Blinded by partisanship

by Russ Roberts on September 3, 2008

in Data, Politics

When I first took economics, I learned from my textbook (Samuelson) the fallacy of post hoc, ergo propter hoc. After this, therefore because of this. Alan Blinder commits this fallacy in this New York Times article (HT: James Gambrell).

After noting that Democrats and Republicans have different economic policies, Blinder argues, drawing on work of Larry Bartels in Unequal Democracy, that Democrats run the economy better than Republicans:

Data for the whole period from 1948 to 2007, during which
Republicans occupied the White House for 34 years and Democrats for 26,
show average annual growth of real gross national product  of 1.64 percent per capita under Republican presidents versus 2.78 percent under Democrats.

That 1.14-point difference, if maintained for eight years, would yield
9.33 percent more income per person, which is a lot more than almost
anyone can expect from a tax cut.

Blinder is aware of the fact that the President doesn’t run the economy. He adds:

Such a large historical gap in economic performance between the two
parties is rather surprising, because presidents have limited leverage
over the nation’s economy. Most economists will tell you that Federal
Reserve policy and oil prices, to name just two influences, are far
more powerful than fiscal policy.

Most economists will also tell you that Presidents don’t even control fiscal policy. Here in the United States we have three branches of government. The Presidency is one of them. Then there is what is called the Congress. They have a say, too. Then there are external events besides fiscal policy and monetary policy and oil prices that affect the economy and that are beyond the President’s control. Demographics. Cultural trends. Technology. None of these are controlled by the President. There’s also war. You can argue the President controls whether we go to war, but I’d argue you’d want to factor in war separately if you want to assess the quality of a President’s economic policies.

Blinder does admit that the future may not be like the past:

Furthermore, as those mutual fund
prospectuses constantly warn us, past results are no guarantee of
future performance.

But then he reassures the reader with this jaw-dropping sentence:

But statistical regularities, like facts, are stubborn things. You bet against them at your peril.

What? Statistical regularities are stubborn things? No they’re not. They are dominated by randomness almost by definition. That’s why they’re called regularities. They reveal a pattern in the data. Nothing more. Nothing less. Without a theory (and saying Democratic presidents are better at running the economy than Republican presidents is not a theory) it’s just a regularity.

And what does that last sentence mean? Bet against the statistical regularities of the past at your peril? So he’s saying that if McCain is elected, you can be pretty sure that he’ll do a worse job than Obama would have. You can bet on it. It’s a result you can rely on.

That is going to be a difficult bet to enforce. I know it’s an idiomatic express to mean it’s practically a sure thing. But thinking about the impossibility of determining the winner of such a bet makes it clear that the whole argument is meaningless.

Be Sociable, Share!

Comments

comments

34 comments    Share Share    Print    Email

{ 34 comments }

Dave September 3, 2008 at 4:17 pm

Any economist that aligns him or herself with a party cannot possibly be taken seriously. Krugman falls into this category as well.

Ike September 3, 2008 at 4:18 pm

Can anyone quickly replicate his study to see how ownership of the House of Representatives factors in?

Jay September 3, 2008 at 4:27 pm

Last time I checked there is a strong correlation between U.S. defense spending and the life expectancy of South Africans.

I propose anyone who wants to reduce the U.S. defense budget authorize the immediate nuking of South Africa. Their life expectancy would plummet causing our defense budget to follow suit.

Statistical regularities, like facts, are stubborn things. You bet against them at your peril.

chris September 3, 2008 at 4:39 pm

There's no doubt that Blinder is engaging in some ideological semantics–what does that say?–but there was a study done by Jeffrey Frankel at Harvard a few years ago that showed that the economy performed better when a Democrat was in the White House than a Republican. He summarizes it here.

Chris O'Leary September 3, 2008 at 4:48 pm

From the stupid question department…

Is the Federal Reserve, with its control over interest rates and the money supply, powerful enough to rate as a fourth, potentially non-constitutional (unconstitutional?) branch of government?

If anyone has "control" over the economy, isn't it the Fed?

Russ Roberts September 3, 2008 at 5:09 pm

chris,

The article you cite has nothing to do with performance per se. Rather it argues that Republicans are more likely in recent years to run deficits and to be less supportive of free trade relative to Democrats. In the first case, Frankel (who served under Clinton's Council of Economic Advisors) ignores (as Blinder does) any role of Congress. In the second, he extols the comparison with Bill Clinton, conveniently forgetting Clinton's refusal to enforce the trucking component of NAFTA.

diz September 3, 2008 at 5:48 pm

If the Democrats would credibly promise to cut spending to Clinton's 2000 levels (plus inflation), they'd probably get my vote. I think this would mean a budget of around 2.4 trillion instead of the 3.0 trillion we have.

But this doesn't seem to be their plan…

The Albatross September 3, 2008 at 6:23 pm

“When I first took economics, I learned from my textbook (Samuelson) the fallacy of post hoc, ergo propter hoc.”

I seem to remember Professor Blinder saying something similar my freshman year.

Methinks September 3, 2008 at 8:24 pm

You can argue the President controls whether we go to war…

Well, I guess you could sort of. On the other hand, if somebody declares war on us and decides to attack, then that's not the president's choice anymore. Also, it's the congress, not the president, that decides whether we go to war or not. Just another item that is not always in the president's control.

anyway..

It's interesting that Blinder chooses to go the route of political hack and discredit himself by insulting people's intelligence with such nonsense. He is one of the relatively few people who can explain the pros and cons of the different proposed economic policies rather than shilling for a particular party. I'm left to believe that the Dem's proposed policies are so terrible that he doesn't dare comment on them.

Flash Gordon September 3, 2008 at 9:07 pm

Since the R's often comes in after the D's have been in power, and vice versa, couldn't it be argued that R's policies are benefiting the D's because of the lag time for those policies to produce measurable results in economic indicators.

And likewise D polices hurt R's who have to spend their time in power living with the consequences of the D's policies and correcting their errors.

For example, might not Bill Clinton have benefited from Reagan tax cuts, ending runaway inflation and lower interest rates. :Likewise, Reagan had to deal with the Carter legacy.

The Albatross September 3, 2008 at 9:36 pm

I always love these simplistic comparisons, but I am not surprised as Professor Blinder taught me that everything could be solved by a simple expansion of the money supply. However, let us address this point with slightly less simplicity, which is still simple (and thusly inadequate). I would classify Eisenhower as sleepy, but illiberal, as he did little but allow the economy to work out the huge disruptions caused by the New Deal and World War II (hence the relatively slow economic growth in the 50s). I would classify Kennedy as more liberal, as he fathered the round of trade liberalization, which bears his name, and the reduction in tax rates that would be passed in his memory. He still did nothing to reduce the 10 percent of GDP committed to national defense (now about half). True, Johnson passed the massive expansion of state power that we now know as “The Great Society”, but it was the illiberal Emperor Nixon that filled it out with an equally massive expansion of government regulation. Furthermore, it was Nixon that pressured the Federal Reserve to inflate the economy to ensure his re-election. Ford mostly followed in this path. However, it was President Carter who would preside over the first deregulations (airlines, among others), which was shepherded through the Senate by Edward Kennedy. I will further credit Carter with appointing Paul Volcker to the Federal Reserve. However, the latter could not have accomplished his goal of destroying inflation at 22.5 percent interest rates without the support of President Reagan. I believe that President Clinton largely continued the “liberal” policies of Reagan (with a little help from a hostile congress), but (like the Labour government in Britain), he piggybacked on the reforms of the previous decade, whose prosperity will be enjoyed by whoever succeeds President Bush. Again, it is a matter of policy. Some Democrats were more relatively more “free market” than Republicans, who were likewise more “free market” than some Democrats. Prosperity is a matter of policy (or at least not screwing it up), more than party. Over the last 50 years, Ill-liberalism (and its opposite) appears a bi-partisan issue.

gappy September 4, 2008 at 12:05 am

I was surprised by the article too. The author is reputable but the thinking is magical. In general, Larry Bartels' book has received a lot of attention: Krugman, DeLong, Rodrick among others. Has anyone read it here? I would like to read it at some point, but is not high on my list.

Lee Kelly September 4, 2008 at 12:44 am

Without a theory (and saying Democratic presidents are better at running the economy than Republican presidents is not a theory) it's just a regularity. – Russell Roberts

That 'democratic presidents are better at running the economy than republican presidents' is 'just a regularity'? Or is the regularity that people say so?

Whatever the case, the statement 'democratic presidents are better at running the economy than republican presidents' is a testable theory, albeit false. The evidence which Alan Blinder provides is consistent with his theory, but perhaps Alan Blinder should have spent more time searching for evidence which challenges his theory — he would not need to search very far to find it.

Ryan September 4, 2008 at 1:07 am

Correlation doesn’t equal Causation!

I think one of the most common errors in logic today can be answered by that statement. Maybe you can buy some billboard space that says this in big, bold letters.

Lee Kelly September 4, 2008 at 1:40 am

If there are no exceptions then correlation does equal causation.

Most people who are accused of committing this fallacy are, in fact, proposing a correlation without exceptions i.e. a universal hypothesis. If they are guilty of anything then it is usually verification bias, since it is easy to find finding "verifications" for almost any hypothesis. Those who are interested in learning, and without any ulterior motives, would do better to search for falsifications.

Ike September 4, 2008 at 2:18 am

Actually Lee, correlation is not causation. Period. Two events may universally share a common origin, yet there is no means to assign a "cause" to one or the other.

The problem outlined above is one of extrapolation and interpolation. With such a small dataset, and so many other potential contributing factors, Blinder just don't know enough to come to the conclusion.

I'll go further and state that predictive modeling is not enough to carry causation. The theory of Epicycles made for great predictions about planetary observation, yet was useless in establishing causality.

Your call to search for falsifications is off the mark as well – the dataset is too small. It would be more productive to find other likely factors that show an even stronger degree of correlation.

Lee Kelly September 4, 2008 at 2:56 am

Ike,

There is always a means to assign a causal relation between two events: imagination. That is all it takes, the ability to imagine that one event causes another and then hypothesise as such. The resulting theory will be a universal proposition, implying a relation between two events with no possible exceptions.

Such a theory is unverifiable, ruling over an infinite domain, but it is not necessarily unfalsifiable. What is not relevent is the small dataset or Blinder's prior knowledge, neither of which has anything to do with the theory's truth or falsity, which presumably ought to be our main concern.

There are, no doubt, many variables involved here, and Blinder's treatment of the topic is notably 'simplistic'. But what is important is not that some other variable might contradict Blinder's theory, but whether it actually does. In other words, it is not a negative mark against a theory that it might be mistaken

Unfortunately, the root of our disagreement can likely be traced to deep metaphysical presuppositions. I would love to discuss them with you, but that is a very tall order and this is the wrong place.

BoscoH September 4, 2008 at 3:17 am

Funny title for this post! Alan Blinder has what's called an aptonym.

Orlando Armaswalker September 4, 2008 at 9:33 am

The data quoted is extremely flawed. I would gather that it would look differently if you compared Republican vs Democratic Congresses, in fact, it would be reversed; something like 3.75% average GDP per capita for Republican Congress and 1.5% for Democratic Congress. It's not the executive who has governmental influence on the economy, it's the lawmaking and size of government powers of the Congress: irrespective of Senatorial control.

Gary September 4, 2008 at 9:41 am

Flash Gordon makes the point that perhaps administrations derive benefit (or detriment) from their predecessors. I've often considered this. My ability to look back is limited, since I only barely remember Bush I being in office, but I do remember discussing in the fall of 2000 (I was in high school) how whoever won the election would most probably be a one termer, since the economy was bound to correct in their administration.

The market did crash in September 2001… but the circumstances actually helped the President to re-election.

While the Bush II administration is charged with this poor market performance, its not as if it was his fault. Had the Clinton administration done some things differently airline stocks might not have gotten crushed. Had the Fed raised rates in a more aggressive manner in the late 1990s, the run up (and resulting fall) might have been less dramatic.

Dallas Weaver September 4, 2008 at 11:18 am

This type of simple minded analysis makes the assumptions that there is no time delays in the system. For example, in the 60's , Kenedy and Johnson printed money to pay for the war and we had stagflation (inflation + slow growth) in the 70's and Johnsons killing of R&D in the physical sciences about 1968 make Japans success in the electronic/solid state physics areas in the 80's inevitable. Time delays measured in decades are common in economics/technology so the other party gets credit/blame for what was set up by the previous party.

Tim September 4, 2008 at 12:58 pm

An alternative theory: We elect Democrats when our economy appears strong, because we can afford to. In the same way that minimum wage correlates with good economic times because in good times we can afford more of a drag on the economy.

Or it could be that the sample size of Democratic presidents is too small for much of a confidence interval. Does Blinder give us any econometric reason to believe this is not the case?

Ike September 4, 2008 at 1:10 pm

I actually thought the same thing, Tim. What if there IS a solid "universal proposition," but that Blinder mistook chickens for eggs.

DKH September 4, 2008 at 1:57 pm

Whenever I see the argument that Mr. Blinder is making, I always want to make Flash Gordon's point in response. The economy in whole (as measured by GDP) is just too complex a system for Mr. Blinder's simplistic point.

Dan September 4, 2008 at 2:22 pm

Flash Gordon has a point. The article says nothing about the conditions of the economy when a president takes office. Every Republican who took over for a Democrat came into office with the U.S. either in a major war (Eisenhower-Korea and Nixon-Vietnam) or on the brink of a recession (Reagan and Bush 43). When a Democrat took over for a Republican (Kennedy, Carter, and Clinton) the U.S. was at peace and with a growing economy. Clinton benefited greatly because he took over just as the Cold War had ended.

As well, the analysis begins in 1948, so the large drop in GDP after WWII is not included in Truman's presidency, but the corresponding drops in GDP at the end of the Korean and Vietnam Wars are included, which conveniently come under Republican presidents.

In addition, Presidents Nixon and Ford could hardly be called believers in free-market economics. "We are all Keynesians now" -Nixon. Two Democrats, Johnson and Clinton signed into law major supply-side tax cuts.

Instead of comparing growth rates of presidents by party, you should compare them based on their economic policies.

jorod September 4, 2008 at 5:07 pm

The Republicans usually come in after the Democrats have screwed things up.

Mesa Econoguy September 4, 2008 at 6:42 pm

Blinder’s argument is largely superficial and economically meaningless:

Are we to assume, as it would seem, that economic outcomes are due solely to the party affiliation of the president? What about what the presidents actually do (Kennedy cut taxes, Bush I raised them)? Doens't congress figure into this at all? And are the results supposed to be instantaneous — is a president responsible for economic outcomes from the very moment he is inaugurated?

[Don Luskin]

Al, you get an F.

Chuck E September 4, 2008 at 7:17 pm

The data quoted is extremely flawed. I would gather that it would look differently if you compared Republican vs Democratic Congresses, in fact, it would be reversed; something like 3.75% average GDP per capita for Republican Congress and 1.5% for Democratic Congress. It's not the executive who has governmental influence on the economy, it's the lawmaking and size of government powers of the Congress: irrespective of Senatorial control.

Orlando, was this an actual study that was done? Or did you do a quick & dirty analysis? Either way, I'd be very interested in reading about it.

I also like Dave's comment that economists should just avoid aligning themselves with political parties. For one thing, people's brains shut off once a party or candidate is mentioned in the middle of an argument.

This actually shaped up to be a very interesting thread. Thanks to everyone who's posted.

Chuck E September 4, 2008 at 7:17 pm

The data quoted is extremely flawed. I would gather that it would look differently if you compared Republican vs Democratic Congresses, in fact, it would be reversed; something like 3.75% average GDP per capita for Republican Congress and 1.5% for Democratic Congress. It's not the executive who has governmental influence on the economy, it's the lawmaking and size of government powers of the Congress: irrespective of Senatorial control.

Orlando, was this an actual study that was done? Or did you do a quick & dirty analysis? Either way, I'd be very interested in reading about it.

I also like Dave's comment that economists should just avoid aligning themselves with political parties. For one thing, people's brains shut off once a party or candidate is mentioned in the middle of an argument.

This actually shaped up to be a very interesting thread. Thanks to everyone who's posted.

Mesa Econoguy September 4, 2008 at 7:21 pm

And Blinder definitely fails on this point alone:

Data for the whole period from 1948 to 2007, during which Republicans occupied the White House for 34 years and Democrats for 26, show average annual growth of real gross national product of 1.64 percent per capita under Republican presidents versus 2.78 percent under Democrats.

[Blinder]

For instance, if you lag the per capita GDP results by two years — not a bad guess of how long it takes policy to show up in output — the difference between Democratic and Republican presidencies vanishes (both are 2.2%). And if we look at it terms of government control — that is, party control of both the presidency and both houses of congress — then the Democrats do worse than the Republicans (2.0% versus 2.1%). But wait! Divided government, when no party controls both the presidency and the congress, does even better: 2.3%!

[Don Luskin again]

Please re-learn policy implementation lag, Mr. Blinder.

floccina September 5, 2008 at 3:58 pm

Considering the complexity of the economy, you would probably need tens of thousands of years of data to have a large enough data set to make that call.

The Dirty Mac September 5, 2008 at 10:29 pm

Thanks to Blinder, I now know that if we had elected Strom Thurmond (D-SC), we wouldn't have had all these problems.

Sam September 7, 2008 at 9:22 pm

This blogger clearly didn't read the book. He is posting in response to this Blinder guy. If he read the book, he would know that Bartel's empirical models control for behavior by the Fed and oil price fluctuation. Second, he would know that even though Bartel's has a weak theory, it is not simply stating that Democrats are better at running the economy than Republicans. It is a bit more nuanced and logical than that. My sense is that this Russell fella is a bit blinded by his own partisanship in his need to disqualify this research based on a second hand accounting of it. It just makes the guy sound stupid. If this guy wants people to take him seriously he should read the literature that he criticizes before bashing it.

kerrin October 8, 2008 at 8:36 pm

After noting that Democrats and Republicans have different economic policies

They are actually quite similar. Spend, spend, and spend. The only difference is where they spend and how much they tax or borrow.

Most economists will also tell you that Presidents don't even control fiscal policy.

Perhaps not directly, but they appoint those that do (e.g. Ben Bernanke).

Presidents along with the legislative branch do have an effect on the value of the dollar, which does have an effect on the health of the economy. Executive and legislative branches contribute with budgets, spending, and fed appointees who make monitory policy.

Consider this:
A presidential appointee to the federal reserve, Ben Bernanke, decides to lower interest rates to "stave off inflation." When the fed reduces interest rates this reduces the value of the dollar. Why? Because U.S. Treasury bonds have a lower yield when interest rates are low. Lenders (largely foreign) are then less willing to buy Treasury Bonds at these lower rates because their return is lower. Plus the $10+ Trillion in federal debt makes them even less comfortable to buy (lend) more.

When foreign countries are not buying Treasury Bonds (lending) the fed must get money from somewhere for the spending. So they essentially have more money "created" or put into the system. The more money in the system the more the value of the dollar falls.

As the dollar falls the price of imported goods increases (we import a sh** load of stuff). When things cost more people buy less and their devalued dollar buys even less imported goods (i.e. inflation). There you have it, a worse economy is the result. Presidents, their fed appointees, and legislators have a big role in this current mess.

Data for the whole period from 1948 to 2007, during which Republicans occupied… and Democrats…show average annual growth of real gross national product

Gross National Product (GNP) is one way of measuring the health of an economy but is less accurate now with nations increasing their production in other nations (i.e. "overseas"). A more popular measure these days is Gross Domestic Product (GDP). But even GDP is not an accurate measure of the health of an economy.

Economist Frank Shostak:

The GDP framework cannot tell us whether final goods and services that were produced during a particular period of time are a reflection of real wealth expansion, or a reflection of capital consumption. For instance, if a government embarks on the building of a pyramid, which adds absolutely nothing to the well-being of individuals, the GDP framework will regard this as economic growth. In reality, however, the building of the pyramid will divert real funding from wealth-generating activities, thereby stifling the production of wealth.

Here's a musing/question for you "right wingers":
As a reflection on GDP are U.S. wars similar to building pyramids?

Or for "the lefties":
As a reflection on GDP would social programs be similar to building pyramids?

Previous post:

Next post: