Thursday, November 12, 2015

Housing, A Series: Part 82 - The Components of Home Yields

Today, I'm going to step back a bit and think about the components of the implied yield on home ownership.  Here is a diagram of the basic outline of home yields.

At the core of an efficient home market is the double trend toward no-arbitrage home prices.  In a market with unconstrained supply, the Real Net Yield / Price of homes in the aggregate should move with other similar discount rates.  Over time, this rate of return on homes appears to move roughly in parallel with risk free real long term interest rates, with a small added spread.

We tend to think of the price as the dependent variable in home valuations, but home prices also have a natural arbitrage tendency with the cost of new supply.  This is especially the case for homes where most of the value of the homes are based on the structures, as opposed to the lot or the location.  In that context, home prices are also roughly tethered to the cost of similar new homes.

Falling interest rates correspond to rising Price/Rent levels.  But, in an unconstrained market, with these dual arbitrage forces, prices should not be able to rise.  The rise in Price/Rent would have to come from falling rents.

This is one of the tragedies of the housing boom.  The main effect of lower long term real interest rates on housing should have been a broad decline in rental expenses, which would be especially helpful for lower income households.  This should have been a period where low inflation reduced de facto income inequality and pushed material quality of life higher across the income scale.  But, instead supply constraints in a few cities undermined the arbitrage factor that normally would moderate home prices, so that in those cities, rents remained level and home prices rose.

One way we might think about the problem of housing supply constraints is that as the value of the property moves above replacement value and as location value becomes a larger portion of the total value, the price of that property will become more sensitive to interest rate fluctuations and less sensitive to the moderating influence of new building.  Here is a scatterplot of the largest 20 metro areas, where rent affordability in 1995 is a proxy for supply constraints, and I compare this to the change in home prices from 1996 to 2005.

There is a strong relationship here.  This is just one more way to view the housing boom through the lens of supply constrictions.

We can think through the effects of different changes in an unconstrained environment vs. a constrained environment.

In an unconstrained environment, the median household tends to keep housing expenses at about 20% of income.  So, if we think about the diagram above, falling long term real interest rates would cause both net and gross rents to drop, relative to prices.  But, since the median household in these cities tends to prefer more housing consumption when the budget allows for it, this will have the effect of increasing demand.  So, households will push total rent expenses back up - but not by pushing home prices up.  This will happen through an increase in the quantity of homes "consumed" or lived in.  So, if nominal rent expenses remain a constant proportion of incomes, lower interest rates will cause rents of existing homes to fall, prices in existing homes to remain level, and new homes to be built to accommodate the growing demand.  In fact, housing demand might be a little inelastic.  If households can rent much nicer homes for the same amount of money, relative to other expenditures they have, they may even increase their total housing consumption, in rent terms, as a reaction to lower long term interest rates, even if the first order effect is for rent to decline.  (By the way, this would be capital intensive, and should push up real long term interest rates.  The fact that we have undercut this potential adjustment from the economy since the crisis could be a significant factor in the unusual decline in long term interest rates since 2007.)

Here is a graph comparing Home Price / Median Income in the main closed access cities to the main open access cities.  Very late in the boom, Phoenix did see a bit of a bump.  Even this was related to at least a temporary supply constraint.  But, I think this generally is a pretty stark picture of two housing booms that had little to do with each other.

And, next is a graph of rent affordability in these cities.  Rent inflation was moderate during the boom in Dallas and Atlanta (and Phoenix until late 2005 and 2006), so the stability in rent affordability reflects real increases in housing growth.  Price/Rent levels tended to rise, at least moderately, everywhere, real rents increased, rents on existing homes were falling relative to other prices, and housing starts were strong.

Source
This is the picture of a normal, functioning, open economy that happens to have low real long term interest rates.  If the national statistics were more a reflection of these cities, would anybody be calling this a bubble?  If out of control banks, if a bubble in AAA rated securities, if a decade-long run-up in private-issue securitizations and subprime loans were the source of the housing "bubble" wouldn't we expect to see broadly similar behavior throughout the country?  Yet, what we see are quite normal behaviors in the majority of the country and very extreme behaviors in a few cities that have notorious housing problems.  How did this ever look like anything but a supply problem to anyone?

Notice that rental affordability and rent inflation are both high across the board now.  Nobody lives in an open access housing market now.

Going back to the beginning diagram, let's think about what happens in a closed access city when long term real interest rates decline.  Nominal rents will still remain stable.  But, since there is little new homebuilding, in those cities the Price/Rent ratio increases through rising home prices.  And that is basically what we saw.  The most significant trend was the rising prices and Price/Rent ratios in the problem cities.

Here is a graph of rent inflation relative to core inflation, for the US, the 8 cities from the Case-Shiller 10 index that have CPI rent inflation data for the period (which includes the largest closed access cities), and an estimate of inflation for the US outside those cities.  Rent inflation was high in the closed access cities, but it was declining.  This is partly because of the aftermath of the internet bubble in Silicon Valley.  But, I think partly what we are seeing is the reaction of households to falling rents in open access cities, which was drawing households away from the closed access cities.

Source
The housing boom coincided with net migration out of the closed access cities because low long term real interest rates provide a relative cost of living savings in housing - but that can only happen in open access areas where there is not a constraint on housing supply.  The relative value of housing was improving in the open access areas, so this naturally created a draw from the closed access areas.  This relieved pressure on rents even in the closed access cities.  The fact everyone notices is the liquidity provided by the housing "ATM", but this seems to be rarely noticed.  The housing boom was disinflationary.

Look at rent inflation since 2008.  That is what a demand shock related to the national banking sector looks like, with rents rising in lockstep - except we have created a negative demand shock where there never was a positive demand shock.

Taking one last look back at the first diagram, I want to think about taxes.  When I first started this project, I expected them to be at the center of the story.  They may push owner-occupier prices up by 10% or more.  I still think they are worth thinking about, but their effect pales compared to the supply problem and our misdiagnosis of it.

Tax breaks - the largest of which is the non-taxability of imputed rent - have the same effect as falling interest rates.  In housing, these breaks tend to accrue only to owner-occupiers.  As with lower interest rates, the long run effect of tax breaks is probably mostly to increase housing demand from owner-occupiers, which is accompanied by rising Price/Rent ratios.  I think this might further bifurcate housing into multi-unit rental housing and single-unit owner-occupier housing, because in efficient markets landlords can't compete with owner-occupiers on an after-tax basis.

Property taxes, on the other hand, don't affect Real Net Yields or Price/Net Rent, but increasing property taxes increases gross rents and decreases Price/Gross Rent.  Property taxes will generally increase rent expense for both owners and renters.  This is regressive for two reasons.  First, low income households spend more on housing than high income households.  Second, because low income households tend to be closer to their minimum comfort level of housing expenditures, which is why they spend more of their incomes on rent, they also have more inelastic demand for housing, so they will tend to adjust their housing consumption less in the face of higher taxes.

I think the only realistic way to remove the tax benefit of imputed rent for owner-occupiers is to stop taxing capital income, or at least stop taxing real estate income.  This could be paired with increased property taxes.  Even though property taxes are quite regressive, I think this arrangement would be much less regressive than the arrangement we have today.

With home prices so far below intrinsic value, a switch of this kind now would not be as much of a shock as it normally would be.  And, if we aren't going to allow middle class households to take out mortgages any more, this would level the playing field to encourage continued expansion of the single-unit rental market.

Ps. It may seem like higher property taxes would help solve the closed access city problem by lowering home values.  But, since it would only lower values by reducing housing demand through higher gross rent expense, it would mainly be making the cost problem worse, leading to more stress for low income urban households and more migration between closed access and open access cities.  On the other hand, if higher public revenues through property taxes or development fees could incentivize cities to allow more building, then maybe higher housing taxes could be part of a grand solution.

No comments:

Post a Comment