Thursday, January 24, 2013

Goodhart on NGDPLT

Two economists from Morgan Stanley, Jonathan Ashworth and Melanie Baker, joined with Charles Goodhart to critique nominal GDP level targeting.   Or perhaps it would be better to say that the goal is to smear a potential competitor of the inflation-targeting orthodoxy.

First, they are confused about the introduction of a nominal GDP targeting regime:
One of the problems of starting an NGDP target system is that the start date for ‘history’ to commence is itself entirely arbitrary. By juggling with the start date, and the desired growth path, one could leave the MPC with an immediate requirement that could vary anywhere from a huge expansion to a severe retraction. For example we show below what the implicit current gap is between the desired path for nominal GDP and the actual path for nominal GDP if history were deemed to have started in 1997 Q2, and growth paths of, say, 5% and 4% were also deemed to have been appropriate, as an upper and lower example, respectively. With the 5% path, the MPC would, assuming we aim to hit the target two years ahead, currently have to expand nominal GDP by around 10% p.a. With the 4% path, the MPC would have to keep nominal GDP growth down to around 2.3% p.a. (these estimates are based from the end of Q3 2012 to end 2014).

For the U.S., anyway, nominal GDP remained very close to a 5.4% growth path during the Great Moderation.   If we were on that growth path, and wanted to shift to a slower growth rate, then it would be foolish to go back to some past date and project that alternative growth rate to the present, and then implement an immediate drop in the target to the new growth path.   

The obvious approach would be to start the new growth path in the present, or better yet, allow for an adjustment period.   Announce now that nominal GDP will shift to a lower growth rate a few years in the future.  

Of course, what happens when an incompetent central bank has allowed nominal GDP to fall far below a long standing trend?     In my view, the obvious solution is to return to the previous trend.   If that trend is deemed to have an excessively high growth rate, then the growth path can be adjusted at some future date.   Just because nominal GDP is below the previous trend doesn't mean that the growth rate of the projected value of the trend cannot be adjusted.  

However, suppose the central bank is so incompetent that nominal GDP has fallen below a long standing trend and has allowed it to remain there for years?     At some point, it becomes sensible to forget about the previous trend.    For example, after a few decades looking to the past trend would seem pointless.    After four years, it is more of a judgement call.   Most Market Monetarists, including me, favor a new growth path somewhere between the current level of nominal GDP and the growth path of the Great Moderation.

The next criticism is that CPI data is monthly, provided a few weeks after the end of the month, and is not revised often.   Nominal GDP data, on the other hand, is only quarterly, finalized near the end of the subsequent quarter, and subject to significant revisions.    This criticism has some merit.   More and better information is always better.     Monthly nominal GDP figures that were so good that the statisticians found little reason to revise them would be great.   However, just because CPI and CEP figures come out monthly and are rarely revised hardly means they are the proper target for monetary policy.   There are daily figures for gold and perhaps there are hardly any errors discovered in the daily closing price.   Does that make a gold standard especially wise?

As a practical matter, Market Monetarists advocate targeting the forecast.   For example, targeting the expected value of nominal GDP one year in the future.    The problem of quarterly data, "final" measures nearly one quarter after the end of the quarter, and possible later revisions of the measure is solely a problem of accountability.     Was nominal GDP actually at the target level?    When we finally observe that it was not, was that because there was some change in spending that no one could have anticipated?   Or was nominal GDP on target, and what appears to be a deviation was just a measurement error?   Perhaps nominal GDP did come on target, but really, a few years later, further analysis suggests that the "true" value of nominal GDP was really too high or too low.  

I think it is fair to say that no Market Monetarist is advocating a mechanical rule mandating changes in a policy interest rate or changes in the quantity of base money based upon the deviation of the most recent past measure of nominal GDP from a target level.    If we continue to target inflation (or the price level,) targeting the forecast remains the appropriate approach.   If the inflation rate one year out is targeted, the 11 measures of the CPI until that time provide no help for determining what should have been done before.

Their arguments deteriorate from there.    They ask:
Other real (e.g. supply-side) factors determine growth; the long-run Phillips curve is vertical. Do those advocating a nominal GDP target now deny that? Do they really believe that faster inflation now will generate a faster, sustainable, medium- and longer-term growth rate?

The short answer is "no."  

However, the formulation here is puzzling.   Will higher inflation now generate a faster sustainable medium term and long term growth rate.

Market Monetarists believe that that a faster trend growth rate of nominal GDP will mostly result in a higher trend inflation rate.  At first pass, mostly can be replaced by entirely.    This would include consumer prices, but also resource prices, including money wages.   Nominal interest rates would be higher too.    For the most part, employment, unemployment, the level of real output, the growth rate of real output, real wages and real interest rates would all depend on "real (e.g. supply-side)" factors.   Most of us are very concerned with these "real" matters, though we do not consider them central to evaluating alternative monetary regimes.

Few Market Monetarists would sign on to an assumption of superneutrality.   And I think most of us are convinced that shrinking nominal GDP is more likely to be disastrous than neutral, mostly due to the implications of a financial system based upon hand-to-hand currency that has a zero nominal interest rate.     Some worry that these same problems spill over to very low trend growth rates for nominal GDP.   For example, constant nominal GDP might not be a good idea.  But none are arguing that higher trend growth rates for nominal GDP always generate higher average growth rates of real output much less higher levels of human well-being.

However, their question wasn't about whether advocates of nominal GDP level targeting believe that a 15% NGDP growth rate would generate more real GDP growth  than a 10% NGDP growth rate.   It was whether more inflation now will generate a faster medium or long run growth rate.      I still believe (and hope) that even if we keep to a 2% inflation target, eventually, nominal (and real) wages will fall to a low enough growth path so  that real output will return to potential output--the productive capacity of the economy that depends on the real, supply-side factors.   From that perspective, the growth rate of real GDP over that entire period would be the same.   The only question then is whether the average level of real output during the period is higher or lower.   

Of course, advocates of nominal GDP targeting are claiming that bringing nominal GDP closer to the growth path of the Great Moderation will raise the level of real output and employment now.   In other words, the recovery of production and employment to the long run growth paths would be more rapid than if nominal GDP continues on its new, lower growth path.   For Market Monetarists, any increase in inflation would be an undesirable side effect.    (The new Keynesians, on the other hand, see the higher inflation as essential to reduce real interest rates and increase real expenditures.)

More importantly, Market Monetarists are arguing that a monetary regime that allows inflation to vary with "supply shocks," will result in more employment and real output on average.   Further, nominal GDP level targeting will generate more rapid recoveries which will also tend to result in more employment and real output on average.   Of course, avoiding unsustainable booms due to efforts to pump up inflation in response to a favorable supply shock or cutting booms short due to level targeting would tend to have the opposite effect on employment and output.   However,shorter and  less frequent booms are unlikely to fully offset the gains in real output and employment from shorter and less frequent recessions and almost certainly would not offset the gain in human welfare.  

Will an improved monetary regime result in a higher sustainable growth rate in the medium term or long run?   I find that doubtful and instead think it is likely that a better monetary regime is one of those things that will raise the growth path of real output.   But more important than any gain in measured real output would be the improvement in human welfare.

They then claim that nominal GDP level targeting will result in more uncertainty regarding inflation:
 If we knew what the future sustainable long-run rate of growth would be, we could set a current nominal GDP growth target that would on average deliver that, plus 2% inflation. But we do not. Moreover, the view is steadily gaining ground that it is more likely, than not, that real growth in the future will be below the average of past decades; technological innovation may slow and demographic developments will be adverse. So, if we wanted to maintain price level stability, with inflation at 2%, we should be considering a nominal GDP growth target of slightly under 4%. That is not what the advocates of such a target propose.
Given our uncertainty about sustainable growth, an NGDP target also has the obvious disadvantage that future certainty about inflation becomes much less than under an inflation (or price level) target. In order to estimate medium- and longer-term inflation rates, one has first to take some view about the likely sustainable trends in future real output. The latter is very difficult to do at the best of times, and the present is not the best of times. So shifting from an inflation to a nominal GDP growth target is likely to have the effect of raising uncertainty about future inflation and weakening the anchoring effect on expectations of the inflation target.

This is entirely correct, but Market Monetarists don't favor anchoring inflation expectations.   Instead, we believe it is much better to anchor expectations about the future level of nominal GDP.   The primary reason is that it is better for inflation to vary with supply shocks than to use monetary policy to offset those supply shocks and keep the growth path of prices on some target.  

If there are many negative supply shocks, so productivity grows more slowly than expected, then having nominal GDP grow as expected implies that prices rise more quickly than expected.   Resource prices, such as wages, will be more likely to grow as expected.  Nominal interest rates are likely to be more consistent with expectations as well.    Equity prices are likely to be more consistent with expectations.   Why is it that keeping consumer prices at the expected level is especially important?

So, if nominal GDP was on a 5% growth path, and real potential GDP varied between 1% and 4% growth rates, then the inflation rate would vary between 4% and 1%.   This doesn't amount to a threat of double digit inflation,  much less hyperinflation.

Finally, I am getting tired of these absurd suggestions  that advocates of nominal GDP targeting are proposing a return to the Great Inflation of the sixties and seventies.

If we thought that we had learnt anything from the travails of the 1960s and 1970s, it was that monetary expansion in the medium and longer run does not bring faster, sustainable growth. If anything, the opposite is true; faster inflation, at any rate beyond some threshold, deters growth. The long-run Phillips curve is vertical. It was on this analytical basis that the case both for Central Bank independence and a specific inflation target was made.

During the 1960s and 1970s, the growth rate of nominal GDP increased, and at an growing rate.   There was no target for slow, steady growth in nominal GDP.   Instead, there was an effort to target unemployment and real GDP growth.    Market Monetarists advocate a slow, steady growth path for nominal GDP.   We do not favor targeting real GDP or the unemployment rate.   Further, slow steady growth of nominal GDP  is nothing like more than a decade of increasing nominal GDP growth rates.  (See Nunes here, or better yet, look at the Cole and Nunes book.)

Market Monetarists argue that nominal GDP level targeting would have avoided both the Great Inflation and the Great Recession.    We further argue that nominal GDP targeting would result in macroeconomic results  similar to those of the Great Moderation all the time, though an explicit target could have avoided the boom and allowed for more rapid recoveries from the two mild recessions.   There are very good reasons to believe that a nominal GDP level target would not only avoid disasters like the Great Inflation and the Great Recession, it could provide at least moderately better results than the Great Moderation.


  1. I have said this over and over, you can easily get these guys on your side, just say:

    no make-up

    2/3's of the value of NGDPLT is in the LT. So even if you don't get the bump right up front, a new floor is in place.

    What is so bothersome is that some MM don't admit the 4.5% functions as a hard cap on growth. And when you have a obvious hard cap on anything WHAT HAPPENS???

    A fist fight over splitting up the pie occurs.

    Now every new regulation raises costs and eats up the pie. Every new raise for public employees eats up the pie.

    SELL THIS. This is what people wat to hear!

    MM is horrible at putting the meat in window, you shouldn't sit around annoyed everyone hates the speech you are giving.

    your speech stinks.

    1. This is also one of my big complaints about NGDP level targeting. What happens if the economy can run better than 1.3% real growth.

      Figure 2% inflation
      figure 1% population growth
      We're talking a MAX long term growth of 2%, all dialed in by NGDP level targeting.

      This is a path to misery.

  2. Michael:

    Don't figure in 2% inflation. If we have a 5% nominal GDP growth path, and 1% population growth, and per capita productivity grows 3%, then we will have 1% inflation.

    The more rapid growth in productivity slows the growth of average and marginal costs. For the typical market, supply curves are shifting to the right. Each firm faces a situation where lower prices and expanding production increases profits.

    This is nothing like a situation were slower growth in demand motivates firms to cut production and only when lower resource demand results in a lower growth path of resources and so costs, is a recovery profitable.

    If nominal GDP growth grows 5%, and the population grows one percent, and per capita productivity grows 4%, then there is zero inflation and rapid real per capita growth in output.