Why is the average U.S. wage still near 2008 levels? This is a great source of confusion to members of the U.S. Federal Reserve Open Markets Committee (FOMC), who have routinely predicted higher wage gains from the corresponding 5.9 percentage point decline in the U.S. unemployment rate from the cycle-peak of 10.0% in 2007. Two data points neatly illustrate a demographic transition underway: (i) job growth of 110,000 per month is required to hold the unemployment rate steady, according to the Federal Reserve of Atlanta, and (ii) monthly payroll has exceeded 200,000 per month during the last 2-years. Taken together, this means roughly two Millennials are being added to the labor force for every single retiree. Now, if we make the simple assumption that retirees, with perhaps 40-years of additional experience, earn higher salaries than new entry-level employees, the overall average wage is effectively lowered. Make sense? Imagine a firm with 20 employees adding two new workers with wages of $50,000/year to its payrolls and subtracting one worker with a wage of $80,000/year. Should we be expecting anything but downward pressure? Simple math works here. What’s more, data from the Bureau of Labor Statistics (BLS) that tracks employer cost of employee compensation shows growth of annual growth of 2.5% through the recent period, so employers are paying more. The growth just has not materialized in the average national figure yet. Key to understanding this data is the demographic shift taking place: ~10,000 Baby-Boomers are retiring each day though 2028, according to Pew Research, as ~10,000 Millennials turn the age of 21 each day, so we have tent poles on either side of the workforce. For better or worse, the FOMC forecasting model is derived from the Phillips Curve that does not account for such demographic nuances. It’s the reason why forecasts continue to overshoot actual figures.
The weak wage growth today is a concern because, when one rolls the forecast period forward, the demographic situation changes dramatically, such that there is a strong up-tick in the average figure when the demographic transition has passed the midway point. Think about it. Historical precedent can be found in the early 1970s, when the ‘Baby Boomer’ generation first entered the workforce, and stagnant wage growth led the U.S. Congress to pass a 25% increase in the minimum wage in May 1974. This should sound familiar because the State of California recently passed the same level of increase and the U.S. Congress is now deliberating the issue. While the topic of hyperinflation in the 1970s remains a source of contentious debate in academia, there should be no doubt that a demographic situation similar to the above contributed to the massive inflation spike of the 1970s. Just look at the wage data. This point was underscored in a recent study from the Bank for International Settlements (BIS), where Juselius and Takats (2015) found “around half of the total average reduction in inflation from its peak” between the late 1970s and early 2000s is attributable to the increase in the relative share of working age population or a “reduction of around six percentage points in the inflation rate” in the period. This argument implies too much credit has been awarded to central banks for lowering inflation. It also makes more sense.
Against the above backdrop, it’s important to recognize that labor represents more than 60% of the cost of goods sold, so any sizable movements in wages will translate directly into inflation, but only after the demographic transition noted above takes-hold. In the meantime, rent is growing at 4x the rate of income, according to the National Association of Realtors, so we can expect upward pressure in shelter costs, which are captured in the owners’ equivalent rent (OER) component of the consumer price index (CPI) at a meaningful ~32% of the total weight, when excluding food and energy. OER not only rose 3.2% in the last year, but the figure was also well below real rent growth of 5.2%, a routine trend in place since 2000, given per annum OER growth of 2.5% and annual rental growth of 3.7% in the same period. First implemented in 1982 by the Bureau of Labor Statistics, the OER formula is based on (i) a survey of homeowners, who are asked the estimated cost of renting-back their homes, and (ii) adjusted for changes in actual rent from a separate survey of landlords. This means it is naturally a lagging indicator, so we can expect higher inflation in the shelter cost category. As a result, the best forward view on inflation is likely achieved by adjusting the reported CPI figure to reflect actual rent increases, something that can be achieved by substituting real rent of 5.2% for the reported 3.2% in OER. That adjustment alone pushes CPI excluding food and energy from 1.9% to 2.7%, which is well above the FOMC 2.0% target. The gap widens too, when single-family house prices are considered; Zillow data showed a 10.0% increase in listed home prices in 2015. Taken together, this means the current market forecast showing inflation near 1.9% for the next 10-years vastly understates the reality of the situation. Inflation will most likely be double that figure in half the period, before higher oil prices are factored in. Any investor buying 10-year notes near 0% levels will, in turn, be losing money for the majority of its duration.