Author: Andrew Berger-Gross
The headline economic indicators reported every month in the news tell us where we are, but not how we got there. State and local labor market watchers know from the North Carolina Commerce Department’s monthly employment releases that job growth in North Carolina is trending upward as the economy shakes off the last vestiges of recession, but average wages have not increased at the pace we’ve come to expect during economic recoveries. Recent research suggests that, in order to sort out the causes for this seeming contradiction, we might need to look beyond the headline numbers to examine the underlying dynamics of the labor market.
The dynamism of U.S. labor markets has been on the decline since the national recession of 2001. By “dynamism” we are referring to particular mechanisms of employment change, such as job creation and job destruction; job hiring and job separation; and job-to-job movements of individual workers. Every one of these mechanisms appeared to slow during the 2001 and 2007-2009 recessions and failed to recover fully during subsequent economic expansions. In October 2014, LEAD published an article showing employment dynamism has declined in North Carolina as well during this period.
Datasets that allow us to track these dynamics at a detailed scale (including Quarterly Workforce Indicators (QWI) and Business Employment Dynamics) are relatively new; as a result, economists have not had much time to fully explore these data series or explain their ups and downs. Moreover, because they are released at a substantial time lag, these data often do not figure into the news cycle-driven economic reporting we get from the news media.
In this article, we take a closer look at data from QWI, which allows us to track employee turnover (defined as hires and separations as a percent of all stable jobs) at the state and local level. Turnover rates in North Carolina started to fall in late 1999, coinciding nearly exactly with the point at which the wage gains of the late 1990s subsided and were succeeded by a prolonged period of little-to-no wage growth. A casual observer might reasonably interpret these movements as indicating that the two trends are related.
One of the most contentious topics in economics nowadays is diagnosing the causes of slow wage growth in the U.S. during the first part of the 21st century. Many economists point to the “polarization” of the labor market, noting that changes in technology and trade flows have caused employment to decline for middle-wage jobs (such as those traditionally found in the manufacturing sector) while employment in low-paying sectors has increased since 1999. However, the August 22, 2014 session of a recent meeting of Fed policymakers in Jackson Hole, Wyo. introduced another hypothesis — namely, that the decline in average wages might be caused (at least in part) by the slowdown in U.S. labor market dynamism.
There are a couple of different reasons why we might expect a decline in employee turnover to negatively impact wages. One explanation looks to the “creative destruction” that occurs as poorly performing businesses fail and high productivity businesses grow, shedding old jobs and adding new jobs in the process. A decline in this creative destruction might harm the productivity of the overall economy, which itself is an important determinant of economic growth and wage increases. Another explanation is based on the notion that it can be hard for employees and employers to know whether they are a good fit for each other prior to the hire date. This theory envisions employee turnover as a consequence of this uncertainty, one implication being that a decline in turnover may reduce the job-to-job changes that help match workers to the “right job” and are important contributors to early-career wage growth.
The jury is still out regarding whether the decline in labor market dynamism is actually playing a role in wage stagnation, and clearly economists have more work to do in this area of research (particularly at the state and local level.) However, in an era when policymakers are increasingly looking beyond the headline numbers to diagnose what ails the economy, data on employment dynamics have the potential to contribute meaningfully to our understanding of how the labor market functions.
Local-level data from QWI allow us to dig even deeper to see how the turnover trend is unfolding in North Carolina’s metropolitan statistical areas (MSAs)1:
These data make it plainly clear that the decline in labor market dynamism is widespread, affecting every one of the state’s 15 metro areas. However, we cannot necessarily say that comparatively large turnover declines are associated with worse labor market outcomes — in fact, a simple regression analysis suggests the opposite. For example, Raleigh — which has experienced explosive job growth and seen the second-largest increase in average wages over the past 14 years — had the largest decline in turnover during this period (4.2 points). The strength of this relationship is such that it explains nearly half of the variation between metro areas, although whether turnover has a causal relationship with wages is still an open question.
This is far from the last word on this subject, as economic researchers are continuing to explore the causes of employee turnover and its impact on the wider economy. In the meantime, LEAD will continue to scrutinize these indicators in order to determine what they mean for labor markets in North Carolina and its various regions and how they can be used to inform economic policymaking in the state.
General Disclaimers:
Estimates from the QWI program are derived from both survey and administrative data, and are subject to sampling and nonsampling error. Any mistakes in data management, analysis, or presentation are the author’s.
Footnotes:
1 These metro-area data from the QWI program use geographic boundaries defined by the U.S. Office of Management and Budget in February 2013. Other statistics (including the unemployment rate for local areas) continue to be produced using previous geographic definitions; these data series will be updated in early 2015.