Sunday, August 24, 2014

NGDP Targeting in Developing Countries

Scott Sumner commented on the article by Pranjul Bhandri and Jeffrey Frankel arguing that Nominal GDP targeting is relatively desirable for developing countries.    Scott disagreed and argued that it would be better to target labor compensation.   He may be correct that labor compensation is the better target, but his argument was faulty.

Scott wrote:

However I've also argued that NGDP targeting is not optimal when countries depend heavily on the production of commodities with very volatile prices. Suppose 50% of Kuwait's NGDP were oil production. If global oil prices doubled, and Kuwaiti oil output remained unchanged in physical terms, then the central bank of Kuwait would have to reduce non-oil production to zero in order to keep NGDP stable. Obviously that would not be optimal.

If Kuwait's GDP (real and nominal) was composed 50% oil production and 50% of other things, and the world price of oil doubles and the number of barrels of oil produced stays the same, and real GDP is held constant, then real GDP from all other activities must drop to zero.  I grant that would not be optimal.

But nominal GDP targeting doesn't require that real GDP remain constant.  The factor Scott ignores is the exchange rate between the Kuwati Dinar and other currencies.   Assuming that the global price of oil is quoted in dollars, the immediate effect of doubling of global oil prices would be a doubling of the dollar price of the Kuwati Dinar.   While the dollar price of oil will have doubled, the price of oil in Kuwati Dinar would be the same, and assuming, for the moment, no other change in Kuwati economic activity, nominal GDP for Kuwait in Dinars is unchanged.

Of course, this large increase in the exchange rate of the Kuwaiti Dinar is going to have major effects on the Kuwaiti economy.   Prices of imported goods in the shops will fall substantially.   The unchanged nominal incomes will now purchase twice as many imported goods.    All Kuwatis benefit from the favorable supply shock.

In a not unrealistic scenario for Kuwait, suppose all the other goods produced by the economy are nontraded.    Kuwait exports oil and operates shops that sell imported goods and have barbers who give haircuts.   Suppose that the nontraded goods are normal.  The Kuwaitis can now buy more foreign consumer goods.   They need and desire bigger and more expensive shops.  They also get more frequent and fancier haircuts.

The increased demand for nontraded goods would tend to raises their prices in Kuwaiti Dinar, and so raise nominal GDP.   This would require a  further appreciation of the Kuwati Dinar.    This reduces the value of the unchanged quantity of oil in Dinars, and so reduces nominal GDP.     The Dinar must rise enough so the Dinar value of oil produced in Kuwait falls enough to exactly offset the increase in demand and so expenditure on nontraded Kuwati goods.

Now, the reduced price of oil in Dinars should result in a lower quantity of oil supplied and so free up labor and other resources to shift over to production in the nontraded goods sector.   Former oil workers shift over to work in shops selling imported consumer goods or else be barbers.  

Of course, Scott's assumption that the the quantity of oil remains the same suggests a lower Dinar price of oil frees up no resources from the oil industry.   That is plausible enough.   It is all extraction of existing oil   The cost of the mechanics repairing the pumps and pipelines is trivial.    It is nearly all rent extraction.   Initially the Dinar rose enough that the nominal value of the rents were unchanged, but the income effect raising the demand for nontraded goods requires a further appreciation of oil so that the nominal rents from oil fall enough to keep total spending unchanged. Perhaps the Emir hires a few less soldiers and instead they go work in the shops and cut hair.   The remaining soldiers get more haircuts and go to fancier shops.   The barbers and shopkeepers also go to fancier shops and get more haircuts.   And everyone enjoys many more imported consumer goods.

However, let's suppose that some Kuwaitis are employed growing dates.   Perhaps they export dates.  Perhaps they solely sell dates domestically in competition with imported dates.   When the world price of oil doubles and the exchange value of the Kuwati Dinar doubles as well, the date growers are in trouble.   The prices they receive from exporting dates fall in half.   The price they can get for their dates from domestic retailers also fall in half.   

This would lead a collapse in Kuwait's domestic date industry.    The appreciation of the Kuwati Dinar would therefore need to be less than in the situation where oil is the only traded good.   Dinar expenditures on oil would rise, and Dinar expenditures on dates would fall.   This would be desirable to the degree it would create signals and incentives to shift resources away from date production to oil production.   Again, the assumption in Scott's scenario that the quantity of oil remains constant implies that no additional labor and other resources are needed in the oil industry.    Still, the lower prices of imported consumer goods increases real income, which increases demand in the nontraded goods sector as before.   For Kuwait, it is easy to imagine that the adverse impact of the change in world oil prices on the date industry would have no significant impact on the entire economy.

Of course, other developing economies might be different.   

Anyway, the problem Scott sees is not really that a developing country has a major product with a volatile price.   On the contrary, what nominal GDP targeting does is make the exchange rate and the prices of imported consumer goods adjust rather than wages and other nominal incomes.   The problem Scott really sees is where the product is one with an inelastic supply.    If expenditure on that product increases, there is little possibility of expanding its production by increasing the employment of resources.   Nominal GDP targeting requires that expenditure be reduced elsewhere in the economy to offset the increase.  The signal and incentive to contract output and employment in the rest of the economy to free up resources to expand the product where expenditure increased is inappropriate because it is difficult or impossible in the sector with increased expenditure.  

So, if there is some developing economy where a large portion of its output is oil, copper, diamonds, or gold, and it has substantial output of other traded goods, perhaps local farmers compete with imported grain, then shifts in the world demand for the good with inelastic supply will impact the exchange rate, spending in export and import competing industries.   Spending on the product with inelastic supply will change in terms of the domestic currency, requiring offsetting changes in spending in the rest of the economy. 

Sunday, August 10, 2014

What are Banks?

There is a NY Times article about Adnat Admadi's view on banking regulation.   Her focus is on increasing capital requirements.   Like most free bankers, I see increased capital as desirable.   The article mentions that most firms don't have capital requirements at all.   Most firms are mostly funded by equity because lenders insist on it.   According to the article, banks are different because depositors are protected by the government from loss.

And that points to why I see increased capital as desirable.   Before deposit insurance, banks did keep much more capital.   At least in the U.S., it was the introduction of  deposit insurance that resulted in substantial decreases in bank capital ratios.

However, even without deposit insurance, banks were mostly funded by deposits.   Why?

It is because banks are financial intermediaries.   They are not simply suppliers of loans.   The deposits that banks issue to fund loans provide services to depositors.   Traditionally, these were monetary services.

Consider a grocery store.   It buys food products from wholesalers and then sells food, mostly to the final consumers.    The grocery store must finance its operation--the store, equipment, inventory, and so on.   The typical grocery story is mostly financed by equity, I suppose.   But there is no notion that the owners of the grocery store must have equity equal to a substantial portion of total sales of groceries.

A bank also has a building and equipment.   But its total assets also include loans and investments.   These are similar to the grocery store's sales of food.   Further, the deposits the bank uses to fund these assets are more similar to the grocery store's purchases of food from suppliers rather than the instruments issued toequity and debt holders that finance the grocery store.

While bank-issued currency provided important benefits in the past, and could do so in the future, for many years these monetary services involved checkable deposits.     While checkable deposits have come to make up a remarkably small portion of bank liabilities, at least part of the reason has been the development of sweep accounts.   By sweeping funds out of checkable deposits and into something else right before the time they must be reported, banks not only avoid regulation, but money and banking statistics are distorted.

However, there is little doubt banks have long issued deposit liabilities whose primary special characteristic is that they are guaranteed by the government.   Certificates of deposit have a lot in common with commercial paper, but the bank liabilities are guaranteed by the government.

While additional capital for banks is desirable, the key problem with capital requirements is that the purpose of capital should be to serve as a buffer.   That means that the buffer should be used when banks suffer losses.   And so, when a bank has exceptional difficulties, its ample capital buffer should be allowed to decrease without interfering with the continued business of the bank in both issuing deposits and making new sound loans.  Obviously, a bank should then rebuild its capital as it recovers.

Giving discretion to the regulators is probably worse than a strict rule.   During good times, when banks have few loses, so what if loose definitions of capital allow requirements to be met on paper.   And then, when banks are losing money, that is when the regulators get tough and make sure that banks rebuild their capital.   Just when banks should be using the cushion they should have built up during good times, the regulator starts strict enforcement.

The problem with a required capital ratio is that restricting new loans and using the funds to pay down deposits (or accumulating "safe assets" with low capital requirements,) is not desirable.   Of course, if it is only a single small bank in a healthy banking system that must shrink its balance sheet to meet the requirements, the effect on the economy is small.   This is especially true because other banks would be in a position to expand deposits and loans.   If necessary, they could issue new equity to take advantage of the profits from the added business.  

But what happens if many banks are in difficulty at the same time?   Having all banks shrink their balance sheets at the same time is not a good thing.

Further, as mentioned in the article, there will be a constant tendency for financial innovation aimed at getting around the regulation.    If those quasi-banks suffer runs, the run will be to the well-capitalized "safe" banking sector.   Those banks should be rapidly expanding in such a scenario.   Requiring that such banks raise capital (or even postpone paying dividends) will interfere with such a needed expansion.

For example, during the financial crisis of 2008, investment banks were issuing quasi-deposits to fund quasi-mortgage loans.   They were shadow banks.   When there was a loss in confidence, and the quasi-depositors ceased rolling over their overnight funds, they received payment in their conventional checkable accounts and simply held the funds at conventional banks.   Did the supply of money decrease?  Yes, if overnight repurchase agreements issued by investment banks are counted.   Or was it the demand for money increased?   Yes, if only checkable deposits issued by conventional banks count as money.   Regardless, what needed to happen is for conventional banks to expand their deposits and their lending.   Capital requirements made that more difficult.   (Of course, the fact that the commercial banks were also under diversified by being over invested in real estate loans made the problem doubly difficult.)

In my view, rather than requiring banks to fund their asset portfolio with some fixed ratio of capital relative to deposits, a better approach is to make the monetary liabilities that provide the rationale for banking more like equity.  

First, there should be an option clause that allows banks to stop a run.   Banks need to be able to postpone payment, though the banks should pay penalty interest.   In other words, depositors should be compensated for any postponement with bonus interest.  Further, the suspended deposits should be negotiable.   In particular, other banks should be able to accept them for deposit either at par or at a discount.  

Second, if a bank is insolvent, then each depositor should suffer a write down of their deposit balance with  compensation by equity--with the reorganized bank being well-capitalized.   And this reorganization of banks should be rapid.   Days, not months.

Kevin Dowd once explained it well.   Closing failed banks for months or more makes as much sense as wheeling out hospital patients into the street because the hospital has financial problems.

If banks have government deposit insurance, then these changes can be required as a condition of continuing that insurance.   Of course, the real point is that with these sorts of reforms, banks might be able to operate without deposit insurance.    And if banks have no deposit insurance, then they can maintain their own liquidity and capital policies in order to attract depositors.

Further, as soon as we explore systematic issues, the nature of the monetary regime becomes paramount.   A desirable nominal anchor--such as nominal GDP level targeting would help.   Further, avoiding a monetary base that has no nominal risk and a zero nominal yield is also desirable.

A century ago, the nominal anchor was the fixed price of gold, and gold served as a base money with zero nominal risk and a zero nominal yield.   Those days are gone.   The banking system does not need to be  able to withstand a massive shift from everything else to gold, which would require a massive deflation of nominal output and nominal income.  

Yes, a 30 percent capital ratio might make sense with a gold standard.   Maybe it is wise when hand-to-hand paper currency is the fundamental monetary base and central bankers insist on using a nominal interest rate instrument to target inflation.    In my view, those are the policies that need some radical revision.

Thursday, July 31, 2014

The Election is Over!

Election day was Tuesday.   I won 68% to 32%.  

There were two "attack mailings."   My "favorite" was the one where my picture looked a bit like a "Captain Planet" villain with text to match.  

Over the last two months, I knocked on about 2,200 doors.   It was slightly more than half of the total in my town.   I spoke to more than 1,000 voters at their doors.    It was hot and tiring work, but it does give a good picture of what people are thinking and the state of the Town's roads and drainage infrastructure.   I think my staff will be happy to stop getting my daily cell phone photos of areas that need work.

Classes start at the end of August.    I hope to do some economics blogging over the next few weeks.

Sunday, May 11, 2014

Nominal GDP Targeting--Some "Micro" Observations

Supply-sider Alan Reynolds wrote a critique of nominal GDP targeting recently.   David Beckworth responded.   Lars Christensen also wrote a post that didn't mention the Reynold's piece, but in some ways was also a response.

Reynolds begins by arguing that nominal GDP growth is simply the sum of real GDP growth and inflation.    He goes on to claim that if the Fed were to target nominal GDP, then if real GDP growth is high, then the Fed would be required to cause deflation and if real GDP growth is low, it would have to then respond by causing inflation.

Beckworth responded by pointing out that nominal GDP is calculated first, and then a price index is calculated.  And then finally, real GDP is calculated by dividing nominal GDP by the price index.   (In truth, several price indices are calculated for different types of nominal output--like consumer goods and services or capital goods.   And then nominal consumption is divided by the consumer expenditure price index and investment by the investment index.   The real amounts of the parts are added together.    The implicit GDP deflator is actually calculated by dividing nominal GDP by real GDP.    However, I am not sure to what degree that complication matters.)

We can imagine a world where the real economy is determined first.   Perhaps a Walrasian auctioneer determines all the relative prices and quantities appropriate to general equilibrium.    And then, the Fed sets the quantity of money, and the Walrasian auctioneer determines the money price level necessary to equate the demand to hold money to the quantity of money.  And finally, everyone uses the already determined relative prices and the price level to determine money prices.   First real output is determined and then the price level is determined.

The real world is nothing like that.   The production of various goods and services, and so, real output is simultaneously determined with the money prices of those same goods and services, and so the price level.   

Of course, I am using a microeconomic model to frame this vision.   The demand curve for some good shifts to the left, and both the equilibrium price and quantity decrease.

However, the simple supply and demand model is literally based upon perfect competition.   Instead, my true framing is monopolistic competition.   The demand for some firm's product shifts to the left, and that reduces marginal revenue.  Given marginal cost, this makes it profitable for the firm to cut production and prices.    There is no first production and then prices to it.   The firm jointly determines the price and quantity to maximize profit.

Some markets might come closer to a production first and then prices later approach.   For example, the farmer plants seed.   Months later, after weather happens, there is a harvest.   Quantity is determined.   And then, many agricultural products are sold on auction-type markets.   Price is determined.

And even for a more typical market, production and prices necessarily depend on anticipated demand.   If demand is less than anticipated, then excess inventory may well be sold at lower prices.

However, nominal GDP level targeting is all about reversing these sorts of decreases in demand.   And so, given these simple models, such a reversal would result in inflation, but also increases in real output.   If the demand curve shifts left now, and then later shifts right to its original position, then the equilibrium price first falls and then rises.   The equilibrium quantity first falls and then rises.   The same is true of the profit maximizing price and quantity combination for a firm in monopolistic competition.  It falls with the decrease in demand and then rises again when demand recovers.

To the degree prices are sticky, then this fluctuation in price is dampened.   This would be both the deflationary effect of the reduced demand and the "inflationary" recovery of price as well.   For simple supply and demand, it is as if the short run supply curve is highly elastic.  

Of course, there is a sense in which this entire analysis is a massive fallacy of composition.   Nominal GDP falls, and so the demands for just about every good or service falls.   And this analysis treats each and every good in isolation.  When the demand for other goods fall, this increases the supply of a good, because opportunity cost for producing it is lower.   And so, the decrease in nominal GDP implies a shift of demand to the left for all goods, but then also a shift in the supply of all goods to the right.  The equilibrium price of all goods would fall, but their quantities would remain the same.

There is a sense in which this is what "should" happen.   And in a world where all markets clear perfectly,  I believe that this is what would happen.   In such a world, real output and relative prices would be independent of aggregate demand, or more fundamentally, imbalances between the quantity of money and the demand to hold it.  

But that is not the way the world works.   Firms do respond to shifts in aggregate demand, that is, to shortages or surpluses of money, as they would if the demand for their particular product had changed.   And so, reversing shifts in aggregate demand result in a reversal of undesirable shifts in real output and employment.   And further, treating the inflation due to a recovery from deflation as if were somehow a "cost" is a mistake.   Prices are recovering to where they belong.

Beckworth and Christensen, focuses more on shifts in aggregate supply.   Given Reynold's reputation as a supply-sider, that makes sense.   If you assume that prices are sufficiently flexible so that shifts in aggregate demand have little or no impact on real output, then what is left other than shifts in potential output?

Suppose various anti-business programs by the Obama administration have reduced potential output.   Does that mean that real GDP falls, so the Fed would have to engineer an increase in inflation for nominal GDP to rise back to target?   Is the extra inflation adding insult to injury?

Again, think about the micro implications.   Suppose the government mandates benefits for workers in some industry.   This raises costs.   Using simple supply and demand analysis, the supply curve shifts left.   The result is a simultaneous decrease in the equilibrium quantity and increase in the equilibrium price.   If this occurs in all industries at once, then real GDP decreases and the price level increase simultaneously.

Considering monopolistic competition, the mandated benefit shifts the marginal cost curve for a firm to the left.   The profit maximizing quantity is lower and the profit maximizing price is higher.   If this happens to many firms at the same time, the result in reduced real output and a higher price level.   There is no need to generate inflation to push nominal GDP back up to target.

Of course, again, this micro analysis is ignoring what is happening to all of the other markets.  If all markets suffer added costs due to the mandated benefits, and they all produce less, so real output and income fall in aggregate.   Lower income should reduce demand.   Further, when the rest of  the firms produce less, that reduces the opportunity cost for any one firm.   These lower opportunity costs are signaled by reduced wages and other resource prices.   Only to the degree lower real wages and other resource prices result in reduced offers for sale, does output fall in the aggregate.

If we imagine that all of price adjustments occur instantly, then we can imagine market clearing at an unchanged price level.   Firms have higher costs because of the mandated benefits.   They have lower costs due to paying lower wages and other factor prices.   The reduced incomes for workers and other resource owners results in lower demand to match the lower supply in each market.   But how realistic is that?

Quite the contrary, the most likely short run effect of increases in costs due to mandated benefits would be higher prices and reduced output.   For the Fed to prevent this from causing inflation, it would have to reduced spending on output.   This would force output and employment down even further, and if the Fed remained committed to this policy, eventually, wages and other factor incomes would fall to a lower growth path, allowing a very gradual recovery of output without any increase in inflation.

Christensen argues that if a central bank responds to adverse aggregate supply shocks by "tighter money," then we will see slower growth in real output due to the supply shock and then slower nominal GDP growth due to the central bank;s response.   It will appear that slow real output growth "causes" slow nominal GDP growth.   He points out that this is due to a regime that targets inflation.  

I would add that if people expect the central bank to restrict nominal GDP growth to combat supply side inflation, then the expectation of slower growth in nominal GDP in the future will immediately depress nominal GDP.   The inflationary effect of the adverse supply shock will be dampened.  And further, the expectation of the future slowdown in spending will exacerbate the reduction in real output immediately.

With "flexible" inflation targeting, where we never know what the central bank is really going to do, then the possibility that they will suppress the inflation from the supply shock will dampen the inflation and exacerbate the reduction in output immediately.   And then, when it turns out that they will in fact seek to dampen the inflation, so that there is no possibility that they will flexibly allow inflation to spike, then the more direct effects of the monetary tightening will be felt on both inflation and real output.   More positively, if they do prove themselves willing to allow higher inflation, there should be some recovery of output as well as more inflation.

Wednesday, April 16, 2014

Interest Rate Targeting and Finanical Instability

David Beckworth linked to slide program by Michael Darda.   I agree with Darda's analysis in general.   I would emphasize that it is like 1936.   Spending on output is far too low, but at least some at  the Fed are worried that "loose" money is causing excessive speculation.   We can only hope that we don't end up with the same result--a steep recession in an economy that is already below potential.

Still, I am more and more concerned that Federal Reserve policy tends to cause problems with financial instability.   The problem, however, isn't maintaining the nominal anchor, but rather targeting interest rates.   More generally, the problem is interest rate smoothing.   However, in the current situation, the problem is an attempt to generate a recovery by promising to keep interest rates low.

Interest rate smoothing is a policy by a central bank to keep money market interest rates stable.   Of course, there is probably no central bank today that tries to keep interest rates fixed beyond a very short time horizon.   Typically, they adjust interest rates periodically in order to keep inflation on a target, though in the U.S. they also most promote full employment.   What interest rate smoothing amounts to is a commitment to keep interest rates unchanged as long as inflation remains on target and real output at potential, and if deviations occur, adjust interest rates in a series of modest steps.   Mainstream macro usually works on the assumption that the changes in interest rates occur once and for all based upon the size of the deviation of inflation or real output from target, and interest rate smoothing then is a modification of the Taylor rule so that the current interest rate is adjusted to the target through a series of modest steps.  This should have little adverse effect because everyone is assumed to understand the underlying rule.  Knowing that the policy rate will adjust soon, longer term rates adjust promptly to their expected long term level.   And it is these longer term interest rates that impact spending decisions.   And expectations about these spending decisions are what determine production and pricing decisions, and so real output and inflation.

So, why interest rate smoothing?   In my view, it is one of those situations where mainstream macro is simply following the lead of central bankers.   They like to smooth interest rates, and mainstream macro shows that it is consistent with macro stability.   Rationalizing what central bankers want to do somehow ends up playing a key role in mainstream macro.  

But why do central bankers want to smooth interest rates?   Surely, the underlying reason is that money center financial institutions borrow on money markets.   A key role of financial intermediation is to borrow short and lend long.  Even traditional investment banking has involved borrowing short to underwrite stocks or bonds.    If short term interest rates spike, this is costly to financial institutions.   Keeping short term rates fixed avoids those problems.   Of course, keeping short term interest rates fixed is a recipe for inflationary disaster, and impossible.   Still, it is possible to manage the short term rates so that the adjustments are slow and steady.   And this allows the financial intermediaries to make needed adjustments.   While mainstream macro makes all of this very mechanical--rule driven--central banks have never implemented such a policy.   And so, needed adjustments in interest rates can be implemented in a way that allows money center banks to make needed adjustments in their balance sheets in a smooth way.  

From a monetary disequilibrium perspective, there are two different sources of interest rate fluctuations.    Focusing on increases, one possibility is an excess demand for money.   By money, I mean the financial instruments used as media of exchange.    An excess demand occurs when the existing quantity of money is less than the demand to hold it.   Those short on money may sell off short term assets they hold for his very purpose or perhaps borrow--effectively issuing and selling new money market instruments.   This will tend to cause money market rates to rise.   From a monetary disequilibrium approach, the solution is to increase the quantity of money to accommodate the demand.   It is also important to consider the possibility of lowing the interest rate paid on financial instruments used as money to reduced the quantity demanded.

The other reason interest rates might spike is because of an excess demand for credit.   This could occur because of an increase in desired borrowing to fund purchases of capital goods or consumer goods.   Or it could occur due to a decrease in desired lending.   It could be that firms choose to lend less and instead internally fund purchases of capital goods.   Or it could be that households are saving less.   From a monetary disequilibrium perspective, the interest rate should rise to bring the credit supplies and demands into balance.   More fundamentally, the interest rates should rise to coordinate saving and investment.   If the monetary regime is such that the quantity of money rises or the demand to hold money falls, creating an excess supply of money, this is a failure.   The result is that interest rates fail to adjust enough to keep saving and investment equal, and instead total expenditure fluctuates.  A central bank with a policy of interest rate smoothing is purposefully causing monetary disequilibrium to keep the interest rate from adjusting enough to clear the supply and demand for credit.   Avoiding an interest rate spike due to an increase in the demand for or decrease in the supply of money is generating an undesirable excessive increase in spending on output.

From the point of view of individual money center banks, it hardly matters whether an increase in interest rates is due to an increase in credit demand--more investment or less saving--or a shortage of money.   Their borrowing costs increase either way.   But from the point of view of maintaining monetary equilibrium and so, overall macroeconomic equilibrium, it makes a difference.

A policy of interest rate smoothing then, keeps interest rates too stable.   This means that financial intermediaries have less risk and can operate with more leverage.    This makes financial institutions more likely to fail due to those increases in interest rates that the central bank finally approves to avoid inflation.   However, as explained above, slowing any needed interest rate increase can give financial institutions time to adjust. 

However, interest rate spikes are hardly the only thing that might go wrong for financial intermediaries.  Bad investments, such as investing into a speculative real estate bubble, can also generate losses.   

Of course, a monetary regime that allows monetary disequilibrium to generate fluctuations in money market interest rates will create even more risk to financial intermediaries from interest rate spikes.   And so, they would tend to have more capital and less leverage than an ideal regime that always adjusts the quantity of money to the demand to hold it.   Further, if the monetary regime allows that disequilibrium to fester, so that spending on output, production, and prices all decrease due to a shortage of money, the solvency issues that creates for financial intermediaries will greatly increase their motivation to limit leverage, have ample capital, and even hold higher reserves.  

But surely, creating fear for financial intermediaries is hardly a sufficient reason to have a monetary regime that periodically results in recessions and deflation.    Still, a monetary regime that seeks to protect financial intermediaries from all short term increases in interest rates to the degree possible, is creating extra financial risk as well as tending to generate booms in output and inflation.

Wednesday, February 12, 2014

More on Negative Interest Rates

Miles Kimball wrote a kind response to my previous post.  However, it was quite critical.   One small advantage that I have over Kimball is that I regularly read his blog, and so am familiar with his arguments.   In fact, I have written posts in response to a number of the points he brought up.

Most of his points amount to criticism of nominal GDP level targeting.   While I favor that  regime, I grant that it falls short of perfection.   However, I have not yet been convinced by Kimball that price level targeting is better.

My approach to supply shocks is fundamentally microeconomic.   It roughly follows the analysis in Selgin's monograph, "Less Than Zero."    I suppose that an "oil price shock" is the sort of supply shock that concerns me, though a shift in supply for any single good or service creates the same issue.

If there is a shift in supply for a single good in a multi-good economy, the least bad option is for the price and quantity of that particular good to change without any response by the macroeconomic regime.   In particular, generating monetary disequilibrium to force all of the other prices in the economy, including nominal wages, to change in the opposite direction to stabilize some price index is an undesirable policy.  

Now, Kimball has that problem solved.  First, he proposes to only stabilize the price level within a broad band.   Make the band broad enough, and then, sure enough, any change in the price level due to a shift in supply of some particular good can be ignored.   Second, he proposes to stabilize "core" inflation, and ignore volatile prices.   Any shift in supply of any of the goods whose prices are counted as volatile will be ignored.    And finally, he proposes that a new price index be stabilized, one that focuses on sticky prices.   Any shift in supply of any good with a sufficiently flexible price will be ignored.

I think a target growth path for spending on output is superior to all of these fixes for price level targeting.   I am very open to the use of bands in monetary rules.   These rules are best understood as constraints on the monetary regime.   However, making the band so broad that changes in the price level due to shifts in the supply of particular goods are likely to stay within the band is not constraining enough.

Since I favor targeting the growth path of spending on output, I naturally would consider some shift in spending away from trend under Kimball's approach to price level targeting.   For example, suppose spending on output falls.   Some prices fall as well.   Almost by definition, these would be the flexible ones.   Some of them might be volatile as well.   The low relative prices of those goods discourages production, even though markets clear.   And firms with sticky prices reduce production to prevent an excessive build up of inventories.   The central bank can congratulate itself on keeping the price level within its wide band, and really it was only the volatile and flexible prices that shifted.   It sounds like 2008 to 2014 all over again.

Of course, if we require that the central bank also manage the output gap, then there would still be a problem with the scenario above.  But we know that there are central bankers that will consider a job half done (price level target achieved, at least) as good enough.  And further, there are plenty of economists who can cast quite plausible doubts on any measure of the output gap.

Keep nominal GDP on a target growth path, or within a narrow band, is a simple rule.   For those of us who believe that the discretion of central bankers should be limited, that is an important criterion.  That it is the same thing as keeping the quantity of money on a stable growth path in the special case where velocity is constant is also important.  The central bank is not using monetary policy to manipulate the economy--it is just accommodating shifts in velocity.

Kimball cites a paper that argues that the price level should be stabilized in the face of  a positive productivity shock  That was in response to my statement, "Negative interest rates to prevent unusually rapid growth in productivity from causing deflation seems like a very bad idea."    

Here I had in mind the arguments made by David Beckworth.   Productivity rises, pushing inflation below trend. But the natural interest rate rises due to the additional productivity growth.  The Fed keeps interest rates low to get inflation back up to its two percent target.   If, like Kimball and I, you favor a trend inflation rate of zero and an ability to have short and safe interest rates less than zero, then this sort of scenario could play out with deflation and negative short and safe interest rates aimed returning the price level to target.

However, I am really more concerned with the following scenario.   Peace breaks out in the Persian Gulf.   Oil prices drop significantly.   The CPI in the U.S. falls significantly.   The natural real interest rate is little impacted, but the deflation is greater than the real interest rate.   To keep the CPI stable, it is necessary to generate increases in the prices of other goods and services, including nominal wages.   The nominal interest rate is pushed below zero to cause the real interest rate to fall below the natural rate, generate a boom, increase the prices of goods and services other than oil, and create a sufficient shortage of labor to raise the growth path of nominal wages.   What a bad idea.

My understanding of Kimball's argument is that the scenario is quite different.    Labor productivity rises.   During the adjustment period, there is a decrease in labor's share of income and an increase in capital's share of income.   In the long run, continued saving and capital accumulation returns the income shares to the initial values, consistent with steady state growth.   If nominal income shifts up to a higher growth path, total nominal wage income can continue on an unchanged growth path and capital income can rise to a higher growth path.   Nominal wage rates can continue on an unchanged growth path.  As the economy adjusts to the long run steady state and the share of labor income recovers, nominal wage income shifts to a higher growth path as do the nominal wage rates.   Capital income shifts down to a lower growth path.

While I wouldn't describe this as a reason why the price level should be kept stable, I might see this as a reason why stabilizing the growth path of wage income is better than stabilizing total nominal income.   Glasner and Sumner both have argued for stabilizing the growth path of a wage index for much the same reason.    These are the sorts of reasons why most of us understand that nominal GDP level targeting is not perfect, just better than inflation or price level targeting.

How important is this problem in the real world?   Was employment stabilized by the Fed's efforts to keep inflation stable in the face of more rapid growth in productivity?   I am doubtful of any overall boom in 2002, but I do think there was a boom in late nineties.  A cursory look at the time series shows a bump in nominal GDP, real output and employment above their long run growth paths.  

As for free banking, I think Kimball has not fully internalized the idea of a monetary regime where hand-to-hand currency is not base money.   If base money solely takes the form of deposit liabilities of the central bank, then currency just isn't that important.   Sure, it has some advantages for some transactions.   Some of those transactions are even legal.   But there is no need to worry about financial collapse if there is a temporary suspension of hand-to-hand currency payouts when the currency is not base money--when the deposits are not defined as promises to pay currency.  

I have been interested in such systems for a long time.   While the term "free banking" is sometimes identified with a system with no central bank, full privatization of hand-to-hand currency is consistent with all sorts of monetary policies directed by a central bank.      In my view, having the central bank provide deposit accounts directly to the public or manage a currency at a varying exchange rate is just a bit too rube goldberg.    Of course, I admit that any government dominated scheme, no matter how convoluted, is more politically feasible than complete privatization of hand-to-hand currency.

Friday, February 7, 2014

Kimball on Ending Recession and Inflation

Kimball has an interview on Wonkblog about electronic money and interest rates.    I am not sure where the "ending recessions forever" actually came from, but that seems a bit oversold.   I don't think all recessions are associated with the zero nominal bound.   Surely, there are some supply side recessions.    Allowing short and safe interest rates to fall below zero would at best make potentially deep recessions that drive the nominal interest rate to zero a bit milder.

However, since central banks can expand the quantity of money enough to avoid a deep recession, negative interest rates really allow the central banks to continue with "business as usual."   A central bank can focus on a short and safe interest rate.  Further, the needed increase in the quantity of money would be smaller.  A negative interest rate on reserves will reduce the demand for reserves.   It should be possible to have a "bills only" monetary policy.

Kimball also says that electronic money will end inflation.   Here he depends heavily on the notion that the reason for having trend inflation is to keep nominal interest rates higher on average and reducing the chance of hitting the zero nominal bound.    In other words, again, the "problem" with zero trend inflation is that it interferes with central banks' "business as usual."  That is, focusing on periodic changes in a short and safe interest rate. 

I favor a stable price level on average, but I think that the price level should rise with adverse supply shocks and fall with favorable supply shocks.   Trying to keep the price level fixed would result in deeper recessions with adverse supply shocks and tend to cause booms with favorable supply shocks, even with electronic money.   Negative interest rates to prevent unusually rapid growth in productivity from causing deflation seems like a very bad idea.   

Kimball is very skeptical of suspending currency payments as a solution to the zero nominal bound.   Perhaps the reason I find it less troubling is that I know that free banks in the 18th century had an option clause to allow the suspension of currency payments.   Governments interfered with freedom of contract, and the option clause disappeared.   But in practice, suspensions occurred regularly in the 19th century.   They were just illegal.  

On the other hand, I favor the private issue of hand-to-hand currency.   As long as private currency isn't government insured, the interest rate on central bank reserves and Treasury-bills might be quite negative before anyone decides that bank-issued currency is a better store of wealth.