Paul Krugman responds to both THE FORGOTTEN MAN and a Bloomberg column that I wrote recently. He’s contesting the smallest of details in a very general picture of the period — TFM is a panorama book, not a data book. Nonetheless, in the spirit of debate, it seems worthwhile to address his questions:
Dr Krugman’s first point is that TFM offers misleading statistics when it comes to unemployment. He actually uses the adjective “misleading.”
There are a number of data sets seeking to capture joblessness in that period. (Remember, they didn’t have all the tools we have today.) The annual data set which I used in THE FORGOTTEN MAN, and which Dr Krugman deems misleading, originates with the government Bureau of Labor Statistics and the great scholar of employment, Stanley Lebergott. Richard Vedder of Ohio University actually addressed/broke out employment month by month using the same data, so sometimes TFM cited him. Anyhow these data show average annual unemployment in the twenty percent range for a number of years – 1933, 1934, 1935. At points in 1937 or 1938, unemployment gets back to 20%, which I mention in TFM.
These “misleading” data, Lebergott/BLS, are widely used. How widely used? Even by the President-elect, apparently. Mr Obama, talking on Sixty Minutes spoke about unemployment in 1932 and 1933. Mr Obama said:
“Well keep in mind that 1932, 1933, the unemployment rate was 25 percent, inching up to 30 percent.”
That’s a BLS number. It’s also a Census Bureau number. Table D-1-10 “Labor Force and its Components, 1900 -1947” . Indeed it looks like the President-elect thinks even these Lebergott/BLS data are too conservative since he says “up to 30%” and the table I am citing does not go that high.
The data set that the President-elect did not cite is the one which Dr Krugman seems to be suggesting is preferable. (Dr Krugman never spells it out, he just says we’re being misleading.) The second set of numbers come from a useful paper by scholar Michael Darby. This set puts unemployment in 1932 and 1933, at 23 percent or so and at 21 percent. Eric Rauchway of UC/Davis likes this data too. It has somewhat lower unemployment numbers for the 1930s, and therefore makes the New Deal look better. Rauchway actually accuses me of “lying” because I don’t do use these numbers. That seems odd. We’re all just trying to figure out what happened, not get personal. None of us knows precisely. This is supposed to be a discussion. Here is the Rauchway post.
Dr Darby’s point, long story short, is that the make-work jobs of the New Deal were not counted in the BLS/Lebergott data. That is true and was intentional. Lebergott and the government believed that it was wrong to count short-term emergency jobs as employment because it was inhumane to pretend that a short-term project that might tide a family over was the same as “regular work.”
Makes sense to me. But suppose we accept Dr Darby’s definition, to which Dr Rauchway refers explicitly and to which Dr Krugman is probably gesturing.
What did I mean when I said that the Darby data painted a somewhat better picture of the New Deal economy than the government data? Just what I said: “somewhat.” The unemployment rates on the Darby chart, which I could find only in subscriber databases like jstor (Journal of Political Economy, Vol. 84, Number One “Three and a Half Million Jobs have been mislaid,” subscriber) range high as well. Unemployment is in the teens from 1931 through 1935, even hitting 20 percent. In 1936, a year in which the federal government spent more than it ever had as a share of the economy on peacetime projects, Washington managed to get the unemployment all the way down to 9.9 percent (Dr Darby’s “corrected BLS” number, Table 3). The next year it got near to 9.1. Then it was up again to 12.5 in the Depression within the Depression. Then 11.3. (Dr Darby has another set where things look slightly worse in some of these years)
In short, these Darby are not numbers to build the Obama presidency on. Today we consider our unemployment, in the six-seven percent range, a disaster. Even by the data that Drs Krugman and Rauchway choose, the Roosevelt Administration didn’t get to that level until 1941 (war buildup and war). President-elect Obama would run in the other direction if someone offered him a policy with that kind of outcome.
What else? Some people got tired of the “do we count the relief workers?” question. One was Lee Ohanian of UCLA. Dr Ohanian sidesteps the entire issue by counting hours worked overall. He summed it up in a note to me: “In 1939, total hours worked per working age person – including the hours of those on government payrolls – were 21 percent below its 1929 level. This is little recovery compared to hours at the trough of the Depression in 1933, when total hours per adult were 27 percent below its 1929 level. My recent work with Harold Cole indicates that hours worked should have been above normal during the mid-1930s, not 20 percent or more below normal.” This is another picture that also shows a scary amount of uenmployment.
Dr Krugman makes a second charge, that I misrepresent John Maynard Keynes by associating Lyndon Johnson’s Great Society with Keynes when the Great Society was a social and not an economic program. In 1964 Johnson pledged to build a Great Society with an emphasis on social improvement in his Great Society speech, just as Dr Krugman said. My point was that the political engine of the 1960s treated any spending, including Great Society spending, as a stimulus. Keynesianism defined the very lexicon of policy – that is why Milton Friedman said “we are all Keynesians now.” TIME even gave Keynes a cover: “The Keynesian Influence on the Expansionist Economy,” read the banner in the corner. Keynesianism was crucial window dressing for the Great Society show. Spending on all sides became permissible, and that only made sense if you cared less about deficits and more about growth – Keynesianism. Thus in July 1965, after many pieces of new legislation, The New York Times was writing headlines such as: “Johnson Policy Will seek to Prolong Boom: Administration Commits Itself to Spur Economy by Tax Cut or New Spending…” The same story has Johnson saying his new budget would “include sharp increases in spending from programs enacted during the past few years.”
President-elect Obama’s just-named chair of the Council of Economic Advisers, Christina Romer, herself hammered home the Keynesian construct of Johnson’s presidency, and the role of social welfare projects such as the Great Society within that construct, in a recent scholarly paper of her own, “Macroeconomic Policy in the 1960s: The Causes and Consequences of a Mistaken Revolution.”
Romer: “The Johnson Economic Reports were even more explicit about the lack of interest in long-run balance. The 1966 Economic Report stated: ‘In focusing on balance of the economy this policy strategy cannot give top priority to balance in the budget.’ (p 180) In listing the main tasks of public policy, Johnson’s section of the same report mentions stabilization, expanding opportunity, social welfare, even equilibrium in the balance of payments, but nothing about long-run budget balance.” (The italics are mine)
Summing up: Keynesianism is about deficit spending. From 1961-1969, i.e., the period that covers the Kennedy-Johnson years, the country ran a deficit every year except 1969, strong consistency for a non-Depression non-World War government. (Table B-79), Economic Report of the President). Then the Seventies came – deficit years. The Sixties made it clear that henceforward deficits would be the rule not the exception. We were indeed all Keynesians now.