Monetary Policy and House Prices: Interest Rates and Timing

“The beginning of the run-up in housing prices predates the period of highly accommodative monetary policy.”  Bernanke January 2010

“Economists who have investigated the issue have generally found that, based on historical relationships, only a small portion of the increase in house prices earlier this decade can be attributed to the stance of U.S. monetary policy.  This conclusion has been reached using both econometric models and purely statistical analyses that make no use of economic theory.”  Bernanke 2010

At the AEA meetings earlier this month Chairman Bernanke gave a speech disputing the Fed’s role in the recent sharp rise in housing prices.  His arguments ran along three key lines:  the timing of policy accommodation and the run-up in house prices do not match, models of interest rates and house prices imply little or no relationship between the two variables, and the international relationship between house prices and interest rates is inconclusive.  The last point I addressed in a previous post (here).  In this post, I will address the first two points. 

Timing

On timing, Bernanke’s main defense is that the beginning of the run-up occurred well before monetary policy became especially accommodative.  While house prices did start to rise in the late 1990s, house price appreciation accelerated sharply in late 2001.  The cumulative rise up to that point was consistent with previous U.S. house price cycles. 

The graph below shows 5-year cumulative real house price appreciation between 1980 and 2009.  The price series is the FHFA house price index deflated by the CPI.  On this basis, house prices began to increase in 1997.  That is in real terms in 1997, real house prices had finally recovered to their 1992 levels.  Between 1995 and 2000, house prices increased slightly more than 10 percent, or an average increase of about 2 percent per year.  High growth but no bubble. 

The Fed began its easing cycle in early 2001.  By the end of that year, the Fed funds rate breached 2 percent, hitting its lowest level since 1960.  By any measure, monetary policy was loose.  But policy was especially loose when compared to the advice given by a Taylor Rule.  Take a look at this picture from Bernanke’s speech.  Whether using backwards looking data or real-time forecasts, the Taylor Rule indicates an increase in policy rates by the end of 2001.  Appropriately, the Fed lowered interest rates substantially following the attacks of 9/11.  The Fed was responding to risks posed by the attacks rather than to incoming economic data. 

The timing of the lower interest rates corresponds to a sharp increase in the growth rate of real house prices.  Beginning in 2001, the five-year growth rate of real house prices rose from 13 percent (the red horizontal line) to a maximum of over thirty percent, a rate more than double its previous maximum. 

Timing does not eliminate the Fed’s culpability. 

Interest Rates and House Prices

“Economists who have investigated the issue have generally found that, based on historical relationships, only a small portion of the increase in house prices earlier this decade can be attributed to the stance of U.S. monetary policy.  This conclusion has been reached using both econometric models and purely statistical analyses that make no use of economic theory.” 

Notice, the Chairman slipped quietly from economic theory guiding the choice of Taylor Rule in the previous section to statistical analysis in this section.  This switch seems odd given the robust theoretical relationship between monetary policy and house prices.  If monetary policy controls the real interest rate, then lower interest rates translate directly into higher long-term house prices. 

To test this thought, I plug the path of interest rates into a completely standard model of house prices and consumption.  (If you want details, post a comment or email me.)  Using data on the housing stock, I find the change in interest rates accounts for most of the change in house prices over the period in question.  One cannot dismiss the possibility that looser monetary policy led to the increase in house prices.  And, it’s this point in its many guises that lead many critics to blame Fed policy. 

However, the same standard model of housing does a very poor job of matching house prices between 1980 and 2000.  The historical relationship between house prices and real interest rates in the United States is quite weak.  Take a look at the longer picture below.  Mortgage interest rates fell from 18 percent in 1981 to less than 10 percent in 1987.  The cumulative rise in house prices was less than 10 percent.  Likewise, between 1990 and 1995, interest rates fell about 2½ percentage points; house prices were about unchanged, on balance, over this period.  I would also note here that VAR models are also unsuccessful in reproducing, out of sample, house prices in any 10 year period between 1980 and 2009. 

A recent paper (here), builds a statistical model that explains the discrepancy.  They show that the interest rate negatively covaries with the risk premium on housing.  In other words, falling interest rates are always exactly offset by some other factor.  The authors give no reasonable explanation of the negative covariance. 

Nonetheless, across countries and across time, house prices have risen when long-term rates are low.  A 2005 Fed Study found a strong, lagging relationship between interest rates and house prices.  Take a look at Chart 3.7 in the linked paper.  There, a decline of about 1 percentage point was associated with a 15 percentage point increase in real house prices, on average.

So, the evidence is mixed:  theory is clear, empirics less so.  Lower interest rates should raise house prices and the lower rates observed in the 2000’s are exactly consistent with the rise in house prices.  But, their historical relationship in the United States belies this effect.  The historical relationship is why the VARs reject policies role in the house price run-up.  I wonder what a cross-country VAR would produce. 

But the Fed believed in the link between housing and interest rates

The Fed tried to bring down long-term interest rates.  During the period of loose policy, the Fed had the specific intention of pushing down long-term interest rates by promising to keep its policy rate low for an extended period.  Even when the Fed began to raise rates in 2004, it emphasized that the increases would be measured (their word, generally taken to mean a slow increase over a long period of time).  As we saw above, this policy seems to have been at least partially successful. 

The Fed also believed in the link between the movements in long-term interest rates and the increase in house prices.  Here is the Fed speaking in the voice of then Gov. Bernanke:

The decision to purchase a home is probably the most interest-sensitive decision made by households … I expect residential investment to continue strong this year. Mortgage rates have risen in the past month but remain low relative to historical experience, while new household formation, improved job prospects, and income growth should ensure a continued healthy demand for housing.  Bernanke April 22, 2004  

These two ideas lead me to believe that the Fed thought it was influencing the housing market. 

My Last Word

I don’t know if loose monetary policy caused the run-up in house prices.  My instincts, however, continue to tell me that monetary policy is not the culprit.  (I tend not to ascribe so much power to monetary policy.) 

Nonetheless, the circumstantial case is somewhat persuasive. 

The Fed was loose; interest rates fell; house prices rose. 

Simply dismissing these facts is not productive.  I would expect a more robust