Numbersense Pros: Interview with Dean Baker

Today's Numbersense Pros features Dean Baker, an economist and co-director of the Center for Economic and Policy Research. His blog, Beat the Press, is my first stop for economics news. What attracts me is his data-first attitude. Every post on his site is about interpreting economic data--there is none of the data-free theorizing that is typical on other economics blogs.

Baker adopts a cynical, almost satirical, tone, which is fun for readers but perhaps unamusing to the targets of his scorn. To give you a flavor, one of his recent posts is titled "Larry Summers Thought the Stock Bubble was Cool and Missed the Housing Bubble". Another post on how to fix unemployment begins with "there are two types of people in the world: those who make complicated things simple, and those who make simple things complicated."

Here is the interview.

KF: How did you pick up your impressive statistical reasoning skills? What sets you apart from other economists?

DB:

I don't know that my skills are so impressive, but insofar as I do things differently from other economists I think it is the result of thinking more about the data, meaning how it is gathered and its underlying patterns, rather than trying to beat it up with complex statistical analysis. Certainly our econometric techniques have improved substantially over the last 2-3 decades, but I think it will be a rare that a new technique will allow us to recognize a trend or relationship that we couldn't see with more simple tools.

***

KF: What is your pet peeve with published data interpretations?

DB:

People often fail to consider the most basic issues about the data they are looking at. I recently saw a news article commenting on how the Case-Shiller data released in July showed that the housing market was not being hurt by the jump in interest rates at the end of May and beginning of June. In fact, the Case-Shiller data released in July told us nothing about the housing market in June. The data was a 3-month average for the period between March and May. Furthermore, the data is based on closings. Since it typically takes 4-8 weeks between a signed contract and a closing, most of the sales that went into the May data were contracted in the period from January to April.

There are many other cases I could cite like this in which people analyzing the data have not even thought for a minute about the nature of the data they are looking at. It really helps to know what the data are.

***

KF: Which source(s) do you turn to for reliable data analysis?

DB:

That one is not easy. There are plenty of economists whose work I respect and when I see them write on a topic I can be confident that they have serious things to say, but in terms of regular analysis of the economy, I can't say that there is anyone who I think is especially good.

KF: Thank you. You are indeed one of the few people who really know the economic data. We are very lucky that you blog.

***