Wednesday, October 8, 2014

Salter and Hogan on NGDP Level Targeting

Alex Salter and Thomas Hogan have a working paper that points to problems with a central-bank directed policy of targeting the level of nominal GDP.   The argument is that while such a policy may helpfully stabilize aggregate demand, it will create undesirable consequences for aggregate supply.

I certainly appreciate the work that Salter is doing on Nominal GDP level targeting   I don't find their argument persuasive   Or at least, I think there is a problem with the example they use to argue that divergent expectations between the central bank and market participants will cause problems.

They consider a situation where some shock has pushed nominal GDP below target.   A central bank targeting nominal GDP will seek to return nominal GDP to its previous growth path.   They assume that the central bank considers the shock "nominal" and so increases the quantity of money to offset the decrease in velocity   Market participants, however, believe that the shock was structural.   Real wealth has been destroyed.   And so the market participants do not believe that the central banks commitment to returning nominal GDP to target is credible.

However, a belief that a shock is "real" and has destroyed wealth doesn't prevent the central bank from returning nominal GDP to target.   It simply means that that the price level will shift to a higher growth path.   The inflation rate would pick up for a time, and then return to its previous rate, but now on a higher growth path.   This "structural" problem would also be reflected in a lower growth path for real output.

The central bank, believing that the shock was solely nominal, would expect instead that real output would recover to its previous growth path, and the price level would also return to its previous growth path.

The divergence in expectations in this situation would be that the market participants, believing the problem to be structural, will expect the recovery of nominal GDP to be a shift to a higher growth path of prices while real output remains on a lower growth path.   The central bank, on the other hand, expects that real output and the price to will return to their initial growth paths once the monetary disequilibrium is relieved.

What is most puzzling about this example is that if the central bank agreed with the market participants and thought the problem was structural, it would still expand the quantity of money enough to offset any change in velocity and return nominal GDP to the target growth path.   Then the central bank, like the market participants, would expect this to involve a shift to a higher growth path for the price level and a lower growth path for real output.

That there is some structural problem that persistently reduces productive capacity does not make returning nominal GDP to target unfeasible.  Further, it is difficult to see how this leads to an excess supply of money.

Salter and Hogan describe how the supposed structural problems have result in a leftward shift in the long run and short run aggregate supply curve.  Then they oddly describe this in micro terms as a inward shift in the production possibilities for various goods and services.  (I would think production possibilities is a macro concept.)   The relevant micro concept is that the supply curves for various goods and services are believed to have shifted to the left.   Firms believing that would tend to raise their prices along with reducing their production.   While this response is undesirable if they are making a mistake, it is consistent with a return of nominal GDP target.   As explained above, this micro response is consistent with the price level shifting to a higher growth path and real output shifting to a lower growth path.

Salter and Hogan are concerned by the excess supply of money generated by the central bank as it pushes the price level higher and point to the stagflation of the seventies and a variety of empirical studies that suggest that increased inflation is disruptive.   In my view, the seventies are not very instructive, since that was a period where nominal GDP was not on a stable growth path but rather had an accelerating growth rate.  

Of course, a study of a nominal GDP level path would most certainly show that higher inflation was associated with lower real growth in the short run.   That is how it works.   When supply side factors lead to slower growth in productivity, inflation will be higher.   As for any long run relationship between inflation and real output growth, this involves setting the growth rate implied by the rule.   Is a 5 percent growth path for nominal GDP better?   That would imply 2 percent inflation if potential output is on a 3% trend.   Or would 4 percent, 3 percent, or 2 percent be better.

There is no doubt that a nominal GDP level target will do worse than an inflation target in stabilizing short term inflation expectations.   However, I believe that expectations that nominal GDP will be at a particular level in future periods provides a better macroeconomic anchor than knowing that the price level next period will be at the same level.  

I do think that divergent expectations between firms setting prices and making production and employment decisions and the central bank could lead to problems.   I believe that nominal GDP targeting avoids problems due to difference in views about whether shocks are nominal or structural--at least to the degree that this simply involves differences expectations regarding the growth path of the price level or real output.   Instead, I would be more worried that entrepreneurs are naive Keynesians and believe something like the paradox of thrift.

So, suppose there is some structural change so that real wealth has been reduced.   Being poorer, people spend less.   That implies that nominal GDP will remain below target.   While the central bank  create more money and lower interest rates, if no one wants to borrow, then they are just pushing on a string.

Now, in reality, a decrease in wealth does reduce consumption and increase saving.    Those who are poorer seek to rebuild their lost wealth.   However, the increase in saving supply results in a lower natural interest rate.   Assuming market rates adjust, investment expands enough so that total spending is not depressed    Thinking of a misallocation of resources--capital specific to housing construction lost, for example, then this added investment can be used to rebuild the other types of capital goods that had been crowded out by the excessive investment in sawmills or cement plants.

What would be ideal is for the reallocation of resources to occur with prices and output based upon on target nominal GDP.    That productive capacity might be permanently reduced doesn't make this impossible at all.   For example, capacity constraints for capital goods more in demand may result in higher spending on them generating only modest increases in production and substantially higher prices.   Meanwhile, the reduction in prices for houses and housing construction equipment and perhaps other consumer goods might be much smaller along with larger decreases in output.   The price level rises and real output falls.    Hopefully, this upward shift in the growth path for the price level and reduction in real output will be partially relieved and reversed as resources shift and bottlenecks ease.

But suppose entrepreneurs are naive Keynesians.   They don't believe that nominal GDP will return to target.   They base their investment decisions on the assumption that spending on output will remain on its current growth path.   With those expectations, they invest less.   Must the central bank create an excess supply of money to force nominal GDP back to target?  

Perhaps.   Of course, these perverse expectations imply a lower demand for investment and so a lower natural interest rate.   A lower market interest rate implies a higher demand for money.   Is the quantity of money necessary to return nominal income to target simply accommodating this unusually high demand for money?  

Perhaps this is what Salter and Hogan have in mind.   My thought is that such entrepreneurial error is possible.   It seems to me that what is an excess supply of money and a market rate below the natural interest rate is ambiguous in this situation.     However, I would also see this as less a persistent problem and more an issue of learning the new regime.   And so, any such problems would become less severe as time passed.  


  1. "In my view, the seventies are not very instructive, since that was a period where nominal GDP was not on a stable growth path but rather had an accelerating growth rate."---Woolsey.

    Excellent blogging.

    The 1970s are a poor indicator of what would happen with monetary stimulus in the 2010s.

    Back in the 1970s, a much larger fraction of the private workforce was unionized, and foreign trade was a much smaller component of GDP.

    In brief, the supply side was domestic in the 1970s, and clogged by regulation and unions. Does anyone remember the regulations of trucking, railroads and airlines? Reg Q? Ma Bell? Airline tickets were regulated by price!

    Today the supply side is global. How does one generate demand-pull inflation when the supply-side is global?

    In the 1970s you had a top MTR of 90%, down to 70% at the latter years. There were capital shortages and not capital gluts. This is important, as today when a price signal (and profits) suggests a bottleneck, there is plenty of private-sector capital to apply to the bottleneck.

    The amount of capital pouring into the energy sector is a sign of that---and a reflection of globalized capital markets, another feature absent from the 1970s.

    The 1970s were characterized by Big Steel, Big Auto, Big Labor, Big (and clunky) Retailers---all destroyed now. Can Ford raise prices when there is a Kia, Hyundai, BMW, Volkswagen, Toyota, Nissan and Honda on the scene, let alone GM and Chrysler?

    There are no major manufacturers or retailers today with pricing power (well, maybe Apple, and even then....).

    Today we have the Wal-Mart global sourcing model, Craigslist and $1 stores.

    You know what the CPI-U is in Dallas, a region with strong labor markets?

    1.2% YOY July. And some people say the CPI overstates inflation.

    Avoiding robust labor markets and growth due to fears of not smart monetary policy.

  2. Thanks for the comments, Bill. It's certainly true that, on the macro level, the asymmetric expectations Thomas and I discuss will result in a given level of NGDP breaking down into more inflation and less real output growth. But to me NGDP targeting only makes sense as an attempt to mimic monetary equilibrium and achieve monetary neutrality. I'm worried that the expectations asymmetry will change the underlying conditions at which the money market would hypothetically clear, and the result would be inflation at the macro level, which is troubling only because the micro-level accompaniment is noise in the price mechanism. Thomas and I are worried that this will have real effects, and permanently lower real income. Whether you buy this depends on what you think about how monetary disequilibrium affects relative prices, neutrality vs. superneutrality in levels vs. growth rates, etc.

    As to the 1970's, we included it simply as a reminder that the attemp to exploit stable, expectations-insensitive relationships in macro variables is a bad idea. Obviously this is not a good description of what would happen *had there been a rule*, due to the Lucas critique. But that's not the angle we were getting at in that section.

    Again, thanks for your attention. I hope we can continue debating this!

  3. It’s tough to say about GDP target level, but we just need to make sure we have proper idea and awareness before doing anything. I am really glad that I trade with OctaFX, it’s all thanks to them that I have been so much comfortable and that’s to do with their rebate program where I am able to gain 50% back on all trading orders which includes the losing trades too, so it’s all really special and allows me to be successful.