Thursday, October 15, 2009

More Fed Cheerleading

The battle for the narrative of the Fed actions in the early-to-mid 2000s continues. The latest salvo comes from a blog post by David Altig of the Atlanta Fed and a Cato Policy Analysis piece by Jagadeesh Gokhale and Peter Van Doren. In both cases the authors absolve the Fed of any wronging doing during the housing boom period. I am not surprised to see Fed cheerleading coming from a Fed insider, but from the CATO institute? These are strange times.

In the first article, Altig conveniently finds a modified form of the Taylor Rule that shows the Fed acted no differently than it had in past 20+ years when monetary policy seemingly worked fine. The first problem with this piece is the obvious problem of data-mining a modified Taylor Rule that justifies ex-post his employers actions. If Altig really wants to be convincing, he needs to explain why the original Taylor Rule, which does show the Fed being unusually accommodative during the housing boom, is suspect and why his modified Taylor Rule is better. As John Taylor has shown, the original Taylor rule goes a long way in explaining this crisis. For example, Taylor shows in the figure below that deviations from the Taylor rule in Europe were closely associated with changes in residential investments during the housing boom there (click on figure to enlarge):

Even if Altig could show that his modified Taylor rule makes more sense, there is still the question of whether monetary policy was truly optimal during the previous 20+ years to the housing boom. This was the period of the Great Moderation--a time of reduced macroeconomic volatility--whose appearance has been attributed, in part, to improved monetary policy. As many observers have noted, though, this also was a period of the Fed asymmetrically responding to swings in asset prices. Asset prices were allowed to soar to dizzying heights and always cushioned on the way down with an easing of monetary policy. This behavior by the Fed appears in retrospect to have caused observers to underestimate aggregate risk and become complacent. It also probably contributed to the increased appetite for the debt during this time. To the extent these developments were part of the reason for the decline in macroeconomic volatility, the Great Moderation and the monetary policy behind it becomes less of a success story.

In the second article Gokhale and Van Doren make the following arguments: (1) detecting asset bubbles is a difficult thing to do; (2) even if the Fed could have detected and popped the asset bubble in the housing market in the early-to-mid 2000s it would have done so at the expense of a painful deflation; and (3) the Fed's ability to reign in home prices was limited. On (1) I agree that responding to an asset bubble after it has formed is challenging. But that is not the the point of most observers who find fault with the Fed during this time. They would say the Fed could have prevented the housing boom from emerging in the first place had monetary policy started tightening before June 2004. On (2) the authors still think the deflationary pressures of that time were the result of a weakened economy. This is simply not the case. As I just recently noted on this blog (here and here), rapid productivity gains were the source of the deflationary pressures, not declining aggregate demand. In fact, by 2003 nominal spending was soaring at a rapid pace. In other words, the deflationary pressures of 2003 were vastly different than the deflationary pressures of 2009. On (3) the authors claim that there was simply no way for the Fed to reign in home prices since the influence of its target federal funds rate on other interest rates declined during the time of the housing boom. While it is true the link between monetary policy and long-term interest rates is more tenuous, the authors argue that even interest rates on ARMs and other subprime-type mortgages were beyond the Fed's influence. A CATO Policy Briefing by Lawrence H. White, however, provides evidence that the supbrime market was in fact very sensitive to the Fed's action during this time. Below is figure that corroborates White's work by showing the effective interest rates on ARM mortgages along with the federal funds rate. Is there any doubt? (Click on figure to enlarge):

Update: To support my claim that nominal spending was soaring by 2003 I have posted a figure below that shows the growth rate of domestic demand relative to the federal funds rate since 2002. The years 2003 to 2004 are marked off by the dashed lines. Note that the growth rate of nominal spending is increasing during the 2003-2004 period while interest rates are kept low for most of the period (click on figure to enlarge):


  1. You could also reference the important empirical work of Tobias Adrian at the NY Fed. He shows a strong relation between fed funds rate and the growth of dealer balance sheets which were used to short-term refinance the bank SIVS and conduits that held the mortgage securities

  2. During the period in question,nominal expenditure was below its 5% trend growth path.

    Low short term interest rates during some period imply slightly lower long term interest rates. Lower long term interest rates make the use of long lived real assets less costly and more profitable. More profit from long lived assets results in more rapid growth in demand and more rapid increase in prices--for a time.

    It doesn't make loans secured by overvalued assets made to people who can't make the payments profitable. I think the problem was entrepreneurial error--the view that overvalued assets were really undervalued. That way, people who can't make the payments aren't bad credit risks because their collateral will soon be greater than the value of the loan.

    I don't doubt that time preference, population growth, and productivity growth are important determinants of the natural interest rate. However, all of these pass through the minds of human beings. The natural interest rate, like any price, must coordinate the actions of real people, rather than reflect some platonic underlying reality.

  3. "Asset prices were allowed to soar to dizzying heights and always cushioned on the way down with an easing of monetary policy."

    This seems to me to offer a solution to the Fed's difficulty in identifying bubbles. Don't try. Simply lean against asset prices, without regard to whether there is a bubble. The logic for this is that asset prices are part of the credit channel, and the Fed needs to take more account of the credit channel in policy making. The Fed has recognized the risks to the economy from declines in asset prices, probably because asset prices tend to fall faster than they rise - there's that nasty credit channel at work. However, when asset prices rise much faster than inflation or the historic trend or whatever measure the Fed chooses to adopt, that probably spurs too much credit creation, or unbalanced credit creation because holders of the rapidly appreciating asset have a faster increase in collateral than other sectors.

    If the Fed could choose which asset to discriminate against, that would make problems of imbalance easier to deal with. The whole China/US trade-and-credit cycle was driven partly be credit innovations which made houses a convenient form of collateral for borrowing. To the extent that the Fed cannot target asset classes, we do need to recognize that other factors were at work and find mechanisms outside of monetary policy to lean against asset bubbles.

  4. Hi. I agree that "the natural interest rate, like any price, must coordinate the actions of real people, rather than reflect some platonic underlying reality". Unfortunately, this didn't happen. And I am afraid that there are another "housing bubbles" to be expected if the government continues pushing the interest rates down as this is certainly not the natural development of economy.
    All the best,