Housing: Part 196 - Observations on the GSEs as a housing subsidy

The GSEs (Fannie & Freddie) are frequently blamed for feeding the housing bubble.  This is strange, since their behavior was countercyclical.  When home prices were rising the fastest and were at their highest, GSE growth was slowing.  From 2003 to 2006, total mortgages outstanding that were issued through the GSE conduit only grew at an annual rate of about 5%.  And total GSE book of business plus private MBS held by the GSEs only grew at 6% annually over that time.  From the end of 2003 to the end of 2006, they lost a fifth of their market share.  That's not even in terms of originations - that's based on mortgages outstanding.



Source

Another cited source of excess from the GSEs is the lower interest rate that is facilitated by the low cost of debt that comes from the government's guarantee - explicit or implicit.  Before the crisis, this amounted to about 0.25% if we compare it to the typical spread between conforming loan rates and jumbo rates.

Let's compare this to income tax benefits to homeowners, which, according to the White House, amounted to about $218 billion in 2015, including tax savings from imputed rent, mortgage interest, capital gains, and property tax deductions.

About $5 trillion in GSE facilitated mortgages were outstanding in 2015.  A rate reduction of 0.25% on $5 trillion is equal to $12.5 billion - roughly 6% of the value of income tax benefits.

In 2015, net income to homeowners after expenses and depreciation but before interest expense was $748 billion.  Comparing these numbers, as a first estimate, income tax benefits were equal to 29% of the net income value of owned homes.  The GSE subsidy amounts to less than 2% of the net income value of owned homes.

When I began this series, which I thought would last maybe a dozen posts or so, that tax effect was the question I was trying to answer.  I ended up chasing this topic down several other rabbit holes, although ironically, I think I have been led back to a confirmation of the distortions created by those income tax benefits.  That confirmation itself backs up the other findings I ended up with.

The owner-occupier from much of the bottom tier of the housing market was effectively locked out of the homebuyer market from 2007 on.  We can basically use the 1st quintile of zip codes, by home price, as a proxy for housing markets that lost their owner-occupier demand, and the 5th quintile as a proxy for housing markets that didn't lose their owner-occupier demand.

Across the board, prices diverged at that point.  The market bottomed in 2012, and since then prices have rebounded in the new normal.  It seems that, relative to 1999, prices in the 1st quintile are tracking about 10% to 20% below 5th quintile prices.

This is probably a very low estimate of the effect of income tax policy on home prices, because 1st quintile homes never internalize much of the potential tax benefits.  Looking across zip codes at Price/Rent ratios, even in 1999, P/R in 1st quintile zip codes was much lower than in 5th quintile zip codes.

Between the collapse of private securitizations and tightening standards at the GSEs, the market for mortgages in the bottom tiers of the housing market dried up in 2007.  The gap of more than 10% between the value of a home for a landlord and the value for an owner-occupier meant that we created a context where the bottom had to drop out of home prices at the low end before a new equilibrium could be found.  We killed the mortgage market and that created the drop in home prices at the low end - not the other way around.

One other point here, looking back at the jumbo spread graph:  The shock in 2007 was not limited to private subprime securitizations.  There was a shock in jumbo spreads at the same time.  This is because it was a liquidity shock.  Not a liquidity shock from a lack of cash, per se.  But, a liquidity shock from a lack of mortgage borrowers.  And the reason is that the country lost faith in the housing market.  Buyers and lenders were convinced that home prices were about to plunge in an unprecedented way, and the Fed confirmed that they intended to let this happen.

This caused credit risk to shoot up across buyers, because the credit risk wasn't coming from buyer characteristics.  It was coming from housing expectations.  Thus jumbo loan markets were just as stressed as subprime.

And, the one set of institutions that were immune to this shock were the federal agencies - the GSEs and FHA - since their investors didn't take on credit risk.  Since the GSEs were semi-private, they were derided as being greedy and reckless.  Since FHA was public, they were able to go about the business of supporting the mortgage market, and continued to expand.