But, as much as I applaud how each of those pieces pushes back against both the idea that the Fed was accommodative in 2007 and 2008 and the idea that QE operated through the reduction of long term interest rates, I still see a tendency to fall back into some of the same assumptions. From Thornton's piece:
If the Fed distorted asset prices between June and December 2003, one can only imagine what the FOMC’s zero-interest-rate policy over the period December 2008 to the present has done. And, of course, Kohn is correct: the intention of the policy is to distort asset prices in an attempt to reduce long-term yields. But such actions produce unintended distortions: a strong and persistent rise in equity prices, a marked change in the behavior of commodity prices, a resurgence in house prices and residential construction beyond what is warranted by economic fundamentals, and excessive risk taking, even by those who are least well situated to take it. These are the unintended consequences of QE and the FOMC’s zero-interest-rate policy.I have written extensively in my ongoing series about how home prices were not high in 2003 because credit or artificially low short term rates had pushed them above levels justified by fundamental valuations. I am pleased to see momentum in opinions about monetary policy that points out the hawkish policy decisions they were making before and during the financial crisis. But, the idea that home prices were unjustified is so powerful and so palpable, that even in the midst of these questions, questioners tend to accept whatever assumptions are required to conclude that the housing and equity markets were overvalued because of monetary policy. Thornton, for instance, in his article, sharply questions the idea that low long interest rates reflect expectations of loose policy that creates persistently low short term rates. I agree with him that that model seems incoherent. But, when it comes to explaining rising equity or housing markets, he seems surprisingly willing to accept that model.
I understand that this is difficult. The idea of a housing bubble seems unassailable. To attempt to argue against it would be reputationally risky - certainly too risky to entertain the idea in a sort of back-of-the-envelope initial line of questioning. I believe that the reason I might get to be a voice in popularizing the idea that there wasn't a bubble may be because I have done such a smashing job of avoiding having any reputation to ruin.
So, in a way, the intellectual barriers to my narrative are strong. Yet, on the other hand, Clougherty, Selgin, and Thornton have already accepted the fundamentals of my narrative in other contexts, so they don't need to be convinced to change their fundamental beliefs; they just need to be convinced to apply them more thoroughly.
I have plenty of posts in the housing series that go into the details of housing valuations. In this post, I want to mention equities, because the same sort of assumptions seem to be feeding ideas about equities. Most observers seem to attribute rising equity prices to monetary accommodation, credit expansion, and "reaching for yield".
Most analysis of equity pricing behavior uses price movements or total returns over a period of time. I think this conflates income, capital gains, expected returns, and real shocks, and leads to a lot of confused analysis. I believe that if we look only at expected (or "required") returns on equities, we find a surprisingly stable real implied yield on corporate equity over long periods of time. Aswath Damodaran at NYU has developed an extensive data set of variables that allow us to analyze equity values back to 1960.
The Capital Asset Pricing Model gives us the following very simple equation for valuing equities:
For returns on bonds, all we need to know is the coupon, because cash flows are fixed and expected inflation is paid as a portion of the interest rate. (A 5% bond might be earning, say, a 2% real yield plus a 3% inflation expectation.) But, equities (and homes) are real assets, so cash flows rise over time with inflation or real growth. This means that to estimate the equity risk premium, we need to estimate the expected growth in cash flows. For equities over the long term, this is the expected growth of net profits.
One problem we have with this question is that we have only had market rates on real bonds for about 20 years, so before the late 1990s, it is very difficult to separate inflation expectations from real growth expectations. But, to the extent that we have data, I interpret Damodaran's work to be turning the CAPM on its head. Real expected returns seem to be very stable. Most of the changes in equity prices come from real shocks to corporate earnings and changes in growth expectations.
This means that "reaching for yield" does not describe market behavior. And it means that if accommodative monetary policy is pushing market prices up, this is probably a sign that policy had been too tight, and accommodation is improving the broad prospects of the real economy.
Here is a graph comparing the required returns to equity (ERP + risk free rate) to treasury bond rates since 1961. There are basically two eras here. Before the mid-1990s, real rates fluctuated somewhat, but they were swamped by fluctuations in inflation. This makes it appear as if
Since the late 1990s, we do have markets for real risk free rates. But, inflation had settled down enough by the late 1980s that we can get a pretty good estimate of real rates back to at least 1990, simply by subtracting inflation from nominal rates. Here I have used the GDP deflator as the inflation proxy. So, we have the era before 1990 where fluctuations in inflation dominated, and we have the era since 1990, where inflation has been fairly stable and real risk free rates have dominated. This means that during this period, if required total returns to equity are, indeed, stable, we should see a negative relationship between ERP and risk free rates.
Here is a scatterplot of ERP and 10 year risk free rates (minus inflation) since 1989, together with the scatterplot of ERP with the market rate on real 30 year bonds since 1999. Here we see this strong inverse relationship.
Since 2008, Damodaran has estimated ERP on a monthly scale. Here is a graph of monthly total required returns on equities (risk free rate + ERP). After the disastrous events of September 2008, there was a brief bump in required returns to about 10%, which subsided back to the stable 8% nominal level by the summer of 2009. So, the brief deep plunge of equity values that bottomed out in March 2009 was part of a brief shock away from the typical required level of returns. But, later in 2009, required returns simply moved back to their normal range, and none of the subsequent gains had anything to do with "reaching for yield". They have mostly reflected real recovery of nominal spending and corporate earnings.
Here are scatterplots, using market rates on TIPS bonds, of monthly ERP since September 2009, compared to both 10 year and 30 year real interest rates. Again, we see a strong inverse relationship.
This suggests that changes in long term real interest rates have little or no effect on equity prices. The ramifications of this are stark. This means that gains in equity markets should be taken as a sign of optimal monetary policy. To the extent that policy makers are calling on monetary policy to tighten up in order to rein in "overheated" asset markets, they are literally calling for monetary policy to tighten up until it has damaged broad, real economic growth. Loose money or low rates do not cause equity markets to "overheat". Even if monetary policy is inflationary, this will not lead to persistent increases in equity values, because the stable required returns on equities are real returns, not nominal returns, so equity prices should be unaffected by inflation expectations in the medium to long term. And, clearly, when monetary policy was overly inflationary in the 1970s, stock prices were tempered.
This also speaks against the idea that low interest rates are related to leveraged corporate risk-taking. If corporate assets have a stable discount rate, then low real interest rates should be taken as a much stronger signal of risk aversion - not risk-taking. And, I do believe that this error in interpretation is one of the elements that led to the misidentification of high home prices with speculative over-reach.
Home prices do behave differently than equities, though. Cash flows and discount rates on homes are more similar to real long term bonds, so home values can move up when long term real interest rates are low. In my next post, I will return to that topic, with more of my analysis of how even that effect would be mitigated in the absence of the supply constraints which were the primary cause of home price appreciation in the 2000s.
PS: These are the scatterplots above, using 1 year or 1 month changes in ERP and real risk free rates unstead of levels. ERP is a necessarily noisy estimate, so first differences have low correlations, but the coefficients are still surprisingly strong.
PPS. On second thought, these scatterplots of the first differences may be adding more heat than light, since inertia in equity prices, and earnings and growth estimates, would naturally lead to an inverse relationship in the short term.