The Federal Housing Finance Agency reported that contract rates on purchases of newly-built homes rose 7 basis points over the month of August 2017 to 4.01 percent. The average mortgage rate on purchases of newly built homes is 47 basis points above the low touched in October 2017. However, it remains 17 basis points below the recent high reached in February 2017. In spite of the month’s increase, mortgage rates remain historically low.
Previous analysis (here, here, and here) has highlighted the role played by changes in the rate on the U.S. 10-Year Treasury note, using shifts in its components to explain both the trend in mortgage rates and the implied macroeconomic conditions. More recent analysis used the real return and the implied inflation compensation to describe the potential impact of the Fed’s balance sheet normalization on mortgage rates. Chair Yellen noted in a recent speech on inflation that “Although research suggests that the fall in financial market-based inflation compensation instead primarily reflects a decline in inflation risk premiums and differences in the liquidity of nominal and indexed Treasury securities, the notable decline in inflation compensation may be a sign that longer-term inflation expectations have slipped recently. Meanwhile, the average mortgage risk premium has remained steady in recent years.
This method of assessing mortgage rates is similar to evaluating monetary policy, or an appropriate level of the federal funds rate. While actual policy is made with a degree of judgement, a rule can help to inform the policy debate and be used by forecasters to predict future moves in both the federal funds rate and mortgage rates. According to the “Taylor Rule”, named after John Taylor, the economist that developed the formula, the federal funds rate (ἰ) is a function of inflation (∏), the natural rate of interest (r*), the distance between inflation and the FOMC’s 2.0 percent target of price stability ((∏ – 2%)), the difference between current GDP and potential GDP ((YActual – YPotential)), as well as a variable capturing any additional, but non-systematically important information (v).
ἰ = ∏ + r* +[.5 * (∏ – 2%)] + [.5 * (YActual – YPotential)] + v
A previous post illustrated that the output gap may be near zero and currently, 12-month consumer price inflation (CPI) inflation is 1.9 percent, although prices of personal consumption expenditures (PCE) are typically used instead of CPI. If inflation returns to the FOMC’s price target of 2.0 percent, a belief of the median FOMC member, and no additional information systematically impacts our model (v = 0 on average), then the simplest form of the Taylor Rule can be reduced to the additive combination of current inflation and the natural rate of interest, similar to assessments of trends in mortgage rates by gauging the inflation expectations and the real return components of the 10-Year Treasury note rate.
ἰ = ∏ + r*
Since the current rate of inflation is known (in this case we have assumed it to be 2.0 percent for simplification purposes), then one rules-based assessment of the appropriate level of monetary policy using a basic form of the Taylor Rule is based on the natural rate of interest, r*, or the real return at full employment. Since the natural rate of interest represents the price of money, then it is directly related to inflation, and supply and demand factors, specifically savings rates, both public and private, and investment demand determine its level.
A lower natural rate of interest is typically attributable to “factors that increase savings, depress investment demand or both”. These factors typically coincide with more moderate economic growth and lower inflation. Research by Kathryn Holston, Thomas Laubach, and John C. Williams find that the “Explanations for this decline [in r*] include shifts in demographics [and] a slowdown in trend productivity growth.”
Unfortunately, r* is unobservable, but economists have made attempts to estimate it. The implied r* taken from the September 20, 2017: FOMC Projections materials, indicates that the median estimate of the long-run r* is 0.80 percent, the difference between the median long-run federal funds rate and the median PCE inflation. Alternatively, important research by the Federal Reserve Bank of San Francisco estimates r* to be about zero currently. Recent research by Ken Rogoff notes that the “real” component of the 10-Year Treasury note rate, a measure of expected r*, is currently 0.49 percent.
Assessments of r* reflect current macroeconomic conditions and have implications for monetary policy and for mortgage rates. Within this form of a Taylor Rule and assuming that inflation reaches 2.0 percent, then an r* equal to 0.80 percent indicates that the federal funds rate should reach 2.8 percent (2.0 + 0.80), if r* is 0.49 percent, then the federal funds rate should reach 2.49 percent, but if r* is zero, than the federal funds rate should converge to 2.0 percent.
Finally assuming inflation expectations return to and remain at 2.0 percent, that the mortgage risk premium remains near approximately 170 basis points, its 1972 to 2016 average, and that the longer-run r* remains the same and doesn’t increase as the federal funds rate rises, then first, the yield curve would flatten as the rate on the 3-month Treasury bill approaches the rate on the 10-Year Treasury note, and second mortgage rates would remain below 5.0 percent.
A previous post indicated that the recession probability rises as the yield curve flattens. Simultaneously, if the longer-run r* of 0.49 is reached and expectations for the future remain unchanged as the federal funds rate rises, then average mortgage rates should rise to around 4.19 percent (2.0+1.70+0.49), and if the longer-run r* of 0.80 is reached, then average mortgage rates should reach 4.50 percent.