Lies, Damned Lies, Statistics and the Employment Situation Report

(A version of this article appeared in TheStreet)

On Wednesday morning the monthly Automatic Data Processing (ADP) Employment Report was released.

Ever since they started doing so there’s been lots of disagreement over how accurately ADP’s numbers parallel or will reflect the government’s Non-farm Payroll statistics that are contained in the Employment Situation Report that comes out two days later. Regardless of the validity of the ADP Report it still has the capability of moving markets.

After last month’s disappointing official numbers the ADP number was pretty much on target for expectations but the market showed some mild disappointment in the pre-open trading, however, it got over that disappointment very quickly.

At this point we’ll have to wait until Friday to see whether there is concordance between the two reports and to see how the market reacts to the official numbers and revisions.

I recently collected some historical data looking at the relationship, if any, between the Employment Situation Report and market performance. Watching the release of that most recent government report and seeing the market react to very disappointing numbers by pushing the market higher had me realizing that it seemed as if the market always went higher with the announcement of those official numbers.

Who knew that in the comfort of my retirement I would actually once again find myself dusting off a statistics program and running t-tests? That is the sort of thing that nightmares are made of, rather than dreams.

Contrary to my supposed realization, beginning with January 2011 there were no associations between the reporting of data, regardless of its content and how the market did, neither in the week leading up to the report nor the week following. Neither did the day before the report release have any predictive value.

That is unless you looked at the past 18 months.

During that time period there was a statistically significant likelihood that on the week preceding the report and the day preceding the report that the market would move higher. The following week simply performed no differently whether preceded by an Employment Situation Report or not.

For those interested in such things, the random chance that the market went higher on the day of the Employment Situation Report release was less than 4% and less than 2% when considering the prior week.

While that seemed compelling, as with all statistics it helps to look beyond the numbers and try to have an understanding of possible confounders or environmental factors that may have played a role.

Perhaps coincidentally the past 18 months reflected the beginning of the third and final phase of Quantitative Easing.

Because of that possibility I wasn’t terribly excited about being further long the market in anticipation of an additional Employment Situation Report fueled run higher, considering that there appears to also be an association between the announcement of the taper and the market’s fortunes. Certainly the past month causes one to rethink a bullish thesis and the environment may now be substantively different with the taper in place.

Additionally, In hindsight, starting the week with a 325 point loss seemed to indicate that playing the market for a weekly advance in anticipation of Friday’s report wouldn’t have been a very good idea.

However, about half of the weekly gains seen in Employment Situation Weeks came on the day of the report, suggesting that there might be some advantage to adding long positions prior to Thursday’s close, even in the face of a market teetring and looking for direction and even in the face of losses earlier in the week.

Based upon the pattern of the past 18 months, while I generally don’t consider index ETF trading, I may look at the possibility of purchasing some SPDR S&P 500 Trust (SPY) shares with the anticipation of closing the position at the end of Friday’s trading by selling slightly out of the money call options on the position, as premiums are beginning to reflect increasing volatility and could enhance any report related advance in the index.

While statistics may have a wonderful ability to confirm whatever thesis one wishes to expound the relative benign reaction to a benign ADP report gives me reason to suspect that optimism may be warranted from 3:59 PM Thursday until 4 PM on Friday.