When Companies are Desperate to Hire, But Don’t (Stats Heavy)

What used to be an average of 15 days to hire workers for open jobs (2009) is now taking an average of 27 days (2015), at least according to one (selective) study in the Wall Street Journal. There’s some evidence that this is a worldwide phenomena. There are a lot of excuses as to why this is happening, but no empirical evidence. My preliminary regression models show that firms are desperate to hire, but taking longer to actually hire, and this is a huge paradigm shift since 2008.

The original study points to the fact that firms are taking longer to do background checks, but fails to operationally define “background checks.” The background checks could range from credit checks, to criminality, to child support payments. If credit checks are being used, well, most people had their credit trashed in 2008/2009, and are still having problems with that. Since there is no operational definition for “background check,” then there is no way to actually get data for that, and it becomes speculation only.

Some economists say that the technological change has shifted the pool of available workers with the proper skill set. Consider this however: the amount of time it takes to hire even the most menial labor (such as retail workers) has doubled. What great technological changes have happened at your local Walmart or grocery store since 2009 that can explain this? The same question applies for the F.I.R.E. sector as well. Besides, if jobs are so hard to fill with skilled workers, why haven’t wages/salaries budged overall?

Then there is this chart, which shows a pretty close-to-even rate for job openings and actual hires, which is the first time this has happened since 2000. Background checks cannot explain this. This smells like some type of organizational behavior inside the firm that is unknown. And with unknown variables, there’s only one place to turn: a regression model.

Hires versus openings1

Using the data we have from the JOLTS survey data from the Bureau of Labor Statistics, the correlation between hiring and job openings was .126 (not much of a relationship). Since 2008 however, that correlation was .935 (very strong relationship). Even with the difference in N sizes for the two time frames (roughly double) there is still about a 90% increase in correlation.



It stands to reason that as more jobs become available, that firms will be more desperate to hire: hiring=f(job openings). How desperate? Here’s a simple econometric (regression) model of hiring rates (dependent) versus job openings (independent).

Hires= β1+β2 (job openings)+u


As expected, there is a huge statistical relationship that appears after 2008. Before 2008, there was no statistically significant slope between job openings and hires. After 2008, that all changes (suddenly) and so do the changes in hiring behavior of firms.

2004-2008 regression results
Regression 2004-2008


2008-2014 regression results
Regression 2008-2014

The “unknown” factors (residuals) are normally distributed in both time models, validating the betas and significance.

Resid Plot 20081


Resid Plot 20141


Plus, for 1 regressor and the N value, the Durbin-Watson is well within the upper and lower limits, which means this is not an autocorrelation “anomaly.”

What this shows is that firms are exponentially more desperate to hire, and that the “unknowns” are not really explaining anything significant. Yet they are taking their time in hiring. This model by no means is the be-all or end-all of the issue. While I could not find labor data to correlated enough to incorporate into a regression, that doesn’t mean that it’s not out there. There may also be a log-linear pattern, which I haven’t tested for yet.

What it does say is that the behavior of firms in their hiring process has shifted significantly since 2008, and that there’s no real reason why they should be dragging their feet. The rest is pure speculation, and thus, not “science.”

So what’s really going on? I suspect a methodological problem with the surveys themselves; and the lack of data. Surveys may not be asking the right questions of firms, and firms may have a bias error built into the question pool. There may also be discrimination going on (both legal an illegal) that is preventing firms from hiring. Firms also typically refuse to acknowledge that when an applicant applies for a job, the need is immediate; workers need jobs now, not 30 days later. In other words, with a low labor force participation rate, and a large number of people under-employed (U6), there’s no reason to suspect an actual labor shortage. Without actual data measuring these things however, there’s no scientific way to prove it.

This entry was posted in Econometrics, Economics, Labor, Sociology, Statistics. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s