NFP Report: Trend vs Single Data Point

The show must go on . . .

Despite the events of the past few days, the never ending stream of data, useless and informative alike, continues unabated.

My longstanding view is that most of the data in these series is not statistically meaningful. The collection process is imperfect; the models used to analyze this data are rather flawed. And the very human tendency to over-emphasize the most recent piece of data tends to increase the total amount of volatility, be it in equity prices or belief systems.

What does matter is the overall vector of a given economic sector. Vectors include the rate of acceleration or deceleration, persistency, direction etc.  Think overall “trend” and changes thereto. For employment, this means: Are we seeing an increase in the factors that lead to hiring? What is the ratio between hires at big firms vs small firms? Are Wages increasing, staying flat, or decreasing; Temp workers getting hired, total hours worked etc. What are the likely data and modeling errors? Collectively, those factors all add up to an issue of the employment situation roughly improving, maintaining a stability, or getting worse.

Hence, each data point should be looked at in terms of whether it is continuing the overall trend, or suggesting a reversal in trend. Everything else is noise.

Consider the transition over the past five years — 2006-2011. In 2007, the entire new job creation pool was a result of Birth/Death adjustments. That was a warning sign about a major change in trend. In 2008, we saw monthly job losses of half to three quarters of a million per month. That reversed after the market collapse and massive liquidity/bailouts of 2009. In 2010, job creation barely kept up with population growth. By the time we hit 2011, even that modest job creation began faltering.

For this month, the projected gain in U.S. payrolls in September is small:  The median forecast of 91 economists surveyed by Bloomberg News was that NFP Employment likely climbed by 55,000 workers after no change in August. Jobless rate is expected to remain unchanged at 9.1%.

Of course, economists have been proven themselves inadequate at making forecasts, which is yet another reason to look at the data, and remain focused on the overall trend.

What's been said:

Discussions found on the web:

Posted Under