A special Barrons edition.
Tyler Cowen did a bad thing in recommending Econospinning, by Gene Epstein:
Marginal Revolution: Econospinning: Imagine lengthy polemics against the use of numbers in the work of Paul Krugman (most of all), the Op-Ed page of The Wall Street Journal, Brad DeLong, Steve Levitt, and Barbara Ehrenreich, among others. Except the vehicle isn't the blogosphere, it is a book! This one is guaranteed to ruffle feathers...
I open the book at random and get page 143. The heading halfway down the page is "The Employment Report and the Bond Market." The text is:
...there is evidence that the employment report misinforms the bond market.... Once the [nonfarm payroll] number is announced at 8:30 AM [on the first Friday of each month], bond prices can react accordingly. But do they react the right way? The recent run-up in short term interest rates does not inspire confidence.
From June 30, 2004, to March 28, 2006, the Federal Reserve hiked the interest rate... 15 times in a row. The yield on the two-year Treasury note rose in response, from 2.64% on July 2, 2004... to 4.89% by April 2, 2006.... 22 employment reports [were] released over this period. How much credit can they be given for alerting the bond market to this 225 basis point rise in the two-year interest rate? No credit can be given at all....
[On] the 22 trading days on which these employment reports were released... the yield rose 11 times, fell in 9, and was flat the other two times...
Idiot. Fool. Must... calm down. Must... adjust medication doses. Must... remain calm. Must... explain things simply and clearly. Ah. I can feel the neurotransmitter balances change. That's better.
Gene Epstein, Economics Editor of Barrons, believes that the monthly "employment report misinforms the bond market" because during the period from mid-2004 to mid-2006 that the Federal Reserve was in a tightening cycle--and thus that bond yields were rising--bond yields did not exhibit substantial jumps in the immediate aftermath of the release of the monthly employment report. Instead, about half the time bond yields jumped. About half the time bond yields fell.
But suppose that bond yields had jumped, regularly, on the day the employment report was released. What then? Then there would have been a lot of free money to be made: sell bonds the day before the employment report is released, buy them back after the release, and so make your fortune.
That kind of thing doesn't happen. Financial markets may not get asset prices right all the time, but they are very good at eliminating easy opportunities to make lots of money via simple trading strategies. Systematic, arbitragible moves in asset prices in one direction only (or predominantly) in response to the monthly flow of economic data simply do not happen, whether or not the statistic released is informative and useful.
The monthly employment reports from 2004 to 2006 did not provide signals that short- and medium-term bond yields were on a rising trend--everyone trading knew that. The monthly employment reports provided signals as to whether employment was rising faster than previously expected--in which case bond yields would rise because that meant the Federal Reserve would be likely to raise the federal funds rate faster than previously expected--or slower than previously expected--in which case bond yields would fall because that meant the Federal Reserve would be likely to raise the federal funds rate slower than previously expected.
You don't have to be a fundamentalist believer in efficient markets to know the following proposition:
About half the time the monthly employment report will be stronger than expected, and bond yields will rise; about half the time the monthly employment report will be weaker than expected, and bond yields will fall.
But it appears that you do have to be a lot smarter than and know more about Wall Street than the Economics Editor of Barrons.