## House Rich – House Poor

Last week I presented a graphic that illustrates the changing average price of homes by state. This week, I want to illustrate something that is more relevant to affordability. FRED provides data on both median salary and average home prices by state. That means that we can create an affordability index. Consider the equation for nominal growth where i is the percent change in median salary (s), π is the percent change in home price (p), and r is the real percent change in the amount of the average home that the median salary can purchase (h).

(1+i)=(1+π)(1+r)

Indexing the home price and salary to 1 and substituting each the percent change equation (New/Old – 1) into each percent change variable allows us to solve for the current quantity of average housing that can be afforded with the median salary relative to the base period:

h=s/p-1

If h>0, then more of the average house can be purchased by the median salary – let’s vaguely call this housing affordability. Both series are available annually since 1984 through 2021 for all 50 states and the District of Columbia. The map below illustrates affordability across states. Blue reflects less affordable housing and green reflects more affordable housing since 1984.

## Concentrating on Housing

Housing has become more expensive. Below is a figure that illustrates the change in housing prices since 1975 by state. By far the leaders in housing price appreciation are the District of Columbia, California, and Washington. The price of housing in those states has increased about 2,000% – about double the national average. That’s an annualized rate of about 6.7% per year. That’s pretty rapid seeing as the PCE rate of inflation was 3.3% over the same period. It’s more like an investment grade return considering that the S&P has yielded about 10% over the same time period.

## Two Decades of Real Estate Data

Total spending on real estate construction has been rising since 2011. By 2016 it had reached its previous 2006 peak. However, total spending on *residential* real estate construction didn’t reach its previous 2006 peak until November of 2020. The graph below also includes the proportion of residential construction spending (Green). It has been rising since 2009. In and of itself, nothing is good or bad about this figure. We might be spending less on non-residential construction because we are getting better at using less land per unit of good or service produced. Or, it could be that our real investment in future production is falling relative to our current residential consumption.  Regardless, the share of residential construction hasn’t been at this level since 2003.

Importantly, the difference in spending has not been driven by different construction costs. Both residential and non-residential construction costs have moved in tandem since 2010. Therefore, the rise in residential construction spending is not merely nominal – a greater proportion of resources are being consumed by residential construction. Indeed, real residential construction is up about 25% from 2019. The figure below illustrates real residential and nonresidential construction.

That figure requires a double-take.

## Housing & The Fed’s Reputation

I am not worried about inflation and I’m not worried about the total spending in the economy. As I’ve said previously, total spending is on track with the pre-pandemic trend and, I think, that helped us experience the briefest recession in US history. When output growth declines below trend, we face higher prices or lower incomes. The former causes inflation, the latter causes large-scale defaults. Looking at the historical record, I’m for more concerned about the latter.

I do, however, want to call special attention to the composition of the Fed’s balance sheets. Specifically, its Mortgage Backed Security (MBS) assets. Having learned from the 2008 recession, the Fed was very intent on maintaining a stable and liquid housing market. Purchasing MBS is one way that it maintained that stability. Its total MBS holdings almost doubled from March of 2020 to December of 2021 to \$2.6 trillion. Should we be concerned?

At first, a doubling sounds scary. And, anything with the word ‘trillion’ is also scary. Even the graph below looks a little scary. MBS holdings by the Fed jumped and have continued to increase at about a constant rate. Is the housing market just being supported by government financing? What happens when the Fed decides to exit the market?

Luckily for us, there is precedent for Fed MBS tapering. The graph below is in log units and reflects that a similar acceleration in MBS purchases occurred in 2013. Fed net purchases were practically zero by 2015 and total MBS assets owned by the Fed were even falling by 2018. Do you remember the recession that we had in 2013 when the Fed stopped buying more MBS’s? Wasn’t 2018-2019 a rough time for the economy when the Fed started reducing its MBS holdings? No. We experienced a recession in neither 2013 nor 2018. Financial stress was low and RGDP growth was unexceptional.

Although there was no macroeconomic disruption, what about the residential sector performance during those times? Here is a worrisome proposed chain of causation:

1. Relative to a heavier MBS balance sheet, the Fed reducing its holdings increases supply on the MBS market.
2. This means that the return on creating new MBS’s falls (the price rises).
3. A lower return on MBS’s means that there is less demand from the financial sector for new loans from loan originators.
4. A tighter secondary market for mortgages decreases the eagerness with which banks lend to individuals.
5. Fewer loans to individuals puts downward pressure on the demand for houses and on the price of the associated construction materials.

The data fits this story, but without major disruption.

Less eager lenders went hand-in-hand with higher mortgage rates and less residential construction spending. The substitution effect pushed more real-estate lending and spending to the commercial side. Whereas residential spending was almost the same in late 2019 as it was in early 2018, commercial real-estate spending rose 13% over the same time period.

But, importantly in the story, the income effect of a Fed disruption should have been negative, resulting in less total spending and lower construction material prices. And that’s not what happened. Total Construction spending rose and so did construction material prices. Both of these are the opposite of what we would expect if the Fed had caused disruption in the housing construction sector due to its MBS holding changes.   Spending on residential construction fell understandably. But spending on commercial construction and the price of construction materials rose.

My point is that you should not listen to the hysteria.

The Fed has a variety of assets on its balance sheet and it pays special attention to the residential construction sector. Do you think that there is a residential asset bubble? Ok. Now you have to address whether the high prices are due to demand or supply. Do you suspect that the Fed unloading its MBS’s will result a popped bubble and maybe even contagion? It’s ok – you’re allowed to think that. But the most recent example of the Fed doing that didn’t result in either a macroeconomic crisis or substantial disruption in the construction markets.

The Fed has a track record and it has a reputation that serves as valuable information concerning its current and prospective activities. The next time that someone gets hysterical about Fed involvement in the housing sector, ask them what happened last time? Odds are that they don’t know. Maybe that information doesn’t matter for their opinion. You should value their opinion accordingly.