Saturday, December 28, 2013

Interest Rates and Investment

Scott Sumner recently responded to a question from a commenter.  Dustin asked:

"An elementary question on the topic of interest rates that I’ve been unable to resolve via google:   Regarding Fed actions, I understand that reduced interest rates are thought to be expansionary because the resulting decrease in cost of capital induces greater investment. But I also understand that reduced interest rates are thought to be contractionary because the resulting decrease in opportunity cost of holding money increases demand for money.

One? Neither? Both? Little of each? Depends?"

Sumner claimed that it was difficult to answer.   He doubted whether most macroeconomists could give a sensible answer.   He made a remarkable claim:

"It’s not at all clear that lower interest rates boost investment (never reason from a price change.)  And even if they did boost investment it is not at all clear that they would boost GDP."

(On the other hand, he immediately followed that with the statement that open market purchases reduce short term nominal interest rates and boost nominal GDP.)

David Glasner found Sumner's response puzzling and accused Sumner of confusing a change in demand with a change in quantity demanded. 

The solution to that puzzle is simple.   Sumner has recently been drawing supply and demand diagrams for base money with 1/P on the vertical axis. Patinkin did the same ages ago.  A change in the interest rate shifts that demand curve. A change in the price level is a movement along the curve.

And so, if  we consider an exogenous decrease in the interest rate, the demand for money curve shifts to the right.  The new equilibrium price level is lower.   Given real output, that implies lower nominal GDP.

Most economists who teach introductory macro (as opposed to Macroeconomists, I suppose,) would deal with the question easily.   I would use it as an opportunity to discuss the difference between a change in demand and a change in quantity demanded using a Keynesian money market diagram–with the interest rate on the vertical axis.  In other words, I would answer the question much like Glasner.

If the quantity of money is given, and starting at equilibrium, a lower interest rate leads to a shortage of money. Those short on money sell securities, forcing the interest rate back up. The lower interest rate that might raise investment and nominal GDP is impossible. Unless of course, the quantity of money increased.

Now, consider a more classical framework. An increase in the supply of saving results in a lower interest rate.  At the old interest rate, desired saving is greater than desired investment.   The lower interest rate causes the quantity of saving supplied to decrease and the quantity of investment demanded to increase, making them again equal.  There is no direct impact on nominal GDP.   Nominal and real consumption decrease and nominal and real investment increase.  

However, this lower interest rate that coordinates saving and investment also raises the demand to hold money.   Sumner's (and Patinkin's) curve shifts right.  If the price level doesn't fall immediately, due to sticky prices (including wages,) the economy goes into recession. That could very well reduce investment (and saving, for that matter.) It also reduces the demand for money. (Sumner’s and Patinkin's curve shifts back left due to lower real income.)
In the long run, the price level falls. Real output recovers, and in the end, the interest rate is lower, real saving and real investment are both higher, but nominal GDP is lower.   Real and nominal consumption are both lower, but nominal investment is ambiguous.

Rather than thinking about a coordinating change in interest rates, we could think of interest rates being lowered through a price ceiling. The quantity of saving supplied decreases and quantity of investment demanded increases. The short side prevails. People are saving less and consuming more. Investment falls to match the quantity of saving supplied. No change in nominal GDP.   Real and nominal consumption are higher.   Real and nominal investment are lower.

The demand for money increases as before   The Sumner/Patinkin demand curve shifts right.  Some of the frustrated savers just accumulate money balances since they no longer find it attractive to fund investment by purchasing new debt and equity.    The economy goes into recession. This reduces real output and desired saving and investment. But, in the long run, the price level falls.

Given the price ceiling on “the” interest rate, we end up with more real consumption and less real investment than the start.   Nominal GDP is lower.   Nominal investment is lower, but nominal consumption is ambiguous.

But, of course, Sumner's real point is that an expansionary monetary policy involves an increase in the quantity of money.   It is not simply a lower interest rate generated by households saving more or tighter regulations on usury.

Is a lower interest rate expansionary?   As Yeager always would say, it depends on what caused it. 

Tuesday, December 24, 2013

How Important Is Shadow Banking?

During the great Williamson internet debate on purportedly deflationary quantitative easing, Nick Rowe mentioned in passing that he didn't understand Izabella Kaminska.   Scott Sumner wrote a post about how macroeconomists have views so different that they are in effect speaking different languages.   He linked to Nick's remark, using Kaminska as an example of someone doing macro that he doesn't understand.  

How was Kaminska relevant?   She has also argued that quantitative easing is deflationary, though not for the same reasons as Williamson.   It rather has to do with the Fed purchasing safe collateral, which interferes with the operation of the shadow banking system.   

Market Monetarists frame quantitative easing as a means of increasing the growth path of nominal GDP.     I think it is fair to say that all Market Monetarists believe that the level of nominal GDP could reach any target without there being a shadow banking system.   And further, that the real economy could survive without a shadow bank system.

On the other hand, I suspect there is more controversy as to whether the shadow banking system adds any value.   I, at least, suspect that it is at best a work around various undesirable banking regulations.   An appropriate deregulation of banking would result in the end of shadow banking and an enhancement of social welfare.  At worst, shadow banking is a net loss to the economy.

Of course, there are many people whose private interests are closely tied to shadow banking.   Kaminska covers them, and I think she may be too influenced by their special pleading.

Still, I think there is a fundamental intellectual error, very common among noneconomists.    It is the distinction between the nominal and the real.    A high level of nominal GDP (or price level) may imply a large nominal credit sector, but is  quite consistent with a small real credit sector.    Failure to make this distinction is sometimes called the money illusion.  That is the significance of Scott Sumner's claim that Zimbabwe proves that it is always possible for a central bank to generate inflation.   More to the point, it is always possible for a central bank to generate a higher level of nominal GDP.   Having a shadow banking system operating as it did in 2006 might be necessary for a full recovery of the real incomes of shadow bankers, but it is not necessary to return nominal GDP to the growth path of the Great Moderation.

Suppose all payments were made with sacks of gold coins.   The nominal interest rate on the coins is zero (and there are heavy storage costs.)   The government has a national debt and funds part of it with treasury bills.    Financial firms and other large businesses begin to hold treasury bills rather than sacks of gold.   The bills coming due every week or so provide extra cash receipts to cover needed expenditures.  

From the point of view of the firms holding the bills, they are saving a bit on their storage costs.   They earn interest on the T-bills rather than nothing on the sacks of gold.   Assuming the government has a modest national debt and is fiscally responsible, there is little credit risk.   And since the bills are claims to gold, the real risk is approximately the same as holding the sacks of gold.

From the point of view of society, the reduction in the demand for gold results in a slightly higher price level and some transitional inflation.   Some resources that would have been used to mine gold instead produce other consumer and capital goods.   And some of the gold that made up the coins now provides services in circuitry, dentistry, and as lovely jewelry.  

The government  will also benefit by funding the national debt at a slightly lower interest rate.   This will allow for slightly more government services or, better yet, lower taxes.   Since all plausible taxes distort, this further enhances welfare.

Private and social benefits.     Of course, those who are unable to economize on their cash balances suffer a capital loss due to the transitory inflation.   It isn't all benefit and no cost.

The benefits of using Treasury bills as a substitute for money is greatly enhanced by the development of a secondary market.   Rather than only purchasing new issues and waiting until the bills mature, cash balances can be used to purchase the T-bills and when cash is needed, the T-bills can be sold.   Of course, each firm using this technique bears transactions costs--they hire traders who are soon picking maturities and even buying and selling to try to profit from slight moves in the prices of the T-bills.   There is a bid-ask spread, which represents the transactions costs for those making the market.    These costs are traded off against the cost of storing the gold coins and the opportunity cost of the interest foregone.

But the private benefits and the social benefits are at least loosely aligned.

Suppose that a sudden massive loss in trust in the market makers causes the use of Treasury bills as money substitute to be much less attractive.    The traders on the secondary market would face disaster.     Instead of constant transactions as payments are received and then used to purchase T-bills and then sales of T-bills to fund purchases, there would be little activity.   The market makers would obviously suffer.  Some of those who had been managing their employers cash balances could even be laid off!

However, there would also be adverse macroeconomic consequences.   The demand for gold coins would expand.  If wages, and perhaps other prices, were sticky, then the increase in the demand for gold coins would result in lower output and employment.   And while a sufficiently lower price level would result in an adequate real quantity of gold coins, the increase in the real burden of private debt would be disruptive.   This would include a higher real government debt.  Not only must the government fund a higher real debt, it must do so without being able to sell T-bills as a substitute for money.  This would require a reduction in the provision of government services or higher distortionary taxes.

The private disaster faced by those trading the T-bills would be matched by a social disaster.   What can be done to recover confidence in the traders, allow them regain their livelihoods, and at the same time, allow aggregate demand and real output to recover?   The recovery of prices would alleviate the pressure on private debtors as well as reverse the increase in real government debt.

Leaving aside the gold, the trading off of interest and transactions costs was worked out by Tobin years ago.   Something that all monetary economists learn early on.

Now, suppose that rather than sacks of gold coins, government-issued paper currency is used as money.   Because a variety of denominations can be issued, including very large denominations, storage costs are insignificant.  Instead, it is simply a matter of trading off transactions costs and interest.   Otherwise, the logic of holding T-bills rather than paper currency is little different from sacks of gold coins from the private perspective.  

But the social point of view is much different.   The paper money is cheap to print and has no other use than money.   The ability to reduce balances of paper currency by trading T-bills provides no social benefit.   If the quantity of paper currency is assumed to track that of the gold coins, then there would be real capital losses due to inflation for those who cannot economize on their holdings of paper currency.  

But that is not necessary.   Instead, the quantity of currency can be reduced to maintain its purchasing power.   In other words, as the demand to hold paper currency falls, the central bank/treasury can reduce the quantity of paper currency   To do so, the central bank would make open market sales.   Alternatively, the treasury funds more of the national debt with T- bills and less with paper currency.

Once these adjustments are made, then there are no social benefits at all.   The firms using T-bills as a money substitute have employees to manage their trades.   There are firms making markets in the T-bills.   While all of these transactions costs are offset by private benefits--reduction in interest forgone, there are no social benefits at all.   All those financial traders are being pulled away from the production of useful goods and services.   There are no savings on the national debt.   Rather than funding it at zero interest with currency, it is necessary to pay interest on the T-bills.   This implies fewer government services or somewhat higher distortionary taxation.

Now, let us suppose that there is a loss of confidence in the market makers and there is much less willingness to use T-bills as a money substitute.   The demand for paper currency rises.   If the quantity remained fixed, or grew too little, the various social calamities described above would still occur.  However, it is possible to increase to the quantity of paper money as much as needed.   The central bank purchases T-bills.   The T-bills no longer held as money substitutes are now being held by the central bank.   Those who would have held the T-bills are instead holding paper currency.   

Unfortunately for those who traded the T-bills, they still suffer a private calamity.    And all of those now holding currency can look back at the interest they had earned when they could trust the T-bill traders.   A journalist who had been covering them might be quite focused on their private disaster.   Open market purchases?  What good will that do them?    The T-bills that they would have been trading over and over would just sit idle at the central bank (or be destroyed by the treasury.)

But there is more liquidity than before.   And if the quantity of money rises to meet the demand, there is no need for a reduction in spending on output or prices. Further, social costs are reduced.  The government is funding more of the national debt with zero-interest currency.   There is no need to mine the paper money out of the ground or pull it out of alternative uses like jewelry.   And, all of those T-bill traders can find some more productive work.

Of course, in the real world, bank deposits play the role of paper currency for most purposes.   And a loss in confidence in those trading T-bills was not a major factor in 2008.   The demand for T-bills has expanded greatly.   It was rather the more exotic money market instruments cooked up in an effort to enhance the yield earned by those managing cash balances that flopped.    But there is no gold standard.    It is not at all clear that shifting business from the conventional banking system to the shadow banking system provides any social benefit.    On the contrary, it is quite possible that shifting back the other way was on net beneficial.   But just not to those in the shadow banking industry.

Sunday, December 15, 2013

Private Money and "Quantitative Easing."

Suppose an single bank in a competitive banking system wanted to expand its lending, perhaps by holding a larger portfolio of bonds.    It would need to obtain funding.   In perfect competition, it would pay the going interest rate on deposits.   It would expand its balance sheet until the marginal cost of providing intermediation services equals to interest margin.

However, I never think in terms of perfect competition.   With imperfect competition, an individual bank expanding its balance sheet would both charge less on loans and pay more on deposits.   The interest rate margin is driven down as the marginal cost of providing intermediation services rises.   When marginal revenue equals marginal cost, profits are maximized.

To me, Williamson's account of quantitative easing is similar.   A central bank wanting to expand its portfolio of long term bonds, perhaps because it wants to drive down the yield on them, must provide a higher yield on its liabilities.   Of course, a single bank is mostly seeking to gain market share at the expense of other banks offering very similar liabilities.   A central bank in a closed economy is competing with other sorts of assets, and so the liquidity premium on its liabilities are central.

What about deflation?

Consider Hayek's model of  banks issuing private currency.    Unlike a conventional banking system where each bank accepts the other banks' monetary liabilities for deposit at par and settle net clearing balances, Hayek's system had each bank's monetary liabilities float.   He also emphasized a scenario where all the banks seek to stabilize the purchasing power of their monetary liabilities.   In other words, the price level in terms of each bank's money would be  stable.    Perhaps because the private and competitive issue of deposits is conventional, Hayek at least seemed to emphasize the issue of banknotes--hand-to-hand currency.   Still, given this picture of the monetary order, any bank could expand it lending--undertake quantitative easing--by paying a higher nominal interest rate on deposits, just as would occur in a more conventional banking system.

Interestingly, Benjamin Klein described the same institutional order as Hayek, but he emphasized the return each bank would provide on hand-to-hand currency by generating inflation or deflation.   From that perspective, a bank wanting to expand its asset portfolio would need to provide a increased yield on its currency.     A bank would need to lower the inflation rate.   In other words, a bank wanting to expand its asset portfolio would generate a higher deflation rate for the currency it issues.

How could a bank actually implement the policy?   Obviously, it would announce and advertise its new policy, generate added demand for its liabilities, and purchase the assets it wants.    Of course, it would need to adjust its policy to generate the promised rate of deflation.   

Consider a bank paying an increased nominal interest rate on deposits.   We could imagine the bank just adding funds to deposit accounts.   It would hope that its customers would notice this occurrence, and come to expect a higher nominal interest rate.   But what bank would do that?   They tell their customers that they are paying a higher interest rate and would advertise what they are doing.

Consider then a central bank.   If it wanted to expand the size of its balance sheet, and it didn't want to pay a higher nominal interest rate on its reserve deposits, it could instead promise a lower inflation rate.    And how would it do that?   Step one would be to promise lower inflation or deflation.

Now, imagine a single bank in a competitive banking system that simply purchases the assets it wants and does nothing to provide funding.   The result would be adverse net clearings.   It would be compelled to take action to pay off its claims at the clearinghouse.   With a Hayek/Klein regime, the result would not be an automatic appreciation of the currency.   On the contrary, the result would be a depreciation of the currency on the interbank exchange.  

And suppose a central bank creates money and buys assets, keeps the interest rate on reserve balances unchanged or even lowers it.  What happens?    Inflation, not deflation.

Finally, suppose a central bank wants to expand its balance sheet, and it promises lower inflation (even to the point of deflation) to raise money demand, and expands its balance sheet beyond that greater demand to hold money and generates inflation.   Suppose it breaks its word?   This, of course, is exactly the problem that concerned Klein.    Why not undertake an inflationary expropriation?

Larry White insists that this is why banks need to issue redeemable banknotes and deposits.   Inflationary default is prohibited by contract.   And for a central bank, James Buchanan always insisted that this is why constitutional restrictions of the issue of money are necessary.

Does Quantitative Easing Cause Deflation?

Steven Williamson developed a model that seems to imply that quantitative easing results in deflation.    Nick Rowe was very critical.  

As Nick and others have pointed out, Williamson showed the weakness of his framing.   Some real world market phenomenon cannot be usefully described by carefully modeling a long run equilibrium.

Williamson's model has an element of truth.   Suppose a central bank wanted to have a large balance sheet.   And further, it wanted to keep nominal interest rates low.   How could it manage that?    All it would need to do is generate expectations of deflation.   This would raise the real interest rate that could be earned on its liabilities for any given nominal interest rate.  That would increase the demand to hold its liabilities.   And so, the central bank could issue more liabilities and hold a larger portfolio of assets.     Why would it want to do that?   Perhaps it wants to funnel more credit into the housing market than private investors consider wise.

None of this is news to Market Monetarists.   But it has a bit of a  "through the looking glass" aspect, because Market Monetarists are always making these arguments in reverse.   Sumner is constantly arguing that low expected inflation results in a high demand for base money, and it is either explicitly or implicitly considered a bad thing.   In Boom and Bust Banking,  Jeff Hummel's contribution is a long attack on what appears to be an effort by the Fed to expand its balance sheet so that it can allocate credit.     The key complaint is the payment of interest on reserves.

So, Market Monetarists would say that trying to generate a large balance sheet for the central bank should be avoided.   Paying interest on reserve balances tends to raise the demand for central bank reserves, which we recognize means that a higher quantity of base money would be needed for macroeconomic equilibrium.   The inference that Market Monetarists would draw is that the interest rate on reserves should be low, zero, or negative.     We don't see such a low interest rate as valuable in and of itself, and so do not favor generating deflation to make the low interest rate consistent with long run equilibrium.  Quite the contrary, we sometimes predict and perhaps just hope that a recovery of spending on output would allow credit markets to clear at higher real and nominal interest rates, including any interest rate paid on reserves.   If anything, the thrust of Market Monetarist arguments is that one reason to target a more rapid trend growth rate of nominal GDP is to generate slightly higher nominal interest rates and a smaller equilibrium balance sheet for the central bank.  

Why have Market Monetarists supported quantitative easing?    The reason is to raise spending on output--nominal GDP.    If the current growth path of nominal GDP is counted as an equilibrium, then the point is to generate a disequilibrium--an excess supply of base money-- that will be cleared up through an increase in the demand for nominal base money balances by an increase in spending on output--a higher growth path of prices, a higher growth path of real output, or some of both.  

Nick Rowe finally emphasized this point, after it was pointed to by Robert Murphy.  Williamson has the central bank seeking to have a large balance sheet in real terms, which must be done by making base money more attractive to hold.   Paying higher nominal interest rates on reserves, or given whatever interest rate is paid, generating expectations of lower inflation, increase the real demand to hold base money balances.   To the degree that quantitative easing is supposed to generate a higher growth path of for the price level, the goal is not to raise the central bank's balance sheet in real terms.   The higher nominal quantity of base money is supposed to be held in the long run because the higher growth path of prices reduces real balances back to equilibrium.

However, pretty much all the Market Monetarists have been more and more inclined to see the current level of base money as more than adequate.   (And how is that different than excessive?)   The problem is the target.   Creating large amounts of base money that will be soon removed has little effect on spending on output.     If the Fed wants more spending on output, it needs to target a higher growth path for nominal GDP.    Lower unemployment on the condition that inflation doesn't rise above 2.5% in the medium run just doesn't make it.

Oddly enough, the target is implicit in Williamson's framing too.   If the "target" were for the central bank to have a large "real" balance sheet and keep nominal interest rates low, then it must generate expectations of deflation.      The problem would be with this odd target.

And, of course, there is Williamson's odd notion that this would somehow actually result in goods and services being exchanged at progressively lower prices.