This week most expected number is probably the employment numbers: nonfarm payrolls and the unemployment rate. As James was saying earlier, the employment situation remains bleak. Consensus expects 175k jobs lost and unemployment precariously balanced at 9.9%. Now, yes, nonfarm payroll is highly volatile and can be revised drastically next month. And yes, the unemployment rate is supposed to be a lagging indicator. But, as I mentioned earlier this year the financial recovery is maybe more than ever linked with the nature of the economic recovery. The latter being dependent on the employment numbers, one understands why tomorrow’s figures are highly expected. Moreover, the market is looking for optimism and gasoline to restart this rally. The market expects 175k jobs losses, is there room for disappointment? Let’s confront this figure with the recent ISM and ADP reports.

Last month ISM reports have proven to be quite accurate for Q3 GDP. But, they were quite contrasted this week. Let’s see if we can use the information on employment in those latest reports.

The week started with positive news from the PMI index. Indeed from the ISM website, one could read:

ISM’s Employment Index registered 53.1 percent in October, which is 6.9 percentage points higher than the 46.2 percent reported in September. This is the first month of growth in manufacturing employment following 14 consecutive months of decline. An Employment Index above 49.7 percent, over time, is generally consistent with an increase in the Bureau of Labor Statistics (BLS) data on manufacturing employment.

As Steve Liesman put it (and I double checked), starting from January 1980, we had a PMI greater than 50 and a negative report (meaning job losses) only three times. That’s to say, there is a probability of 98.9 % to get a positive report (i.e. no job losses). Wahoo!  And, unsurprisingly, this lifted the market up. Now, if you want to compare apples with apples, i.e. the manufacturing index with employment in the manufacturing sector, the picture is a bit less rosy. We have a 94.4% probability of getting a positive growth for construction employment and 90.5% for durable goods. So much for the 5% confidence level that statisticians are fond of. But still, on Monday, the odds were really pointing to a much better-than-expected number.

And then came the non-manufacturing report (NMI index) on Wednesday. This is what one could read from the ISM website:

Employment activity in the non-manufacturing sector contracted in October for the 21st time in the last 22 months. ISM’s Non-Manufacturing Employment Index for October registered 41.1 percent. This reflects a decrease of 3.2 percentage points when compared to the 44.3 percent registered in September.

What a contrarian piece of news! When one looks at the historical data and counts the number of times the index was that low, coupled with a job report better that -175k, this is what you get: 0. Each time we had such a bad NMI employment report, we also had more than 175k job losses. Two observations though. First, we have data from 1997 only and not a lot of points below 41.1, so we are in unknown territory. Second, again you may want to compare apples with apples, so if you focus on services employment, the last time we had such a bad NMI index was last May, and we “only” lost 99K in the services industry.

Finally, yesterday, the ADP report, which is for me the worst news of the week (the market realized this only late Wednesday). Nonfarm private employment decreased 203k from September to October 2009 on a seasonally adjusted basis, according to the ADP National Employment Report. Digging into the report, you see that the ADM figures and the BLS figures are strongly correlated (more than 90%). Now, the ADP estimated that the employed population in the US decreased by 0.19 % from the previous month (see figure below), which then leads to two conclusions. First, one may indeed expect job losses tomorrow (which contradict then the PMI index). Making a back-of-the-envelope calculation, i.e. transforming the figure to be in line with the BLS data, I found that the ADP estimation would mean 245K job losses. Ouch….

Take your pick. As far as I’m concerned, after all the better-than-expected numbers this week (including today new jobless claims), I expect the music to stop and believe that tomorrow’s report will show more than 200k job lost.