The Primary (and Hidden) Risk to U.S. Inflation?

The U.S. labour force participation rate just hit its lowest level in thirty-five years, dropping to 62.8% in December. (As a reminder, the U.S. participation rate measures the percentage of working age Americans who are either employed or who are actively looking for work.)

There is an ongoing debate about whether or not this labour force withdrawal is caused by cyclical or structural factors, and the answer to this question has direct implications for Canadian mortgage rates.

If the U.S. participation rate is dropping because of cyclical factors, then this trend will reverse when there is  acceleration in U.S. economic momentum. The U.S. Fed believes this to be the case, and as such, it is using aggressive monetary policy stimulus to try to fuel economic growth and spur the U.S. participation rate higher.

If the U.S. participation rate is dropping because of structural factors, however, this means that the decline is being primarily driven by long-term changes in the labour force, for example, by shifting demographics and by a growing mismatch between the skills that businesses need and those that the labour force can supply. A structural shrinking of the U.S. labour pool would be far less susceptible to changes in the U.S. business cycle and far less prone to influence from Fed policy.

Most of the experts I read appear to believe that the drop in the U.S. participation rate is caused by cyclical factors, but is this wishful thinking? Forgive me while I scratch my contrarian itch with some recent statistics from one of my favourite contrarians.

In a report released last week, economist David Rosenberg offered the following insights about the current state of the U.S. participation rate (with my comments in italics):

• Almost 80% of Americans who withdrew from the labour force were over the age of fifty-five. The aging of the U.S. labour force is a structural, not a cyclical, phenomenon.
• The first of the 78 million American baby boomers retired in 2013 and there will be an average of 1.5 million retiring boomers each year for the next fifteen years. The impact of retiring boomers on the U.S. participation rate is only beginning to be felt.
• The number of people who withdrew from the labour market but who actually wanted a job fell by 600,000 in 2013. If the drop in the U.S. participation rate were caused by a cyclical slowdown in the business cycle, this number would be rising because people would be involuntarily pushed out of the active labour force. Whereas if the participation rate is falling at the same time that there are fewer disenfranchised workers, this implies that structural, rather than cyclical, forces are at work.

These data started me wondering. Was the drop in the U.S. participation rate primarily caused by the start of the Great Recession or might history show that this was actually the mother of all misunderstood coincidences? Is it possible that demographic shifts and the long-term decline of blue-collar, middle-class U.S. jobs came to a head at the same time as the U.S. economy slowed for different reasons?

If the Fed is right that the falling U.S. participation rate is primarily caused by cyclical factors, then pouring monetary stimulus on the problem should help reverse that trend. And the Fed can expect to be able to manage rising labour costs by gradually dialing back its stimulus as the U.S. economy picks up steam and the supply/demand balance between the labour force and the economy reverts back to its longer-term average.

But if the falling U.S. unemployment rate is in fact caused primarily by structural factors, inflation will not be nearly as easy to control because the U.S. economy will not have an over-supply of labour to absorb, over time, as the Fed fundamentally believes. If the U.S. labour force does not, in fact, have a pent up supply of available and qualified workers in its back pocket, even a small increase in demand will create a shortage of labour. Since rising labour costs are arguably the single biggest drivers of inflation, the Fed’s monetary stimulus experiments could end up triggering higher labour costs, which choke off the very economic momentum that it is trying to create.

This high-inflation low-growth scenario is referred to as stagflation, and we last saw it in the 1970s and early 1980s. For those too young to remember, mortgage rates then were a multiple of what they are today.

Consider these employment statistics that were also shared by David Rosenberg in the same report last week:

• The jobless rate for those with a college degree fall from 4% to 3.3%. There is lots of demand for white-collar workers.
• Of the 1.4 million net new jobs created in 2013 according to the Household Survey, 85% of them required a college degree or some facsimile thereof. At the same time, demand for workers who do not fit this profile is virtually non-existent. This growing problem is referred to as a skills mismatch and it is structural, not cyclical, in nature.
• According to Fed Beige book the unemployment rate for “management, professional and related occupations” plunged from 3.7% to 2.3% in 2013. This segment accounts for a total of 55 million workers, roughly one third of working Americans. If demand continues to exceed supply for this category of worker, it is easy to envision higher-than-expected labour costs fueling higher-than-consensus inflation, referred to as ‘cost-push’ inflation, over time.

If the Fed is wrong in its assumption that the falling participation rate is caused by cyclical factors, its assumptions about future inflation, and its assumptions about the future path of both short- and long-term interest rates, will be wrong as well. Given that the Canadian and U.S. economies are so intertwined, if the U.S. experiences higher inflation, we will import it, and our mortgage rates will end up being higher than the consensus is forecasting today.

Even if such a scenario unfolds, it should take some time to do so. And given that the Fed remains steadfast in its approach, I still expect it to keep short-term U.S. rates at ultra-low levels for long enough to make our variable-rate mortgages the odds on favourite for saving money over the next five years. But any responsible interest-rate forecast must acknowledge contingent risks, and the evolution of the U.S. participation rate and unemployment statistics is the primary one that I see today.

Five-year Government of Canada bond yields were down three basis points last week, closing at 1.70% on Friday. Five-year fixed rates are coming down slowly and a good market rate is now offered in the 3.29% to 3.34% range, depending on the features, terms and conditions that are important to you.

Five-year variable-rate mortgages are available in the prime minus 0.55% range (which works out to 2.45% using today’s prime rate of 3.00%). I can also now offer a five-year variable at prime minus 0.60%, but with slightly more restrictive terms.

The Bottom Line: Is today’s ultra-low U.S. participation rate the cause of either cyclical or structural factors? While most experts are betting on the former, there are compelling trends in U.S. labour statistics to suggest that it may be the latter. We’ll keep an eye out as the answer to this question becomes clearer because it will play a major part in determining how fast our mortgage rates will rise in future.