Negative Real Interest Rates

Tim Duy commented on a quote from Scott Sumner.   Sumner wrote:

I don’t think Keynesians should be arguing that lower real interest rates are the key to recovery. A bold and credible monetary stimulus that was expected to produce much faster NGDP growth might well raise long term risk-free interest rates.

Duy then drew a diagram:

According to Duy, the Keynesian view is that an increase in expected inflation will shift the IS curve to the right and increase the (nominal) interest rate.     However, he recognizes that if the Fed commits to increasing nominal expenditure, this will improve "confidence" also shifting the IS curve to the right.   

Rather than show nominal interest rates on the vertical axis,  an alternative approach shows real interest rates.   While there is no zero bound on real interest rates, the zero nominal bound shows as a horizontal LM curve at a real interest rate equal to the negative of the expected inflation rate.  

The real market interest rate is at r1, the negative of the inflation rate.   It crosses the IS curve at y1, leaving real income below potential income, yp.    The result is an output gap.   The natural interest rate, rn is well bellow the real market interest rate.   At that interest rate, the IS curve shows a level of real expenditure equal to potential income.

The way market monetarists see it, Keynesians propose creating sufficient expected inflation so that the real interest rate will fall to the natural interest rate.   The LM curve shifts down and to the right.   (The increase in expected inflation shifts it down, and the real quantity of money expands to shift it to the right. )  The nominal interest rate, stays equal to zero. This will move along an unchanged IS curve, raising real expenditure until it equals potential income, closing the output gap.

Market monetarists instead favor an increase in expected nominal expenditure.   While this could result in higher expected inflation and a lower real interest rate as above, it will also result in higher real income and output.   By increasing future real income and output, this results in increased real expenditure now, shifting the IS curve to the right.   

In this case, the real interest rate rises above zero to the new natural interest rate, rn2.   Real expenditure and real income both rise, to the potential income, closing the output gap.   The nominal interest rate was zero, and now rises above the natural interest rate by the unchanged expected inflation rate.

From a market monetarist perspective, this is the preferred result.    Increasing expected inflation and lowering real interest rates is not in any sense a goal.    Unfortunately, the increase in expected nominal GDP might in fact cause an increase in expected inflation.    If this were to occur, the result would be intermediate.   The IS curve shifts to the right because of expectations of increased real income and output, and the LM curve shifts down and to the right, because of an increase in expected inflation.    With this intermediate scenario, the real interest rate might fall, but it might rise.

In this scenario, the expected inflation rate increases, shifting the LM curve down.   The IS curve shifts to the right because of the increase in expected future real income and output.  The natural interest rate rises to a positive level and the real market interest rate rises to that same level.  The nominal interest rate rises above the new, natural interest rate by the new, higher expected inflation rate.

This diagram represents a scenario where the expected inflation rate inceases, so the LM curve shifts down.   The IS curve shifts to the right because of an increase in expected real income and output.   The natural interest rate remains negative, but it is much higher than its initial level.   The real maket interest rate has risen to this new, higher natural interest rate.    The nominal interest rate, which was zero, is now positive, shown as R2.    Again, the increase in the expected inflation rate was not needed. 

In this scenario, the expected inflation rate rises, and the LM cuve shifts down (and to the right.)   The IS curve again shifts to the right due to the increase in expected real income and output.    The natural interest rate does rise to the new, more negative market real interest rate.   The only differrence from the "Keynesisan" scenario above, is that the expected inflation rate rose by less.   So, while a smaller increase in expected inflation was necessary, because of the small impact of expected future income and output on current real expenditures, some increase was necessary to bring the real interest rate down to a now higher natural interest rate.   An increase in expected inflation was necessary to close the output gap.   The nominal interest rate remains zero.

Which of these scenarios is more realistic?   Market monetarists argue that it is unnecessary to choose, and that instead the sole goal should be an increase in nominal spending on output.   If the better scenarios actually occur, then nominal interest rates should be allowed to rise, and we should be pleased that the recovery occured with no increase in expected (or, hopefully, actual) inflation.    On the other hand, if the increase in nominal spending on output generates higher expected inflation, and this reduces real market interest rates, then so be it.   But the goal is not higher expected inflation and lower real interest rates.    The goal is to close the output gap, so that real income equals potential income and the real market interest rate equals the natural interest rate.

Nominal GDP Targeting and Import Prices

Suppose that the price of an imported consumer good rises significantly.  This could be due to a decrease in the global supply or else an increase in foreign demand.   Further, suppose there is no import competing industry for this particular good.  

Suppose demand for the imported good is unit elastic.   Quantity demanded falls in proportion to the increase in price, so that total spending on the product remains unchanged.   Foreign exchange markets continue to clear.    The total amount spent on imports is unchanged as is the total amount earned from exports.   Spending on domestic product is unchanged as well.    Spending on domestic product generates an equal nominal income.

Real income has fallen.   The price level, including the imported good, is now higher.   And so, deflated by that index, real income is lower.   Less of the imported good is enjoyed.
  
With nominal GDP targeting, this is all that happens.   Spending on domestic product and nominal income were not impacted.  

On the other hand, if the consumer price level is targeted,  then the increase in the price of the imported good pushes the price level above target.   This would trigger a monetary contraction.    For the most part, the result would be reduced prices of domestically produced goods and services, with their prices falling enough to offset the inflation in the price of the imported consumer good.   Of course, since falling demand for any particular good often signals the need to contract production, it is likely that output and employment would fall as well.   Hopefully, in the long run, wages and other nominal incomes would shift to a lower growth path, leading to lower unit costs, and so a recovery of production and lower prices of domestic goods.

To the degree that the monetary contraction triggers higher interest rates and a net capital inflow, the exchange rate will rise.   This could directly dampen the increase in the price of the imported consumer good, lower the prices of other imported goods, and reduce the foreign demand for exported goods.  One the prices of exported goods and imported goods fall, the appreciation of the currency will be justified by purchasing power parity.

With inflation targeting, if the increase in the price of the imported consumer good comes as a surprise, (like the realization of a random variable,) then it is ignored.   Bygones are bygones.   There has been a temporary spike in inflation.   Unfortunately, if the price of the consumer good is foreseen, and part of an ongoing process in real time, then inflation targeting would require a monetary contraction as above.  

Of course, with "flexible" inflation targeting, the central bank can allow inflation when it thinks it best.   But then, that isn't much of a rule.

Could it be that a large run up in oil prices, partly due to supply problems in the Middle East and Africa, but also due to growing gasoline demand from India and China, combined with "flexible inflation targeting," has led to an economic contraction that aims to force down other prices, including wages and other nominal incomes, enough to keep inflation on target?

The Phillips' Curve

Marcus Nunes recently quoted Lawrence Meyer regarding his view of the Phillips curve.    While Meyer granted that money growth determines inflation in the long run, he thought that shifts in the inflation rate are caused by movements in the unemployment rate for labor away from the natural rate of unemployment.   This effects the growth rate of wages which impacts price inflation.   

Meyer explains that while he initially thought he had a good idea of the actual value of the natural unemployment rate, he came to see his specific estimate as being mistaken.     Hopefully, the implication is that Meyer would no longer favor using monetary policy to target the unemployment rate. 
Meyer's approach to the Phillips curve is that changes in unemployment cause changes in inflation.   This was the old Keynesian approach, and new Keynesians look at it much the same way.   (Though old Keynesians were skeptical about the long run adjustment of unemployment to a natural rate.)

An alternative framing is that changes in inflation cause changes in unemployment.   This is the approach that is usually associated with the old monetarists.  The usual story was that an unexpected increase in the inflation rate creates some kind of confusion.   Perhaps workers fail to recognize lower real wages, and so accept job offers more quickly.    And what about unexpectedly lower inflation?    One common story is that those seeking employment fail to see that many of the available job offers reflect acceptable real wages, and so continue to search.

Rather than focus so much on unemployment, like Meyers, Phillips curve analysis frequently just shifts to a discussion of real output and potential output.   The Keynesian approach would be that if real output exceeds potential, then this will cause inflation to rise.   And if real output falls below its potential level, then this will cause inflation to fall.

Focusing on real and potential output, the old monetarist framing is that if the inflation rate rises above its expected level,  then this causes real output to rise above potential.    If, on the other hand, inflation falls below its expected level, this causes real output to fall below potential.   

Market monetarists have a third framing.     An increase in spending on output causes more rapid inflation and more rapid growth in real output.  And a decrease in spending on output leads to slower inflation and slower growth in real output.

Further, market monetarists have come to emphasize growth paths.   If spending on output shifts to a higher growth path, then the price level and real output shift to higher growth paths.   If spending on output shifts to a lower growth path, then the price level and real output shift to lower growth paths.  

From a market monetarist perspective, the Keynesian framing of the Phillips curve is that more rapid growth of spending on output leads to real output rising above potential which then leads to higher inflation.   Slower growth of spending on output leads to real output falling below potential which then leads to slower inflation. 

It seems plausible enough. Firms respond to growing shortages of goods and resources by raising prices (more quickly.)   Firms respond to growing surpluses of goods and resources by raising prices more slowly (or cutting them.)

From a market monetarist perspective, the old monetarist framing of the Phillips curve is that more rapid spending growth leads to prices rising more quickly, beyond what was expected, which leads to more rapid growth in production and employment, pulling it above potential.   Less spending growth leads to slower inflation, below what was expected, which results in slower growth in production and employment, so that it falls below potential.

Again, this seems plausible enough.   Higher prices create an incentive to produce more and higher wages an incentive to work more.   Lower prices usually signal a need to slow (or cut) production and the employment of resources.   (Of course, only if the changed opportunity costs implied by changes in everyone else's selling prices are ignored.   The old monetarist approach is assuming, plausibly enough, that expected inflation reflects those true opportunity costs.)

However, the market monetarist approach would be that rather than place output or prices in second place in the train of causation, they simply occur together.   More rapid growth in spending on output leads firms to expand production and employment (more rapidly) and raise prices (more rapidly.)  Slower growth in expenditures on output lead firms to expand production more slowly (or cut it) and raise prices more slowly (or cut them.)

The market monetarist approach is consistent with scenarios where only prices or real output respond to changes in spending growth.   And it is also consistent with scenarios where one of the two changes at first, and then later both respond.

In fact, most market monetarists favor the natural rate hypothesis where "in the long run," only the price level changes in response to changes in spending on output, and real output remains equal to potential.   This long run equilibrium price level is equal to spending on output divided by potential output.    Long run equilibrium then, requires that spending on output, the price level, and potential output are on growth paths where at each point in time, spending on output divided by the price level is equal to potential output.   Over time, the growth rate of spending on output less the inflation rate is equal to the growth rate of potential output.   

Market monetarists are quite aware that it is possible that the growth rate of spending on output minus the inflation rate equals the growth rate of potential output, but real output remains below potential output.   In such a scenario, the price level is on a growth path that is too high relative to the growth path of spending on output.  Or, in other words, the growth path of spending on output is too low for the existing growth path of prices.

So far, only the old monetarist approach has included anything about expected inflation.   Certainly new Keynesians and market monetarists believe expectations are important.   However, any remaining old monetarists, and the new classical and real business cycle theorists who have adopted their framing, have a tendency to use this simple Phillips curve to understand the real world.    If inflation is equal to expected inflation, then real output must be equal to potential.   If there is an inflation target, and inflation is on target, then real output is equal to potential.   If it were somehow possible for expected inflation to rise, and actual inflation remained unchanged, then this would force real output below potential.   However, by far the most likely result of an increase in expected inflation would be for inflation to match expectations, and real output remain equal to potential.    Observed changes in real output are almost certainly due to changes in potential output.  

From the Keynesian perspective, especially the new one, a Phillips curve shows that expected inflation, along with the output gap, determines actual inflation.    If there is an inflation target, and it is credible, the actual inflation rate will vary about the expected one with the output gap.   If real output is less than potential, inflation will be below target.    If real output is above potential, then inflation will be above target.    Given expected inflation, changes in nominal interest rates cause proportional changes in real interest rates, which impact real expenditure and so real output.    An increase in the inflation target, if credible, will raise the actual inflation and have no effect on the output gap, at least in terms of the Phillips curve.    For new Keynesian economists, given the nominal interest rate, higher expected inflation leads to lower real interest rates, more spending on output, and a smaller output gap.      Plenty of inflation expectations are involved.

And what of market monetarists?    What is the monetary regime?   Growth path for the price level?   Growth path for nominal GDP?   Inflation target?   Flexible inflation target?   It depends.

However, it is safe to say that market monetarists do not believe that real output can only be below potential as an ephemeral condition where expected inflation has failed to fall enough to match actual inflation.   Instead, it occurs when spending on output is too low relative to the price level, which  is the same thing as the price level being too high relative to spending on output.

Further, an increase in expected inflation, given expected growth in spending on output, is likely to be contractonary.   And still further, a given increase in expected spending on output will be more expansionary, the lower the increase in expected inflation.  Keynesian claims that we "need more inflation," or "we need to raise expected inflation" are wrong headed.
Finally, consider a scenario where inflation targeting is so credible, that everyone always raises prices and wages in a way consistent with the target--no matter what.   Actual inflation is always on target.    A central bank whose sole mandate is an inflation target would be completely successful.

Considering the old monetarist view of the Phillips curve, real output would necessarily be equal to potential output.   Of course, they would be assuming that the quantity of money would be just right at each point of time so that real spending on output equals potential at the target price level.   In their framing, if real spending on output was wrong, the price level, and so inflation, would deviate from target.

From the Keynesian perspective, there is a puzzle.   If there appeared to be an output gap, that would signal the need to lower interest rates.   Still, the model of the Phillips curve would suggest that inflation should fall too.   If inflation does not fall, then perhaps there really is no output gap.   Maybe it is better for the central bank to just be pleased with its success in keeping inflation on target and not worry about the output gap.   Imagine the debates.   Maybe they should just wait and see.

From the market monetarist perspective on the Phillips curve, if spending on output falls to a lower growth path, then it is certainly possible for inflation to be growing at the target growth rate while real GDP grows, but on a path well below potential.    It might be puzzling as to why firms raise prices (and give pay increases) at a rate consistent with what the central bank tells them should happen, when there are surpluses.    But the basic model doesn't imply that this is impossible.  
With a nominal GDP target, even if prices become very sticky at the targeted trend, there is no problem.   For example, suppose there is a 3% growth path target for nominal GDP and potential output has a trend growth path of 3%.   This results in a constant trend price level of 100.   Suppose that price level expectations become so strong that the price level never deviates from 100.   However, if spending on output falls too low, then even if the price level is stuck, real output falls, so that nominal GDP falls belwo target, signaling a need for an expansion in spending.

But suppose the central bank's target for nominal GDP  is so credible that  production continues to rise 3% and prices remain at 100?   Then what is the problem?   

In my view, with nominal GDP targeting, the expected price level is equal to the target for nominal GDP divided by expected level of potential output.   An increase in nominal GDP above target is an unexpected increase in spending on output, and this will cause the both the price level to rise above its expected level and the inflation rate to rise above its expected level.    The increased spending on output would also raise real output above potential.

However, the expectation that nominal GDP will return to target in the subsequent period will dampen any increase in spending on output.    In particular, expansions of consumption due to temporary increases in income and or investment due to temporary increases in sales would be dampened to the degree that people perceive that the economy is "booming."

Further any increase in the price level would be especially dampened by the expectation that nominal GDP will return to target.  This is especially easy to see if the trend price level is constant.  (Nominal GDP growing at the same rate as potential output.) The "boom" would be a time when prices are high for buyers, motivating them to postpone purchases.    It is a time when prices are high for sellers, which would certainly motivate them to release inventories, and perhaps expand production.   However, gearing up for a permanent increase in production would be a mistake, with nominal GDP returning to target the next period. 

If nominal GDP would grow more slowly and fall below target, then nominal GDP would be below its expected level and prices would fall (or grow more slowly.)  The price level would be below its expected level and the inflation rate would be below its expected level.    Real output would grow more slowly (or perhaps fall.)   Real output would fall below potential.

Fortunately, the expectation that nominal GDP will rise to target in the subsequent period will dampen any decrease in spending.   Long term investment would be little impacted.   Worries about future job losses would have little impact on saving.   And, in fact, reduced saving to maintain consumption levels is sensible, when real output and income is expected to recover rapidly.

Again, the effect on inflation would be very dampened, as is easy to see if the price level is stable on average.   Any deflation would be pushing prices to a temporarily low level, making this an especially good time to buy.   It also makes accumulating inventories attractive, selling the goods later when prices recover.    Because of the nature of nominal GDP targeting, where negative inventory investment takes away from nominal GDP, keeping nominal GDP on target in the subsequent period provides an outlet for selling off accumulated inventories.

Suppose instead that potential output increased more than expected.   This would cause the price level to fall below the expected level and cause unexpected disinflation (deflation rather than stable prices.)   Would actual output to fail to rise with potential?    While this is a possibility, an alternative is that those particular firms whose productivity has increased will see their unit costs fall, and so lower their particular prices to expand their sales.    While it is true that productive capacity is above its expected value and the price level ends up below its expected level, the rational for deviations of real output from potential that occur from slower spending on output don't apply.

If potential output increased less than is expected, this would cause the price level to rise above its expected price level and inflation to rise above its expected inflation.  (Inflation rather than the expected stable price level.)    Would this cause real output to rise above potential?   While possible, it is much more likely that the particular firms and markets where productive capacity has grown more slowly (or decreased) would raise their prices to reflect their higher unit costs.    While prices would be higher than expected, there is nothing like the process by which unexpected increases in spending leads to more rapid growth in production as well as prices.

In my view, neither the Keynesian nor the old Monetarist framing of the Phillips curve are able to capture these relationships.

A Scenario Where a Price Level Target is Superior

In my view, nominal GDP targeting is superior to a price level target when there is a decrease in the supply of some particular good (the cotton blight.)   It is also better when there is a shift in demand, and the prices (including resources like wages) are sticky in the sector losing demand.  

Are there any scenarios where a price level target is superior to a nominal GDP target?

Suppose there is a 10 good economy.  All goods are produced by self-employed sole proprietors.   There is a uniform reduction in supply for all goods.   Perhaps the economy is all agricultural and there is bad weather.   More plausibly, there is a government regulation that has more or less the same adverse effect on all economic activity.

While the decrease in supply of each and every good would seem to raise their prices, the reduction in the real income that can be earned will reduce the demands for each and every good.   If prices remain the same, all markets clear.    The price level is the same.     With the quantity of each and every good reduced, real output has fallen.   Lower quantities at constant prices implies less expenditure on output.   This results in a lower nominal income.   Lower nominal income at constant prices is lower real income, matching the lower real output.

If one of the 10 goods serves as medium of account, and its supply is reduced in proportion to all the other goods, then the result is as above.    On the other hand, if there is a fiat currency in fixed supply, (or a gold standard where the existing stock of gold is taken as fixed,) and money is a normal good, then the reduction of real income will result in a lower real demand for money.   Given the quantity of money, there will be an excess supply of money.   As that money is spent on the various goods, the price level will rise.   This will raise nominal expenditure and nominal income, but deflated with the higher price level, real income and and real output remain the same.

A nominal GDP target would have the same effect as a fixed nominal quantity of money (or a gold standard with a fixed stock of gold.)   To the degree the sole proprietors understood the nature of the regime, they would not expect lower nominal demands for the products, and would respond to the decrease in supplies by charging higher prices.   Oddly enough, this is what they would do if they were truly ignorant of the nature of the problem, and simply took the decrease in productivity to be peculiar to their own market.   

A price level target, on the other hand, would allow the price level to remain stable.   With a fiat currency serving as medium of account (and medium of exchange,) its quantity would need to be reduced in proportion to the decrease in the supplies of the other goods.   But then, nominal income would fall with the reduced real output and real income.    Interestingly, if the demand for credit is positively related to real income and real output, and interest rates are pegged by the central bank, then it might "automatically" reduce the quantity of money in line with the reduction in real income and output. 

This scenario has approximately no connection to the real world.   While uniform productivity shocks are possible, self-employed sole proprietors?   Unfortunately, it does have a more than passing resemblance to the models used by mainstream macroeconomists.

Nominal GDP Targeting and Shifts in Demand

How do various monetary regimes perform when there is a shift in demand between two goods?

Consider an economy with ten goods.   The demand for cotton rises and the demand for burgers falls.   If the prices of both goods are perfectly flexible, and assuming the supply and demand curves have the usual shapes, the price of cotton rises and the price of burgers falls.   The quantity of cotton rises and the quantity of burgers falls.   There is little or no impact on the price level or aggregate real output. 

If the reason for spending less on burgers was to be able to spend more on cotton, then it is reasonable to assume total spending has not changed at all.     This implies an unchanged nominal income.   With the price level more or less unchanged, real income is unchanged along with real output.

Suppose that the supplies of both cotton and burgers are more elastic in the long run than in the short run.    As time passes, the initial changes in the prices of the two goods are at least partially reversed, while the changes in the quantities grow.    Since the prices and the quantities moves in opposite directions for each good, the effects on spending for each good is  ambiguous, but likely to be small as is the effect on total spending.   With the price of cotton rising and the price of burgers falling, partly reversing the initial change, there is little effect on the price level.   

Now suppose that the supply of cotton is inelastic in the short run, but the supply of burgers if very elastic.    While the cotton farmers earn greater profits due to the higher prices and would like to expand production, it takes time to obtain the necessary machinery, hire and train workers, and so on.   On the other hand, the burger industry can quickly close down marginal stores and lay off workers.    In effect, the price of cotton appears flexible in an upward direction and the price of burgers appears sticky in the downward direction.  

With these assumptions about the elasticities of supply, the price of cotton rises a lot, and the price of burgers falls just a little.   The price level is higher.    Further, the large decrease in the quantity of burgers, evaluated at nearly unchanged prices, is a large decrease in real output.   And while the increased spending on cotton may exactly match the decreased spending on burgers, deflating the greater nominal value of cotton by the higher price suggests only a small increase in the real production of cotton.

The shift in the demand away from a good with an elastic supply and towards a good with an inelastic supply results in a higher price level and a lower level of real output.     As time passes, the elasticity of supply for burgers might rise even more, but there is little room for change.   On the other hand, the great transitory profits for cotton production provides a signal and creates an incentive for an expansion of production.   Supply becomes more elastic in the long run.    And so, as time passes, the price of cotton falls and its production rises, while the price and production of burgers change little.    Therefore, the price level falls back towards its initial value and real output rises back up to its initial value.   Because cotton prices are falling at the same time cotton production is rising,  the effect of this long run adjustment of spending on cotton and also of total spending is ambiguous, but likely small.

Clearly there are many second order effects of these changes.   The supplies of subsittutes in production for cotton fall.   The demands for cotton substitutes also make adjustments--rising then falling.   The demand for resources used in cotton production rise and so does their prices.    The demands for resources used in burger production falls.   Presumably the demands for complements for burgers fall.  The slight transitional change in real income would impact the demands for luxuries and necessities in a different fashon.   Still, it is likely that the largest effect would be on the prices of cotton and closely related products.

I have managed to do this two-sided micro analysis with no assumption about the monetary regime.   (I did the same with the initial analysis of the cotton blight.)     I wonder how many other economists, particularly with a micro approach, would do the same?   It is, of course, unrealistic.   Somehow this simple shift in demand between two goods is done with some implicit monetary regime in the background.

Suppose that one of the other eight goods, call it gold, serves as medium of account.   The dollar is defined as a fixed amount of gold.    This regime appears consistent with the analysis above.   The price of burgers fell relative to the price of gold and the price of cotton rose relative to the price of gold.   There is no particular reason to expect that the price of gold relative to the prices of any other the other of the seven goods changed much.    In the scenario where the supply of burgers was elastic in the short run and  the supply of cotton inelastic in the short run, then real output and real income decreased.   If gold is a normal good, this would reduce the demand for gold, but the same would be true of all normal goods, and the price of gold, like all the other goods, did fall relative to the price of cotton.

If gold serves as medium of exchange, in the initial scenario, the price level and real output remained unchanged.   There was no change in the real demand to hold money or the real quantity of money.   There is no reason to expect monetary disequilibrium.  

In the second scenario, where the price level increased, because the price of cotton rose more than the price of burgers fell, this would reduce the real quantity of money.   However, the reduction in real output and real income would reduce the real demand to hold money.   No shortage of money would develop that would prevent the temporary increase in the price level.   Further, as the production of cotton adjusts in the long run due to the growing elasticity of supply over time, the lower price of cotton and higher quantity cotton implies a lower price level and higher level of real income.   The real quantity of money rises, but the real demand for money rises as well.

Suppose that instead of  some "third good," like gold, cotton served as medium of account.   With the the dollar price of cotton fixed by definition, the increase in the demand for cotton cannot raise its dollar price.   But with an increase in demand, and particularly in the scenario with an inelastic short run supply, the relative price of cotton must rise.   This can only occur by a decrease in the dollar prices of the other nine goods.   While a Walrasian auctioneer could manage this easily, in a decentralized market economy, the signal to those selling the other nine goods would be fewer sales at existing prices.   The burger market, already suffering lower demand, would be hit again, but more importantly, other markets, that are not directly related to the change, would have to make adjustments in their prices because of the shift in demand between burgers and cotton.   Of course, as the long run adjustment in the quantity of cotton occurs due to the greater long run elasticity of supply, the relative price of cotton would again fall, requiring an increase in the prices of other goods.   The increase in the demand for cotton, when cotton serves as medium of account, results in first a perhaps sharp deflation, and then an inflation as the production of cotton makes its long run adjustment.  

While the second order effects of these adjustmetns would impact many markets, and prehaps the market for every other good, it is hardly likely that the decrease in demand for those other goods would accurately signal a need to cut back the production of all of them.   While the production of burgers most certainly needs to be reduced due to the lower demand, freeing up resources to produce more cotton, any adjustments in the optimal production of the other goods is ambigous and almost certainly not a decrease in all of them.   At first pass, the production of those goods best that can best absorb the resources freed up by the contraction of the burger industry should expand, while the production of those industries with resources that can also be used to produce cotton should contract.

Interestingly, if burgers served as medium of account, and the supply of burgers was highly elastic, then a decrease in demand would require only a slightly lower relative price and so only a slight inflation in the prices of the other goods    The price of cotton, however, would still rise sharply due to the difficulty of expanding production quickly, and then fall.   This doesn't necessarily mean that  burgers would be the ideal medium of account.   It isn't obvious that if there were an increase in the demand for burgers, production could be ramped up quickly.

If gold served as medium of account, shifts in the demand between other goods, like cotton and burgers, could result in short run changes in the price level and real output.   On the other hands, shifts in the demand between other goods and gold would likely cause very sharp changes in the price level and real output.   An increase in the demand for gold would result in deflation, and given the nature of gold production, only a very slow adjustment in long run production.   Further, given the large stock of gold relative to current production, a decrease in the demand for gold would not have an effect like burgers, but rather would be be highly inflationary, though the price level should settle at a higher level.

Suppose a fiat currency is used instead some good like gold to serve as medium of account and medium of exchange.  If there was an increase in the demand for cotton and decrease in the demand for burgers, the effects would be much the same as if gold served as medium of account and medium of exchange.    In the first scenario, there is little or no effect on the price level and so the real quantity of the fiat currency would be unchanged.   With real output and real  income unchanged, there would be little impact on the demand for the fiat currency.   

In the second scenario, in the short run, the price level increases due to the significantly higher price of cotton and only slight decrease in the price of burgers.   The real quantity of the fiat currency would fall.   However, because of the reduction of real output and real income, the real demand to hold money would fall as well.   There would be no shortage of money preventing the temporary increase in the price level.   And, as cotton production expands and the price of cotton falls again, the real demand to hold the fiat currency would rise with real income as the real quantity of the fiat currency rose with the lower price level.   A fiat currency with a fixed quantity would not prevent temproary changes in the price level due to a shift in demand between goods.

Unfortunately, if the real demand for the fiat currency should change and the quantity remains fixed, then the result is similar to the effects of an increase in the demand for cotton when it serves as medium of account.   It would be almost exactly the same as the effects of a change in the demand for gold when it serves as medium of account.  Fortunately, the quantity of a fiat currency can be adjusted according to changes in demand.

Since shortages or surpluses of money lead to changes in the price level, it would seem plausible that maintaining a stable price level is the proper rule for its management.   However, consider the scenario where the demand for burgers has decreased and the demand for cotton has increased.   If the prices are equally flexible, or alternatively, the elasticities of supplies and demands result in offsetting impacts on prices, then the price level would remain unchanged, and the regime of price level stabilization would require no adjustment in the quantity of money.

On the other hand, in the scenario where the supply of cotton was inelastic and the supply of burgers elastic, the price level increased.   A regime of maintaining a stable price level would require a decrease in the quantity of money sufficient to force the price level back down.   While there would be some slight dampening in the increase in the price of cotton, and perhaps a slightly greater decrease in the price of burgers, the larger effect would be modest decreases in the prices of the other 8 goods in the economy.  

As the production of cotton gradually expands, due to supply being more elastic in the long run, the lower price of cotton would result in a lower price level, which would require an expansion in the quantity of money.   This would dampen the decreases in the price of cotton,  reverse any added decrease in the price of burgers due to the monetary contraction and, most importantly, allow all the prices of the other goods to recover.  

To the degree that the other 8 goods respond like burgers, especially to a temporary decrease in demand, a sharp monetary contraction might be needed to reduce their prices enough.  Unfortunately, that implies the simultaneous a sharp contraction in the production of all of those other 8 goods.

 It is instructive to consider the impact of the monetary contraction on the process by which the quantity of cotton recovers.    Without the monetary contraction, the price of cotton rises further and so the profitability is immediately greater.   Resources are pulled into the cotton industry, though with the contraction of the burger industry, there is some pushing of resources towards cotton as well.   With the monetary contraction, the price of cotton rises less, and so there is a bit less pull, but the rest of the economy contracts more, so there is much more push.   In other words, it would be lower resources prices, including wages, that would provide a more significant part of the impetus to expand cotton production.    If resource prices, and especially nominal wages, were sticky, there could be very little impetus from that source.   

In the limit, suppose that all prices and wages are completely rigid downward.   The monetary contraction must be severe enough to prevent the price of cotton from rising, that being the only way to keep the price level from rising.   While production and employment in the rest of the economy would fall, there would be no signal to those producing cotton that it is profitable to expand.    Fortunately, it is unlikely that prices would all be perfectly rigid, but hopefully considering that extreme scenario illustrates why simply allowing the increase in the price of a good with increased demand to cause a higher price level is the least disruptive approach.   Forcing other prices down to keep the price level stable is simply an undesirable disruption in the the adjustment process.

Suppose that nominal GDP is targeted rather than the price level.   In the first scenario, the price and quantity of burgers fall and the price and quantity of cotton rises.   Even if spending on each good changes, total spending was unchanged.   A target for nominal GDP has the same consequence as a target for the price level.   There is no monetary change necessary.

In the second scenario, however, where the supply of burgers is elastic in both the short and long run, but the supply of cotton is inelastic in the short run and becomes more elastic over time, there is a difference.   Like a gold standard and a fixed quantity of money, a rule for stable nominal GDP requires no monetary change.    While the price of burgers falls little, the substantial reduction in the quantity of burgers imply a substantial decrease in spending on burgers.   Similarly, while the production of cotton rises little in the short run, the price rises substantially.  Spending on cotton rises.   While the price level increases, real output decreases, leaving nominal GDP little changed.   As the production of cotton expands, and the price falls, there are offsetting impacts on spending on cotton.    The lower price level and expansion in real output have an offsetting impact on nominal GDP.  

Backward looking inflation targeting would have a similar consequence as nominal GDP targeting.   If the increase in the price of cotton is observed, then no effort would be made to reverse it.    Unfortunately, forward looking inflation targeting would require a monetary contraction that prevented the shift in demand from causing a higher price level.    And, of course, flexible inflation targeting allows the monetary authority to do as they choose.  Unfortuately, it appears that this very flexibility creates a motivation to "maintain credibility" and "keep inflation expectations anchored" so that a contractionary monetary policy is implemented when there is a shift in demand between goods.

A rule keeping nominal GDP on target allows those prices of those particular goods with increased demand to rise, generating appropriate signals and incentives for those firms to expand production and employment.    As production expands, prices then fall again, returning the price level to its intitial level.   Nominal GDP targeting is much superior that stablizing the price level when there is a shift in demand between various goods.   Slow, steady growth in spending on output provides the least bad macroeconomic environment for microeconomic adjustment.

The Cotton Blight Problem

Like most market monetarists, my approach to monetary policy emphasizes alternative regimes.  Which is best, (or from my rather pessimistic turn of mind, which is least bad?)

One test for monetary regimes is the cotton blight test.  What happens when there is a decrease in the supply of a particular good?   

Suppose we have an economy with ten goods, one of which is cotton.   Unfortunately, there is a cotton blight that ruins Mississippi for growing cotton.   Initially, the value of the cotton crop makes up about 10% of the entire economy.  

The effect of the blight is so severe that the price of cotton doubles and the production of cotton falls in half.   With cotton initially making up 10% of the economy, the price level has risen by 5% and aggregate real output has decreased by 5%.

Because the demand for cotton happened to be unit elastic, spending on cotton remains the same.   At least initially there is no reason to suppose that the prices or quantities of other goods change.   Total spending on those other goods and  total spending on all goods together remains the same.

However, real income has fallen.  The same total spending generates an equal nominal income, but with prices 5% higher, real incomes are all 5% lower.   This, of course, matches the 5% reduction in real output.

The increase in the price of cotton provides an appropriate signal and incentive to conserve on the use of cotton.   It also provides an appropriate signal to cotton farmers in South Carolina to intensify their production efforts (and to work hard to keep the blight from spreading.)  It creates an incentive to come up with a cure for the blight.

Of course, there would likely be second order effects.   The demand for substitutes, like synthetic fibers, would rise.   The demand for complements, like bleach might fall.   The demand for products preferred by cotton farmers in South Carolina could rise, while the demand for the products preferred by the cotton farmers of Mississippi might fall.  With real income falling by 5%, the demand for luxuries would fall more than 5%, while the demand for necessities would fall by less than 5%.   But it is difficult to see why the higher price of cotton and the need to pay more for cotton would fail to provide the proper signals and incentives to make these adjustments.

Suppose one of the other 9 goods, call it gold, serves as medium of account.   The unit of account, the dollar, is defined as a fixed amount of gold.    The result is approximately as described above.   The price of cotton doubles in terms of gold.   The prices of the other 8 goods are little changed, and the dollar price of gold remains fixed.   Cotton production is halved, and real output falls by 5%.   The price level rises 5%.   While total spending on output and nominal income is unchanged, real income falls by 5%.

Of course, it is usual to think of a gold standard as including more than using gold to define the unit of account.   Gold might also be used as a medium of exchange.    Does that make any difference?

The increase in the price level by 5% reduces the real supply of money by 5%.    However, the lower real income reduces the demand for money.   If the real demand for money were exactly proportional to real income, then even if gold served as money, the effect would be the same.   The price level rises 5%, spending on output remains the same, nominal income remains the same, and real income falls 5%.  

Now, suppose instead that cotton is the medium of account.   The cotton blight still reduces the supply of cotton, and its relative price doubles and the quantity falls in half.    With the dollar defined as an amount of cotton, the dollar price of cotton cannot rise.   Instead, the dollar prices of all the other 9 goods must fall by 50%.    With the price of cotton unchanged, the equilibrium price level falls by about 45%.   (No change for 10% of the economy and a 50% decrease for 90% of the economy.)   Nominal spending falls by 50%, with 50% less cotton purchased at the same price and assuming prices are sufficiently flexible, 50% lower prices for the same amounts of the other goods.  Nominal incomes fall the same amount.   But with the price level falling 45%, real income falls about 5%, matching the 5% decrease in real output.
With a Walrasian auctioneer, this adjustment in prices could be accomplished easily.   An economist doing these calculations can solve for the 50% lower prices for each of the 9 other goods just as easily as he can solve for the 50% higher price of cotton.    In a decentralized market economy, the signal each seller is going to receive for the need to set lower prices is lower sales at the previous prices.   Frequently, lower sales imply a shift in demand away from a particular good towards other goods and so the need to reduce production and release resources for the production of the other goods.   The likely effect of the cotton blight would be a general depression of production throughout the economy and at best only a gradual recovery as prices fell.

To avoid problems associated with picking a commodity as a medium of account, suppose that instead a fiat currency is issued and its quantity is fixed.   The fiat currency is solely demanded as a medium of exchange, and it serves as medium of account.   The dollar is defined as unit of the fiat currency.
Now, return to the cotton blight.   As before, the immediate effect is a 50% increase in the price of cotton and a 50% decrease in the quantity of cotton.   With cotton being 10% of the economy, that is a 5% increase in the price level and a 5% decease in real output.    It would appear that nominal expenditure would be unchanged in aggregate as would nominal income.   Again, with the price level rising 5%, real income falls with real output.

But what of the fiat currency?   The 5% increase in the price level is a 5% reduction in the real quantity of money.   However, if money is a normal good, the reduction in real income also reduces the real demand for money balances.   If the demand for real money balances is exactly proportional to real income, then the real quantity of money falls exactly in proportion to the decrease in the real demand for money.   There is no shortage of money to force a return of  the price level to its initial level.    The 5% increase in the price level is consistent with monetary equilibrium.

A "blight" on currency should be easy to avoid.    Assuming the demand for the fiat currency hasn't changed, don't decrease the quantity of currency.   Print more to replace any that wears our or is damaged.

On the other hand, it is possible that the demand to hold the fiat currency might change, and the effects of those changes might be roughly similar to using a good that is subject to a blight as medium of account.    In particular, the quantity of the fiat currency could be managed so that is perfectly elastic--it changes according to the demand to hold it.  

What happens if the quantity of a fiat currency rises beyond the demand to hold it?   Those with excess money balances spend them, increasing the demands for the various goods in the economy.    Other things being equal, this tends to raise their prices, and so the price level.

On the other hand, if the quantity of the fiat currency is less than the demand to hold it, those receiving money in payment build up their holdings by refraining from spending it.     Other things being equal, the reduced demands for various products would result in lower prices for them and so a lower price level.

It appears that the proper rule for a fiat currency is to stabilize the price level.   If the price level rises, this suggests that too much money was created, outstripping the demand to hold it.  If the price level falls, this suggests that too little money was created, failing to meet the demand to hold it.

Unfortunately, appearances can be deceiving.  Return to the scenario of the cotton blight.   If a fiat currency was being managed to stabilize the price level, then the immediate effect of the 50% increase in the price of cotton is a 5% increase in the price level.   To reverse this inflation, the quantity of the fiat money would need to be reduced by approximately 5%.   This would result in the prices of all goods falling about 5%.   The price of cotton would fall 5% as well, but that would leave it about 47.5% higher than its initial level.   The prices of the other 9 goods in the economy would fall about 5%.    Nominal spending in the economy would be approximately 5% lower.   Nominal income would be 5% lower.    The 5% lower nominal income with the price level remaining the same implies 5% lower real income, matching the 5% reduction in real output.

This could easily be managed by a Walrasian auctioneer.  And, of course, an economist can do these calculations quite easily.  It is really not much more difficult than calculating what happens to the price level when the price of cotton simply rises 50% and stays there.   Unfortunately, in a decentralized market economy, the only way all of those selling the other 9 goods get a signal that they need to cut their prices by 5% is lower sales at current prices.  Lower demand usually means that prices perhaps should drop, but production should drop so that resources can be freed up to produce other goods.   In this situation, it is the wrong signal.

If the quantity of the fiat money is managed according to nominal GDP targeting, then when the blight increased the price of cotton by 50% and the price level by 5%, no change in the quantity of money would be necessary.   Because spending on output  and nominal income are unchanged, there is nothing to be done.   Nominal GDP targeting has the same effect as a gold standard (assuming there is no blight on gold itself) or a fixed quantity of fiat money.   Nominal GDP targeting simply avoids an inappropriate monetary intervention to force the prices of other goods to fall enough to offset the effects of increase in the price of the good with reduced supply.

Backward-looking inflation targeting has the same consequences as nominal GDP targeting.   Suppose the inflation rate is targeted at zero percent.    When the cotton blight raises the price of cotton by 50% and so the price level by 5%, nothing is done.   The rule just seeks to avoid any further change in the price level.    Unfortunately, forward-forward looking inflation targeting still requires a monetary contraction in anticipation of the cotton blight.   In order to keep the price level from rising, the quantity of money would be decreased by 5%, so that the price of cotton will only rise 47.5%, and all the other prices be cut 5%, leaving the price level stable.

Generally, a forward looking posture would seem sensible.   For example, if the demand for money were to fall, it would be better to reduce the quantity of money before it had a chance to cause prices to rise.   With a backward looking inflation target, nothing would be done until after the price level had already risen, with the aim of preventing further increases.

Flexible inflation targeting allows the monetary authority to allow inflation whenever it believes it is appropriate.   In that case, it would have discretion to decide whether it should reduce the quantity of money to partly offset the effects of developing cotton blight on the price level.   Unfortunately, with this discretion comes a need to build credibility.   Allowing a developing cotton blight to generate a 5% increase in the price level appears to create worries that the quantity of fiat currency will be allowed to greatly surpass the demand to hold it, resulting in "unanchored inflation expectations."

Worse, a flexible rule is hardly a rule at all.

Nominal GDP targeting looks to be the least bad response to the cotton blight problem.   It is much better than a price level rule, forward looking inflation targeting, and it is an actual rule, unlike "flexible" inflation targeting.

Hamilton: Should the Fed Do More?

James Hamilton is skeptical that the Fed should do more.  By doing more, he has in mind targeting more rapid inflation or more rapid growth in nominal GDP.  He then asks what concrete measures the Fed could use, and suggests further quantitative easing--purchases of more government bonds, particularly longer term ones.    
Market monetarists favor targeting a slow, steady growth path for nominal GDP.    If nominal GDP should fall below that growth path, then "more rapid" growth in nominal GDP is appropriate.   That is the only way for it to recover to the target growth path.

During the Great Moderation, nominal GDP was on a stable growth path, though there was no clear target rule.   Still, most market monetarists believe that it was the drop in nominal GDP far below its trend growth path in 2008 that generated many of today's economic difficulties.     While most have given up their calls from 2009 for a commitment by the Fed to return to that prior trend, most favor some catch up growth, defining the initial value of the new growth path for nominal GDP targeting well above its current value.

Not only are market monetarists among those calling for more rapid growth in nominal GDP, they also call for whatever amount of quantitative easing is needed to reach the target.     In particular, market monetarists reject putting some specific limit on asset purchases as a constraint.    If base money has to be 50%, 80%, 100% or 120% of the target for nominal GDP, then so be it.

Now, market monetarists don't expect that such large increases would actually be necessary, but what is essential is that the Fed be willing to do whatever is necessary.   It is that commitment that will generate market expectations that will cause rapid increases in spending up to the target.  (And the commitment must be equally strong to sell off assets and reduce base money however much is needed to reverse any deviation of nominal GDP above target.)

Hamilton's concern is that the heroic open market operations will fail to get nominal GDP growing faster and cause the government to fund the entire national debt by what is in effect overnight borrowing.     The problem with funding the entire national debt "short" is that if there is a loss in confidence by lenders, the government could be called upon to pay off the entire national debt more or less immediately.    If that were to occur, then either explicit or inflationary default would be unavoidable.

If, instead, the government borrowed in a more balanced fashion, then a smaller portion of the national debt would come due each year.    For example, it might be possible to have 5% come due each year.     While running a primary surplus sufficient to pay that off would be difficult, it isn't literally impossible.

Hamilton's ineffectiveness argument has been made before.   He cites Cochraine.     The argument is that having the Fed purchase long term government bonds is functionally equivalent to having the Treasury fund the national debt with fewer long term bonds and more Treasury bills.

Hamilton claims that when the Fed purchases long term bonds, it pays for them by creating reserve balances for banks.   The reserve balances pay low interest to the banks.  (In fact, slightly higher than the yields on very short T-bills.)   Since reserve balances at the Fed are safe and short government liabilities just like T-bills, there would be no difference than if than if the Treasury sold T-bills and use the proceeds to pay off long term government bonds.

(I have a small quibble with Hamilton's discussion.   When the Fed buys long term bonds from someone other than a bank, it directly increases the balance in some firm's or household;s checkable deposit.    Assuming primary dealers aren't solely providing bonds to the Fed by permanently reducing their inventories of government bonds, then quantitative easing has a direct impact on the quantity of money held by the nonbanking public.)

If Treasury bills are effectively money, then the Treasury can create money by refinancing the national debt, funding it with T-bills rather than long term government bonds.   So?   Hamilton seems to appeal to intuition, just as others have before.   Obviously, having the Treasury refinance the national debt with T-bills can't be a "panacea."    Well, obvious to whom?

The argument is that because the interest rate on T-bills is so low, they are a perfect substitute for money.    Whatever intuition one might have about  changing the financing of the national debt by selling short term rather than long term bonds under "normal" conditions, when T-bills have flexible market prices that equate their supplies and demands, would hardly apply.  Under normal conditions, yields might fall on long term bonds as they are paid off and yields might rise on T-bills as more are issued.   Any effect on aggregate expenditure on output would surely be minimal--under normal conditions.

But if there is some special situation where T-bills are a perfect substitute for money, then financing the entire national debt with T-bills is a massive increase in the quantity of money--more than a doubling in 2012.   My intuition from normal conditions is that such a huge increase in the quantity of money would cause a substantial increase in nominal expenditure on output, and be highly inflationary.

Unfortunately, I suspect my intuition is as wrong as Hamilton's.   I have no confidence in the ability of even massive quantitative easing to greatly increase nominal expenditure unless it is combined with an explicit commitment to an appropriate target.    Even vast increases in the quantity of money that are expected to be withdrawn in the near future will have little or no effect in expenditures on output or the price level.

Interestingly, the quantity of this quasi-money (T-bills as perfect substitutes for money) will automatically decrease when the demand for it  falls.   Their prices will fall and their yields will rise.   They will cease being "money."    (I am sure that the Divisia advocates can explain this better.)

When the Fed purchases the long term bonds, there is nothing "automatic" about the reserves ceasing to serve as money when the demand for them falls.   Of course, if the Fed is expected to sell off the bonds and decrease reserves when the demand for money falls, then quantitative easing would have much the same effect as having the Treasury refinance its debt.    The solution is for the Fed to commit to sell off the bonds only if nominal GDP rises above an explicit target growth path.

I agree with Hamilton that with a government central bank, creating money involves shortening the term to maturity of the national debt.    Considering the crisis in Greece, and the years of government financial crises described in Reinhart and Rogoff's This Time is Different, there is reason to believe that having the government borrow a lot and very short is not such a good idea.

Those who favor high interest on reserve balances to "flood the economy" with liquidity or favor 100% reserves for monetary liabilities should think about what that means for public finance.   Sure, if the regime is one where the quantity of money never falls, and its purchasing power is just allowed to decrease when the demand to hold money falls, there is no problem. (Well, there is that hyperinflationary collapse scenario.  Why demand this paper currency anyway?)  Or, if you assume that the government always  has the ability to borrow at reasonable interest rates because there is no risk of default (this time is different, right?) then any decrease in the demand for government money can be matched by a decrease in the quantity of money and an increase in the amount of interest-bearing  government debt.   

In my view, Hamilton's worry about a shorter term to maturity for the national debt is sensible.    This suggests that the interest paid on reserve balances should be lower than the interest rate on short term government debt, so that banks conserve them.  Flooding the banking system with liquidity is a bad idea.  Further, while reserve requirements are not too relevant right now with banks holding way more reserves than legally required, they should go.   Finally, privatization of hand-to-hand currency should be considered.     These sorts of reforms would minimize the need for government to "borrow short" as part of the ordinary operation of the monetary order.

It should not be the government's role to provide lots of short and safe assets for investors.   If the market clearing yield on those sorts of assets is less than zero, then people should pay to hold them.   The government is in no better position than anyone else to "liquidate" its assets and pay off huge debts.   The amount of primary surplus it can generate is small in any one year, and so, it should keep its total liabilities low and its current liabilities lower.

Of course, as long as the entire financial system is based upon government guaranteed, zero nominal interest, hand-to-hand currency,  no other short and safe asset can have a nominal yield less than the cost of storing currency.   (And if currency is issued in "convenient" dominations, say $100,000, $1,000,000 or $10,000,000, how high are the storage costs?)    It is this constraint that requires that the central bank (or Treasury) stand ready to issue whatever amount of short and safe assets people want to hold, or else face a contraction of nominal expenditure on output and possible deflationary spiral of prices and wages.

Fortunately, keeping nominal GDP from falling below trend prevents avoidable financial problems for government.    Further, keeping nominal GDP from falling below trend avoids conditions where large increases in the demand and decreases in the private supply of short and safe financial assets occur.   In particular, a clear commitment to a target growth path for nominal GDP both in the U.S. and the EU would improve government finances, largely through improved tax collections, and reduce the demand for base money, allowing for a contraction in central bank balance sheets.   Further, a commitment to keeping nominal GDP on a steady growth path would  allow for appropriate fiscal austerity without contractions in aggregate real output or a deflation of prices. 

The paradox is that short of reforms that allow negative nominal interest rates on short and safe assets, the Fed must promise to buy a huge amount of assets if  necessary, so that in reality it will not have to do so.   

But the key is a clear commitment to a target growth path for nominal GDP.

Empirical Evidence for closing the nominal GDP gap

David Beckworth has done cross-sectional analysis regarding nominal GDP gaps and unemployment and credit market risk.   Country's with nominal GDP further below the trend of the Great Moderation have higher unemployment rates.   A simple estimate suggests closing the nominal GDP gap would reduce the unemployment rate by 2 percent.

Great News!

Matt O'Brien reports at The Atlantic that Chicago Fed President Charles Evans has called for nominal GDP level targeting.   That is one out of twelve Presidents, and then there are the seven governors.   But it is a start.

HT  Marcus Nunes

More on Andofatto's Challenge

There have been several efforts by market monetarists to answer David Andolfatto's question.   While I thought the answers quite good, they mostly were explanations as to why a regime of nominal GDP targeting is superior to inflation or price level targeting.  

Nick Rowe gave three reasons for nominal GDP targeting, all of which were very good.   1) Stable nominal GDP is a better baseline for contracts than a stable price level or inflation because it shares the risk of unexpected productivity shocks between borrower and lender.  (David Eagle has done much work on this topic.   2)  Nominal GDP targeting performs as well as inflation or price level targeting for aggregate demand shocks and depending on the exact nature of the shock, sometimes much better than price level targeting for aggregate supply shocks.  3) Nominal GDP level targeting does as well as price level targeting in self correcting aggregate demand shocks, and better than inflation targeting.    All key points, but the framework is comparative institutional--what regime is best?     Inflation targeting is weak on self-correction.   Price level targeting is disastrous for some supply shocks.   Nominal GDP targeting looks like the least bad option, and by the way, it is probably the better base line for contracting.   If someone really wants to be shielded from economy-wide real risk, then a special inflation-indexed contract would be appropriate, assuming there are lenders willing to make that commitment.

David Glasner added another regime-oriented plus related to nominal and real interest rates.   To the degree that the natural interest rate depends on expected future real output, then the negative relationship between real output growth and inflation with nominal GDP targeting tends to stabilize the nominal interest rate.   For example, expected slow real output growth will lower the natural interest rate, and while this will tend to lower the nominal interest rate, the expected slow real output growth will raise expected inflation, tending to raise the nominal interest rate.   I suspect Glasner puts no more weight on stable nominal interest rates than I do as a general proposition.   However, he is very concerned with avoiding situations where the equilibrium nominal interest rate is negative.   Nominal GDP targeting makes that less likely.       

Scott Sumner responded as well, with a grab bag of 9 arguments for "more" nominal GDP and two opposed to inflation targeting.   Those last two arguments are clearly based on "comparative institutions," as were many of the 9 arguments in favor of nominal GDP.     I agreed with all of Sumner's arguments.   

But what about Andolfatto's question?   He wasn't asking whether or not we should have a regime of targeting a growth path of nominal GDP rather than a growth path for the price level or inflation rate.   He was asking whether higher nominal GDP is desirable for the economy in 2012.

Sumner argues that the price level should return at least part way to its 1984-2008 trend because the shift from debtor to creditor is undesirable in the midst of a global debt crisis.   He adds that financial markets appear to think that higher inflation is desirable at this time (citing Glasner.)   He argues that a drop in real wages would desirable in at least some labor markets--people that "should" be getting nominal wage cuts will get appropriate real wage cuts.   He argues that the Europeans might respond to growth in U.S. nominal GDP by expanding their own nominal GDP.   And finally, he argues that a monetary policy that raises nominal GDP will "crowd out" inefficient public policies, like debt-financed government spending, that are justified by the supposed need to raise aggregate demand.  

Sumner gave first place to an argument that is a call for a regime shift:

One reason I favor more NGDP now is that even a belated shift toward level targeting would help to restore Fed credibility in future recessions, which is especially important if they again run up again the zero bound (as Woodford has emphasized.) It’s easier to prevent NGDP from plunging in the short run if markets expect it to return to trend in the not-too-distant future. The second advantage is that it makes policymakers more accountable. They can’t keep sweeping past errors under the rug (as the BOJ does.)

Face it, market monetarists are all about rules.   Andolfatto's question, "why is increasing nominal GDP a good idea right now" is difficult to answer.   
From 1985 to 2008, nominal GDP remained very close to a 5.3%  growth path.  Market monetarist "theory" has come to emphasize the benefits of expectations generated by clear communication of a goal for future levels of nominal GDP.    While there are some hints that Greenspan actually cared about nominal GDP, there was no clear communication of a nominal GDP growth path.  The actual "rule" was something like "mumble, mumble" something about price stability, and no long run trade off between inflation and unemployment.   

The dual mandate requires price level stability in the long run and high employment.   Somehow, a stable price level was transformed into persistent 2% inflation, and high employment to slowly closing any output gap between real GDP and potential.    The Taylor rule suggests a formula relating the policy intereset rate to deviations of inflation from 2% and the output gap.     Mainstream macroeconomists, anyway, assume that everyone in the real world takes that formula very seriously.
Market monetarists like to give the actual stable growth path for nominal GDP credit for the relatively mild fluctuations in real output and employment and the low and stable inflation rate during the Great Moderation.    The disaster of 2008 and 2009, and the current low level of output compared to its previous trend and low employment relative to its previous peak are blamed on the empirical reality that nominal GDP fell well below the growth path of the Great Moderation and remains far below it today.

The other empirical reality is that the price level fell to only a slightly lower growth path and the shift in the growth path of wages was smaller still.   The theory is that with nominal spending on output falling to a growth path 12% below the previous trend, and prices and wages falling only about 2% below trend, real expenditure fell to about 10% below trend.   Firms responded to the shrinking sales by cutting production and employment.  

After the trough, nominal GDP began to rise again, but prices and wages continued to grow and on an only slightly lower growth path.   Real sales have expanded, and firms have responded by expanding production but at a slow rate.   Real sales, and so, real output and employment, remain far below the trend growth path of the Great Moderation.

 The prediction is that if nominal GDP grew more quickly, promptly returning to the growth path of the Great Moderation, prices and wages would perhaps rise a bit, also returning to their growth paths of the Great Moderation.   However, real sales would grow quickly, and firms would respond to the growing real sales by expanding production and employment rapidly.    Production and employment would then recover to something close to their growth paths of the Great Moderation.

Market monetarists are aware that there is an alternative explanation.   On that view, in 2008, the productive capacity of the economy fell, and shifted to a lower growth path.   Happily, nominal spending on output fell along with the productive capacity of the economy, so that only modest adjustments in the growth paths of prices and wages were needed.   Apparently about a 2% reduction in the growth path of prices was necessary and slightly less for wages.  

What is the evidence for this?   It is pure theory--market clearing.   If nominal expenditure had fallen to a growth path that was more than 2 percent below the growth path of the productive capacity of the economy, then theory "proves" taht prices and wages would have fallen by more.   In theory terms, the theory is that if  real expenditure is below the productive capacity of the economy, then there are gains from trade by lowering prices and wages.    Optimizing agents will not sacrifice the greater utility they could receive by utilizing any idle productive capacity to produce more consumer goods in the present and future.  

I find this theory plausible enough.     It is strengthened by the empirical reality that firms are raising prices and wages rather than simply leaving them unchanged.   If there is plenty of excess capacity--idle land, labor, and capital--then the current low levels of nominal expenditure growth would generate larger gains from trade in the form of more real output and employment if prices and wages were left unchanged.    Perhaps cutting prices and wages is a challenge, but why increase them in the face of what are supposedly wide ranging surpluses?   (That's not to say that firms are continuing to build up unsold inventories of goods, but rather that they supposedly have the capacity and willingness to expand production if they could sell more at current prices.   Why raise prices and wages unless there would be shortages at unchanged prices?)

Andolfatto doesn't discuss this, perhaps because it is all supposed to be obvious.   In his second post on his challenge, he does link to an earlier post where he critiques "sticky prices."    To me, it seemed like a criticism of periodic price adjustments due to menu costs.  

I think it is fair to say that market monetarists don't have some unique explanation as to why changes in nominal expenditure, even given unchanged productive capacity, don't result in rapid adjustments in prices and wages, but instead result in changes in real sales, and firms adjusting production and employment to those real sales.   I am more than willing to replace periodic price adjustments due to menu costs (Calvo pricing) with some alternative.    Search theory and game theory might provide better explanations.   I am doubtful, however, that prices and wages do smoothly adjust to keep real expenditure equal to productivity capacity, and that "general gluts" are just an illusion after all.    (Game theory considerations seem to be a likely candidate for explanations as to why the trajectories of prices and wages have such inertia.    Could a central bank's commitment to 2% inflation generate a Schelling point of everyone raising prices and productivity-adjusted wages by 2% from their current levels?)

Andolfatto continues to return to the issue of debt.   I was a bit puzzled by his notion that market monetarists believe that debt overhang is restricting aggregate demand, and that inflation is desirable to reduce this debt overhang and so allow additional real expenditures.     If more real expenditures are needed, presumably because prices and wages have failed to fall enough and left real expenditures below potential output, then expanding nominal expenditure would increase real income and also reduce the debt overhang.   Sumner, anyway, often points out that it is the increased nominal GDP that directly fixes the debt overhang--to what degree it involves higher prices or higher real output doesn't much matter on that particular count.  

Andolfatto focuses on the small shift in the growth path of the price level (about 2 percent,) when it is the large shift in the growth path of nominal GDP that is important (more like 12%) for preexisting debt contracts.   Of course, if prices and wages are all sufficiently flexible so that this reflects a very large decrease in productive capacity, then raising nominal GDP back to its previous trend would result in substantially higher inflation for a time.   While this would count as a benefit of a nominal GDP targeting regime, creditors didn't make their contracts under such a regime.   Most of the contracts were made under the Greenspan mumble regime or the unspoken 2 percent inflation regime.   We had a regime where the government (the Fed) promised creditors that they would be shielded from any loss due to "supply shocks."   All the losses would be suffered by equity investors and workers.

Of course, with a "flexible" inflation target, you really do get "mumble, mumble."   Flexibility means that inflation will be on target unless the Fed decides that it is best for it not to be on target.   Not much of a rule.   So, mumble, mumble, flexible, low inflation...what should be expected?
My own view is that the Fed should immediately implement rapid growth in nominal GDP approximating the Volcker-Reagan recovery of 1982-1984.   And then, we should start a new regime--a 3% growth path for nominal GDP.    Given that regime, if we have some failure by the Fed to hit the target, I certainly will favor prompt and rapid action to return to target.