Thursday, June 21, 2018

Upside Down CAPM: Part 5 - Returns on real estate investments

To review: upside down CAPM is the idea that CAPM models should start at the expected rate of return on diversified at-risk capital and the rate of return on risk free savings is the at-risk rate of return minus the premium savers pay for the service of protecting savings for deferred consumption.  Changes in actual returns for at-risk capital come from cyclical changes in profit and from changes in expected long term real returns, but expected real returns at any point in time tend to remain fairly stable at 7-8%.  Fluctuations in real risk free returns come from shifts in the premium savers are willing to pay to avoid risk with deferred consumption.  The equity risk premium doesn't come from changes in expected at-risk returns. It comes mostly from changes in the premium for safety that determines the risk free rate.

This model has led me to realize that I have probably been approaching real estate markets with some misplaced hubris.

Mortgage rates generally follow risk free rates with a small spread, so that mortgages basically fall in the category of low risk saving - deferred consumption.  To the extent that we can measure returns on home equity systematically, real rates of return seem to generally follow risk free rates. Owned home equity is also basically a low risk fixed income security.

But, that doesn't seem to be the case on institutionally owned real estate.  Equity REITs seem to have a fairly stable cap rate and dividend rate.  I had chalked this up to an undeveloped marketplace, where there aren't aggregate market measures that change everyday like in bond markets.  So it seemed like this asset class was like a fixed income asset class, but with a real yield that didn't fluctuate as much.  That didn't seem rational.

However, thinking with an Upside down CAPM framework, having a more stable expected rate of return is a characteristic of equity.  There isn't a mystery here. Institutionally owned real estate is basically a low-beta equity, and it has a low and stable expected rate of return, just as we would expect low beta equity to have.  Its risk comes from cyclical and long term shifts in real returns.

Keep in mind that I am talking about the entire market, so that cap rates are similar to expected returns on all equities, conceptually.  For developers, expected returns on individual projects, or even on all new projects at a given time, may fluctuate more, just as the market for IPOs or private equity seems more volatile and cyclical than the equity market as a whole.

The reason it is more equity-like than owned real estate is because it comes with management costs and vacancies.  This makes net margins lower and more cyclical than owner-occupied real estate.

So owner occupied real estate acts more like low risk fixed income, and the level of debt on a property is mostly related to the owner's need for capital.  Young owners tend to be more leveraged and older owners are less leveraged. Investor owned real estate is like low beta equity and the use of credit is more a product of the market and the property. To the extent that a property can offer the service of safe savings because of relatively certain cash flows, equity holders can optimize profit by using credit as a source of capital.  In fact, market forces will cause profit to be bid down to the point where real estate equity holders have to use credit to achieve a market rate of return.  This is similar to the utilities sector.

This would mean that real long term interest rates would moderate real estate activity between owned and rented properties.  When interest rates are low, at the margin some households would be induced to ownership. There are several ways to think about this. Low real rates reflect demand for low risk savings and deferred consumption. Owner equity serves this function in a way that investor owned equity doesn't. Another way to think about it is to think in terms of cap rates. Cap rates on investor owned properties, being more equity-like, remain stable.  But "cap rates" on owner occupied properties fluctuate with long term real interest rates, so as rates decline, owner-occupied property is worth more than investor-owned property.

It happens that homeownership rates (controlling for age demographics) were high in the late 70s and the 2000s.  In the 70s, inflation (and nominal rates) was high, so rising ownership was explained as an inflation hedge.  In the 2000s, inflation (and nominal rates) was low, so rising ownership was explained as a result of cheap and easy credit.  (Most of the rise in homeownership happened in the late 90s when rates weren't particularly low, so that's not a great explanation.)

Low long term real rates can explain why price/rent ratios were high at both times, but I haven't felt that comfortable explaining why that would lead to higher ownership rates.  Maybe this relationship between owned and rented properties, and the difference between equity and low risk savings, could be part of the explanation for why ownership rates rise when long term real interest rates decline.  (This isn't the case today because housing markets are dominated by credit repression which keeps millions of households out of the market.)

This framework also suggests that multi-unit building should have been much stronger in the 1990s when long term real interest rates were relatively high.  The bias toward equity exposure at the time should have translated to a bias for investor owned real estate.  Maybe some of the increase in homeownership rates in the 1990s was already coming from limits on multi-unit developments in the Closed Access markets, which was blocking this natural shift to rented property when real interest rates were high.

This is all still speculative.  Please provide conceptual or factual criticisms in the comments.

1 comment: