Too Much Information

The much-maligned 'Efficient Market Hypothesis' proposes that market prices already incorporate all known information, and that only new information can make prices change, other than in a completely random fashion.  There are some problems with this idea, but in the very short term, market participants sure act like it's true.  That's why they sit huddled in front of screens for every data release, and then furiously trade the market as soon as the data comes out.  As a strategist, I always found the flurry to be pretty amusing - just look at how much data there is between this morning (August 26th, 2008) and next Friday's employment report:

8/26  09:00     S&P/CaseShiller Home Price Index JUN (167.2)
8/26  09:00     S&P/CS Composite-20 YoY JUN (-16.2%)
8/26  10:00     Consumer Confidence AUG (53.0)
8/26  10:00     New Home Sales JUL (525K)
8/26  10:00     New Home Sales MoM JUL (-0.9%)
8/26  10:00     House Price Index MoM JUN (-0.4%)
8/26  10:00     House Price Purchase Index QoQ  2Q (-1.5%)
8/26  10:00     Richmond Fed Manufact. Index AUG (-10)
8/26  14:00     Minutes of Aug. 5 FOMC Meeting
8/26  17:00     ABC Consumer Confidence AUG 24
8/27  07:00     MBA Mortgage Applications AUG 22
8/27  08:30     Durable Goods Orders JUL (0.0%)
8/27  08:30     Durable Goods Ex Transportation JUL (-0.7%)
8/28  08:30     GDP (Annualized) 2Q (preliminary) (2.7%)
8/28  08:30     Personal Consumption 2Q (preliminary) (1.6%)
8/28  08:30     GDP Price Index 2Q (prliminary) (1.1%)
8/28  08:30     Core PCE QoQ 2Q (preliminary) (2.1%)
8/28  08:30     Initial Jobless Claims AUG 23 (425K)
8/28  08:30     Continuing Claims AUG 16 (3390K)
8/29  08:30     Personal Income JUL (-0.2%)
8/29  08:30     Personal Spending JUL (0.2%)
8/29  08:30     PCE Deflator YoY JUL (4.5%)
8/29  08:30     PCE Core MoM JUL (0.3%)
8/29  08:30     PCE Core YoY JUL (2.4%)
8/29  09:45     Chicago Purchasing Manager's Report AUG (50.0)
8/29  10:00     U. of Michigan Confidence AUG Final (62.0)
8/29  10:00     NAPM - Milwaukee AUG
9/02  10:00     ISM Manufacturing AUG (49.5)
9/02  10:00     Construction Spending MoM JUL (-0.4%)
9/02  10:00     ISM Prices Paid AUG
9/02  17:00     ABC Consumer Confidence AUG 31
9/03                Domestic Vehicle Sales AUG (9.4M)
9/03                Total Vehicle Sales AUG (13.0M)
9/03  07:00     MBA Mortgage Applications AUG 29
9/03  07:30     Challenger Job Cuts YoY AUG
9/03  10:00     Factory Orders JUL (0.4%)
9/03  14:00     Fed's Beige Book
9/04  08:15     ADP Employment Change AUG (-19K)
9/04  08:30     Nonfarm Productivity 2Q (final) (2.9%)
9/04  08:30     Unit Labor Costs 2Q (final) (0.7%)
9/04  08:30     Initial Jobless Claims AUG 30
9/04  08:30     Continuing Claims AUG 23
9/04  10:00     ISM Non-Manf. Composite AUG (49.0)
9/04                ICSC Chain Store Sales YoY AUG

and finally...
9/05  08:30    The Employment Report, including...
                       Change in Nonfarm Payrolls AUG (-70K)
                       Unemployment Rate AUG (5.7%)
                       Change in Manufacturing Payrolls AUG (-35K)
                       Average Hourly Earnings MoM AUG (0.3%)
                       Average Hourly Earnings YoY AUG
                       Average Weekly Hours AUG (33.6)
  

(Shamelessly copied from Bloomberg, LP)

Whew!  I did say too much information, didn't I?  Two full weeks, starting on a Tuesday, and including the Labor Day Holiday!  I've included the average of published estimates in parentheses, where there is one, and that's the key.  When the data is released, nobody cares what the actual numbers are, it's all about how far away the data is from the estimate.  Wall Street bows down to the efficient market hypothesis at the time of data release, even though they know full well that the data is often dramatically revised in the following weeks, months, and even years.  It's only later on that the economists have a chance to piece the data together into a coherent story, usually by looking more at the trends in the data than the actual levels themselves.  And that's where the really interesting opinions seem to come from - taking a step back and looking at a large collection of data developed over a long period of time - right in the face of the efficient market hypothesis, which says you shouldn't be able to do that.

To me, the best analogy is that there is an information pipe like the gradient shown below, and the closer you are to the front, the quicker you get the information.  The other side of the coin, though, is what to do with it once you have it.  That's why Wall Street firms have such an advantage over the public.  They get the information first, and they have squadrons of professionals whose job is to pore over the information and put together a coherent strategy.  By the time the public gets the information, it has been chewed over by enough brain and processing power to put a man on Alpha Centauri.



Of course, all of this begs the question as to why all these geniuses have lost so much money this year, and that comes down to what you do with the information.  Do you slowly and steadily try to use your advantage to outperform the market, or do you get lured into the opportunity to make massive short-term profits without worrying too much about what's going to happen a couple of years down the road?  I think the answer to that question is fairly obvious, all things considered.  Which allows me to say that a detailed explanation of how Wall Street sank the island of Manhattan - again - is a good subject for another post.

(Oh, and you may be wondering about the lines in the list that are struck through... those are data releases that are on my personal shit list - the data that I think shouldn't be released and really don't need to be.   You can add producer price index to the list, but I think I'll save the explanations for another post, too.)