A Sharp Expectation Shock is Needed

Via the FT's Alphaville we learn that Goldman Sachs is discussing some "radical" options the Fed could use if economic conditions deteriorate further. One of the options discussed is a nominal GDP level target. It would be a radical change from way the Fed currently operates, but such a shock is exactly what the public now needs.  For the past few years the economic outlook of households and firms has been dismal and consequently they have had been accumulating a large stock of money assets.  If the Fed were to announce a nominal GDP level target it would provide a big expectation shock that would reverse much of this buildup.  

One of the ways this shock would play out is through the many more observers who would be wailing about the reckless course of monetary policy, the horrors of debasing the dollar, the end of Western Civilization, and other hard money concerns.  Similar concerns were raised when FDR effectively did the same thing in 1933 with his own QE and price level target program.  These concerns about FDR's program provided a much needed shock to nominal spending and inflation expectations.  As a result, there was robust recovery from 1933 to 1936.   A nominal GDP level target would provide the same kind of shock today if it were properly signaled and followed through by the Fed.

It would require the Fed to buy up assets until the nominal GDP level target was reached. In other words, the Fed would be committing if needed to a permanent increase in the monetary base, something it has yet to do as recently noted by Michael Woodford.  The effectiveness of such a shock would not be contingent on increased financial intermediation (though it ultimately would be affected too). Rather, this shock would work two ways. First, by raising inflation expectations, it would increase the cost of holding money assets for the non-bank public. This would create a hot-potato effect for non-bank holders of money assets and that, in turn, would lead to a mother-of-all portfolio rebalancings. Ultimately, this rebalancing would end with higher nominal spending. Second, it would simultaneously cause nominal spending forecasts to rise and this too would lead to more current nominal spending. Given the large slack in the economy, the increase in nominal spending would mean a rise in real economic activity.

But enough from me. Here is what Goldman Sachs has to say about nominal GDP targeting:

[W]e have again received some questions about the possibility that the FOMC might move to a nominal GDP target ... It’s important to note that depending on the interpretation, the Fed’s dual mandate (in which policy responds to both employment/real GDP and inflation) already has some similarities with a nominal GDP target (in which policy responds to the product of real GDP and the price level). The key differences are that (1) an announced nominal GDP target is much simpler and therefore more powerful than the hazier dual mandate, which is interpreted differently by different people; and more importantly, (2) the dual mandate is defined in terms of rate of change of prices, while a nominal GDP target depends on the level of prices. ...The implication is that a nominal GDP target, Fed officials attempt to “make up” for past undershooting of inflation via future overshooting. In other words, a move to a nominal GDP target is tantamount to a temporary increase in the inflation target.

I am glad to see Goldman Sachs telling its clients about nominal GDP targeting.  The fact that nominal GDP targeting is now being discussed by Goldman Sachs, The Economist, The Financial Times, The Wall Street Journal, The Telegraph, The National Review, and other media outlets means this idea is gaining traction. The public needs a sharp expectation shock and a nominal GDP level target is radical enough to do it.

The Economist Magazine Takes a Closer Look at NGDP Targeting

Here is the article.  It does a fair job discussing the pros and cons of such a rule.  Among the advantages for NGDP level targeting is that it better handles supply shocks:

They could also react more appropriately to supply shocks. Take the example of an economy that is hit by a negative supply shock through high oil prices depressing output and raising inflation. An inflation-targeting central bank may feel compelled to tighten policy, worsening the slump in output, whereas one mandated to hit NGDP could be more flexible. There could be advantages, too, in the opposite case where a positive supply shock through productivity-enhancing new technology boosts real GDP growth while lowering inflation. An inflation-targeting central bank would respond by easing monetary policy, which could produce asset bubbles, whereas an NGDP-targeting central bank would hold steady. Certainly inflation would be more volatile, but the overall economy would not be.

One of the disadvantages is how to adopt and implement a new rule when it is not widely known:

 For all its theoretical merits, a switch to NGDP targeting would throw up some new problems—and old ones. The Fed has not exactly sat on its hands since the financial crisis began in 2007, so it is far from clear it could easily reach the new goal.

The short answer is that the Fed would announce (1) its targeted growth path for NGDP and (2) commit to buying up as many securities as needed to reach it.  Knowing that the Fed would be willing to buy up trillion of dollars of assets if necessary to hit its target would cause the market itself to do much of the heavy lifting.  That is, the public would adjust their portfolios in anticipation of the Fed buying up more assets and in the process cause nominal spending to adjust largely on its own.   I go into more detail here how this would work, but the key point is the Fed would be better managing nominal spending expectations.  Such a rule would have better contained nominal spending expectations in 2008 and avoided the worst of the crisis. Just look at Sweden who effectively does something like a nominal GDP level target.

Update: Bill Woolsey provides much more commentary on The Economist article.

Does Higher Expected Inflation Really Spur Spending?

I will let the data answer the question.  To do that, I took the Cleveland Fed's monthly 10-year expected inflation rate series and plotted it against the subsequent growth in nominal consumer expenditures.  I looked at a one, two, and three-year horizons for subsequent consumer spending growth.  Below are the scatterplots created from this exercise. (The data starts in 1982:1 since that is the beginning of the Cleveland Fed's expected inflation series.) 

Not only is there a strong relationship, but it gets stronger at longer horizons for consumer expenditures.  So changes in expected inflation do affect nominal spending.  This is nothing new and is a central tenet of modern macroeconomics.  I only bring it up now because some observers have questioned whether there really is this relationship.  Reviewing this relationship also reminds us why it is important for the Fed to be clearer about the future path of monetary policy.

Update: I used expected inflation above because that has been the focus of recent debates. As an advocate of nominal GDP targeting, I believe a better perspective is to look at nominal spending expectations.  To do that, I  went to the Survey of Professional Forecasters and took the average annualized quarterly growth rate forecasted for nominal GDP over the next year and compared it to the actual growth of nominal GDP over the next year. Here is the figure:

Once again, expectations matter.

Michael Woodford Explains the Problem with Fed Policies

Michael Woodford, one of the top monetary theorist in the world, has an Op-Ed today that does a great job explaining why the Fed's policies have failed to gain traction in the economy.  His key point is that the Fed has failed to clearly communicate the path of future monetary policy.  In so doing, the Fed has failed to shape expectations forcefully enough to make a dent in nominal spending.  In other words, the problem is not that the Fed cannot do anything, but that the Fed has failed to act properly.  Woodford says a price level target would solve the problem (and by implication so would a nominal GDP level target) of properly shaping nominal expectations. 

Here is Michael Woodford making his point by showing the flaws with the Fed's QE programs:

The economic theory behind QE has always been flimsy...The problem is that, for this theory to apply, there must be a permanent increase in the monetary base. Yet after the Bank of Japan’s experiment with QE, the added reserves were all rapidly withdrawn in early 2006. The Fed has given no indication that the current huge increases in US bank reserves will be permanent. It has also promised not to allow inflation to rise above its normal target level. So for QE to be effective the Fed would have to promise both to make these reserves permanent and also to allow the temporary increase in inflation that would be required to permanently raise the price level in that proportion.

The one time QE did work was for FDR during the 1933-1936 period.  The big difference between then and now is that FDR's QE program was supported by an explicit price level target that FDR himself promoted.  Also, the monetary base creation supporting FDR's QE program was permanent as can be seen by the figures in this post. FDR knew how to do QE right. 

Michael Woodford also addresses the portfolio channel of monetary policy that been used as an argument to support the Fed's QE programs (my bold below):

A more recent argument for QE stresses not the increase in reserves, but the change in the composition of assets held by the Fed. But asset trades of modest size by a central bank are unlikely to affect prices, unless the markets in question have seized up. 

Once again, the point is not that portfolio channel does not work but rather that the Fed has to be fully committed to using it. The Fed needs to unload both barrels of the gun.  By only engaging in asset trades of modest size, the Fed is not doing enough to meaningfully shock and shape nominal expectations.  Here too an explicit price level or nominal GDP level target would be useful.  For such a monetary policy target would tell the public that the Fed intends buy up as many assets as necessary to hit the level target. It would not stop at some arbitrary amount like it did with the $600 billion of QE2.   Knowing that the Fed would be willing to buy up trillion of dollars of assets if necessary to hit its level target would cause the market itself to do much of the heavy lifting.  Investors would adjust their portfolios in anticipation of the higher inflation and nominal spending. This seems to have worked in Sweden where larger asset purchases (25% of GDP versus the Fed's 15% of GDP) combined with an explicit inflation target (that works more like nominal GDP level target) has brought a robust recovery.

Michael Woodford ends his piece with these recommendations:

Mr Bernanke can and should use his speech today to explain how his policy intentions are conditional upon future developments...A clarification could help the economy in two ways. First, he could signal that a temporary increase in inflation will be allowed, before policy tightening is warranted. This would stimulate spending by lowering real interest rates. Second, specifying the size of any permanent price-level increase would avoid an increase in uncertainty about the long-run price level. This in turn would ward off an increase in inflation risk premiums that might otherwise counteract the desirable effect of the increase in near-term inflation expectations.

In other words, he thinks the Fed should adopt a price level target. I wonder what he thinks about a nominal GDP level target.

The Washington Post Is Confused About Monetary Policy

Fortunately, Ramesh Ponnuru is here to clear up the confusion.

The Other Side of Household Balance Sheets

A popular explanation for the ongoing economic slump is that the United States is in the midst of a balance sheet recession.  This view holds that the vast amount of household debt built up during the housing boom is now being unwound and that this deleveraging is creating a drag on the economy.  Though intuitive, this balance sheet recession view is inadequate because one, it ignores the potential offset in spending by creditors and two, it misses a more fundamental problem: the elevated demand for liquidity.  I believe one of the reasons for this confusion is that advocates of the balance sheet recession view tend to focus on the liability side of household balance sheets while ignoring the details of the asset side.  A close look at the asset side reveals that despite the collapse in overall household assets, there has been a inordinately large buildup of liquid assets.  It is this accumulation of money and money-like assets rather than the deleveraging itself that has kept nominal spending from experiencing a robust recovery.  Here are the numbers.

From the peak of household asset values in 2007:Q2 to the latest data for 2011:Q1, households have lost around $9.4 trillion worth of non-liquid assets. Despite these large losses and the subsequent slump in personal-income growth, households have somehow increased their holdings of money and money-like assets by a staggering $1.6 trillion as seen in the figure below:  

The composition of the increase in liquid assets is also interesting. Most of the increase has come in the form of time and saving deposits, though treasuries have been important too.  Money assets alone (cash, checking account, time and saving deposits, money market accounts) have remained elevated and close to their peak value in late 2008.  And, as John B. Taylor notes, money demand appears to be growing even more. 

I personally place more importance on the growth of the money assets  in slowing down nominal spending for two reasons.  First, they are fixed in nominal value and thus only grow by households actively acquiring them.  Treasuries, on the other hand, may be increasing through acquisition and valuation change.  Second, the demand for liquid assets in general can only affect nominal spending by influencing the medium of exchange, money.  It is the only asset on every other market and thus is the only asset that can directly affect nominal spending.  An elevated demand for treasuries matters, then, by by spilling over to the demand for money.   

As a share of total assets, household's liquid assets have risen and remained elevated as seen in the figure below.

Now some may note that the rise in the share of liquid assets is not all that large in terms of percentage points.  A rise in the share of liquid assets, however, does not need to be terribly large to impede nominal spending.  This is especially true with money assets because they are the medium of exchange.  One would expect nominal spending to be sensitive to changes in money demand.  The figure below confirms this.  It shows the relationship between the growth rate of money assets as a share of total assets and the growth rate M3 velocity for the period 1951:Q4 through 2011:Q3.  (The M3 data comes from NowandFutures.) The relationship is surprisingly strong, given that money velocity can also be affected by factors like innovations in financial transactions.  

Observers, therefore, need to be paying more attention to the build up of money assets in household balance sheets.  Until this development changes, there will be an ongoing drag on nominal spending.  One way to address this problem is to introduce a nominal GDP level target. This is how it would work in practice.

Update:  In the comment section I further explain why excess money demand rather than deleveraging is the reason for the weak aggregate demand.

Central Banks Still Have Much Ammunition

Ambrose Evans-Pritchard writes there is still much monetary policy can do to support the economy:

[W]ith fiscal policy exhausted, the burden must fall on monetary policy. Here we have barely begun to use our atomic arsenal even at zero rates. As Milton Friedman taught us – though nobody in Frankfurt -- it is a fallacy to think that low rates are loose. Zero can be extremely tight. 

That may be the case now with US Treasury yields signalling deflation and M2 velocity collapsing as it did pre-Lehman. 

To those who argue that the Fed is pushing on the proverbial string, David Beckworth from the University of Texas replies that the Fed showed between 1933 and 1936 that it could deliver blistering growth of 8pc a year despite debt deleveraging in the rest of the economy.

He is referring to this post where I noted the following:

[H]ouseholds were also significantly deleveraging during the Great Depression.  This experience would fit the standard definition of a balance sheet recession.  Below is a table from his paper that shows household balance sheets in real terms.  Note that between the 1933 and 1936 U.S. household underwent a cumulative deleveraging  in real terms of 20%. This is far more in percentage terms that has happened over the past few years. And yet between 1933 and 1936 the U.S. economy had a robust recovery.  Real GDP averaged almost 8% growth during these years. 

   Source: Mishkin

The balance sheet recession view cannot easily reconcile the large deleveraging by households and the rapid real economic growth that occurred between 1933 and 1936.  What can explain it is a more nuanced view that acknowledges creditors with excess money demand were confronted by FDR's original quantitative easing program.  This QE program was far better than recent ones in that FDR clearly signaled a price level target and backed it up by devaluing the gold content of the dollar and allowing unsterilized gold inflows. In otherwords, FDR signaled that he was going to allow a significant and permanent increase in the monetary base and followed through on it.  This change nominal expectations and caused creditors to start spending their money balances.  

Unfortunately, the 1933-1936 recovery was cut short by a tightening of monetary and fiscal policy.  Consequently, many observers overlook this great natural experiment of monetary policy that shows monetary policy does not push on a string when there is significant deleveraging.  

Another natural experiment is present-day Sweden.  Its housing sector also acquired much debt and its economy was also hit hard by the economic crisis.  However, it has had a robust recovery and the reason appears to be a much more aggressive monetary policy.  All of this shows that monetary policy still can pack a punch.  The question, then, is whether the Fed has the desire and political capital to do so.

Update:  The Mishkin table above is in real terms.  To see what was happening to nominal household debt, I constructed the two tables below.  The first one converts the Mishkin table into nominal growth rate terms using the CPI.  The second one uses current dollar data from the 1940 statistical abstract on "individual and other noncorporate" debt.  Both tables indicate nominal household debt was falling through 1935.  So the recovery of 1933-1936 largely coincided with a reduction in household nominal debt too.