Monday, July 27, 2015

Housing Tax Policy, A Series: Part 47 - The Devastated "Expectations Channel" in 2007 and a note on rent inflation

Bankruptcies of mortgage originators began building up in early 2007 and funds with leveraged investments in subprime-based securities were collapsing throughout 2007.  This was when delinquencies were really only beginning to rise from boom-time levels.  Delinquencies in mid 2007 were in the ballpark of delinquencies in 2001 which was not considered to be a particularly poor housing market.  But, one thing that had begun to happen by mid-2007 was an unprecedented collapse of home prices.

I think everyone has generally assumed that the collapse of hedge funds and investment funds in 2007 was due to the combination of (1) high default rates on collateral and (2) high leverage in the funds that left no room for error.  Default rates were rising at the time, but they weren't outrageously high yet.  But, these were securities with market prices.  Actual cash flows on securitizations from 2006 and 2007 would have been low because of defaults.  But, because the defaults were being triggered by unprecedented widespread nominal drops in home prices instead of by more typical causes of defaults, the effect on securities prices may have been especially pernicious.  Potential buyers of an MBS experiencing high defaults because of a poor labor market would price in an expected recovery.  But, in mid 2007, someone valuing an MBS with high defaults would look at home prices off 10% and accelerating downward, and they may have modeled very high future default rates into their valuations.  Markets are forward looking.  So, even before defaults reached their high levels, the market values of MBS's may have been collapsing because of expectations of the ongoing collapse in home prices.

There has been extensive discussion about how the models used to rate securities backed by subprime loans were flawed because they didn't account for the potential for correlated defaults.  But, I suspect that the collapse in the market values of those securities happened before most of the actual defaults.  And, if that was the case, then an expected recovery in home prices in mid 2007 would have had tremendous benefits.

And, here we reach the circular problem with the crisis.  If we don't assume that home prices required a massive correction, then large-scale monetary and credit market support in 2007 seems reasonable, even obvious.  Markets in 2007, in effect, may have pulled the imminent collapse in home prices back in time from the near future to the present, reflected in the collapsing prices of MBS's, and a change in those expectations would have been significantly helpful.  But, if we do assume that home prices required a massive correction, then that support seems dangerous.  And, in this way, the crisis became self-imposed.

At the August 2007 FOMC meeting, after many bankruptcies and fund collapses, including the Bear Stearns subprime hedge funds, the FOMC statement was still reporting to anyone who might have been valuing an MBS that  "the housing correction is ongoing" and the "Committee's predominant policy concern remains the risk that inflation will fail to moderate as expected."  Home prices were nearly 10% from their peak in nominal terms, already unprecedented in modern American financial experience, the Fed Funds rate was still pegged at its high of 5.25%.  And the 2.1% core CPI inflation that the Fed was worrying about consisted of 1.2% non-shelter inflation and 3.4% shelter (i.e. rent) inflation, which had jumped up in 2006, apparently in reaction to the supply shock when homebuilding collapsed.


The legend is incorrect.  The scale for housing starts is noted on the axis.
If we break out inflation by tenure, I think we get a picture of the problem.  From the mid-1990s until the early 2000s, rent inflation was high.  Renter inflation tended to be higher than owner equivalent inflation (OE Rent).  This reflects the sharp supply constraints we have seen in the core metro areas since the 1980s.  As the recovery in housing starts built up in the late 1990s, multi-unit starts failed to recover as they had in previous decades.  So, as housing expansion recovered, I think what we see in the inflation data is the rising cost of renting driving households from multi-unit housing in the cities to single family homes outside the cities.  In the 2003-2005 period, single family home building finally rose to a level that accommodated all of this housing demand.  So, we still see the pricing pressure in rental housing, but total housing inflation is moderate.

Now, keep in mind, this period does look like a boom in single family homes, because in order to meet aggregate housing demand, the SFH market had to expand enough to accept all of the households who have been locked out of metropolitan multi-unit housing.

In 2006, housing starts collapsed and rent inflation shot up.  This looks like a clear indication of a supply shock, and the trend shifts are sharp.  This also happened to coincide with the inversion of the yield curve, which correlates negatively with bank credit growth and has been a leading indicator, with about a 1 year lag, of recessions.

In 2007 and 2008, housing starts continued to collapse, but now so did rent inflation.  This suggests that demand was suddenly collapsing more strongly than supply. During this period, renter inflation pushed far above owner equivalent inflation as the breakdown of mortgage credit and home ownership markets pushed many households back into rental housing.

Since 2011, we have the worst of all these factors.  Housing starts remain at extremely low levels, and shelter inflation is again somewhat high.  But, extremely low demand is keeping rent inflation from exploding.  If demand recovers, I think we should expect single family home construction to recover with it, keeping owner equivalent rent from jumping.  The significant expansion of a rental market in single family homes is helping to keep renter income low and is probably causing more of the renter inflation to be reflected in owner equivalent rent, but continued supply constraints in the major metropolitan areas are continuing to put upward pressure on renter inflation.


  1. "and the 2.1% core CPI inflation that the Fed was worrying about consisted of 1.2% non-shelter inflation and 3.4% shelter (i.e. rent) inflation"

    Don't recall the exact numbers, but weren't the 10 year TIPS spreads still 2.5%+ at this point in time? I seem to recall that the Fed started raising rates when the 10 year TIPS was approaching 3%, and a low current inflation rate with a normal 10 year rate implies the market thought the recession was going to be minor.

    1. That's a good point. Although, the idea of TIPS spreads as a recessionary signal itself becomes complicated if half of the inflation premium is coming from a negative supply shock.

      While treasury yields were level, breaking them out to look at forward yields shows that forward real yields were declining as the Fed began to raise rates. Now that you have brought this up, I'm thinking that this is another way that the housing supply problem undermined monetary policy. Forward inflation expectation looked healthy, but that is because they incorporate the supply problem. Real estate investors were still bidding up homes in 2004 and 2005 because of expected rent inflation. Why wouldn't they be bidding up TIPS bonds for the same reason?

    2. Oh, here is the post I did on the forward yields.

    3. You end up in some tricky waters there. If the market is killing MBSs because they expect a decline in home prices then there is only a very short window for a rebound where it makes sense to have long term inflation expectations that are based on supply issues keeping rents high but short term price problems that cause MBS to drop in price prior to defaults.

    4. I don't think so. Even as late as early 2008, there is no reason that this had to be anything more than a mild recession with a housing supply problem. We can't expect bond markets to predict discretionary Fed behavior. And the worst Fed discretionary decisions came in the fall of 2008.

      Here are 5 year inflation expectations and some CPI indicators. Shelter inflation had begun to moderate, but was still above 2% going into 2008, but it was offset by a significant rise in food and energy inflation. By July 2008, food and energy expectations peaked, and inflation expectations collapsed. By the fateful September 2008 Fed meeting, Shelter CPI inflation was down to 2% and 5 year inflation expectations were falling quickly, nearly down to 1%.

      A strong support of liquidity could probably have still helped. Real estate delinquencies (single family mortgages booked at commercial banks) had only risen from 1.5-2% to 4-5%. (They would later top out at around 11%). Unemployment was a little over 6%. Even then, after most of the housing price declines were behind us, liquidity might have kept the recession mild. But, there would have been outrage if even then it looked like the Fed was propping up the housing market.

      By the end of October, they were down to minus 1%, briefly falling to minus 2% by the end of November. When the Fed finally announced QE1, inflation expectations quickly rebounded back to around 2%.

  2. that was a run on sentence, I need to word it better.

  3. The Fed should get a lot more blame than it does. When Lehman failed, the Fed actually held the Fed Funds rate at 2%. It could have gone straight to zero or even negative (and it surely didn't need to start paying IOR). That failure to act greatly increased the demand for money making matters even worse. The Fed should have responded to the collapsing inflation expectations. That would have mitigated the housing price moves.