Monday, January 26, 2009

Can we get a little objectivity, please?

Is economic analysis driven by science or by ideology? That seems to be a hot topic this week.

In the New York Times today, Paul Krugman claims that many anti-stimulus arguments are not made in "good faith":
"Some of these arguments are obvious cheap shots. John Boehner, the House minority leader, has already made headlines with one such shot: looking at an $825 billion plan to rebuild infrastructure, sustain essential services and more, he derided a minor provision that would expand Medicaid family-planning services — and called it a plan to 'spend hundreds of millions of dollars on contraceptives.'"
On the opposite end of the ideological spectrum, Russ Roberts argues that economists' positions on the stimulus have less to do with their reading of the empirical data than their feelings about markets in general:
"As far as I know, no prominent market-oriented economist has come out in favor of a trillion dollar increase in government spending as a way to improve the economy. Every market-skeptical economist that I have heard is in favor of it on the grounds that it will improve the economy. Each side claims to have empirical support for its position."*
Additionally, the focus of this week's episode of EconTalk, Robert's weekly economics podcast, is the role of empirical evidence in economics.

How pervasive is the issue of bias within economics? Nate Silver over at FiveThirtyEight.com has some evidence on the subject. Looking at the Wall Street Journal's monthly economic forecasting survey, Silver produces the following chart of growth forcasts by economists in different sectors, including finance, consulting and academia:



It's a small sample (55 economists), but it illustrates the point. The most optimistic growth forecasts come from economists in the financial sector, who perhaps have institutional incentives pushing them toward optimism, while the most pessimistic forecasts come from academia. Typical liberal academics, some might say!

This is not entirely new. In 2006, David Lereah, former Chief Economist for the National Association of Realtors, wrote "Why the Real Estate Boom Will Not Go Bust - And How You Can Profit From It". This was hardly an unbiased work. An economist representing the interests of realtors has a strong incentive to promote the real estate industry. Of course, less then a year after the book was published, the market started to go south and... well, you know the rest.

It's entirely likely that ideology clouds the minds of even the best economists. But is it just economists who suffer from a lack of objectivity? A recent survey of scientists' views on global warming suggests not:
"In analyzing responses by sub-groups, Doran found that climatologists who are active in research showed the strongest consensus on the causes of global warming, with 97 percent agreeing humans play a role. Petroleum geologists and meteorologists were among the biggest doubters, with only 47 and 64 percent respectively believing in human involvement." (emphasis added)
Scientists and economists are people too. People are affected by biases in subtle and unconscious ways. This, however, does not mean that there is no objectivity out there. The peer review process, for example, is designed for just this purpose: to reanalyze and replicate results and to ferret out unsubstantiated claims. As long as everyone gets to look at the same data, we should be able to find some areas of common understanding.

That being said, it's wise to listen to more than one expert and more than one perspective.

____________________________________________________________
*It should be noted that Martin Feldstein, a prominent center-right economist at Harvard, has come out in favor of fiscal stimulus.

2 comments:

Gingi said...

Alas, idealogists on either side will continue to maintain their view regardless of the outcome. It is the nature of complex systems like the economy and the environment, compounded by our inability as a species to forecast the future. When the economy recovers eventually, the pro-stimulus camp will attribute it to an effective government bailout program at work; the anti-stimulus camp will argue that the economy could have recovered much more quickly without the package. Likewise, future temperature trends, no matter what they are, will reinforce the arguments by both sides: a warming planet in the face of large-scale mitigation efforts is evidence for the skeptics that human intervention is trumped by Mother Nature; conversely, a cooling planet would serve the AGW (anthropogenic global warming) argument that just as humans curbed the heating trend, so have they caused it in the first place.

Really, if you're an ideologist, you can't lose! And because the systems under examination are so complex, it is fairly easy to cherry pick data that supports your views.

Tim Dzurilla said...

In our econometrics class yesterday, our prof brought up the very issue.

His example was boiling water with a lid on. He said him and his buddies from a range of disciplines, including a chemist who said it was an engineering problem, boiled a lot of pots of water to figure out if it was faster to have the lid on or not.

Issues were with who decides what "boiling" means, how do you know if there is a lid on if the water is boiling, and everyone had a sound theory to support either prediction.

By the end of the day, it was inconclusive. What was concluded was that noone changed their mind whether or not having a lid on made water boil faster. Your perception going into an experiment effects your interpretation.

As Gingi says "You can't lose!" just as long as you're generating sound theories that are influencing policy in the best way you believe.