Thursday, October 15, 2009

Did Low Interest Rates in 2003 Cause the Crisis?

Cato has a new policy analysis by Jagadeesh Gokhale and Peter Van Doren. Among other things, they argue that a tighter Fed policy in the early 2000s would not have been a sensible approach to avoid the housing bubble and current financial crisis. David Beckworth disagrees in a recent post on Macro and other Market Musings. He cites both Larry White and John Taylor, arguing that by keeping the federal funds rate too low for too long in 2003 and 2004, the Fed planted the seeds of the current crisis.

By today's .25% standard, a federal funds rate of 1% doesn't seem shockingly low, but perhaps it was too low.


I believe that the Fed should keep nominal expenditures--total final sales of domestic product--on a slow steady growth path. I prefer a 3% growth path. With a 3% trend growth rate in productive capacity, that should result in a stable price level on average. However, the Fed has preferred an inflation rate of close to 2%, which implies a 5% growth rate of nominal expenditure. How have they performed?

The following chart shows total final sales of domestic product as well as its trend growth path.


Up until 2008, the big picture looks good. I suppose that is why it was called the great moderation.


Take a closer look at the period of 1998 to 2006.


During the period when Beckworth, White, and Taylor claim that interest rates were too low, nominal expenditures were well below the 5% growth path and didn't get back to that growth path until 2006. Of course, the Fed didn't really commit to maintaining a 5% growth path of nominal expenditures, or anything too specific. They appeared to be aiming at increasing the CPI at a 2% annual rate from its current level, where ever that might be. Still, if the Fed was trying to keep nominal expenditure on a 5% growth path, a more rapid expansion of the quantity of money would have been appropriate in 2003.

I don't favor targeting any interest rate, however, it is certainly possible that a more rapid growth rate of the quantity of money could result in lower interest rates. The Fed would be purchasing larger quantities of something, presumably T-bills. Simple supply and demand analysis suggests higher T-bill prices and lower T-bill yields. On the other hand, as Scott Sumner often points out, current nominal interest rates depend on expected inflation and current real interest rates depend on expected real output. What would have happened to market interest rates if there had been a commitment to adjusting the quantity of money enough to keep nominal expenditure on target? We will never know.

16 comments:

  1. This comment has been removed by the author.

    ReplyDelete
  2. Bill:

    If we look at the growth rate of your series then nominal expenditures grow above 5% by mid 2003 and continues to accelerate through mid 2005. With the federal funds rate at 1% through June 2004 and only incrementally increasing thereafter does not that appear to be a highly accommodative monetary policy? This figure here graphs the series in growth rate form--and includes my favorite, final sales to domestic purchasers or domestic demand--along with the federal funds rate

    ReplyDelete
  3. You show GDP dipping below trend, then catching up. That catching up begins right around '03 when the Taylor rule would have tightened policy, but that is just so under the short-run version of your rule too.

    Your presentation seems to assume a belief that GDP is purely trend stationary. I think you need to squarely confront the empirical evidence that real-shocks are pervasive and the underlying series is difference-stationary under a true inflation target.

    In this sense, I think your claims are a bit misleading. You purport to embed a price-level target, but your use of trend-stationary assumptions belies that criteria.

    With this in mind, I'd really prefer to discuss a short-run NGDP target where no effort is made to accelerate back to an arbitrary trend.

    ReplyDelete
  4. It seems like the question you two are considering (David and Bill) comes down to how to think about the starting point for a nominal spending (or final sales) path when trying to create the least distorting price environment. Presumably, David wouldn't think the Fed was being too accomodative today if nominal spending grew by 6 or 7% in 2010, and I'm sure Bill would argue that the high growth rates you're (David) referencing simply put nominal spending back on path after too-low growth in 2001-2002. Seems like he has a strong case here.

    On the other hand, I'm not sure I fully understand the extent to which Bill is willing to push the nominal spending path argument (i.e. if we have a 50% drop in capacity because a meteor hit the US, we should still target 3% growth in NGDP to keep nominal spending in line but allow substantial inflation -- wouldn't this mean that the inflation we accomodate would be expected to come with serious distorting rigidities and some prices rising much more quickly than others, along with nominal contract distortions?).

    ReplyDelete
  5. Jon:

    Thank you for your comment.

    I am not making any assumption that real GDP is trend stationary (though I believe that it is.) I think that the least bad response to a permanent change in the growth path of real productive capacity is a permanent change in the price level. Trying to force changes in the growth path of nominal incomes to reflect the change in real income, permanent or not, is disruptive and pointless.

    I think there is such a thing as nominal shocks (like now) and targeting the level of nominal expenditures is by far the best solution in a world of sticky prices, including resources prices and so nominal incomes.

    I can think of a list of productivity shocks that are trend stationary. For example, the impact of weather on agriculture. The impact of strikes. The impact of a hurricaine on regional production.

    David:

    Thank you for your comment.

    I favor keeping nominal expenditure on a stable growth path. The primary reason for this is for expected nominal expenditures to be predictable.

    I do not favor having nominal expenditure grow at a certain rate (say 3%) from where ever it happens to be. For example, suppose it is 1933 and nominal expenditure is 50% below where it was in 1929. I would not favor a 3% growth rate from the 1933 levels. I would favor a roughly 100% growth rate. Admittedly, there is some point where one must recognize that the adjustments already made to the disaster will be disrupted, and you should just start over.

    Today, in October of 2009, I do not favor a 3% growth rate of nominal expenditure from its current value. I favor a much more rapid growth rate over the next year, returning nominal expenditure to its past growth path.

    If we are thinking about price level or inflation targeting, level or growth path targeting causes problems when there are price level shocks and sticky prices. A 2% inflation target from where ever the price level happens to be may be sensible. I think targeting a growth path of nominal expenditure is better. Sticky prices and price level shocks have a more tolerable effect with nominal expenditure targeting.

    P.S. I think the goal is to stablize total expenditure on domestic output. Why should we be focused solely on how much domestic buyers spend on it? Why don't foreign buyers count?

    The way I see it, the problem of monetary disequilibrium is a disruption of the use of resources to produce output--income, expenditure, and output. Exports are produced with resources. The sale of goods for export generates income.

    Further, one of the pathways by which changes in the quantity of money impact expenditure is through changes in exchange rates and so on exports. Why ignore that?

    Thanks again for the thoughtful comments. Perhaps growth path targeting is a mistake, but I still think it is the least bad option.

    ReplyDelete
  6. dlr:

    Thank you for your comment. As you see from my comment, you understood what I had in mind.

    Isn't it odd that I used the Great Depression and a 50% drop in nominal expenditure (over 3 years) and you also went with a 50% shock, but rather a huge supply shock.

    Maintaining the 3% rule after the disaster would be trivial compared to the huge reduction in the demand to hold money because output and resource availability has fallen in half. (Of course, the quantity of money would presumably drop as well.)

    Oddly enough, I would be willing to accept a 100% increase in prices to reflect such a huge drop in output. That would be better than dropping nominal income in half. But, of course, doubling the nominal incomes of the surviving half of the population, so that they can pay prices twice as high, would hardly make sense.

    Suppose half of the U.S. splits off to form a new country. Should the rump section try to keep nominal expenditure on its previous growth path? I think not.

    Generally, I have been concerned about more gentle changes. For example, if the 3% growth path of real GDP changes, then, after a time, perhaps a different growth path for nominal expenditure should be adopted. My view is that the monetary authority should not be making these sorts of decisions on the fly, but rather such changes should be treated differently by the monetary constitution. I suppose a meteor proviso would be appropriate too.

    ReplyDelete
  7. David:

    Thanks for the diagram.

    Looking at both final sales of domestic product and domestic purchases of domestic product together is pretty cool, really. The difference is exports?

    Growth from one year ago? I calculated quarterly growth rates. They show the rapid growth during the period that concern you as well.

    Does the St. Louis Fed allow trend lines to be added? This excel graphs look ugly on the blog.

    ReplyDelete
  8. Bill:

    You have given me a lot to think about. The returning to trend via higher nominal spending growth for the Great Depression makes a lot of sense. I just have a hard time reconciling this undersanding with the all the other evidence that indicates loose monetary policy fueled the housing boom period. There must be some qualifiers somewhere. Your arguments for using final sales of domestic product are compelling too. I need to switch.

    I like using year-on-year since it gets away from the noisy quarter-on-quarter series.

    On the graph, there are no trends but you can put several series together like I did. Regarding your concern for the excel graphs, you could try modifying the figure (changing colors, font size, gridlines, etc.).

    ReplyDelete
  9. This is interesting. I have been arguing since the end of last year that the Federal Reserve held interest rates too low. Perhaps they didn't, but it occurs to me that they might still have created the housing bubble. It is not enough to simply get the quantity of money correct (for if it is on the moon it would be no good to anybody). A further issue is where the money enters into the economy, and whether or not relative prices are distorted by it.

    Monetary equilibrium isn't just about aggregates, since the specific location of supply and demand changes are important, too. Ideally, we want to see increased supply directed toward people generating the increased demand. Now, I can quite imagine how a free banking system would achieve this, but can the Federal Reserve do it just as well? I suspect not.

    Should increased supply be focused too intensely on one sector, it would seem to create conditions favourable to the emergence of a bubble.

    What would you even call this? It would be a mismatch of the increased money supply with increased money demand, or in other words, sector specific inflation and deflation with aggragates seeming to be in equilibrium.

    ReplyDelete
  10. Lee:

    A year ago, I would have argued that it is nearly entirely about aggregates. And isn't because I am unaware of inject effects. I am less certain with the policies of the Bernanke Fed, which are clearly aimed at directing credit to "weak" areas.

    David:

    In my view, malinvestment can be generated when there is a change in regime. That is because I think it takes time for people to learn the new regime. Hey, I have a Virginia school background, and our economics isn't about perfect optimizers.

    Second, malinvestments can be created when there is a mistake under the existing regime, but only if entrepreneurs and the monetary authority make the same error. (Kling had a recent post where he said something about "great minds think alike," along those same lines.

    While I think the best regime (which is always "least bad" to my way of thinking) is a stable growth path of nominal expenditure, that wasn't the Fed's regime exactly. The actual regime is do good regarding prices and employment, with some hint that this means targeting a 2% CPI inflation rate from where ever the CPI happen to be.

    While I think the situation could use more thought, it seems possible that malinvestment could result because of a mistake by the Fed, shared by entrepreneurs, in the context of its actual regime. Still, I think care must be taken when complaining about "excessive" growth of nominal expenditure. It depends on the policy regime.

    ReplyDelete
  11. Bill:

    I am still thinking about our discussion. For now I will note that if we use the Great Moderation as our regime period and therefore use data (from the Laubach and Williams paper [2003]) for the period 1983:Q1 - 2007:Q4, the average growth rate of the potential GDP is 3.03% on a year-over-year basis or 3.06% on an annualized quarter-over-quarter basis. These number line up nicely with your 3% growth in potential GDP. Now if we look at final sales of domestic product either in year-over-year growth rate form here and or in annualized quarterly growth rate form here they both are usually growing above 5%. The exception is during recessions. Even then, though, if one takes the average of either approach for the recession years one gets just above 3% total growth in nominal spending for the recession year. If 1983:1 is our starting point, then I fail to see any need for nominal spending to catch up during either the 1990-1991 or 2001 recession and certainly not during 2003-2005.

    ReplyDelete
  12. David:

    Thank you for your comment and the citation from Laubach and Williams.

    I will look at these alternative starting and ending points. I used first quarter 1984 to 2nd quarter 2008. The trend grows at 5.2%. The recession in 1990 was a reversal of a deviation of expenditure above trend as was the 2001 recession, though as noted in the post, it went below trend during the recession. It it a bit worrisome that it would be so dependent on exact starting and ending points.

    ReplyDelete
  13. Here’s a rudimentary question: how do these nominal expenditure metrics, which seem to show steady growth, account for asset/commodity booms? If they don’t fully capture the growth in the prices of oil, food, etc… perhaps that’s the missing the excesses created by loose monetary policy.

    Taylor isn’t coming at his analysis from an Austrian or monetary equilibrium perspective from what I understand.

    I’m an amateur so maybe this is just an obvious thought.

    The other possibly redundant thought is that the real problem isn’t in the aggregate growth metrics but in the composition of that growth. The excessively low interest rates played a major role (along with the various gov’t tax, GSE and bank regulatory policies) in shifting investment into capital goods like housing and durables in an unsustainable way. That dumped scarce capital into unproductive lines that are now being devalued. I guess that’s just Hayekian triangle misallocation in action.

    These metrics seem to be hyper-aggregate to the point of missing the, well, mechanics of change…

    ReplyDelete
  14. The idea of exchangeing large sums of dollars for pounds, yen etc... and wait for the prices to rise or fall is not a new concept.

    But people think you need hundreds of thousands to join this game. This is not true.

    You can for a tiny fee have brokers buy currency for you and then to sell them later when you have made a profit.

    ReplyDelete
  15. No, I don’t think so that’s the only reason, as there were many other parts that played the role, but of course it was one of the main reasons. I don’t take tension about all that since it does not matter to me. I am trading with OctaFX broker and they got fabulous service with daily market news and analysis setup, it really puts me in good shape for doing well regardless of the market crisis and I am able to make regular profits with ease.

    ReplyDelete