Wednesday, March 24, 2010

Credit Boom 2003?

Many economists blame the Great Recession on a credit boom instigated by the Federal Reserve in 2001-2004. John Taylor blames a failure by the Fed to follow the rule that bears his name. The Fed's target for the federal funds rate was left too low for too long. San Francisco Fed President Janet Yellen claims that higher interest rates would have slowed housing appreciation and restrained excessive credit creation in securitized markets.

"Austrian" economists sometimes take comfort in these arguments, seeing them as at least partial acceptance of their long standing view about recessions being the aftermath of a credit boom generated by a central bank. In their view, the expansion of the quantity of money by the Fed caused interest rates to fall too low, which resulted in a vast expansion in spending on new single family homes. Rising prices and production in this sector were unsustainable, and resulted in the recession. David Beckworth, of Macro and Other Market Musings, appears to take this approach.

That interest rates were remarkably low during part of the "naughties" is certainly true.


During 2001, the Fed cut the federal funds rate 4.75 percentage points, from 6.5 percent in late 2000 to 1.75 percent in early 2002. After a pause, the reductions continued, hitting a low of 1 percent in mid 2003. The remarkably low level of the federal funds rate was continued until mid-2004. At that point, interest rates began a long series of regular increases.

However, showing that some price is lower (or higher) than usual, hardly proves that it is too low or too high. The role of prices in a market economy is to coordinate the decisions of market participants and changing conditions may require large changes in prices. So, were interest rates "too low" from late 2000 until sometime in 2005 or after?

I believe that the least bad policy for the Fed is to keep nominal expenditure on a slow, steady growth path. Ideally, I favor a 3 percent growth path, which is consistent with a stable price level on average.

However, the Fed aimed at 2 percent inflation, which implies more rapid growth in nominal expenditure. Looking at the actual growth path of nominal expenditure during the Great Moderation (dated from first quarter 1984 until the third quarter of 2008,) its trend growth rate was 5.4 percent. However, there were substantial deviations of nominal expenditures around that growth path. The diagram below shows Final Sales of Domestic Product from 1984 through the end of 2009.


The Great Moderation can best be characterized by nominal expenditure remaining very close to its trend growth path. Of course, the most striking aspect of the diagram is the catastrophic decrease in nominal expediture in the fourth quarter of 2008. While the Great Recession is officially dated as beginning at the end of 2007, the mild reduction in spending growth was "normal" for the Great Moderation until disaster struck a year later.

Other milder deviations include nominal expenditure being above its trend growth path in the late eighties and the late nineties. Of course, the deviation most relevant is during the supposed credit boom--nominal Final Sales of Domestic Product was below its trend growth path during that period. The following diagram shows Final Sales of Domestic Product from 1998 until 2006.


Total final sales fell below trend in the first quarter of 2002. It has remained below trend ever since, but by the second quarter of 2005 and up through the first quarter of 2008, the shortfall remained less than 1 percent. So, there was a shortfall of nominal expenditure from trend of more than 1 percent from the second quarter of 2002 until the first quarter of 2005. The largest shortfall was in the first quarter of 2003--2.6 percent.

The Fed began its rapid cuts in the federal funds rate in late 2000, while nominal expenditure remained well above trend. The rapid decrease had already occurred when nominal expenditure fell below trend. And the further decreases to 1.5 percent and then to 1 percent occurred while nominal expenditure was approximately 2 percent below its long run growth path.

If the goal of the Fed were to keep nominal expenditure on a stable growth path, (and it should be) and a decrease in the federal funds rate is the method by which the Fed expands nominal expenditure (which is plausible,) then the decreases in the federal funds rate were appropriate.

There are economists who accept the goal of stable nominal expenditure, but claim that the Fed's policy was "too expansionary." Larry White made this argument, as has David Beckworth. They focus on the growth rate of nominal expenditure. The following is the growth rate of Final Sales of Domestic Product from 1998 until 2006.

The growth rate of nominal expenditures fell below its trend growth rate of 5.4 percent in the third quarter of 2000. And soon after, the Fed began the long and rapid reduction in the federal funds rate. In 2003, the growth rate of nominal expenditure rose close to its the trend growth rate of 5.4%.

While first quarter growth was a bit below trend, second quarter growth was very close. The third quarter of 2003 had a remarkable 8.6 percent growth rate. But perhaps more troubling than that one remarkable quarter is the trend of accelerating growth starting mid 2004 and continuing through 2005.

During 2003, the federal funds rate fell to 1 percent and remained at 1 percent not only during the large spike in nominal growth, but also at the beginning of the sustained acceleration! Clearly, excessively low interest rates would causing a boom. The growth rates show it clearly.

However, as explained above, nominal expenditure was below trend during this entire period. If nominal expenditure begins on its targeted growth path, and if the growth rate stays constant, then nominal expenditure remains on its targeted growth path. But if the growth rate is too low or too high, then nominal expenditure will move below or above the target growth path. The only way for nominal expenditure to return to the target growth path is for it to grow more quickly or more slowly than its trend growth rate.

Nominal expenditure was above trend in the late nineties. The Fed increased the federal funds rate and the growth rate of nominal expenditure slowed. This caused nominal expenditure to return to its trend growth path. When it reached that path, nominal expenditure did not immediately return to its trend growth rate, but continued to grow slowly, less than trend, and so now nominal expenditure fell below its long term growth path.

During the period of the supposed "credit boom," nominal expenditure was growing faster than trend, but it was returning to its long run growth path (or at least approaching it.) And that was the only way it could return to that growth path.

I have been advocating that the Fed target nominal expenditure one year into the future. I favor a 3 percent growth rate, but suppose a 5.4 percent growth rate of nominal expenditure is continued. If nominal expenditure is currently on the targeted growth path, then the rule would be for nominal expenditure to grow 5.4 percent over the next year. The target for nominal expenditure one year in the future would be 5.4 percent greater than its current value.

However, if nominal expenditure is below its current targeted value, then returning to its 5.4 percent target growth path entails a more rapid growth rate for the next year. It isn't that the target of nominal expenditure is changing. The target for the level of nominal expenditure for each future quarter is determined by the rule. The growth rate, however, varies by the current value of nominal expenditure.

Using the one year rule, it is possible to construct a graph showing what the proper target for the growth rate for the next year should have been. These "targets" are not cumulative. If the policy was followed and effective, then nominal expenditure would return to its growth path and the targeted growth rate would be the trend of 5.4 percent.

Note that the only quarter when the growth rate of nominal expenditure was above the target for the year was the third quarter of 2003. That wouldn't have been a problem as long as it wasn't repeated. It was not until the first quarter 2005 that the growth rate of nominal expenditure was greater than the proper target. Further, keep in mind that nominal expenditure never reached its long term trend.

Given the norm of targeting a stable growth path for nominal expenditure, Fed policy was too "tight" during the entire period when interest rates were unusually low. If the method by which the Fed raises nominal expenditure is to lower the federal funds rate, then the federal funds rate must have been "too high."

Yet even if that is true, just because the federal fund rate must have been too high during part of that period doesn't mean that it should have been lower, or even as low as it was, during most of the period. Perhaps even lower interest rates in the middle of 2002 and then higher interest rates once nominal expenditures returned to the targeted growth path by mid 2003 would have been appropriate.

However, adjusting the federal funds rate is of doubtful utility when the goal is keeping nominal expenditure on a stable growth path. If nominal expenditure is currently on the growth path and nominal expenditure is expected to grow too rapidly, then perhaps higher short term interest rates are equilibrating. Similarly, if nominal expenditure is expected to grow more slowly, and fall below the growth path, lower short term interest rates are appropriate.

On the other hand, if nominal expenditure has already grown too slowly for a substantial period of time, perhaps even falling, and nominal expenditure is now substantially below its target growth path, (as much as 2.7 percent in 2003 or more like 15% in 2010,) then having the Fed commit to low, or falling short term interest rates is almost certainly not an appropriate approach. The depressed level of nominal expenditure reduces credit demand at any given nominal interest rate. The equilibrium level of short term interest rates can easily be quite low because of the destructive effects of the drop in nominal expenditure.

If the Fed had a commitment, not to a series of changes in interest rate, but rather to change the quantity of money however was much is needed to get nominal expenditure back to target, rising credit demand could result in increases in nominal interest rates. And so, if the only way for the Fed to increase nominal expenditure back to a target for the growth path is to lower short term interest rates or keep them low, then the low interest rates were appropriate. However, since the Fed can expand base money by purchasing a variety of financial assets, the lower interest rates are simply an artifact of the Fed's approach to policy.

Of course, certeris paribus, if the Fed purchases financial assets, this tends to raise their prices and lower their yields. But if the result is an expectation that nominal expenditure will have returned to its target growth path in the subsequent year, the direct effect of the Fed's purchases might be swamped by sales from the private sector, leading lower prices and higher yields.

For example, expectations of recovery may lead firms to issue new corporate bonds, and the increased supply of those bonds lowers their prices and raises their yields. Bond investors purchase those bonds, more willing to hold them because the firms issuing them are less likely to fail once the economy recovers. Those investors in corporate bonds sell government bonds that they were holding because of fear of default by private firms. Their sales lowers government bond prices and raises their yields, more than offsetting the direct effect of the Fed's purchases.

Now, if this effect had not occurred, then the Fed's purchases of bonds would have lowered interest rates. And this would result in lower funding costs and result in higher nominal expenditures. But, foreseeing the higher nominal expenditures that response to those higher nominal expenditures could make nominal interest rates rise.

If the Fed is committed to interest rate targeting, then it will interfere with this process. Recovery should move interest rates up, but the Fed will purchase securities in order to keep them down. Of course, as the economy begins to recover a shrinking output gap or higher inflation would cause the Fed to gradually raise its target for interest rates.

The Fed was clearly committed to making modest periodic changes in the federal funds rate. They had some general commitment to price level stability and high employment. However, price level stability meant low inflation. "Taylor rule" reasoning suggests that they were making gradual adjustments in a target for the federal funds rate based upon observations or forecasts of inflation and the output gap. Low inflation meant that the core CPI should be expected to rise 2 percent from its current level, wherever that happened to be.

In other words, the Fed was certainly not explicitly committed to a target for the growth path of nominal expenditure. If the Fed had that commitment, it is not at all clear that the extraordinarily low interest rates would have existed in 2002 through 2004. Further, if the Fed had not been wed to interest rate targeting, it is not at all obvious that an expansionary monetary policy would imply lower nominal interest rates.

In conclusion, nominal expenditure was too low during the period of the supposed credit boom. On other other hand, that does not necessarily imply that interest rates were not too low. It is certainly possible that excessively low short term interest rates were one of the many factors resulting in excessive increases in home prices. So the Fed's approach to monetary policy could be a contributing cause of the speculative bubble in home prices. And that speculative bubble certainly resulted in a misallocation of resources that can only now be gradually corrected by shifting labor to more productive avenues and producing new capital goods.

Worse, the end of the speculative bubble, and the losses of the financial firms that had lent into that bubble, were the key trigger for the disastrous drop in nominal expenditure in late 2008. And now, the absence of a clear commitment to returning nominal expenditure to its previous growth path is resulting in even more severe difficulties with production and employment. A policy approach of gradual decreases in short term interest rates, and further, dealing with the zero bound by promising to keep short term interest rates low for an extended period of time, not only is again not getting nominal expenditure back to its trend growth path, it might also be responsible for a future speculative bubble.

12 comments:

  1. Good post Bill. But how many people will understand it? It's so much simpler just to yell "Look how low the Fed set interest rates!"

    Actually, most readers should get your first point, that monetary policy was not too loose judging by what happened to nominal income.

    But your second point is more sophisticated. It's a version of the Lucas Critique? The statistical relationship between an instrument and a target may not be structural, but may depend on the choices of instrument and target path. Right?

    Actually, re-casting your argument in terms of the Lucas Critique may help people understand it better. And it's the same point you me and Scott have been making about the interest rate not being a good measure of the stance of monetary policy.

    ReplyDelete
  2. (A minor quibble:)

    "Of course, the most striking aspect of the diagram [the first chart in your post] is the catastrophic decrease in nominal expenditure in the fourth quarter of 2008." This remark does not match the diagram, which shows--besides the noticeable decrease from trend starting around November, 2007, running to about May, 2008--a catastrophic decrease thereafter, covering not just the fourth quarter of 2008 but the third as well, and even a sliver of the second.

    ReplyDelete
  3. I just want to say that you're doing an excellent job in your posts to be as clear as possible and add relevant exhibits. It's a pleasure to read them.

    ReplyDelete
  4. Nick:

    Thank you for the comment.

    I hadn't thought about this as fitting in with the Lucas critique. It is obvious now that you mention it. Let's think about that more.

    This post is, of course, informed by Sumner's arguments about interest rates being a poor measure of the direction of monetary policy. And your elaboration of those points, particularly how expectations shift the "IS curve" is more exactly like I think about the issues.

    Of course, the other worry in the article is whether or not the Fed's policy approach leads to malinvestment. The "keep the nominal interest rate at zero for an extended period of time" to fix the zero bound is the real source of my concern.

    ReplyDelete
  5. Anonymous:

    Perhaps there is some error on the chart, but nominal expenditure dropped in absolute terms in the fourth quarter of 2008 at an annual rate of 5.3 percent. It also dropped in the first quarter of 2009 at an annual rate of 2.4 percent. It grew in all the other quarters since December 2007, though all but the second quarter of 2008 were well below the 5.4 percent trend growth rate.

    By the way, the current deviation from trend (4th quarter 2009) is 10.4 percent, not 15 percent. So that was an error. 15% is how much it would need to grow over the next year to return to the old growth path.

    As for gap you see on the diagram before 4th quarter 2008, the deviation was already more than 3% in the 3rd quarter of 2008, which is one of the highest during the Great Moderation, and it had already passed 1 percent in the first quarter of 2008.

    But the recession in 2001 was part of the Great Moderation, and there were several quarters where nominal expenditure was below 2% from trend. If the gap had begun to close in, say, 4th quarter 2008, (rather than an absolute drop,) I would have counted the Great Moderation as continuing.

    In my view, there was a qualitative change between the third and fourth quarter of 2008. In my view, interest rate targeting broke down in the face of a financial panic.

    I now agree with Sumner that it was recognition that it was breaking down that greatly exacerbated the decline.

    ReplyDelete
  6. Don:

    Thank you for your kind words.

    ReplyDelete
  7. Bill:

    I have and continue to wrestle with the growth rate versus level targeting approach for stabilizing nominal expenditures. If I had to pick one approach now I would go with level targeting for all of its virutes, but I have had a hard time reconciling it with what by almost every other measures is clearly a period of excessively loose monetary policy: the early-to-mid 2000s. And I am not basing my view on the "interest rates were too low for too long" argument (though even this argument has merits in terms of the "risk-taking" channel of monetary policy). Rather, I am basing it on the "interest rates were too low relative to the neutral (or natural) rate of interest for too long" view. I have documented this development in this picture which comes from this post. Others have pointed out the growing gap at this time between the actual and natural interest rate too, such as this ECB paper.

    I would also note your charts above use quater-on-quarter annualized growth rates. I prefer using year-over-year growth rates which shows nominal spending growing at a far faster clip. It is also worth noting that maybe the 5% rate all along was inappropriate. Selgin's productivity norm, which is a special kind of nominal income target, does not set a long-run target rate but instead has NGDP grow at the expected growth rate of factor inputs. I am not certain that having a fixed long-run NGDP path is optimal.

    ReplyDelete
  8. David:

    Thanks for the comment.

    Thanks for the citation to the ECB paper. It will take me a lot of time to decipher it.

    I am very skeptical of the "risk" channel of "monetary policy." How does this differ from people choosing to take more risk when the equilibrium safe interest rates fall? Why isn't this an optimal result?

    I know you use year over year growth rates. Since I currently favor targeting nominal expenditure one year in the future, I suppose these figures should be informative.

    Personally, I take Selgin's approach as being an argument that changing nominal expenditure to reflect changes in factor productivity is undesirable. However, it doesn't provide an argument that changing the growth path of nominal expenditure to reflect changes in factor supply is a bad idea.

    So, for example, perhaps more rapid growth of nominal expenditure would have been desirable during the period when the LFPR increased. Or perhaps changes in immigration or baby booms and busts should result in changes in nominal expenditure.

    I don't think this provides a reason to move to targeting the growth rate.

    It is like saying that in 1932, nominal expenditure should have increased 2 percent.

    The hard issue is that at some point there is no reason to "reflate." People will eventually adjust to the new growth path of nominal expenditure and returning to the old one would be pointlessly disruptive.

    ReplyDelete
  9. Bill:

    Glad you agree with the "Lucas Critique" interpretation. Yes, it does seem obvious now, but it's funny that none of us had thought of looking at it that way before.

    Yes, it's the same point that you, Scott, and I have been making, but in a different context. You are making it to talk about the historical policy counterfactaul.

    I think the Lucas Critique interpretation is a point worth making. I have a draft written, and will post it when Worthwhile Canadian Initiative gets a little quieter.

    ReplyDelete
  10. Bill:

    The 1932 example is a great example of why level targeting makes more sense. And you raise a good point that it gets tricky when the public's expectations change regarding the target path for nominal spending.

    Regarding Selgin's rule, you are right that it says productivity changes should not be reflected in nominal spending (i.e. price level should reflect changes in productivity). As a result, though, it also amounts to stabilizing factor prices. For example, with his labor productivity norm version monetary policy effectively stabilizes the nominal wage.

    ReplyDelete
  11. Bill:
    Good post!
    I agree that nominal stability of short interest rates is harmful for the stability of nominal expenditure growth. But I am afraid that alternative policy of Fed purchasing securities to increase the money supply has the same risk of creating bubbles (but perhaps in different sectors).

    ReplyDelete
  12. Great post! I was looking online for credit counseling and my interest in how credit works kinda snowballed. Great info!

    ReplyDelete