September 02, 2007
The Taylor Rule and the housing boom
Stanford Professor John Taylor presented another very interesting paper at the Jackson Fed conference.
Among Professor Taylor's many lasting contributions to macroeconomics is development of the Taylor Rule. As with many issues in macroeconomics, Ben Bernanke's exposition (hat tip: David Altig) is as fine as any:
In its most basic version the Taylor rule is an equation that relates the current setting of the federal funds rate to two variables: the level of the output gap (the deviation of output from its full-employment level) and the difference between the inflation rate and the policy committee's preferred inflation rate. Like most feedback policies, the Taylor rule instructs policymakers to "lean against the wind"; for example, when output is above its potential or inflation is above the target, the Taylor rule implies that the federal funds rate should be set above its average level, which (all else being equal) should slow the economy and bring output or inflation back toward the desired range.
A well-known research paper by Richard Clarida, Jordi Gali, and Mark Gertler is one among many that suggested that much of the reduced volatility of U.S. GDP since the mid-1980s might be attributed to an improved tendency of the Fed to follow such a rule. In particular, the argument is that since 1982, the Fed was able to raise the nominal interest rate faster than increases in inflation to help stabilize the economy. In his remarks in Jackson Hole, Professor Taylor noted that housing investment in particular seemed to be more stable after 1982, perhaps due to precisely this improvement in monetary policy.
However, Taylor was concerned that the Fed deviated from this favorable record during the 2003-2005 period, in that it increased the fed funds rate more slowly than his rule would have suggested. I remember (well, Google helped me find it exactly) that Dave Altig had commented on this two years ago:
In his paper at Jackson Hole on Saturday, Professor Taylor considered two paths for the fed funds rate, the actual path (shown as the solid line below) and a counterfactual path (dashed) that Taylor believes would have been closer to the optimal response.
Professor Taylor then estimated a simple model to try to predict housing starts on the basis of past values of interest rates. The solid line in the following diagram displays the actual path for housing starts, which exhibited a phenomenal boom followed by even more dramatic bust. The figure also displays, in the higher, short-dashed curve, housing starts as they would have been predicted by the model using the actual time path chosen for the funds rate by the Federal Reserve. This predicted path is able to account for a boom and subsequent bust in housing, though both are more modest than what was seen in the actual data. Taylor then conducted a counterfactual simulation for what the model would have predicted had the fed funds rate been raised more quickly (the lower long-dashed curve). These counterfactual simulations suggest that, if instead of doing what it did during 2002-2005, the Fed had instead simply adhered to the Taylor Rule, the result would have mitigated both much of the boom in housing starts as well as the subsequent bust.
There was a lot of discussion at the conference about whether the Fed should have leaned against the accelerating real estate prices during 2003-2005. Taylor's simulations suggest that the Fed should perhaps have been thinking of itself as one important cause of that phenomenon in the first place.
Posted by James Hamilton at September 2, 2007 07:09 PMdigg this | reddit
Listed below are links to weblogs that reference The Taylor Rule and the housing boom:
» Tackling financial distress: BIS calls for 'speed limits' from New Economist
Considerable attention has been devoted to the annual Federal Reserve gathering at Jackson Hole this Labor Day weekend. For a good summary see Brad de Long's post I Am Not at the Jackson Lake Lodge This Labor Day Weekend, James Hamilton on The Taylor R... [Read More]
Tracked on September 3, 2007 12:37 PM
Prof. Hamilton, Thanks for your great comments about the Jackson Hole papers. Is it possible to post the web page where the papers are available?
Thanks a lot
Posted by: Pedro at September 2, 2007 07:51 PM
Pedro, the papers are for the most part not yet available-- they may be up at the Kansas City Fed's website on Tuesday.
In the meantime, here's a link from Greg Mankiw to the Mishkin paper:
Posted by: JDH at September 2, 2007 08:12 PM
Its possible to see the feds actions as part of the credit super cycle. My, no doubt, gross over simplification of the dimly remembered "Bank Credit Analyst's" hypothesis is that every time there is some problem in the USA or even the world economy the Fed lowers interest rates. That defers the problem but at the expense of a greater one somewhere further down the track.
The Fed lowered interest rates in response to the to the "Asian crisis" and then to the LTCM "crisis". Even at the time I can recall commentary around our office that those actions would produce a powerful, housing boom in the USA.
Its interesting to note that most forecasts of underlying demand for housing starts (for instance Joint Centre at Harvard, or NAHB), based on household formation (1.35 m) plus teardown plus unoccupied home place demand are around 2m. Over the past 7 years there have been 12.6 m houses for an average of 1.8 m and a peak in 2005 of 2.07m. Those numbers exclude mobile homes. Despite starts being below underlying demand there is something like 0.5m excess of supply of houses.
If interest rates were set based on whether housing starts were above or below underlying demand where would they be right now?
Posted by: David Leitch at September 3, 2007 01:22 AM
Great stuff. I'm sorry that David Altig's chart didn't go back further, since lots of analysts think that the Fed made the same mistake with the stock market in 1998-99 that it perhaps made with housing in 2002-2005. In the Fall of 1999, though, I wonder if a simple analysis of the FF rate would show the full impact of the astonishing amount of liquidity that the Fed term-repo'd into the market to forestall a Y2K liquidity crunch that never transpired? [As an aside, my understanding is that the Fed first loosened its rules on the types of securities it could and could not purchase to prepare for dealing with the Y2K crunch and then kept the looser rules in place 'caused they liked them].
Greenspan was a great head of the Fed and it's a hard hard job - but the 1998-99 episode and the 2002-05 episode are parts of his legacy that he'd probably like to forget.
Btw, I've worked at firms that have both subscribed and not subscribed to the Bank Credit Analyst, and I don't think that either their "Supercycle" theories or their economic forecasts have been all that useful. And for a typically inconsistent forecasting firm, they seem to lack the appropriate amount of humility.
PS: A good friend of mine in the investing business used to claim that it was critically important for stock market strategists and economic forecasters to have great jokes and deep senses of humor, 'cause if they didn't they had absolutely no economic utility whatsoever.
Posted by: Anarchus at September 3, 2007 07:13 AM
Thanks for sharing papers like this with us.
Taylor's chart deals with housing starts. Is it possible to get something similar with prices?
Also, here is a link to a new article on News Week by the same journalist whose "bubbles are great" view was echoed by Gramlich.
That supports the view that subprime is not contained--that it is not a subprime issue per se, but a bubble based on an easy credit.
Posted by: TedK at September 3, 2007 07:18 AM
Some notes on co-incident history.
In 2004, we faced a US Federal Election. What this evidence suggests is that Saint Alan Greenspan kept the Federal Funds Rate lower than prudent policy may have indicated until George W. Bush was re-elected.
The whole point of having an independent institution to run monetary policy is to minimize the degree to which short term political considerations dictate policy action. Not clear it's working.
Posted by: Paul G. Brown at September 3, 2007 10:21 AM
PGB: "The whole point of having an independent institution to run monetary policy is to minimize the degree to which short term political considerations dictate policy action. Not clear it's working."
How can the FedRes be considered independent(of the Gov't)?
I would think, at the minimum, that if it was, it would have no need of "legal tender" laws.
Also, Could someone eplain why the U.S. Economy even needs the "Federal Reserve"?
Posted by: Mark E Hoffer at September 3, 2007 11:22 AM
Mark E. Hoffer -
Answers to your questions are to be found in any decent history of money and monetary policy. I suggest you start with Wikipedia.
In a nutshell, and as in so many situations, the basic answer is that we have institutions like the Federal Reserve because the alternative (not having a central fiat money authority) is far worse.
Posted by: Paul G. Brown at September 3, 2007 03:36 PM
I wish I knew why John Taylor's version of the Taylor rule produces such a different path for the federal funds rate than Dave Altig's version of the Taylor rule. Without having read Prof. Taylor's paper, I'm rather skeptical as to whether the Fed would have come up with a similar path if it had been applying the rule in real time. With 20/20 hindsight it's easy to come up with a set of parameters that might have produced good results in the past.
Posted by: knzn at September 3, 2007 05:06 PM
This is all interesting, but I wonder that the housing bubble can be discussed without mentioning the exascerbating effect of foreign central banks (esp China) in lowering long-term interest rates, which are after all more closely related to mortgage rates.
Posted by: algernon at September 3, 2007 05:47 PM
On one hand you think that the FedRes, via its control of monetary policy, influenced the '04 election cycle, yet, on the other hand, you still think that the alternative to "a central fiat money authority is far worse" (?)
Isn't our current currency, the Federal Reserve Note, nothing but an evidence of debt?
Maybe we should understand the difference between Money and Currency?
Posted by: Mark E Hoffer at September 4, 2007 05:37 AM
MEH: "Isn't our current currency, the Federal Reserve Note, nothing but an evidence of debt?"
The "debt" status of Federal Reserve notes (and deposits at the Federal Reserve, which have the same status but are more important) is an accounting fiction, because nobody expects those debts to be repaid. (Instead they will be rolled over forever.) Technically, when the Fed purchases Treasury securities, it is lending money to the Treasury, but for practical purposes, it is just giving newly created money to the Treasury, that is, paying part of the government's expenses by "printing" money.
Posted by: knzn at September 4, 2007 07:37 AM
I might imagine that part of the reason that the Taylor Taylor rule and the Altig Taylor rule produce different results is that the Taylor version feeds back the simulated results of earlier changes (whereas the Altig version, I presume, just plugs in actual data). That might help explain why the divergence is able to get wider over time: because both rules are smoothing interest rates, but the path of those rates is different, so the two paths have little tendency to revert toward each other. But I still don't understand why they diverge in the first place: the Altig version actually has Taylor rule rates below the actual in early 2002, while the Taylor version has rates above the actual at the time they diverge.
Posted by: knzn at September 4, 2007 08:37 AM
knzn, Taylor used a coefficient of 1.5 on inflation and 0.5 on output in the equation determining the fed funds rate. In his remarks he described the counterfactual path as "one example" of how a Taylor Rule might work. I'm not sure what coefficients Altig was using.
Might be a good topic for somebody (your blog?) to try to sort these out.
Posted by: JDH at September 4, 2007 09:16 AM
Paul and Mark,
A friend of mine asked the question since the FED is the only institution with any control over money why has currency policy been so abysmal? The answer that is obviously because FED policy has been abysmal.
I do not believe necessarily that the FED should not exist but its original intention was to be the lender of last resort. Today it has become the lender of first resort effectively distorting the value of money up and down, up and down without even knowing what target they are attempting to hit. The FED attempts to manage money long before money needs managing and then they have no bench mark to even manage the money to.
The best success the FED has had since its creation has been in periods where it was virtually inactive. Notice how successful Greenspan was when he first took over at the FED and compare that success to the number of changes in the FFR.
Often the FED creates a mess because they are afraid that if they do not do something people will think they are irrelevant.
Posted by: DickF at September 4, 2007 10:49 AM
So the Fed should have taken the punch bowl away earlier?
I still have little sense of a major, structural problem in housing. Markets are ALWAYS adjusting but they always have stiction. Here, we have TWO interacting markets - one for money and one for housing. Of course they will oscillate and gyrate in a chaotic manner.
In the real world, some lenders will take a haircut, lowering the industry average rate of return. Some home owners will lose the houses a more restrictive lending regime would never have financed. Many home owners not in default will bemoan that they are less wealthy than they had imagined as market prices and unrealized equity shrink.
However, few will be kicked out undeservedly.
Posted by: Joseph Somsel at September 4, 2007 01:10 PM
this: "Technically, when the Fed purchases Treasury securities, it is lending money to the Treasury, but for practical purposes, it is just giving newly created money to the Treasury, that is, paying part of the government's expenses by "printing" money."
"it is just giving newly created money to the Treasury"-- isn't that 'giving' done at interest-payable to the FedRes, at the minimum?
Posted by: Mark E Hoffer at September 6, 2007 02:35 PM
"isn't that 'giving' done at interest-payable to the FedRes, at the minimum?"
Yes, but the Fed always has a tendency over time to increase the size of its portfolio of Treasury securities, so in effect, it rolls over the interest into new principal. Plus which, I think the Fed's profits go to the Treasury, so in a sense, the Treasury is just paying interest to itself.
Posted by: knzn at October 1, 2007 06:52 AM