Structural Shifts: Comparing the Last Two Recessions

by Karl Smith, Modeled Behavior

Dave Altig offers the following

If you try, it isn’t too hard to see in this chart a picture of a labor market that is very close to “normalized,” excepting a few sectors that are experiencing longer-term structural issues. First, most sectors—that is, most of the bubbles in the chart—lie above the horizontal zero axis, meaning that they are now in positive growth territory for this recovery.

Second, most sector bubbles are aligning along the 45-degree line, meaning jobs in these areas are expanding (or in the case of the information sector, contracting) at about the same pace as they were before the “Great Recession.” Third, the exceptions are exactly what we would expect—employment in the construction, financial activities, and government sectors continues to fall, and the manufacturing sector (a job-shedder for quite some time) is growing slightly

First, a clarification. Its not immediately clear from Altig’s post but the chart indicates that  employment in non-motor vehicle manufacturing is expanding far, far faster  than would be expected given the last recovery.

This is important for thinking about any kind of structural story you might be inclined to tell. Manufacturing employment had been falling rapidly for roughly 15 years and then suddenly stopped falling in the wake of this recession and started growing very slowly.

This may be the structural shift that folks are looking for but its important to note that this is basically a “reshoring” shift. Although, its not so much reshoring as a rapid slowdown in offshoring combined with growing underlying demand.

In addition, the dynamics of the dot-com recession were very different. Though it felt mild to most folks and the employment rate only barely peaked above 6% it was actually slightly more harsh for non-financial corporate profit growth.

Here are non-financial corporate profits on a log scale. You can see the divot from Dot-Com was proportionately the largest on record.

That recession behaved in many ways like what you would expect if you were experiencing a demographic slowdown. Capital took a huge beating. Employment growth was slow and unsteady, but unemployment was mild.

Lets look at private sector employment in the two recoveries

The bottom of Dot-Com was long and drawn-out while the Great Recession was sharp, returning almost immediately to growth rate during the last recovery.

If we look at compounded annual rates of growth, private sector payroll growth during the Great Recession return to above long-run trend (2%)  sooner than either of the last two recessions, despite being deeper.

So private sector employment growth has been quite strong. GDP, less so.

There has been a divergence in Okun’s Law, even when looking at long run patterns

Though notice the deviation is even more pronounced using a nominal version of the law, suggesting that nominal spending growth exceeding maximum production growth is not yet the issue

Based on both the growth in payrolls and the decline in unemployment  historical patterns would lead us to expect real GDP growth of around 4% and nominal GDP growth of around 6.5 – 7%.

I think the strongest clue in explaining this discrepancy comes from looking at non-durable goods output.

The big unexpected surge in employment is in manufacturing ex-motor vehicles. Though again its not so much job gains as the absence of job losses. Motor vehicles make up a large part of the durable goods sector, so we can get a proxy for non-vehicle manufacturing by looking a non-durable goods.

Despite the above average employment growth – which is to say the absence of employment shrinkage – production in non-durable goods is way down and indeed declining.

So the GDP discrepancy looks like it may be due to a sectoral shift and possibly the closing of opportunities to offshore portions of the non-durable supply chain.

In general I think looking to GDP introduces unnecessary opaqueness in the modern age. We have easy access to much finer data. In this case GDP would provide a potentially confusing measure of the state of the economy as compared to payroll growth or unemployment.

If it is the case that GDP growth is slowing because non-durable production is slowing despite rapidly increasing employment then it is likely that we will soon see increased investment in the sector boosting productivity and creating a following period where GDP rises faster than employment.

Posts by Karl Smith

analysis blog opinion

Karl Smith

Karl Smith is Assistant Professor of Public Economics and Government
at the School of Government at the University of North Carolina at
Chapel Hill.  He works primarily in the areas of Public Finance and Public Revenue and Expenditure forecasting. Prior to joining the School he was a graduate fellow at the Institute for Emerging Issues, where his work was focused on state and local tax reform. Smith holds a BA and a PhD in economics from North Carolina State University.  He writes at the blog Modeled Behavior and at Seeking AlphaVita.