Reader JohnH critiques my post on real mortgage rates thusly:
With rates changing rapidly this month, the results can change overnight, making the date of the data critical. I provided my inflation assumption and my mortgage rate assumption. None of that was provided in the above chart. And no actual data was provided.
In addition, the after tax calculation depends a lot on the tax bracket assumed. All information about the tax bracket came in the statement about “even for the highest tax brackets.”
In this spreadsheet, I include all the series used to construct the the following three graphs. The spreadsheet is laid out simply enough that I believe an idiot can follow along.
First, calculating the after tax rate for the highest income bracket. If one is trying to show that all real rates are above zero, one wants to use the highest marginal tax rate which pushes down the after tax rate the most. In Figure 1 below, I show the highest marginal tax rate (black line, right scale), and the 30 year fixed rate (blue line, left scale). Applying the adjustment for the maximum tax rate yields the tan line below. If one uses any other tax bracket, the after tax rate will be higher.
Figure 1: Before tax 30 year fixed mortgage rate (blue, left scale), and after tax (tan, left scale), all in percent; maximum marginal tax rate (black, right scale, %). Source: Fannie Mae via FRED MORTGAGE30US, Treasury via FRED, author’s calculations.
Now, consider the 15 year after tax mortgage rate. I only have one measure for 15 year expected inflation (the Cleveland measure). I proxy the 15 year with the 10 year measures from Treasury-TIPS breakevens, adjusted breakevens, and SPF 10 year median. This gives the following picture.
Figure 2: 15 year after tax mortgage rates adjusted by Survey of Professional Forecasters median 10 year expected inflation (blue +), adjusted by Treasury-TIPS 10 year breakeven (tan), adjusted by Treasury -TIPS 10 year breakeven adjusted for risk, liquidity premia (green), and adjusted by Cleveland Fed 15 year inflation forecast (red). Source: For mortgage rates, Fannie Mae via FRED series MORTGAGE15US; Treasury and TIPS from FRED (GS5, FII5), for adjusted breakeven KWW (accessed 4/8), for SPF Philadelphia Fed, for 15 year forecasted inflation Cleveland Fed, and NBER.
For 30 year mortgages, we can match exactly using 30 year breakevens, and 30 year Cleveland Fed forecasts.
Figure 3: 30 year after tax mortgage rates adjusted by adjusted by Treasury-TIPS 30 year breakeven (tan), and adjusted by Cleveland Fed 30 year inflation forecast (red). Source: For mortgage rates, Fannie Mae via FRED series MORTGAGE30US; Treasury and TIPS from FRED (GS30, FII30), for adjusted breakeven KWW (accessed 4/8), for 30 year forecasted inflation Cleveland Fed, and NBER.
If it is not clear to the casual observer, I will note that the real mortgage rate, even after assuming the highest marginal tax rate which yields the lowest nominal rate, is now above zero.
Reader JohnH asserts that using one year lagged inflation is appropriate. I document that the lagged one year CPI inflation rate is a poor predictor for subsequent five year inflation; it is likely even worse for the subsequent 15 or 30 year inflation.
Finally, the assertion that rates change quickly so that monthly characterizations are not apt is not validated. In Figure 4, I show weekly nominal pre-tax mortgage rates, and the inflation expectation adjusted weekly rates. No drastic swings are exhibited.
Figure 4: Nominal 30 year fixed mortgage rate, weekly ending Thursday (black), and nominal yield adjusted by maximum tax rate, and subtracting off 30 year inflation breakeven (blue), both in %. Source: Fannie Mae via FRED, Treasury via FRED, and author’s calculations.
Hence, I conclude: ex ante mortgage rates are currently positive.
