Wednesday, April 30, 2014

The CATO Institute and NGDP Targeting

Guess which think tank gives this advice in its official hand book for policymakers:
The intent of Congress would be better served and monetary policy would be more effective if Congress instructed the Federal Reserve to establish a monetary policy that reflects both their concerns in a single target. The best such target, I suggest, would be the nominal final sales to domestic purchasers—the sum of nominal gross domestic product plus imports minus exports minus the change in private inventories... Congress is best advised (1) to specify a target rate of increase of final sales and (2) to instruct the Federal Reserve to minimize the variance around this target rate. The target rate of increase of final sales may best be about 5 percent a year, sufficient to finance a realistic rate of economic growth of 3 percent and an acceptable rate of inflation of about 2 percent.
So the think tank is advising policymakers to do something like nominal GDP (NGDP) targeting. It may surprise you to learn that it comes from the CATO Institute. But it should not be a surprise. Some of CATO's top monetary experts, like Lawrence H. White and George Selgin, also support versions of a NGDP target. They would prefer a free banking system, but given we have a Fed they would prefer it target NGDP. CATO Senior Fellow Steve Hanke also shares this view as does former CATO scholar Timothy B. Lee. The reason for their support of a NGDP target is because they see it as the most conducive to monetary stability. 

Here is why. A NGDP target aims to stabilize total dollar spending. It is one target that has embedded in it both the supply of and the demand for money (i.e. total dollar spending = money supply x velocity of money). The beauty of a NGDP target is that the Fed does not need to know what is exactly happening to the money supply or money demand. All the Fed only needs to worry about is the product of the two components. There is no need to track the money supply or estimate money demand. By focusing on total dollar spending, the Fed will be fostering a stable monetary environment where movements in money supply and money demand are offsetting each other. 

Another great feature of a NGDP target is that it allows supply shocks to be reflected in relative price changes. No attempt is made to offset them or their effect on the price level. For example, assume there is a technological innovation that raises a firm's productivity. Such productivity gains mean lower per unit production costs that, in turn, should translate into lower output prices given competitive pressures. Here, the increase in a firm’s output from the productivity gains is matched by a decrease in its sales price. For an economy-wide productivity innovation that affects many firms, this response would manifest itself in rising real GDP growth alongside a declining price level and vice versa. But note that the price level times real GDP is simply total dollar spending. Consequently, if the Federal Reserve directly targeted the growth of total dollar spending it would by default be allowing the price level to move inversely with productivity-driven changes in real GDP. This amounts to a monetary policy regime that ignores supply shocks. This feature is conducive to financial stability.

So there are good reason that many folks at CATO support some kind of NGDP target. I bring this up because CATO Senior Fellow Alan Reynolds had a recent article in Investor's Business Daily that bashed NGDP targeting. Unlike his colleagues at CATO, though, Reynolds shows an incredible amount of confusion in his article. So if you are looking to the CATO institute for guidance on monetary policy I recommend you turn to their other experts or its official handbook for policymakers.

Let me detail a few of the many problem with his piece. First, try this:
Prominent economists of all stripes have proposed that the Fed should focus instead on keeping the growth of nominal GDP (NGDP) growing at a steady rate. But growth of NGDP is simply the inflation rate added to the real GDP growth rate.
No, we directly observe NGDP and from it is derived real GDP based one some estimated index of the price level. Reynolds should listen to CATO monetary expert George Selgin on this issue:
The occasional, if tacit, treatment of NGDP as a value that is “derived” by taking the product of two directly observable magnitudes, real output (y) and the price level (P), is as mischievous as it is wrong. We must understand the behavior of both P and y to depend, the first in the long run and the second in the short run, on that of Py, rather than the other way around. That is why it is also important to insist that stabilizing NGDP is not just a rough-and-ready way of minimizing a loss function in which fluctuations of P and Y are separable components. No and no again: if the natural rate of y plummets (natural disaster or war, say), what is desirable is not that we should minimize both P and y movements subject to the supply-shock “constraint. It is rather than we should see P move the opposite way from y, which is done by stabilizing Py.
Here Reynolds esposuses an atheoretical model of inflation, something akin to trend analysis:
[T]he trouble with basing future policy on past inflation news is that inflation is always lower before it moves higher. PCE inflation rates of 0.8% in 1998 and 1.3% in 2002, for example, were followed by 2% inflation in 2003, 2.4% in 2004, and 2.9% in 2005.
What? Inflation is inherently an oscillating process? So the very thing Reynolds is criticizing, Fed policy, is not at all responsible for the path of inflation?

This is just a sample of the confusion in the article. So again if you are conservative and looking for guidance on monetary policy I encourage to look at the actual monetary experts at CATO.

P.S. You could also look to the monetary experts at the Mercatus Center like Scott Sumner. He has several articles there worth exploring.

20 comments:

  1. When banks are free, they follow the rule that practical bankers developed over centuries of experience: Issue money in exchange for short-term real bills of adequate value. But along come some academic scribblers with a narrow rule like NGDP targeting, based on tautological relationships like MV=Py, and guess which rule the herd chooses!

    ReplyDelete
  2. That's great news. I doubt we'll ever get a libertarianish president, but if it happened, it'd be a travesty to watch him blow it with an insane Fed Chair pick.

    ReplyDelete
  3. The late William A. Niskanen, a long-time chairman of the libertarian Cato Institute, favored something like domestic final sales targeting in his book, Reaganomics, which was published in 1988. See page 187.

    ReplyDelete
  4. Anyone that actually follows the numbers knows better. Rates-of-change in monetary flows = roc's in aggregate monetary purchasing power. I.e., roc's in MVt = roc's in nominal-gDp (which is a proxy for all transactions, or as you prefer "total dollar spending", in Irving Fisher's "equation of exchange" (where PT = MVt & not PY = MV).

    The Fed's technical staff & other Keynesian economists simply don't know the differences between money & liquid assets. Otherwise there wouldn't have been large sell offs on Feb 27 2007, Sept 29 2008, May 6 2010, Oct 19 1987, etc. And there never ever would have been a recession after the Great-Depression.

    I.e., all of these blockbuster errors were the result of the roc in MVt, bank debits, & bank legal reserves suffering gargantuan absolute declines.

    ReplyDelete
    Replies
    1. Flow5, I agree it would be better to do MV=PT but for the longest time we have been limited by the availability of data. Now with the release of gross output we will be closer to the PT identity.

      Delete
  5. Dr. William Barnett, Dr. Paul Spindt, and Dr. Richard Anderson had it wrong. The biggest error involves the distributed lag effect. “Divisia – aggregates”, the “debit-weighted-money-... & the Monetary Services Index (MSI) data series ignore this overriding scientific evidence.

    The odd thing is that the truth is simply politically incorrect.

    ReplyDelete
  6. David, I dug up the IBD version here:

    http://news.investors.com/ibd-editorials-brain-trust/041614-697445-feds-powers-to-control-the-economy-are-overated.htm?ntt=fed%20can%20print%20mney%20but%20it%20can%27t%20print%20jobs&p=3

    I thought for sure I'd see a lot of comments from MMists disputing or or questioning aspects of Reynold's article, but I didn't. Why not?

    ReplyDelete
    Replies
    1. Tom, I missed that one. Now David has "put it in it´s place". Next time, if you see something so "ghastly", send us (or just me) a note!

      Delete
    2. Sure thing. I was only aware of it in this case because of David's post.

      Delete
  7. David
    On Niskanen´s "Demand Rule":
    http://thefaintofheart.wordpress.com/2012/08/23/if-a-demand-rule-had-been-conducted-by-design-and-not-by-accident-the-lesser-depression-would-likely-have-been-only-your-run-of-the-mill-recession/

    ReplyDelete
  8. I was hoping for some serious criticism, and still am. Dropping names and semantic trivialities won't deal with the several points I raised as to why it is unlikely to work. I don't speak for Cato, and some Cato folk doubtless disagree with me. The late Bill Niskanen favored targeting nominal final sales to domestic purchasers, which is clearly better than GDP but harder to explain.

    Doesn't my graph suggest the Fed TRIED to target nominal GDP (that is, to boost either real GDP or inflation when both were low)? But policy by trial and error leave us with a lot of trials and errors. Fiscalists try to boost NGDP with budget deficits. That doesn't work either.

    My IBD piece is sympathetic with the effort, just skeptical. If the Fed knew how to hit any target, then we could simply debate targets. They can't, so we can't.

    ReplyDelete
    Replies
    1. Alan, I´m sorry to say but when you write:

      "A typical proposal, from James Pethokoukis at the American Enterprise Institute, is for the Fed to keep nominal GDP growing at a steady rate of 5%. But trying to keep NGDP growing at a 5% pace means inflation would have to speed up during recessions and decline in booms."

      It shows you have no idea about what it means to target NGDP-Level target. Targeting NGDPLT, which the Fed did with some success in 1987-07 ("bad" things happened when it generated some NGDP instability), does NOT mean it has to boost EITHER RGDP OR inflation. It means it keeps NGDP on a stable level path, and if inflation FALLs due to positive productivity shocks, for example, it doesn´t care, it keeps NGDP on the stable path. If inflation rises due to a negative (oil, for ex.) supply shock, it doesn´t react, like the Bernanke Fed did in 2008, by tightening monetary policy (despite "low" (2% FF rate)). It strives to keep NGDP on the stable path.

      The rest of your piece just cumulates conceptual errors.

      Delete
  9. Alan, thanks for stopping by. I encourage you again to note that NGDP targeting is not about boosting real GDP or lowering inflation. It is about targeting total dollar spending, and allowing real GDP and inflation to go wherever the fundamentals take them. If you are looking for a serious criticism of your piece than this is it. NGDP targeting is about creating a stable monetary environment and letting relative price changes and supply shocks be sorted out by the market. I would think you would find this appealing.

    Also note that targeting final sales to domestic purchasers is practically identical to targeting NGDP. This is demonstrated in this figure: http://research.stlouisfed.org/fred2/graph/?g=zj0 As noted to Flow 5 above, there may be some meaningful difference if we start talking about the new BEA measure, Gross Output

    ReplyDelete
  10. David, note that the quoted chapter from the Cato Handbook, 7th ed. (2009), was written by Bill Niskanen.

    ReplyDelete
  11. Lawrence H. White and George Selgin, also support versions of a NGDP target. They would prefer a free banking system, but given we have a Fed they would prefer it target NGDP.

    David, may I ask: what is your own stance? In an ideal world, which regime would be better and why?

    ReplyDelete
    Replies
    1. Free banking isn't an alternative to NGDP targeting. Free banking alone doesn't imply anything about the behavior of nominal magnitudes; that depends on the base money regime. Suggestions to the contrary are a sign of some very fundamental misunderstanding.

      Delete
  12. Excellent blogging.

    I cannot fathom the resistance to NGDP targeting, either on the left or the right. The peevish fixation on inflation is especially unproductive.

    Reading certain economists, an economy that developed a very long trend of real growth and inflation both at 3 percent would be a failure----inflation was too high. Some economists have become deontologists, and the rule is that zero percent inflation is the reason for macroeconomics.

    Forgotten today that the Great Inflation Fighter Volcker got inflation down to 4 percent---and then said "good enough." BTW, the right-wing was in hysterics in those days that Volcker was too tight.

    You know, 3 percent inflation is not TEOTWAWKI. It is 3 percent inflation. Big whoop.

    I will tell you what is a big whoop. Zero percent inflation. The BoJ obtained that in Japan, and the results were ugly and stayed ugly for 20 years.

    Of you want to be hysterical about something, be hysterical about zero percent inflation.

    ReplyDelete
    Replies
    1. My long pattern of learning a new word from a typical Benjamin Cole post continues: "deontologists" ... Lol
      Plus I had to look up TEOTWAWKI even though you'd used it before.

      Delete
    2. Not sure the Fed has the instruments and skill to do what you hope, given $2.4 tn. excess reserves. But nearly any nominal target (including exchange rates and commodity prices) beats Fed staff hunches about "job market conditions" and "slack" https://twitter.com/AlanReynoldsEcn/status/698672801507622912

      Delete
    3. https://twitter.com/AlanReynoldsEcn/status/698674724835090433

      Delete