Sunday, October 30, 2011

Targeting the Growth Path vs. the Growth Rate.

Scott Sumner has commented on Bennett McCallum's article on nominal GDP targeting. Summer comments on McCallum's support for growth rate targeting. While he defends his support for growth path targeting, he also says:
Perhaps I’ve been overly influenced by the 2008 period, when the advantages of level targeting seem relatively large. I would also point to Michael Woodford’s work on liquidity traps. Woodford argues that level targeting is especially important when a central bank hits the zero bound (as he is even more skeptical about QE than I am.) McCallum may be right that when the central bank is doing its job, growth rate targeting is as good as or even better than level targeting. By “doing its job,” I mean targeting the forecast.
In my view, growth rate targeting works great, as long as it works perfectly, and is expected to work perfectly. If the growth rate remains constant always, then nominal GDP remains on a constant growth path.

In my view, changes in the growth rate of nominal GDP that return it to the target growth path are coordinating. Any notion that it is the changes in the growth rate of nominal GDP that cause problems are mistaken.

This is easiest to see when the growth path of nominal GDP has a growth rate equal to the growth rate of potential output, so that the trend price level remains stable. If an excess supply or demand for money causes nominal GDP to shift to a higher or lower growth path, reversing the deviation is less disruptive than accommodating the new growth path and continuing at the old growth rate.

Consider the following numerical example of an economy in equilibrium:

Y* is the target growth path of nominal GDP, Y is nominal GDP, P is the price level, y is real income, yp is potential output, gY is the growth rate of nominal income, inf is the inflation rate, gy is the growth rate of real income, and gyp is the growth rate of potential output.

The economy is in equilibrium with nominal GDP on target and growing at the 3 percent target growth rate. Real output equals potential output and they are growing together at a 3 percent annual rate. The price level is 100 and the inflation rate is zero.

Now, suppose an excess supply of money causes nominal expenditure to grow more rapidly in period 2 at a 4% rate. Further suppose that the short run aggregate supply curve implies that firms respond to the more rapid growth in sales with a split of 40% in prices and 60% in output.

The calculations for period two are simple enough.

Inflation is .4% and real growth is 3.6%. There is a boom in the economy, with the price level rising to 100.4 and real income rising to $10,360 billion, approximately $60 billion beyond potential.

Suppose growth rate targeting is used. Does period 3 then allow a return to equilibrium as follows?

3 10,712 100.4 10,712 10,712 3% 0% 3% 3%

No, it doesn't. Because potential output is $10,609. It rose 3% rather than 3.6% with real income.

Assuming growth rate targeting, the least disruptive possibility would be slower growth in real output during the next period so that it returns to potential. Unfortunately, that implies another period of inflation.

The growth rate of real income is 2.4% so that it returns to potential output and the inflation rate is the difference between that growth rate and the growth rate of nominal GDP, .6%.

Only then can there be a return to equilibrium.

4. 11,033 101 10,927 10,927 3% 0% 3% 3%

With growth path targeting, there is also an adjustment in period 3.

The growth rate of real income again slows as before so that it returns to potential output. However, there is a .4% deflation rate. (The assumption here is that the short run aggregate supply curve implies a similar split for output and prices when growth slows.)

So, is an additional .6 percent inflation more or less disruptive than a .4% deflation? Real output growth is going to slow, one way or another.

If prices were perfectly and instantly flexible, it would make no difference. But suppose prices are sticky?

If the most flexible prices adjust first to more rapid growth in nominal GDP, reversing the inflation of those prices won't be very disruptive. Growth path targeting has exactly that consequence. It is growth rate targeting that would require that the most sticky prices adjust as well.

Suppose that product prices are flexible, and it is wage rates that are sticky. The more rapid growth in nominal expenditure in period two raises product prices, reduces the growth rate of real wages, lowers real unit labor costs, and makes increased production profitable. With growth path targeting, the price level falls back to its initial level, nominal wages continue to grow at trend and real wages grow more rapidly. Unit costs return to equilibrium and so does production and prices. With growth rate targeting, on the other hand, sticky nominal wages must grow more rapidly and rise to a higher growth path, so that real wages and real unit costs return to equilibrium.

Further, considering the reverse shock, where nominal GDP grows more slowly or even shrinks, the most flexible prices may fall or grow more slowly at first. Nominal GDP level targeting reverses the decrease in spending, raising sales, production, and causes the most flexible prices to rise back to their initial growth path. Growth rate targeting requires that the more sticky prices, perhaps including wages, move to a lower growth path in order to return to equilibrium.

What possible benefit would growth rate targeting have? As discussed in an earlier post, it involves a scenario with a supply shocks when the particular goods being shocked have other than unit elastic demand.

The Long Run Equilibrium Price Level

With nominal GDP targeting, the equilibrium price level at any future date is the target for nominal GDP divided by the level of potential output. While the level of potential output cannot be observed either now or in the past, much less perfectly forecasted, the Congressional Budget Offices does provide estimates of past, current, and future values of potential output. Certainly these estimates could be criticized, and future values forcasted for 10 years from now must be taken with a grain of salt, but it provides a good starting point to explore price level and inflation performance.

In the diagram below, the price level, as measured by the GDP chain-type index, is shown in blue. The trend price level from the Great Moderation is shown in red. The Goldman-Sachs/Reagan-Volcker equilibrium price level is shown in green. And the equilibrium price level from the trend growth path of nominal GDP from the Great Moderation is shown in black.

The price level is currently 1.8 percent below the trend of the Great Moderation. An inflation rate of 3.6 percent would return the price level to trend after one year. An inflation rate of 3 percent would return the price level to trend after two years. An inflation rate of 2.6 percent would return the price level to trend after three years. Last quarter's inflation rate of 2.5 percent is close to what would be needed for a three year adjustment path.

The equilibrium price level implied by the trend growth path of nominal GDP for the Great Moderation is found by dividing nominal GDP by the CBO estimate of potential output. It is currently 6.4 percent above the trend price level of the Great Moderation. This equilibrium price level is 8.3 percent above the current price level. An inflation rate of 11.2 percent would be necessary to return the price level to equilibrium in one year. An inflation rate of 7.2 percent would allow the price level to adjust to equilibrium after 2 years. An inflation rate of 5.5 percent would allow the price level to adjust to equilibrium after 3 years.

The equilibrium price level implied by the Goldman-Sachs/Reagan-Volcker alternative growth path is currently 1.9 percent above the trend price level of the Great Moderation and 3.8 percent above the current price level. An inflation rate of 5.8 percent would allow the price level to adjust to equilibrium after one year. An inflation rate of 4.1 percent would allow the price level to adjust to equilibrium after 2 years. And an inflation rate of 3.2 percent would allow the price level to adjust to equilibrium after 3 years.

After these adjustments, the inflation rate implied by the trend price level of the Great Moderation would remain 2.3 percent. The equilibrium price level implied by the trend growth path of nominal GDP from the Great Moderation would imply an average inflation rate of 2.9 percent. The Goldman-Sachs/Reagan-Volcker growth path would generate a 2 percent inflation rate after the adjustment to equilibrium growth path for the price level.

The diagram below shows the inflation rates assuming the gradual, three year adjustment from the current price level to the equilibrium price levels.

Of course, "supply-side" reforms would raise potential output growth and reduce the equilibrium price levels and the inflation rates shown here. One implication of nominal GDP targeting is that price level and inflation depend critically upon supply-side economic performance.

Saturday, October 29, 2011

Christina Romer for Targeting the NGDP Growth Path

Christina Romer came out for targeting the growth path of nominal GDP in the New York Times.

Excellent article. Great news.

The Goldman Sachs and the Reagan Volcker Alternative--About the Same?

Nominal GDP, the flow of money expenditures on current output, is approximately 14 percent below the growth path of the Great Moderation. The norm of nominal GDP growth path targeting implies that this was an error and it should have been prevented or reversed long ago. However, the Fed has yet to adopt the norm. Given where we are today, what should be done? What is the appropriate growth path for nominal GDP?

The recent proposal from Goldman-Sachs includes a growth path that apparently was generated by starting at the growth path of the Great Moderation at the end of 2007 and then lowing the growth rate from 5.4 percent to 4.5 percent. Presumably this is the sum of a trend inflation rate of 2 percent and estimates of the growth rate of potential output over the last decade of 2.5%. During the Great Moderation, the trend growth rate of potential output was 3 percent, and the growth rate of nominal GDP of 5.4 percent resulting in an inflation rate of 2.4 percent as measured by the GDP deflator.

The Goldman Trend is shown in green below.

The current level of the Goldman-Sachs alternative is $16,905 billion and the current value of nominal GDP, $15,198.6 is 10.1 percent below. The usual Market Monetarist approach is to specify a target one year in the future, so the target for the third quarter of 2012 is $17,679 billion. To reach that level in one year, the growth rate of nominal GDP would need to be 16.3 percent.

The target for two years into the future, for the third quarter of 2013 is $18,488 billion. The annual growth rate of nominal GDP from its current value to that level two years from now would be 10.83 percent. Reaching the new growth path over two years suggests a target for next third quarter of 2012 of $16,824 billion.

While 11 percent nominal GDP is quite high, the annual growth rate of nominal GDP for each quarter from the second quarter of 1983 to the second quarter of 1984 was over 10 percent. If Reagan and Volker could manage it, certainly it is possible.

In a previous post, I had suggested that the Fed copy Reagan/Volcker and target nominal GDP growth for the next two years identical to that of 1983 and 1984. In the last few days, I thought of a slightly different approach. Rather than start now, instead, suppose that the Reagan/Volker growth rates had occurred at the trough of the Great Recession, the second quarter of 2009. Nominal GDP is now 10.7 percent below that growth path. If the new growth rate were 4.5 percent, then the similarity to the proposed Goldman-Sachs alternative is striking.

In the diagram below, nominal GDP is represented in blue. The growth path of the Great Moderation is in red. The Goldman-Sachs alternative is in green. If the growth rate of nominal GDP had matched that from 1983 and 1984 from the second quarter of 2009, the resulting "Reagan/Volcker" growth path is shown in black. The dashed line shows a two year adjustment growth path. The growth rate for those two years would be 10.7 percent.

The difference between the Goldman Sachs and Reagan Volcker alternatives is so small, it hardly makes a difference. It is pretty much the situation that Goldman Sachs proposed a new growth path that approximates what would have happened if Obama and Bernanke had done their job as well as Reagan and Volcker. All I can say is--better late than never.

Friday, October 28, 2011

The GDP Report--What's New?

Nominal GDP remains far below the trend of the Great Moderation, and the gap continues to grow. It is now up to 13.87 percent. Nominal GDP for the third quarter of 2011 was $15,198.6, and if it had continued to grow on the 5.4 percent growth path of the Great Moderation, it would now be $17,646 billion.

The current level of real GDP, $13,352 billion in 2005 dollars is 6.72 percent below the CBO estimate of potential output, $14,315.1 billion. The output gap shrank! From 6.84 percent in the second quarter! Of course, it is higher than the quarter before that.

The price level, as measured by the GDP chain-type deflator, remains below the trend of the Great Moderation. Its value came in at 113.8 which is 1.8 percent less than the trend value of 115.88. The gap has slightly decreased.

And so, that is the GDP news from the "level" or "growth path" perspective. But most of the world looks at growth rates. Nominal GDP? Who looks at that? Real GDP grew 2.5 percent. Not too good, but not too bad. The inflation rate was 2.5 percent too. Higher than the trend inflation rate for the Great Moderation. But that is just part of the story.

Thursday, October 27, 2011

Nominal GDP Grows 5%!

Nominal GDP grew 5% in the third quarter. According to the preliminary estimate, anyway.

Are market monetarists ready to rejoice?

While this growth rate is right according to most market monetarists, (though I am one of those who supports a slower 3% rate,) the problem is that the current growth path is way too low.

What the U.S. needs is a couple of years of nominal GDP growth like occurred under Reagan and Volker in 1983 and 1984, adding up to a near 20% increase over two years. Then the growth rate should slow to something like 5% (or 3%.)

Excessive focus on the growth rate of real GDP (2.5%) is a terrible mistake. If real GDP was close to potential, then the growth rate of real GDP would be very informative. According to the CBO, real GDP was about 7% below potential before and it is about the same now. Nominal GDP was about 14% below the trend growth path the Great Moderation, and it is about the same now.

Look at the growth paths, don't just focus on the growth rates.

Wednesday, October 26, 2011

Nominal GDP Targeting and Stagflation

John Carny wrote a post critical of nominal GDP targeting on the CNBC blog.

The headline was pretty awful: Nominal GDP Targeting: A Policy for Stagflation?


Carny seems to think that the purpose of nominal GDP targeting is to raise inflation. Then he makes a series of arguments that are best understood if nominal expenditure remains unchanged and inflation rises.

In reality, the goal of the policy is to raise the flow of expenditure on output. More dollars spent on goods and services produced in the U.S.

It is possible that firms will respond to this increase in the dollar volume of sales by raising prices. To the degree that they do this, the increase in the actual volume of products that can be purchased with that growing dollar expenditure will be smaller. The increase in what the firms actually produce will be smaller. And the expansion in employment will be smaller.

If there were no increase in prices, then the volume of production and employment would grow in proportion to the increase in nominal expenditure. If, on the other hand, prices rise in strict proportion to the rise in nominal expenditure, then production and employment would not be influenced at all.

Advocates for nominal GDP targeting are not simply saying that the Fed will be "loose" for an extended period of time, even if the result is inflationary. Advocates are saying that the Fed will purchase or sell whatever amount of assets are necessary to get nominal GDP to the target growth path and keep it there. Inflation will be controlled by the limited increase in spending along the target growth path. Those who propose a 5% target for nominal GDP are proposing a 2% trend inflation rate. My own preference for a 3% growth path implies long run price level stability.

Carny recognizes that the policy involves an aggressive program of asset purchases, and then begins to discount the ability of the policy to impact expectations. However, he makes a error. He thinks that the expectations of the policy of asset purchases will lower long term interest rates. On the contrary, to the degree the policy is expected to work, expectations of higher future expenditures on output will result in increased credit demand and reduced credit supply, and so higher long term interest rates.

It is rather when the "Chuck Norris" effect fails to work that long term interest rates fall because the Fed is buying assets and paying higher prices for them. The possibility that the consequence of the policy would be expected and result in higher interest rates--short term and long term, isn't necessary. It is just a possible benefit of having the Fed make an explicit commitment to target for a nominal GDP growth path. The explicit commitment reduces the needed increase in the size of the Fed's balance sheet or decrease in short term or long term interest rates.

Carny claims that the policy just won't work when households have too much debt. Again, he imagines that the goal of the policy is to cause inflation and reduce the real value of the debts. But that isn't the point of the policy. The immediate effect on households earning wages is a reduction in worries about future unemployment. Or, more importantly, reduced worries about extended future unemployment. Increased spending in the economy implies increased sales, and increased employment opportunities.

Firms seeing these increased sales will be more inclined to add on employees when they understand that this is the first step in new policy to raise the volume of money expenditures to a target growth path. The most important source of expenditure is by newly employed workers.

Further, for every debtor there is a creditor. Creditor households faced with inflation have an incentive to reduce lending and instead spend on real assets. For example, if the price of a new car or washing machine is going up, it is time to purchase one now rather than wait.

Still further, to the degree that rapid growth in spending impacts prices and not wages, then profits earned by small business will be flush. This should result in increased consumption expenditure by small business owners. Even dividend payouts to stockholders should not be dismissed.

Carny claims that inflation will not cause firms to expand unless they expect "real" profits. Supposedly, they will passively accept a real losses on holding "cash," but will only spend if they earn real profits. Actually, it is entirely sensible for firms to spend cash on real assets to reduce real losses.

I think the best way to understand Carny's confusion is to think about a basic supply and demand problem. Suppose the demand for apples rises. The result of this is a higher price of apples as well as a higher quantity of apples. But doesn't a higher price of apples lower the demand? So, it must be that a higher demand for apples cannot raise the price or the quantity of apples.

The correct analysis is that an increase in the demand for apples causes a shortage at the current price. As the price rises, the quantity demanded falls and the quantity supplied rises until they match. The end result is a higher price and a higher quantity. The most likely process is that the increase in demand results in both a higher price and quantity. The higher price doesn't lead to a decrease in demand.

Higher spending in the economy leads to higher sales for firms, which respond by both expanding production and raising prices. But doesn't the inflation cause people to reduce their purchases? So prices and production cannot rise. No. The higher inflation dampens the increase in real output--it doesn't prevent expenditures from rising.

Finally, Carny's final statement has to be quoted:

In short, its not at all clear that nominal targeting will work as promised—much less generate real economic growth. And it could set off a deflationary spiral that would lead up into low growth and rising prices. In other words, stagflation.

A deflationary spiral that leads to rising prices? What a contradiction!

But stagflation is a possibility. The problem won't be a deflationary spiral. It is rather that if the productive capacity of the economy has been greatly depressed, and the current level of real output is approximately equal to productive capacity, then a shift to a higher growth path for nominal GDP will generate higher inflation. Real GDP will remain at its current low level (relative to the trend of the Great Moderation.) Presumably employment would remain low and unemployment would remain high. Inflation would be high for a time, before settling back to a slower trend rate once nominal GDP reaches the target growth path. High unemployment and temporarily high inflation--stagflation.

However, the problem wouldn't too much debt or odd arguments where people accumulate money that is losing value. The problem won't be that increased demand reduces demand and creates inflation.

The problem would be that firms cannot expand production because they will face bottlenecks. Perhaps there are key employees they cannot find, or special machinery, or some other kind of resource. And so, if the impact of nominal GDP targeting is simply more inflation and little or no real growth or employment, what would be observed is shortages of output as firms struggle unsuccessfully to meet rising sales.

It is important to understand that the problem wouldn't be excessive debt. The problem would be that the productive capacity of the economy fell in about 2007. Nominal GDP targeting keeps the flow of spending on output steady in the face of shifts in the productive capacity of the economy. The result is a higher price level and temporarily higher inflation. Advocates of nominal GDP targeting believe that this is the least bad environment to make the needed adjustments to such a decrease in productive capacity.

Tuesday, October 25, 2011

McCallum and Targeting the Growth Rate of Nominal GDP

Bennett McCallum has written a short article showing his continued support for targeting nominal GDP. It was especially good to see it published by the Shadow Open Market Committee, a traditional outlet for "old monetarism."

McCallum describes some benefits for nominal GDP targeting, but comes down in favor of targeting the growth rate rather than the growth path. He does note that this view is controversial among advocates of the NGDP target. I must admit that in some ways I feel more kinship with advocates of price level targeting than with advocates of targeting the growth rate of nominal GDP!

McCallum explains that targeting the growth rate of nominal GDP can result in level drift. However, as long as the shocks to nominal GDP growth are symmetrical, then this shouldn't be much of a problem.

The disadvantage to targeting the growth path of nominal GDP is that if growth rises too fast, then a contractionary monetary policy must shift it back down to the target. The assumption is that the monetary authority would be creating an additional disturbance, causing a mild, or maybe severe, recession.

Oddly enough, McCallum didn't speak of the reverse situation, where nominal GDP falls below its trend growth path. Is it the case that the monetary authority would be creating an additional disturbance to force nominal GDP back to its trend growth path?

It is exactly this scenario that has made me adamantly in favor of targeting the growth path. Admittedly, the Fed isn't really targeting nominal GDP growth, but the growth rate since the recovery began hasn't been bad. It is slightly above 4 percent, almost exactly the growth rate proposed by the economists at Goldman-Sachs. Presumably, it is equal to an estimated 2.5 percent growth rate for potential output and a 2 percent trend inflation rate. The problem with the current growth path is that it is way too low. Rather than a sharp "V-shaped" recovery, we are languishing with a large output gap and an unemployment rate well above the natural unemployment rate.

More importantly, growth path targeting creates expectations that generate market forces that tend to keep nominal GDP on target. (Of course, on the target growth path, the target growth rate implies remaining on the target growth path.) If the economy appears to be slowing, and nominal GDP is likely to be below target, then sales will be expected to growth extra fast in the subsequent period. This reduces (or hopefully, prevents) positive feedback loops where reduced current growth in real output results in greater saving, reduced investment, or, more fundamentally, additional money demand, so that future nominal growth is also reduced.

Considering prices, to the degree reduced nominal growth results in disinflation (and deflation if the trend inflation rate is very low, say zero or even slightly negative already,) then the reduction in the price level will be expected to be reversed. Current deflation results in expected inflation. This prevents the disastrous positive feedback loop where current deflation creates expectations of future deflation, increased money demand and reduced nominal expenditure and so more deflation. Instead, if the price level falls, it is now low relative to its expected future value, and so motivates an increase in spending now.

To the degree prices are sticky, then prices hardly drop at all. However, reversing the decrease in nominal GDP means that prices can return to equilibrium without needing further disinflation to adjust to the new, lower growth path of nominal GDP. With GDP targeting, (rather than final sales targeting,) future final sales rise in response to an inventory buildup, which provides an incentive to maintain production even in the face of slower sales.

The same stabilizing market forces apply when nominal GDP grows too fast. With nominal expenditure targeting, if the economy appears to be overheating, the expectation will be that it will slow.

Again, looking at prices, if more rapid growth in spending results in higher prices, then their future growth will be slower. If the trend inflation rate is zero, the effect is obvious. Prices rise now, but will be expected to fall. Current prices are exceptionally high, and there is an incentive to refrain from making purchases now, slowing the economy. (With a trend inflation rate, the effect can be better described using nominal and real interest rates. Any given nominal interest rate implies an higher real interest rate if inflation is expected to slow.)

What is the disadvantage? If we imagine an upward shock to the price level, that is not associated with a decrease in potential output, then nominal GDP rises. Targeting the growth path of nominal GDP requires that somehow an excess demand for money develop, forcing a reduced flow of spending, so that prices are forced back down. Targeting the growth rate of nominal GDP, on the other hand, would accommodate this increase in the price level, and allow the flow of money expenditures to grow at the target rate from that point on.

Suppose that instead we imagine a downward shock to potential output, and there is no associated increase in prices. This reduces nominal GDP. Targeting the growth path of nominal GDP would require that an excess supply of money be generated, forcing an increased flow of expenditure, so that prices are forced up. Targeting the growth rate of nominal GDP, on the other hand, would accommodate this decrease in potential output, and allow the flow of money expenditures to grow at the target rate from that point on.

On the other hand, suppose that adverse aggregate supply shocks involve a decrease in supply in some particular market. For example, a drought hits the Iowa corn crop. The decrease in supply of corn raises the price of corn and reduces the quantity of corn. As a matter of arithmetic, the price level rises and real GDP falls. The impact on nominal GDP is ambiguous.

More exactly, as George Selgin has explained, the immediate impact depends on the elasticity of demand for corn. If the demand for corn is unit elastic, then expenditures on corn remain the same and expenditure on everything else also remains the same. If the demand for corn were inelastic, then total expenditure on corn rises, and the demand for everything else falls. And if the demand for corn is elastic, then the expenditure on corn falls, and spending on everything else rises.

Of course, since it is the future level of nominal GDP that is actually targeted, then unexpected supply shocks push nominal GDP away from target. If they are temporary, then they don't effect future nominal GDP. Further, there is no need for monetary disequilibrium to force nominal GDP back to its target growth path. If the corn harvest recovers next year, then the price of corn falls and the quantity of corn rises.

Further, supply shocks expected to persist, or more generally, expected future supply shocks, have impacts that are at least qualitatively appropriate regarding resource allocation. For example, suppose the United Auto Workers gets very militant and demands pay hikes. Their employers acquiesce and raise prices and reduce production. If car demand is inelastic and this problem is likely to persist, total expenditures on cars rise, and spending in the rest of the economy falls. This reduction in demand tends to free up resources to produce other things. But what other things? Perhaps to produce more cars in nonunion shops in South Carolina. Perhaps to produce more export goods to exchange for imported cars. Of course, this is about the long run elasticity of supply of cars. It is certainly likely that slow growth in real incomes for everyone else in the economy would be at least part of the impact of the militant unions. Targeting the growth path of nominal GDP would require slower growth in nominal incomes in the rest of the economy.

If, on the other hand, the demand for cars were elastic, then total spending on cars would fall, and expenditure in the rest of the economy would rise. The demand for other goods would rise, requiring more resources. Where could those resources come from? Perhaps from those who would have been hired by the car companies if the unions were less militant.

It is at about this point where I just go with--a stable growth path of nominal expenditure is the least bad environment for microeconomic coordination. However, there is a bit more to say regarding the distinction between targeting the growth rate or growth path of nominal GDP. Expected future supply shocks to goods whose demands are not unit elastic has the exact same potentially undesirable effect whether the growth rate or growth path of nominal GDP is targeted.

So, it comes down to persistent unexpected supply shocks for goods with other than unit elastic demands freeing up too many resources or demanding too many in the rest of the economy, versus the ability of the economy to better avoid undesirable shifts in aggregate expenditures and to rapidly recover from any adverse effects.

The last three years has convinced me that avoiding and rapidly reversing changes in nominal expenditures--especially decreases--is very important.

Monday, October 24, 2011

WSJ on Nominal GDP Targeting

Kelly Evans, of the Wall Street Journal, wrote a column critical of nominal GDP targeting.

There are at least three problems with this strategy, however. First, it assumes that the Fed can sensibly determine the "right" trend for nominal GDP. Second, it isn't clear that it can actually achieve any such target. And third, doing so would run a huge risk of conflicting with the Fed's congressional mandate to promote "stable prices"—something that can't unilaterally be rewritten. This is because any boost to nominal GDP may well come more from higher inflation rather than from faster growth in underlying GDP, which Goldman acknowledges. After all, the economy's real potential growth rate has been slowing for decades.

What should we make of these three points?

1. Can the Fed find the "right" trend for nominal GDP?

In my view, the right trend growth rate for nominal GDP is the trend growth rate of the productive capacity of the economy. This generates zero inflation over time. If the trend shifts, then the result will be inflation or deflation, but unless the shift is extreme, it will be modest.

More challenging is determining the correct level at which to start the trend. The obvious choice would be the current value of nominal GDP. However, the U.S. spent 24 years on one growth path, and suddenly shifted to a new one that is 14% lower. If the growth path of prices and wages had also shifted to lower growth paths--14%, or even to something consistent with plausible estimates in changes to potential real GDP, say 7%, then going forward from where we are now would be sensible.

I am still confident that if the U.S. continues on the current low growth path of nominal GDP, eventually real output will recover to potential. It might even help if the Fed came out and explained that they want the growth path of nominal GDP to be permanently lower and that prices and wages need about 7% deflation to adjust to it. In other words, maybe the Fed should stop promising 2 percent inflation if we really need a year of 7% deflation and then 2% inflation going forward.

However, I think that it is much better to shift to a higher growth path. The notion that this is some kind of out of control inflation is absurd as long as the new growth path remains below the growth path of the Great Moderation. As far as I know, no one, and certainly no market monetarist, has advocated that nominal GDP rise above the growth path of the Great Moderation. The limit of these proposals is a return to the 5.3% growth path that existed from 1984 to 2007, and most of us have been advocating something less--at least modestly lower growth rates (say 5%,) and some kind of upward shift to reverse part of the drop.

For example, I recently suggested matching the growth rates in nominal GDP for the Reagan/Volker recovery of 1983 and 1984. After a 19% increase in nominal GDP over two years, then return to a 5 percent or 3 percent growth rate.

An alternative methodology is to return to 2007, and start with a new growth rate from that point. Apparently, the Goldman-Sachs proposal selects 4.5%, which would be the supposed inflation target of 2% plus an estimate of productivity growth of 2.5%. I have used the same approach with a 3% growth rate for nominal GDP starting in 2007. Oddly enough, nominal GDP is currently about 7% below that path, and closing that would be consistent with closing the output gap while leaving the price level stable.

This ambiguity about the proper growth path should not be exaggerated as an weakness. The key idea is that once on a target growth path is selected--stay there. When errors occur, reverse them and return to the path. And let everyone know that is the plan.

Evans doesn't discuss any of these issues. Her arguments are:

First, it is tempting, but probably mistaken, to assume the Great Recession came along and knocked the U.S. off an otherwise sustainable growth track. It wasn't an external shock, but internal weakness, that led to the economy's collapse.

One worrying aspect of GDP growth prior to 2007 was that it came even as real household incomes stagnated. Assuming that boom-era growth rates were sustainable, and not fueled by a surge in house prices and a credit boom that simply pulled forward demand from the future, is a huge leap in logic.

This is a very serious error. The problem, according to Evans, is that households couldn't afford to keep nominal expenditure growing on a 5 percent growth path. Apparently, only the appreciation of the houses and borrowing provided the funds needed to keep nominal GDP growing.


The basic identity of macroeconomics is that income is equal to output. This is true both in real terms and nominal terms. There is always enough nominal income to afford the nominal value of output. There is never any need to borrow in order to afford to buy nominal output. Home or other asset appreciation isn't needed for people to afford a level of nominal expenditure equal to nominal output.

If real household income did stagnate, with a regime of nominal income targeting, nominal household income (when added to business income) would grow exactly the needed amount to provide the means to purchase the growing nominal output. Sadly, unless this is about growing retained earnings, the result would be higher inflation.

There is an issue, but Evans is mistaken about its nature. C-

2. Can the Fed can actually achieve its target.

This is certainly a concern. With the Federal Reserve's target interest rate at .25 percent, skepticism that reducing it to zero--supposedly the lower nominal bound for interest rates--would raise demand, seems plausible. Market monetarists advocate cutting the interest rate the Fed pays on reserve balances to something slightly less than zero and then purchasing as many assets as necessary to reach the target. To the degree it isn't expected to work, then interest rates on these other assets the Fed purchases would fall.

But how does Evan's explain the problem?

Consider how recent gains in the consumer-price index, particularly in food and energy, have outstripped any increase in wages. This has hurt real income growth, undermined consumer confidence, and weakened, not strengthened, the economy.

For the Fed to generate inflation, it needs households to believe the central bank is fueling not just higher prices but wage gains, too, so that they start spending more. Otherwise, households will simply tighten the purse strings instead.

She is back to this problem about people being too poor to afford the level of expenditures consistent with the target for nominal GDP.

Should I play the Zimbabwe card.? The people in Zimbabwe were and are very poor, but their nominal GDP in Zimbabwe dollars far surpassed the growth path of nominal GDP being proposed--probably for the next century or so.

Of course, Zimbabwe is hardly a model economy. But it illustrates the error. A failure to distinguish between real and nominal values.

An increase in consumer prices, ceteris paribus, is an increase in nominal GDP. If an expansionary monetary policy raises consumer prices, then nominal GDP increases. This does not create a difficulty for targeting nominal GDP. It would be a way, (not particularly desirable,) by which nominal GDP would increase.

The notion that households expecting higher prices in the future will reduce nominal purchases now, so they can wait and pay the higher prices in the future, is implausible.

Further, while wages form a significant portion of income, there are the other forms of income too--rents, interest, and profit. If prices rise and wages, interest and rents don't rise, then profits rise.

Finally, there is a confusion between real wage rates and total wages. If demand rises, and prices rise and wages don't rise, then it becomes profitable to expand production and hire more workers. While the real wages of those who were already employed may fall, the real wages of those who were unemployed rise.

Unless the increase in nominal GDP is matched by an exactly proportional increase in prices, then there will be at least some increase in real incomes. Regardless, the ability to spend depends on nominal income, not real income.

Moreover, the level of general inflation it would take to transform housing, the thorniest problem facing the economy, would be huge. Boosting home prices by the 15% to 25% that Barclays Capital reckons many households are underwater "would in all likelihood be prohibitively expensive in economic and social terms," says the firm.

The goal of nominal GDP targeting isn't to "transform housing." An increase in nominal GDP to its target growth path would presumably raise housing demand to some degree. This would raise housing prices. Some of those who are "underwater" on their homes would have higher incomes--perhaps some family members would be employed. Perhaps some of them own businesses that would earn more profit. This would make them better able to make their payments. However, there is nothing in nominal GDP targeting that requires that no one be "underwater" on a home mortgage or that total nominal consumption, much less the consumption of each and every person, rise to the growth path of the Great Moderation.

I suppose Evans can be excused for confusing nominal GDP targeting with proposals to raise inflation. Krugman, for example, has said that the "economics" of GDP targeting is increasing expected inflation to reduce the real interest rate and so generate more real expenditure. Rogoff has been calling for inflation to help the housing sector. Krugman treats nominal GDP targeting like a public relations ploy to promote inflation.

No. Nominal GDP targeting is a monetary regime. If nominal GDP falls below target, then it is the responsibility of the monetary authority to get it back to target. If this results in inflation, then the inflation must be suffered. The purpose of the policy is not to generate inflation.

What about Evans' second concern? Again, some justified confusion with proposals to raise inflation. But some very bad economics. D-

And that leads to the third concern.

3. Is nominal GDP targeting consistent with the Fed's mandate for price stability?

The Fed actually has a dual mandate:

The Board of Governors of the Federal Reserve System and the Federal Open Market Committee shall maintain long run growth of the monetary and credit aggregates commensurate with the economy's long run potential to increase production, so as to promote effectively the goals of maximum employment, stable prices, and moderate long-term interest rates.

Well, that is more than two things, but maximum employment and stable prices are generally considered the two important bits.

Nominal GDP targeting changes the monetary aggregates in such a way as to accommodate changes in the demand to hold those aggregates, (offsetting shifts in velocity,) so that nominal expenditure on output grows with the long run potential to increase production. That looks good.

When there are short run shifts in that "potential" output, nominal GDP targeting does nothing to reverse the consequent short run changes in prices, allowing the price level and inflation to fluctuate in the short run. Given all of the focus on the "long run" above, that looks consistent.

And what is the alternative? Have the Fed institute a contractionary or expansionary monetary policy to reverse those changes in the price level. Since a contractionary policy aimed at reversing a supply-side shock to the price level is likely to force employment lower, that looks inconsistent with one aspect of the Fed's mandate.

The CBO estimate of the productive capacity of the economy has increased in each and every year. The growth rate has been less than the 3 percent trend of the Great Moderation for the last 5 years, and so, nominal GDP targeting at the 5.4% trend of the Great Moderation would have resulted in slightly higher inflation than the 2.4% trend--close to 3% for many quarters and near 4% for a few quarters.

If the slowdown in productivity is permanent, then perhaps a shifting the growth path to a slower growth rate would be sensible. However, such an adjustment should be deliberate and with plenty of warning. A shift to a 14 % lower growth path, mostly over a period of one year, is not the best way to proceed.

So, what of this final criticism? Complete failure to look at the Federal Reserve's actual mandate. And, more importantly, a focus on the short run consequences of the regime rather than its long run effect--which is the criterion of the Fed's mandate. F.

Saturday, October 22, 2011

Clark Johnson on the Great Recession and the Goldman-Sachs Toy Model

Clark Johnson has a pamphlet challenging a series of myths regarding the Great Recession. He is especially good on how low interest rates are consistent with "tight" money. I thought there were a few places where money and credit were a bit confused. (Though I certainly would grant that having the Fed pay banks to hold reserves reduces the supply of bank loans. While low bank loans doesn't necessarily mean an excess demand for money, it probably isn't very helpful.)

What I found most puzzling was his adoption of McKinnon's call for the Fed and other major central banks to jointly raise their target interest rates. While it would be interesting to see the Fed commit to make open market purchases until the federal funds rate rises from .25 to 2 percent, (as opposed to the conventional approach of making open market sales until that happens,) I generally think making any commitment to the future path of interest rates is a mistake. The sole commitment should be to a series of levels for nominal GDP. Interest rates now and in the future should just be wherever they must be to hit the target for nominal GDP.

However, I have continued working with the Goldman-Sachs toy model. I sent out an email to various market monetarists asking about the instabilities I found (and the Svensson paper that pointed out the problem years ago. Thank you Peter for pointing this out in a comment on the earlier post. I just found your blog, Josh Hendrickson and David Beckworth both replied, pointing out that McCallum had made a reply, also years ago. (Thanks guys.) Hendrickson's was especially helpful with a suggestion about "the intuition." Models where the flow of nominal expenditure on output don't appear are unlikely to show much benefit from nominal GDP targeting. The instability comes because the interest rate set today effects the output gap next period, and the output gap next period effects the inflation rate one period later.

Anyway, I modified the "IS curve" by replacing the GDP gap with a nominal GDP gap (between nominal GDP and the target, rather than between real GDP and potential) and then added an equation relating the GDP gap to that nominal GDP gap. It twisted the phillips curve out of recognition, with the expected future price level being the target for nominal GDP divided by potential output and the expected inflation rate being the difference between that price level and last period's price level. The current Nominal GDP gap also impacts the current inflation rate. (Thanks Marcus Nunes for the data showing that inflation responds with output and not solely after output.) Rather than the interest rate being based upon the size of the nominal GDP gap, it is set to close the expected nominal GDP gap in the next period.

What does this have to do with Clark Johnson?

Just this morning I set the coefficients in the quasi-IS function so that the current nominal GDP gap solely depends on the expected future one and not at all on the past one. To close the gap next quarter, the Fed needs to set the interest rate at...

47.6 percent!

And while next quarter there is a 35% inflation rate, the output gap closes and the unemployment rate falls to the natural rate. (After that the inflation rate slows to 3% and then gradually slows from there. This is a return to the nominal GDP trend of the Great Moderation, rather than the modified growth path used in the Goldman-Sach's simulation.)

So, 2%? My "model" says that is way too low. (The "target" interest rate rapidly falls to something closer to 6%.)

The more last period's nominal GDP gap impacts this period's nominal GDP gap, the lower the "target" interest rate. For example, a 40-60 past future weighting requires a negative nominal interest rate to close the gap and so, massive quantitative easing with the Goldman Sachs assumption that $1 trillion equals -1%. Also, if the weighting is the opposite, like in the Goldman-Sachs toy model, stability problems still appear.

Thursday, October 20, 2011

Williamson on NGDP targeting

Stephen Williamson has noticed nominal GDP targeting, but clearly hasn't paid much attention. Apparently, he did read the McCallum paper some years ago.

He tries to make sense of the proposal by discussing it in the context of the Taylor rule.

He writes:
Presumably the advocates of NGDP targeting think that standard central banking practice works, i.e. that a sensible approach to policy over the very short term is to specify an intermediate target for the fed funds rate, with the target set according to the current state of the economy relative to the NGDP target.
Well, no. While it is true that some market monetarists have advocated NGDP targeting for some time, we have become much more vocal about it because we don't believe that standard central banking practice "works." In particular generating expectations about future policy rates and their relationship to output gaps and inflation looks to have failed. Nominal GDP targeting isn't about creating expectations about future short term rates and how those will impact output gaps and future inflation. It is about creating expectations about the future level of nominal GDP.

He then writes an alternative version of the Taylor rule:

Thus, we could specify the implementation of the NGDP target as a rule

R(t) = p(t) - p(t-1) + c[y(t) + p(t) - y* - p*] + r*,

where y*p* is the log of the nominal GDP target and c > 0. We can then rewrite this rule as

R(t) = p(t) - p(t-1) + c[y(t) - y*] + c[p(t) - p(t-1) - p* + p(t-1)] + r*

What's the difference between this and the basic Taylor rule? Not much.

Here Williamson makes another error. What he has described is a proposal to stabilize the growth rate of nominal GDP. The proposal is to target the growth path of nominal GDP.

He adds:

(i) The coefficients on the terms governing the response of the fed funds rate to the "output gap" and the deviation of the inflation rate from its target are constrained to be the same.

This is correct. While the proposal is really about a policy reaction function for short term interest rates, targeting the expected level of nominal GDP means that the expected future price level is inversely proportional to the expected future level of real output. At first pass, the expected future price level would be the target for nominal GDP divided by the expected level of potential output.

He then adds:

(ii) The interpretation of y* may be different.
y* may be different. In the NK literature y* is the efficient level of aggregate output ground out in the underlying real business cycle model. The NGDP targeters seem to think of y* as the trend level of output. For practical purposes it does not make much difference, as the people who measure output gaps tend to think of trend GDP as potential GDP.

"NGDP targeters" do not assume that y* is the trend level of output, and are quite aware of the possibility of "productivity shocks." And so, the potential output we have in mind would be something like the efficient level of aggregate output ground out in the underlying real business cycle model. Nominal GDP targeting doesn't require that output gaps be measured. Nor does it require choosing a measure of inflation. (By the way, the CBO estimate of potential output has been running below trend for more than five years. Williamson should get around a bit more.)

Williamson then comes to his conclusion:
Could the Fed actually achieve such a target, even if it wanted to? No. Under current circumstances, there are no actions the Fed can take that could necessarily achieve such an outcome. Indeed, it is possible that the Fed could promise to keep the policy rate at 0.25% for five years in the future, and NGDP growth could fall below the target.

Market monetarists don't generally favor keeping the Fed's policy rate at .25% for five years into the future. We don't favor targeting interest rates at all. I suspect that in Williamson's model economies, the Fed really would have to purchase all assets, and if representative agents have the wrong preferences and technology, the price level would stay the same. But the problem is with his model.
There is no magic in a NGDP target. I know people look at the state of the economy, and think that the Fed should keep trying things. Maybe something will work? Well, I'm afraid not. Even the FOMC dissenters, and their supporters are not quite ready to say that there is nothing the Fed can do under the current circumstances that could increase employment. But they should.
Nominal GDP targeting isn't about increasing employment. While I believe that it would increase employment under the current situation, the reason to use it is that it is a superior regime to inflation or price level targeting.

Anyway, Williamson's "current conditions" includes odd things like taking the interest rate the Fed pays on reserves as constant or making currency endogenous. Market monetarists strongly favor changing the first and would have no problem changing the second if it would help.

DeLong on Nominal GDP targeting

DeLong commented on Krugman's support of nominal GDP targeting. (He had already advocated the regime change, along with quantitative easing and negative interest rates on reserve balances.)

The thrust of his comment is that money creation and fiscal stimulus should be used together. Either monetary policy alone or fiscal policy alone have doubtful consequences, but by creating money and having the government spend it, there is no doubt it can work.

I am much more confident that monetary policy can do it alone.

What are his doubts? What is the market process he describes?

If you are--as we are right now--in a liquidity trap, with extremely interest-elastic money demand, then expansionary monetary policy that involved the Federal Reserve buying financial assets for cash:
  1. will have next to no effect on the short-term safe nominal interest rate--it's already zero.
  2. will decrease the long-term safe nominal interest rate to the extent that your open-market operations today change people's expectations of what your target for the short-term safe nominal interest rate in the future.
  3. will decrease the long-term safe real interest rate to the extent that it decreases the short-term nominal interest rate and changes expectations today of what inflation will be in the future.
  4. will decrease the long-term risky real interest rate to the extent that it decreases the long-term safe real interest rate and to the extent that the assets purchased for cash by the Federal Reserve free up the risk-bearing capacity of private investors and lead to a reduction in risk spreads.
  5. will increase spending to the extent that it decreases the long-term risky real interest rate and to the extent that private spending responds positively to decreases in the long-term risky real interest rate.

Lots of steps here, some of which may well be weak.

Other than 5, all of these steps involve a decrease in the real interest rate, and then 5 is the idea that a lower real interest rate would cause firms and households to spend more. At least some of the steps involve creating expectations about future interest rates and inflation.

An alternative process is that expectations of a higher future flow of money expenditures on output will result in an increase in the profit-maximizing level of output and employment. The increase in the future profit-maximizing level of output increases the demand for capital goods now at any given level of real interest rate.

The increase in the future profit maximizing level of employment decreases the risk of future involuntary unemployment. This results in a decrease in saving now at any given level of real interest rate. (Or, we could simply say that an increase in expected future real income lowers current saving.)

In other words, the increase in present investment demand and decrease in saving supply results in an increase in the natural interest rate. Given the level of real interest rate generated by the first 4 steps described by DeLong, the present flow of money expenditures on output rises.

Using IS-LM analysis, the IS curve shifts to the right because of an increase in expectations of future real output and income, so that it crosses the full employment level of real output at a higher real interest rate. (Nick Rowe goes with a positively sloped IS curve to get at this.)

Why is this process ignored? It does, of course, have a "confidence fairy" element. But so does the process based upon higher expected inflation. Expected inflation can only be created if, somehow, nominal expenditure is going to increase in the future.

Suppose that no one believes that nominal expenditure will rise. Sadly, this forecloses both the higher future real output and higher present natural interest rate path, as well as the higher expected inflation and lower real interest rate path.

Does that only leave us with the pathways described above? Expectations about future policy rates and "free up the risk-bearing capacity of private investors and lead to a reduction in risk spreads?"

I believe that DeLong's approach is too deeply tied to the new Keynesian modeling strategy. The market monetarist approach is that the Fed must commit to purchase whatever quantity of assets needed to reach and stay on the target growth path.

That does not involve convincing anyone that policy rates will be lower in the future. They can expect what they want. The Fed will lower long rates by purchasing long term to maturity assets at higher prices. Sure, if the private sector continues to hold some too, and they don't expect nominal GDP to rise, those particular investors holding them must be expecting that future short term rates will be low. But that isn't what the Fed would be trying to do by purchasing the long term to maturity assets.

And, if necessary, the Fed would start purchasing risky assets (and yes, seeking changes in the law if necessary.) Freeing up the risk bearing capacity of private investors is probably not the best way to look at the situation. The Fed would be taking however much risk as is needed to generate sufficient consumption and and investment to get nominal GDP to target.

But just because the Fed would be committed to increase the quantity of base money to an amount equal to the nominal GDP target (or twice that amount,) if necessary, doesn't mean that there is a need to worry about base money being $15 or $30 trillion. It is not realistic to expect nominal GDP won't rise to target, which means that the higher natural interest rate and/or the lower real interest rate through higher expected inflation paths will kick in. And when they do, nominal and real interest rates can rise.

That Woodford and other neo-Wicksellians/new Keynesians developed simple models that show how a central bank can manipulate a short and safe interest rate according to a rule, and thereby manipulate long term risky real interest rates in a way that stabilizes inflation and closes output gaps is very interesting. Those rules very much depend on everyone understanding how the central bank will manipulate future short term interest rates in response to output gaps and inflation.

However, it is a mistake to assume that the only way that central banks can operate is by creating expectations about how short term interest rates will behave in the future in response to future output gaps and inflation. Since telling people that the Fed expects its policy rate to stay low for an extended period of time and even until 2013 hasn't generated a strong recovery, perhaps there is a problem with the model. But more importantly, imagining that heroic quantitative easing works by creating expectations about future short term rates is very implausible.

Why is it that the effect on the "IS" curve is ignored? If the new Keynesian approach works, then we would never get into a situation where saving rises and investment falls because people don't believe that the levels of short and safe interest rates that the central bank will set in the future won't generate a recovery. The only way that an output gap can be generated is if real interest rates are too high. The only way to fix it is to generate expectations of lower real interest rates.

But we know that strong recoveries create strong credit demand which raises equilibrium real interest rates. In my view, the new Keynesian models are just too tied to a regime of a central bank that adjusts a short and safe interest rate according to output gaps and inflation. They are poorly suited to considering an alternative regime where current and future interest rates can be at any level, and what is stabilized is the growth path of nominal GDP.

I don't deny that if no one expects the policy will raise nominal GDP, then lower nominal interest rates are what make massive quantitative easing work despite perverse expectations. But this hybrid approach where real long term interest rates must be lower is wrongheaded. And it matters from a practical perspective, because if and when people expect the policy to work, nominal and real interest rates should rise. And no one must expect that future "policy rates" will be low.

Playing with the Goldman-Sachs Toy

I was very pleased that two economists at Goldman Sachs have advocated nominal GDP targeting. They used a "toy" model to simulate the effects.

I spent most of yesterday (between classes,) playing with their toy.

Perhaps I am making some error, but it seemed to me that their model simulates a rapid recovery over the next year given the status quo policy of a target federal funds rate of .25 percent and a level of base money at $2.6 trillion. The output gap is mostly closed in a year, and inflation remains moderate, gradually rising towards the Fed's implicit 2% target.

On the other hand, when I tried a target for nominal GDP, my results were awful. With the best case, a boom develops after about one year, with very high positive output gaps. While inflation does slowly pick up, the price level stays well below what I would think is the "equilibrium" value, the target for nominal GDP divided by potential output. Worse, it didn't take much tinkering to generate explosive results.

This doesn't make me less supportive of nominal GDP targeting. It does make me doubt the value of the "toy" model, which is really very standard. First, the economy is not recovering rapidly, which suggests there are some problems with the model. Second, that such models simulate a rapid recovery explains why the Fed keeps holding policy steady. It is just about to work. And finally, these sorts of models do a very bad job in modeling a very different regime--nominal GDP targeting.

Unfortunately, it is hard to escape the conclusion that two models are necessary. One to capture the process of a shift to a new regime. And then the second to model the operation of the new regime.

Krugman Advocates Nominal GDP Targeting

Krugman now supports nominal GDP targeting, citing market monetarists, Scott Sumner and David Beckworth. His view of the market monetarism, however, was somewhat critical.

My beef with market monetarism early on was that its proponents seemed to be saying that the Fed could always hit whatever nominal GDP level it wanted; this seemed to me to vastly underrate the problems caused by a liquidity trap. My view was always that the only way the Fed could be assured of getting traction was via expectations, especially expectations of higher inflation –a view that went all the way back to my early stuff on Japan. And I didn’t think the climate was ripe for that kind of inflation-creating exercise.

Seemed to be saying? Sumner, especially, has been emphasizing the importance of expectations since the day he started his blog, and long before. Further, he probably writes about how it is important to generate expectations of higher inflation more often than he writes about how it is important to generate expectations of higher level of nominal GDP. I know, because I take him to task every time he makes that mistake.

While it is true that expectations of higher inflation in the future can raise nominal GDP now, and could also create inflation now, expectations of a higher level of real output in the future can also raise nominal GDP now, and create more real output now. Sumner's actual view is that expectations of an increase in the flow of expenditure on output in the future will generate an increase in the flow of expenditure on output now. To what degree firms either now or in the future respond to that by raising either prices or production is not essential to the process.

But, of course, an increase in production would be better than an increase in prices. My own view is that an increase in production would be good and an increase in prices would be bad--a necessary evil to maintain the regime of nominal GDP targeting.

On the other hand, I was much more guilty of ignoring the key role of expectations early on. The relationship between expected future nominal GDP and current nominal GDP played little role in my thinking. I certainly downplayed the problems associated with a liquidity trap.

On the other hand, I have never assumed that modest changes in the quantity of money would generate whatever nominal GDP the Fed wants. I have always thought that quantitative easing in heroic amounts might be needed to keep nominal GDP on target in a situation what would otherwise develop into a Depression. Something causes a large drop in the money multiplier and velocity, so a large, perhaps very large, increase in base money would be necessary.

What is the relationship between expected inflation and nominal GDP targeting? There is no need for the Fed to say that it wants more inflation. However, it does need to be willing to accept higher inflation if that is the consequence of nominal GDP rising to the target growth path. The expectation that the Fed would respond to any increase in inflation that occurs by giving up on the target for nominal GDP would make it difficult and perhaps impossible to reach the target growth path.

The Fed cannot play at nominal GDP targeting. It must adopt the new regime. It should adopt the new regime--it is better.

Tuesday, October 18, 2011

Goldman-Sachs on Nominal GDP Targeting

DeLong has posted the Goldman-Sachs newsletter that advocates a target for Nominal GDP.

On the whole, Jan Hatzius and Sven Jari Stehn have done a good job.

They propose a target for nominal GDP that starts growing at 4.5% in 2007. This is less than the trend growth rate of the Great Moderation, which they estimate as 5.3 percent. Nominal GDP is 10 percent below that target. This is very similar to my methodology for generating the 3 percent target growth paths. It starts at the trend of the Great Moderation when nominal GDP was on that trend, and then goes forward at a new target rate. Presumably, they use the slower rate because estimates of potential output growth are below the 3 percent trend of the Great Moderation. If the estimate is 2.5 percent going forward, then 4.5 percent nominal GDP growth implies 2 percent inflation. If and when potential growth recovers, the result would be a slight disinflation.

They use their model to simulate the effects of the policy. Their model is a standard new Keynesian model, based upon the assumption that the Fed shifts the Federal Funds rate depending on current inflation and unemployment, which impacts real expenditures, real output, and the output gap. The output gap determines unemployment, and unemployment, along with expected and past inflation determines current inflation.

They account for quantitative easing by making it equivalent to a change in the Federal Funds rate. According to their estimate, $1 trillion of asset purchases is equivalent to a 100 basis point decrease in the Federal Funds rate. The current base is about $2.6 trillion. Increasing it to $3.6 trillion would be equivalent to reducing the federal funds rate 1 percent. (I am not sure how "global" this relationship is supposed to be. The target for the Federal Funds rate is .25 percent, and while I am sure that a base money of zero isn't really the same thing as a Federal Funds rate of 2.85 percent, perhaps the "extra" $1.8 trillion is supposed to make the current Federal Funds rate equivalent to -1.55 percent.)

Their "rule" tying the target for the Federal Funds rate to nominal GDP is .6 of the gap between nominal GDP and target. (I find this a bit puzzling. Their "Taylor" rule implies a Federal Funds rate of 4% if inflation is on target and the unemployment rate is at the natural rate. It looks to me like they have the Federal Fund rate set at zero when nominal GDP is on target. Maybe it is supposed to be i* = 4 + .6 (y - y*).)

They propose increasing the Fed's balance sheet to $5 trillion over the next year. That would be an increase of about $3.4 trillion. The level of base money would be approximately 1/3 of nominal GDP. But they also propose that the Fed sell off all of those assets by 2013 and return base money to its current (high) level. At that point, the Fed would need to start raising the Federal Fund rate.

From a Market Monetarist point of view, the key problem with the argument is that they claim that the effect of the policy will operate through its effect on real long term interest rates. The asset purchases lower the term premium, while getting back to target implies a lower Federal Funds rate in the future and higher expected inflation compared to a policy based upon their version of the Taylor rule.

The "IS" relationship does take expected real growth into account, with the current output gap depending on future output gaps. (If it is expected that next year's shortfall of real output from potential will be smaller, then this will raise real expenditure this year, and so real GDP, bringing real GDP closer to potential.)

xt = 1/3xt+1 + 2/3 xt-1 - .09(it-1 - inflation t+3 + qt-1) + ft

The "f" represents private sector balance sheet impacts and fiscal policy. (I suppose the "f" stands for "fiscal," but maybe "financial.") The q represents their adjustment for quantitative easing, -1% for every trillion.)

I suppose the simplest way to account for the Market Monetarist view that this policy can result in higher real interest rates would be for the new regime to have a higher coefficient on next period's output gap. An alternative approach would be to include nominal GDP expectations as an additional term. Or, of course, there is always Sumner's radical approach of replacing future inflation with future nominal GDP growth.

I hope someone will put this in front of Bernanke, (and Obama and his Republican opponents.)

Tuesday, October 11, 2011

Is DeLong a Market Monetarist?

Is DeLong a Market Monetarist?

He sure sounds like one here:

It is not Republicans in Congress that kept and are keeping the Federal Reserve from charging for reserve balances or engaging in more quantitative easing or targeting the nominal GDP growth path.

He can use IS-LM all he wants.

Yes, the Fed is Breaking the Law!

The Federal Reserve pays .25 interest on reserve balances. While the payment of interest on reserves is a desirable policy, the current rate is too high. In my view, the Fed should float the rate it pays, setting it a fixed number of basis points below short T-bills. It my view, the Fed should be "paying" a rate that is less than zero under current conditions. In effect, the Fed should be charging banks a storage fee.

It turns out that while the Fed is legally permitted to pay interest on reserves that is less than other money market rates or to pay the same rate, it cannot pay more. David Pearson, commenting on David Glasner's blog quoted the low:

Balances maintained at a Federal Reserve bank by or on behalf of a depository institution may receive earnings to be paid by the Federal Reserve bank at least once each calendar quarter, at a rate or rates not to exceed the general level of short-term interest rates.

Steven Williamson noted this as well, and suggested that the Fed needs to break the law to "do its job." Williamson claims that it must be that T-bills are more liquid than reserves. Still, it is hard to believe that the Fed's job is to pay interest on reserves at a level that reflects the liquidity value of T-bills relative to reserves. With Williamson, it is hard to tell.

I would argue that if the Fed is going to use the interest rate paid on reserves as a policy instrument to restrain nominal expenditure on output, then it must sometimes lead the market. If banks are holding the relevant money market instrument, then it would seem that raising the rate paid on reserves would cause banks to sell the money market instruments, lowering their prices and raising their yields. But if the banks sell off all of them, then that avenue for adjustment no longer exists. What happens instead is that the higher rate paid on reserves pulls up the rate that banks pay on deposits, so that bank depositors also shift out of the relevant money market instrument and into deposits.

It is also possible that when banks respond to higher rates paid on reserves by raising bank loan rates, some borrowers might sell commercial paper rather than utilize bank loans. That would tend to raise the rates on money market instruments too. This would also involve the Fed "leading" the market.

However, I don't favor the use of the interest rate on reserves as a policy instrument. If there is an excess supply of base money, the Fed should reduce the quantity through open market sales. I don't think that the Fed should be using the composition of its asset portfolio to try to direct credit. (Keep nominal GDP on the target growth path and let the market allocate credit.)

More importantly, the Fed doesn't need to be restraining nominal expenditures on output now. Quite the contrary. If the Fed were to use the interest rate on reserves as a policy instrument, it would be "leading" the market lower. And there is nothing in the law that prohibits that--at least until the interest rate hits zero.

Could the Fed "get away" with charging the banks fees based upon the size of their reserve balances without some kind of authorization by Congress? Probably, if they could point to some financial cost to them when banks hold reserve balances. For example, if the yields on T-bills are so low, the Fed is earning next to nothing, or perhaps, literally nothing, or perhaps, receiving a negative yield like everyone else, then any operating costs would have to be covered. The Fed has to keep track of the banks' balances, collect on T-bills and reinvest the funds, and so on. The Fed could buy longer term Treasuries, but that involves a risk. The banks should share in that potential cost.

Of course, if the interest rate on reserve balances falls so low that the result is a currency drain, with either banks accumulating vault cash or the nonbanking public filing safes or stuffing mattresses, then it is too low. Here again, the Fed would need to "lead" the market by "paying" sufficient interest on reserves, regardless of whether market rates are lower, to avoid the cost of printing currency.

Sunday, October 9, 2011

Monetary Disequilibrium and Interest Rates

Nick Rowe has argued that Keynesian and New Keynesian models assume that people always borrow money to hold, and that they fail to explain that people borrow money to spend. Since I think that pretty much no one ever borrows money to hold and instead everyone borrows money to spend, this would suggest that Keynesian and New Keynesian models should be rejected.

However, I cannot believe that Keynesians or New Keynesians really believe this. Instead, I think that these models focus on one particular avenue by which monetary disequilibrium can be manifested and emphasize a plausible temporary equilibrium as a way of describing the adjustment process.

It depends on some plausible institutional assumptions. First, at least some people are involved in “cash management,” trading securities on organized exchanges when they have an excess supply or an excess demand for money. Second, the interest rates paid on money balances are “sticky,” and do not immediately adjust with changes in interest rates on other financial assets, particularly the securities that are being traded by those managing their cash positions. Neither of these assumptions is particularly fundamental. Other than that, there is the assumption that the demand to hold money depends on the opportunity cost, the difference between the interest rates that can be earned on other assets, particularly those being managed, and the interest rate on money itself.

If there is an excess supply of money, those who don’t manage their cash positions just spend it on goods, services, or financial assets. Those receiving the money don’t necessarily want to hold the money. They have excess money balances and spend them as well. The hot potato is jumping about. But this hot potato is going to start hitting the money balances of those who manage their money positions, and they will purchase securities on organized exchanges. The prices of these “bonds” rise, their yields fall and the opportunity cost of holding money falls. The interest rate on money remains unchanged, so those that manage their cash choose to hold more money. The demand to hold money rises until there is no more excess supply of money.

There is no borrowing going on at all. What is happening is that there is an increase in lending. Instead of holding money, those who manage their cash positions, lend more by holding more short term bonds. If holding money counts as a type of lending, either to banks or the government, then those managing cash positions are lending less by holding money and more by holding short term securities.

Consider the opposite scenario. If there is an excess demand for money, those short on money may reduce expenditures out of income or perhaps sell assets. Those buying assets don’t necessarily want to hold less money, and it is highly unlikely that those receiving fewer money receipts want to hold less money. They reduce money expenditures as well. However, before too long, those who manage their cash positions end up with less money and sell financial assets to rebuild money holdings. The interest rates on those assets rise. Since the interest rates paid on money are sticky, the opportunity cost of holding money rises, and the demand to hold money falls until there is no longer an excess demand for money.

In this scenario, there is no borrowing. There is a decrease in lending. Instead of lending by holding bonds, they hold more money. If holding money counts as lending to the banks or the government, then they are lending more by holding money and lending less by holding bonds.

But surely this is a peculiar type of credit market where interest rates change without there being any change in the quantity of credit demanded. If credit demand were perfectly interest inelastic, then the story above makes sense. For example, suppose the securities that are used for cash management are Treasury bills, and the amount outstanding doesn’t respond to changes in interest rates. The government doesn’t expand or contract its budget deficit in response to changes in interest rates.

On the other hand, suppose that some of those issuing the securities that some people use for cash management change the quantity issued based upon interest rates. The demand for credit and the supplies of these securities are less than perfectly inelastic. Since those borrowing are unlikely to do so simply to hold money, the changes in the quantity of credit demanded are simultaneously changes in the demand for goods and services or perhaps still other assets.

And this leads us to Nick’s point. If there is an excess supply of money, and lower interest rates clear it up, if there is any increase in the quantity of credit demanded, and those borrowing spend the money, the excess supply of money still exists. Similarly, if there was an excess demand for money, and higher interest rates clear it up, but this results in a decrease in the quantity of credit demanded, and the reduced borrowing involves less spending, then the excess demand for money was not cleared up after all.

Keynesians (new and old,) include this effect as part of the “IS” curve. I suppose it might be rationalized as a rate of adjustment. Short term interest rates adjust very rapidly to absorb any excess supply or demand for money. And then, those new interest rates gradually result in changes in credit markets and spending on output. Assuming money is a normal good, changes in real output and real income cause adjustments in the demand to hold money. If we assume that output adjusts faster than prices (a heroic assumption, though if we “know” that it is prices and wages that need to adjust and that the full price adjustment will fail to occur before there are changes in output, then perhaps looking at what happens to output with given levels of prices and wages is sensible.)

Aside from questions about how interest rates actually change with monetary disequilibrium, which Nick has also emphasized, I would emphasize how this depends on the interest rate paid on money being sticky.

If there is an excess supply of money, and those managing their cash positions buy securities, and this lowers the yields on those securities, and the interest rate paid on money drops in proportion, then the opportunity cost of holding money doesn’t fall, and the excess supply of money and the “hot potato” continues, regardless of the interest elasticity of the demand for credit. Similarly, if there is an excess demand for money, and this raises security yields, but the yield on money rises in proportion, then the opportunity cost of holding money doesn’t rise, and the “musical chairs” problem continues.

So, sticky interest rates on money itself, closely managed cash positions, trading with securities with very low short run elasticity of supply, and we can treat the short term interest rate keeping the quantity of money demanded equal to the quantity of money supplied over some short term time horizon. Of course, if the whole point of the exercise is to help central banks achieve their revealed preference to keep short term interest rates as stable as possible rather than to keep nominal expenditure on output on a slow, stable growth path, the advantages of this approach are clear enough.