As fundamentals and transaction values pull further away from underlying job growth trends, it becomes increasingly important for commercial real estate market participants to keep tabs on the latter. With July’s Bureau of Labor Statistics report less than two weeks away, the reliability of alternative metrics deserves some attention.

Primary amongst the non-government measures of monthly job creation, the ADP report has assumed a high profile in recent years. Released two days before its government foil – the first Wednesday of the new month – the ADP report covers changes in private payrolls, specifically. In the brief interlude separating the measures, the leading gauge influences many market analysts’ immediate expectations for the private component of the official payroll statistics.

For commercial real estate research analysts, there are two key questions. Is the ADP report a good predictor of the official private payroll statistic? And, if the vectors diverge, is the ADP number a better or worse predictor of space absorption? Discrepancies in the reports are sometimes large, resulting from differences in methodologies and data sources. Depending on the statistical task, the shorter time-series of the ADP data may be a consideration, as well.

There are three basic answers to the question of the ADP report’s utility:

Well Correlated Over Time, Less So During Recoveries

Over the full period for which the ADP series is available, the two series are very well correlated. The extent of correlation varies over time, however. During periods of recovery in the labor market, the two series track further apart. The ADP series lacks the necessary detail to quantify the reasons for this definitively. The methodology suggests an easy explanation; feel free to email me if you are curious.

During the Most Recent Recovery, BLS Tracks National Office Absorption Better

At least over the last year, the BLS report has correlated better (albeit still weakly) with office space absorption statistics for major metros. Don’t place too much stock in this result. There is considerable variation in the degree of correlation between the BLS reporting of private employment, office-using employment, and office absorption across metros.

ADP is an Interesting, But Not an Instructive Monthly Indicator

While there is some utility in reviewing the ADP release each month, the long-term correlation with the BLS report does not mean that it is a good predictor month-to-month. From a statistical point of view, think of the ADP release as a kind of estimator. The following chart shows the error in the estimator, defined here as the difference between the estimator and the actual value. The ADP report is unbiased, meaning that the differences have a mean clos to zero. That is expected given the long-term relationship we have observed.

The cautionary note relates to the very large error that can be observed in the histogram. While the estimator is unbiased, it has a large standard deviation and variance (square of the standard deviation). As shown here with simple differences, it is not atypical for the ADP report to be off by +/- 100,000 jobs during any given month; by up to 200,000 jobs in the extreme. On Wednesday afternoon and Thursday, should the ADP report then be shaping real estate investors’ expectations for Friday? This result suggests that you can afford to wait.