, in case somebody shows up saying that Reagan is pulling the Republican statistics down, or that if we exclude Truman the Democrat advantage vanishes.
"Obama has not yet behaved like a Democrat; his policies on the bail-out have been essentially the same as those of his predecessor who most certainly was no Democrat."
What a bizarre argument. He's essentially using group amalgamation and the fact that most presidents aren't uniformly either to obfuscate how Obama is in fact a standard economic liberal. Many of Bush's economic policies (healthcare spending, education spending, industry bailouts, and stimulus spending) were in fact much closer to liberal than conservative ideal, and those are the items Obbama has continued. The major Bush policy Obama isn't continuing is tax cuts, but that's the one that proves Bush isn't a Democrat. So in fact every one of these items shows Obama is behaving as a liberal. If anything this shows how unconservative Bush was.
If you track the system to monetary policy you don't need any epicycles to explain economic performance.
It's just the author's skull would probably split if he ever conceived of economic policy being driven not by D or R.
Even if we consider only these two variables - economic growth and who is in the White House - there's still no way to determine causality, since
1) there are cycles in economic growth, and
2) economic conditions may affect who is put in the White House.
For example, perhaps people are more likely to vote for a Democrat when economic growth is low, and perhaps the economy then tends to recover regardless of who actually wins the election. This would generate a pattern of higher growth under Democrats even if the President's economic policy never makes any difference at all.
No Sebastian H , the Democratic record is not driven by Clinton. Under JFK-LBJ growth averaged 4.9% and even under Carter it was 3.3%, better than every Republican but Reagan.
average
real gdp
growth
(%)
JFK-LBJ........ 4.9
CLINTON ........3.7
REAGAN ........3.4
CARTER......... 3.3
IKE............ 3.0
NIXON-FORD..... 2.8
BUSH I......... 2.1
BUSH II........ 2.1
TRUMAN......... 1.3
Truman had 3 years at the end of WWII when growth averaged -4.4% and five years of +4.8% growth.
But to address Meagan's argument. I could have made exactly the same argument she is making now about Reagan. When Reagan took office the unemployment rate was 7.5% and it peaked at 10.8% in December 1982, almost two years into the Reagan administration. In the first two years of the Reagan administration growth averaged about 0.5%.
In other words at this point in the Reagan administration the economic record was horrible. Anyone could have written exactly the same story Meagan just did about Obama. Yet, Republicans now make a minor deity of Reagan and worship at the alter of his economic policies. Her jumping to the conclusion that Obama has failed in retrospect could look just as foolish as an article about the failure of Reaganomics written in the first years of his presidency. The odds are very strong that 3 or 4 years from now Meagan will be very, very embarrassed when anyone brings up this article.
Economic policy is simple. Republicans try to use the levers of government to transfer income up the income scale and democrats try to use the same levers to transfer income down the income scale.
Despite all the rhetoric about "supply-side" economics the economic record is overwhelming that "trickle-down" economics does not work very well. Even the supposed beneficiaries of trickle-down economics, the top percentile of the income distribution, do no better than they do under Democrats, and everybody else does significantly worse.
The really ridiculous Republican policy is the "starve the beast" strategy. Deficits do not harm the federal government. Rather they harm the private sector. Starve the beast is a strategy of cutting off your own nose to spite your face.
"Economic policy is simple. Republicans try to use the levers of government to transfer income up the income scale and democrats try to use the same levers to transfer income down the income scale."
This is certainly what Democrats believe. But the majority of Republicans believe government shouldn't do either. Instead it should set the base rules that foster the best economic growth. But if you're a Democrat believing Republicans are out to screw the poor sure justifies screwing them first.
"Economic policy is simple. Republicans try to use the levers of government to transfer income up the income scale and democrats try to use the same levers to transfer income down the income scale."
I'd say that it's your view of economics that's simple. You seem to think that it's all about taking money from one group to give to another, as if there's a fixed, exogenous supply of wealth that simply appears (from the sky?). Who creates the wealth to begin with?
MJ-- You may believe what you want to believe, but the evidence is overwhelming against your belief.
Ann -- who creates the wealth? Everyone, capital alone can not create anything and labor alone can create very little.
spencer,
Actually, the evidence against my belief mostly consists of the fevered imaginings of people who have no idea how human systems work. But if you think you have some be sure to highlight it.
mj,
If that's the case, shouldn't there be more, hum, economic growth under Republican presidents?
Gotta love this.
mj, like any good fundamentalist, starts with the conclusion: he knows how "human systems" work. So, not only does he not feel any necessity to enumerate some facts in support for his case (because it's all so obvious), he is essentially impervious to any evidence that contradicts his dogma. The quality of evidence is judged by the degree in which it supports his initial conclusion, not the other way around.
This was a pretty good post. Megan builds a good case against this sort of analysis: few data points, lack of a plausible causal mechanism for the observed differences (though cactus advanced differential tax burdens upthread and, apparently, in previous posts in Angry Bear), a very blurry line dividing Democratic and Republican economic policies in the last 80 years.
But what these data do very well is show just how untenable it is to make the reflexive assumption that Republican presidents implement better economic policies.
People who continue to insist on this point are not to be seen as stupid; they are showing signs of mental illness.
Nimed,
Anyone who considers the economic growth during the period of presidency as evidence of anything is a fool since
(a) economies are vastly more effected by events other than politics,
(b) the lag between policy and effect is both significant and variable,
(c) the effect of most policies is dependent on the circumstances in which it is enacted, and
(d) economic policies are not consistent between presidents of the same party, among other weaknesses.
I don't need to explain these things to educated, intellectually honest readers, they're obvious. Nor is it my responsibility to educate those who fail either of these standards in, of all things, a blog comment. But I see how you can be confused as this data seems so compelling among those who (a) have no idea what they are talking about, and (b) have only a political agenda.
If you want to learn how human systems work, here's a convenient place to start. http://www.gmu.edu/ They have many of these wonderful places to learn around the country.
Further, I cannot help with your inability to read correctly, as when you say this: "But what these data do very well is show just how untenable it is to make the reflexive assumption that Republican presidents implement better economic policies."
I did not support the idea that Republic presidents implement better economic polcies. I said that Republicans generally do not agree with using "the levers of government to transfer income up the income scale", as held by spencer.
Republicans still have to explain why Republicans generally do so much worse when it comes to real GDP per capita growth rates. The BEA has only been calculating real GDP per capita since 1929, and yet the best performing Republican, Reagan, is beat out by four Democrats, and all but one of the entire bottom half of the sample is taken up by Republicans
Bull. Republicans have to explain no such thing, because it is not true.
Everybody who's taken an economics course, or has an ounce of common sense, knows that economic policy takes effect with a lag.
Put in the standard textbook lag for policy of 18 months (six months from inauguration to enact policy, then a year for it to take effect -- when is Krugman saying we should expect all those new jobs from the Obama stimulus??) and the supposedy inferior Republican economies totally disappear.
In fact, Republican-president economies become slightly better, real growth of 14.7% per term for Republicans versus 13.9% for Democrats, with Republicans having five of the seven best growth terms.
What kind of ninny think Presidents determine economic results from the first calendar day of the year they take office (starting 20 days before they are inaugurated?). And that the results of their policies end on the very day they leave office?
The kind of ninny who wants to claim "Republican presidents produce weaker economies", that's the kind!
(Although some pretty prominent economists and pundits have made this claim -- so it shows how far some pretty smart people will ninnify themselves for partisanship.)
Now of course if we want to get realistic and accurate, and leave the polticial rhetoric-ninny arguments behind, we'll realize that "presidential" GDP growth rates are determine by the business cycle and the stage that each president steps into it, and out of it, which no president can control.
For instance Reagan stepped into a recession worse than today's (so far) that was engineered by the Fed to break the double-digit inflation left by Carter. So Reagan & the Repubs are supposed to get punished by the "GDP scorers" for paying the price of fixing the big problem Carter left??
Reagan, Nixon and Bush II all took office with recessions running ... Johnson and Clinton stepped out at the very top and end of historic booms, leaving recessions dropping on their successors.
And there's only been seven swings between the parties in office since WWII, which is a very small sample size.
That's the whole story: A small sample of presidents stepping into and out of the business cycle, at random points in it -- the full explanation of everything. Chance. (Now, for the first time, a Democrat has stepped in during a bad year. Chance evens out.)
Not to mention Clinton had a Republican Congress, the Republicans mostly had Democrtic Congresses -- how come only the President's party affiliation matters?
Reality: Presidents have very little effect on the business cycle, politicians and pundits only want to endlessly say they do to claim credit and lay blame for partisan reasons in the next election.
History: The Democrats tried to get Herbert Hoover to run as their candidate in 1928. (Even FDR was for it!) If he had, and had won, does anyone think the Depression wouldn't have occurred because Herbert Hoover the Democrat was residing in the White House?
Sure there's a lag in economic measures. But
18 months for monetary policy? I would sure like to see the quote from the textbook he is talking about. "Longer for fiscal policy"? What about tax breaks? How come they take 18 months to have an effect?
Isn't the behavior of economic agents supposed to be immediately altered just by announcing such measures?
And let's not forget that people around here are saying that the slowest type of fiscal policy - government spending - should have produced noticeable results after as little as 4 months.
The President does not control the economy. This is not the Soviet Union. Why is this difficult to understand?