Monday, July 18, 2011

Negative Real Interest Rates and a Gold Standard

Suppose the natural interest rate is negative.   The level of the real interest rate necessary for saving to equal investment given a level of real income equal to the productive capacity of the economy is less than zero.

Further suppose that the monetary order is based upon gold.   The dollar price of gold is fixed by definition.   The nominal interest rate can be no more negative than the storage cost of gold.    Potential lenders will simply hold onto gold rather than lend at any nominal interest rate lower than that.

How does the economy return to equilibrium?    Assuming the natural interest rate is below the near-zero nominal bound for the nominal interest rate, how can the real market interest rate turn negative?   Or is there some market process that will raise the natural interest rate back up?

Similar to the scenario where there is no gold standard and the real interest rate is negative, the shift from negative yield assets to gold, raises the demand for gold.  The nominal and relative price of gold rises enough to clear the market.

With a gold standard, the relative price of gold can still rise.    Further, it can rise to a point where it is expected to fall at a rate equal to the (negative) real interest rate.   However, it is more conventional to describe this as a decrease in the price level to a point so low that it is expected  to rise again from that low level.    The rate of recovery in the price level implies expected inflation.   This expected inflation reduces the real market interest rate.    This process can make the real market interest rate sufficiently negative so that it keeps saving and investment equal given a level of real output equal to the productive capacity of the economy.   Given this negative real interest rate, real consumption and real investment together add up to enough real expenditure to purchase everything firms can produce.

This process requires that there is some expectation of  a "normal" price level.    The decrease in the current price level must be understood to be temporary, and so the lower the price level falls, the more rapid or more persistent the expectation of future inflation.  (If the nominal price of gold were free to rise, market clearing based upon expectations of  future declines in the price of gold similarly requires an expectation of the long run market clearing price of gold.)

One important characteristic of the gold standard involves risk.    Only in the very special  situation where "the" natural interest rate is negative and the price path of gold is certain does the price of gold immediately rise and then fall along that path at the natural interest rate.  In a more realistic scenario, where it is only safe and short financial assets that have market clearing yields that are negative, and gold is a speculative commodity with a highly uncertain price, then the price of gold may rise to a higher level, but then rise at a slower rate resulting in a slightly lower real yield, making it less competitive with other risky investments.

With a gold standard, the relative price of gold may be subject to uncertain fluctuations, but leaving aside bonds indexed to the price level, short and safe financial assets (government bonds, government insured deposits) will be subject to the same uncertainties.   An unexpected inflation will impose losses on those holding government bonds tied to gold just as it imposes losses on those holding gold.   So, rather than gold appreciating (or depreciating) like risky financial assets, it must perform like the least risky assets, and so, by assumption, depreciate at the negative natural interest rate that applies to short and safe assets.

The process by which the price level falls to a level such that it is expected to rise again to some kind of long run equilibrium value of the price level is temporary by its very nature.    If the natural interest rate is temporarily negative, then this process creates a sufficiently negative real market interest rate to return to equilibrium.   If the problem is some sort of perverse expectations, so that expectations of low real income and output in the future lead to increased saving and reduced investment, and so a perversely low natural interest rate, then this process should solve the problem.   The price level falls to a point where its expected future increases generate the inflation necessary to push the real market rate down to that natural rate.   Real income and output recover, the perverse expectations are falsified, saving and investment and the natural interest rate return to normal.

However, if there is some more persistent problem, where even with expectations of real income and output consistent with reality, preferences and realistic investment opportunities result in a negative natural interest rate, this process is more than a bit paradoxical.   The price level must fall to a point where it is expected to rise at a rate equal to the negative natural interest rate forever?  In other words, in secular stagnation scenarios, a sharp deflation with the expectation that the price level will sooner or later rise back to normal doesn't really help.

Fortunately, there is a second process at work that increases the natural interest rate.    If there is no gold standard, and negative real interest rates increase the demand for gold, the resulting increase in the price of gold results in capital gains for those already owning gold.   This increase in wealth results in an increase in consumption, which is equivalent to a decrease in saving.    The decrease in the supply of saving raises the natural interest rate.    With the natural interest rate below zero by assumption, it becomes less negative.  At a sufficiently high price of gold, the natural interest rate rises above the storage cost of gold, and the economy returns to equilibrium.

With a gold standard, this is called the "Pigou effect."    At a lower price level, the real value of gold holdings rise.   Gold being wealth, this is an increase in real wealth.   As wealth rises, consumption rises as well.   That is a decrease in the supply of saving.   The natural interest rate rises.  At a sufficiently low price level, the natural interest rate will rise enough so that it is no longer lower than the cost of storing gold.   The real market interest rate matches the natural interest rate, saving equals investment, and real expenditure equals the productive capacity of the economy.  

Sometimes the "Pigou effect" is confused with the real balance effect.    They are related.   At a lower price level, real money balances are higher, and when real money balances are greater than the demand to hold them, then the excess money balances will be spent on something, resulting either directly or indirectly in an expansion in the real demand for output.  

However, the "real balance effect" includes the possibility that excess money balances will be spent on financial assets, like bonds, raising their prices and reducing their yields.   The lower real interest rate, then, results in an expansion in real expenditure--either consumption, investment or both.  

The "Pigou effect," strictly speaking, is a different pathway that works through increased real wealth, increased consumption, and reduced saving rather than lower interest rates.    And while that effect should exist given the proper institutional framework, (some money is of the outside variety and not a debt to anyone as is gold with a gold standard,) it is of special interest when nominal interest rates have hit the zero nominal bound, so any "excess" real balances spent on financial assets cannot reduce their nominal yields.

Interestingly enough, if "the" interest rate assumption is loosened, so that it is only short and safe financial assets that have reached the zero nominal bound, and longer and riskier financial assets still have positive nominal yields, then the real balance effect can still impact those yields and so result in increased consumption and investment.   Considering a lower price level, a constant real quantity of financial assets is a lower nominal quantity.   At a lower price level, some of the excess real money balances (or, for that matter, the real holdings of zero nominal yield short and safe assets) should be shifted to longer and riskier assets, lowering their real yields.   The nominal quantity of those securities then, would fall by less than the price level, resulting in a higher real quantity, funding increased real purchases of consumer and capital goods.

Outside of a gold standard, the parallel process would be that at a sufficiently high price of gold, some of those holding gold will switch to other risky financial assets--say equities or long term corporate bonds.   This reduces their expected yields and so provide funding opportunities for firms or households to purchase capital or consumer goods.

And finally, without a gold standard, a higher demand for gold and price of gold stimulates gold mining.   With a gold standard, the higher demand for gold leads to a higher relative price of gold and a lower price level.   This includes the prices of resources like labor.    Given lower wages and lower prices of mining equipment, gold mining becomes more profitable and expands.

With a gold standard, any increase in the relative price of gold involves a lower price level.   Rather than specialized gold traders buying and selling gold, perhaps on an organized exchange, people specialized in producing and  selling other products must add to all of their peculiar concerns specific to their markets a further concern--what is happening with the supply and demand for gold.   And while demands for dental work or jewelry or technology in the mining industry would all be included, the emphasis in this post has been "macro" concerns.   How much does the relative price of gold need to rise (the price level need to fall) so that expected future depreciation (inflation) makes holding gold no more attractive than holding other sorts of financials assets?     How much must the price level fall to expand real wealth enough to expand consumption or the demand for risky financial assets enough to bring recovery?

And, of course, leaving aside indexing contracts for inflation, all of these changes in the relative price of gold--in the price level--disrupt all existing contracts.   Unless anticipated, deflation shifts wealth from debtor to creditor, and inflation does the opposite.   All contracts become partly speculations on the supply and demand for gold..

However, there is a further complication.    All the analysis above assumes that the gold standard will be maintained continuously.    Until restrictions were imposed by the government, free banks generally had suspension clauses that allowed them to stop redeeming in gold when necessary.   Further, in the U.S., anyway, it was not at all uncommon that the ban on suspension clauses was ignored, and banks suspended gold payments anyway.    Of course, "necessary" from the point of view of the banks was when there was a large increase in the demand for gold.    In other words, what would happen if the natural interest rate turns negative.

How does the suspension of gold redeemability impact a scenario with a negative natural interest rate?

Sunday, July 17, 2011

Negative Interest Rates and Gold

I have long advocated privatized hand-to-hand currency, and over the last few years, I have emphasized one benefit--to allow for negative nominal interest rates.   While this might be necessary to allow for low, but positive, real interest rates in the face of an expected deflation, I have generally been concerned with the situation where negative real interest rates are necessary.    By necessary, I mean needed to coordinate saving and investment at levels of real output and income consistent with the productive capacity of the economy.

Similarly, while it might be conceivable that all real interest rates might need to be negative into the indefinite future, I have yet to worry much about such a scenario, being much more concerned by situations where the real yields on short and safe assets need to be negative for a limited period of time.    In other words, I am generally concerned with a scenario where market clearing involves positive real yields on risky and long-term to maturity assets (BAA corporate notes or 30 year Treasuries) but negative real yields on six month T-bills or FDIC insured checkable deposits.

Perhaps most advocates of privatized currency favor some version of the gold standard.  (Silver or some other precious metal would do as well, of course.)     While such a  system would provide many of the benefits of privatized hand-to-hand currency, it would not allow for negative nominal interest rates.    As nominal interest rates fall, at some point, storing gold (or whatever metal is used) would become less costly than lending.   The zero (or near zero) nominal bound would apply.

With a managed fiat currency, one solution to a zero nominal bound is to generate sufficiently high expectations of inflation.   At the  zero nominal bound, expectations of rising prices reduces the real interest rate consistent with a zero nominal interest rate to whatever level is necessary.    If the shortest, safest asset needs a -5% real interest rate, then a zero nominal interest rate and 5% expected inflation will clear markets.   The yields on longer and riskier assets can have positive nominal rates consistent with the appropriately higher real rates.

If monetary policy is based upon some kind of inflation target, keeping expected inflation rate at 2%, for example, then this policy is not possible.   The zero nominal bound translates into a -2% real bound.

But leave those issues aside.   Suppose that privatized hand-to-hand currency breaks the zero nominal bound for interest rates or else, expectations of future inflation can be freely manipulated so that a zero nominal interest rate is consistent with negative real interest rates.

What happens to the gold market when real interest rates on financial assets become more negative?

Certainly, it seems likely that the demand for gold would rise.

Outside of a gold standard, the increase in the demand for gold would result in a higher nominal and real price of gold.   Basic supply and demand would suggest that the quantity of gold supplied would rise and the quantity of gold demanded would fall.   The market for gold would clear.

Why would the market for gold clear?    First, consider a processes narrowly tied to the relative price of gold.

Suppose there is "the" real interest rate and it is negative.   The future nominal and relative price of gold is certain. .

If the relative price of gold is expected to be stable or rise, and the real interest rate remains negative, then the demand for gold would rise, raising its current price.   If the relative price of gold is expected to fall, but at a rate less than the real interest rate, then the demand for gold must rise, and so the price rises further.   If the price of gold is expected to fall faster than the rate of interest, then holding gold is less attractive than holding financial assets, and the demand for and price of gold will fall.  

This suggests a simple process leading to equilibrium.   The price of gold rapidly to a point sufficiently high that its future rate of price decrease is equal to negative real rate of interest.     The price of gold rises to a point where the real rate of return on holding gold is no better or worse than holding financial assets.

In a more realistic scenario, where the real interest rate on short and safe financial assets is negative, and risky financial assets have positive real yields, and further, the nominal and real price of gold is uncertain, then a decrease in real interest rates on short and safe assets should raise the nominal and relative price of gold, but its expected price should be expected to continue to increase at a rate equivalent to financial assets with similar risk characteristics.

More exactly, the assumption would be that a lower (more negative) real rate on short and safe assets causes a substitution into longer and more risky assets, which would include gold.   Its price would rise until future expected appreciation matches the return on similar assets.

A second sort of process involves changes in the real interest rate that is needed to clear markets.   Again, the lower real interest rates cause some of those who would have held financial assets to instead buy gold, increasing demand, and raising its price.   Those who held that gold earn a capital gain and are wealthier.   Because of their greater wealth, they increase consumption.   In other words, they reduce saving.   The decrease in saving raises the natural interest rate.

In other words, this is a process by which a decrease in the real interest rate (perhaps to negative levels) reduces the quantity of saving supplied, and matches it to the quantity of investment demanded.   Assuming that some market force or policy rule adjusts the market real rate to the "natural" or market clearing level, this process is one that also reduces the quantity of gold demanded to the quantity supplied as the relative price of gold rises.

Finally, there is the impact on the gold mining industry--the quantity of gold supplied.     Thinking in terms of saving supply and investment demand, it is a bit difficult to characterize the impact on the natural interest rate.   The gold being mined out today will be available for future industrial purposes.   Is it investment?    On the other hand, if people had sold off bonds to pay to have their furniture covered with gold leaf, it would look like consumption and reduced saving.   Looking at the relationship between the "IS" curve and potential output, this looks to be a rightward shift anyway--added demand for current output.

Of course, the increase in the quantity of gold supplied helps clear the market for gold, but there is also some small effect on the interest rate needed to clear markets.   This demand for new gold production is a demand for real output.

And so, looking at the gold market, when nominal and real interest rates fall, the demand for gold may rise.   And the result will be a higher nominal and real price of gold.   This reduces the expected real yield on gold, but it may also tend to increase the real interest rate that clears markets.    It is even conceivable that these effects could keep the real (and ignoring expected deflation,) the nominal interest rate above zero.

One final note--if some investors observe a rising price of gold, and project past price increases into the future, a speculative bubble could develop.    Clever investors should sell short into that bubble, and stop it.   But, some speculators might buy into the bubble, hoping to sell at the peak, stripping as much wealth as possible from "greater fools."  

Even so, the nominal and relative price of gold adjusts to clear the market for gold.

Now, what happens if the nominal price of gold is fixed because gold services as medium of account?

Tuesday, July 12, 2011

The Debt Limit

I have been puzzled by claims that a binding debt limit  somehow implies default on the existing debt.

Some pundits imply that it is somehow natural to assume that interest on the national debt is funded by additional borrowing.   If no additional borrowing is allowed, then interest cannot be paid.   If interest on the national debt cannot be paid, then there is a default on what is owed.

Other pundits correctly note that interest could be paid out of current revenues, but they sometimes add that principal could be paid out of current revenues as well.   While it is correct that principal could be paid out of current tax revenues, which is how the national debt might one day be decreased, this is little connected to the debt limit--a prohibition on additional borrowing.

The basic outline of the problem is that if no additional borrowing is allowed, then as bonds come due, the actual national debt falls by the amount of principal that must be paid.   That means that funds can be borrowed to repay principal while remaining at the limit.    On the other hand, the interest on the national debt is a current expense and must be paid out of current revenue.

Since current revenue is less than current expenses, which includes interest on the national debt, there is a budget deficit.   To fund that budget deficit, the debt limit must be increased.   If it is not increased, then the budget must be balanced.   Either revenues must be increased, say by tax increases, or else current expenses must be cut.

To repeat, a binding debt limit requires that the budget be balanced.   Current revenues must match current expenses, and that includes the interest expense of any existing debt.

Default on the national debt, in particular, the interest payments owed, would occur only if other current expenses are prioritized over the interest payments.     While some might think this is desirable, it is hardly necessary.

Some pundits have argued that because appropriations from Congress are laws, the money must be spent by the President, and so, the President must borrow the money, regardless of any debt limit.     Perhaps, but municipal governments, for example, also have budgets that are passed by ordinance--that is, by law.   And there are other laws, generally set by the state, that restrict borrowing and taxes.   Further, some expenditures involve contractual obligations that "must" be paid.   Other expenditures may be authorized by ordinance, but there is nothing very specific about the timing.    If tax revenues come in less than expected, and there is no authorization to borrow, then the contractual obligations must be paid (interest and principle on debt, for example,) and then other expenditures must be postponed until more funds are received.

Of course, there is no "higher" authority, other than the Constitution, that limits what Congress may do.   But suppose there was no national debt and Congress authorized no borrowing.   Yes, imagine that the U.S. government operates with no borrowing at all.

Further suppose that Congress also sets up a schedule of customs duties, and generates revenue.    Congress appropriates expenditures, say, building forts to protect the coast from invaders.    The customs duties are expected to bring in $200 million, and forts cost $50 million.   Congress appropriates funds to purchase four forts.

Well, as funds arrive at the Treasury, the President starts having the forts constructed.   Construction only becomes possible as funds arrive.   Unfortunately, only $180 million arrive over the year.   At the end of the year, one of the forts is left incomplete.    Has the President violated the law?

I think it is evident that the President has not violated the law.

Now, suppose Congress had prioritized the forts, and at the end of the year, a lower priority fort is complete and a higher priority fort is incomplete.   Here, the President has violated the law, and impeachment may be appropriate.   Of course, a reasonable defense would be that the President planned to complete all of them, but lack of funds, or some other hold up in construction, resulted in the higher priority fort not being completed.

Only if one assumes that Congress has authorized borrowing does it make sense to say that a lack of tax revenues can never be a reason to restrict expenditures.    So, for example, if Congress has authorized borrowing subject to a limit, and the national debt is below that limit, and the President refuses to spend money because he doesn't favor deficit spending or believes that the increase in the national debt (still below the limit) is already too high, then this would be breaking the law.   But that doesn't mean that limiting expenditures to the funds received once the limit is reached is "illegal."

Personally, I don't think that default on the national debt is the same thing as repudiation.   Default happens all the time, and those that default promise to resume payments as soon as possible.   And I certainly don't think that default on the current national debt involves questioning the repayment of the particular funds borrowed by the Lincoln administration to fund its effort to conquer the Southern states or the pensions promised to the soldiers who invaded those states.    I am pretty sure there are no remaining Civil War pensioners and little of today's national debt was issued for that purpose.

Monday, July 11, 2011

Redeemability without Currency

Worries about the zero bound on nominal interest rates are solely an artifact of basing the monetary system on hand-to-hand currency.    The nominal interest rate on such currency is generally zero.    Why would anyone lend when they can just hold onto their "money?"

In practice, nearly all payments are made by check or electronic equivalent using some kind of deposit account.   These deposit accounts are cleared using central bank deposit accounts (reserve balances at the Fed in the U.S.) or automatic clearinghouses.   The automatic clearinghouses, however, create net clearing balances that are also settled up with central bank reserves.

There is no reason why the bulk of the payments system cannot pay interest, and much of it does.   If interest can be paid, generally, the nominal interest rate can be negative.

Of course, banks are obligated to redeem their deposits with hand-to-hand currency.   Similarly, the Fed redeems balances in reserve deposit accounts with hand-to-hand currency.    It is these redemption obligations that makes a monetary order that is mostly about changes in deposit balances that can potentially pay or "charge" interest into one that is based upon the zero-nominal interest hand-to-hand currency.

Suppose this characteristic of the monetary order is dropped.    Obviously, hand-to-hand currency is convenient for many purposes, but it is hardly essential for most people.  

As long as checks and electronic payments clear through the central bank, each individual bank is limited to issuing the amount of deposits its customers want to hold.    Any individual bank that creates excessive deposits will suffer adverse clearings as the deposits are spent or transferred to some other bank.   Overall monetary equilibrium operates as usual, depending on how the central bank chooses to manipulate the quantity of reserves and the interest rate paid upon them.

However, the system is a bit odd in that those holding deposits in banks would have a debt instrument that is payable in something that they cannot directly access.    From a bank's point of view, it must pay off these debts using its balance in its reserve account at the central bank.  

From the depositor's point of view, all that can be done with the funds is to shift them to another bank.    As long as only banks can hold balances with the central bank, from a depositor's point of view, he or she is owed something that is only directly payable into something that he or she cannot directly hold.

One way to avoid this difficulty is to make deposits redeemable into something other than hand-to-hand currency.     For example, suppose deposits are redeemable into T-bills at their current market price as well as reserve balances at the central bank.    An individual depositor can "withdraw" T-bills.   Funds are then removed entirely from the banking system.

Since T-bills are not priced in terms of the deposits of an individual bank, but rather in terms of the deposits of the banking system, all of which are tied to the balances of the central bank, this constrains the individual bank.    While an individual bank could purchase T-bills to meet redemption obligations, this would only be possible if the bank can settle the check or electronic payment it uses at the central bank.

Of course, this "redemption" provides nothing more than what the individual depositor can do anyway.   Any individual depositor can remove funds from the banking system by spending the funds on something.     Making deposits redeemable in T-bills (or gold) at market prices provides no real benefit to the depositor.   A depositor who wants to reduce money holdings can always buy T-bills or gold.   Still, perhaps it will help overcome puzzlement about a debt that is only payable in something than cannot be directly held.

With such a system, where there is no obligation to redeem deposits with hand-to-hand currency, there is nothing keeping the nominal interest rate greater than zero.   There would be no zero nominal bound on interest rates.    (Of course, if market conditions are such that a negative nominal interest rate has inflationary consequences, then avoiding such consequences would require a positive nominal interest rate.)

If the monetary order were changed so that it is no longer based upon hand-to-hand currency, but the central bank continues to provide such hand-to-hand currency, then if nominal interest rates need to be negative, the central bank will face large demands for currency.   Once the interest rate on reserve balances falls below the cost of holding vault cash, banks will demand vault cash.     Similarly, if the interest rate "paid" on deposits becomes more negative than the cost of storing currency in vaults by the general public, the demand for currency by the general public would rise.

In order for the nominal interest rate to become negative, the central bank would have to say "no."   It would have to quit issuing currency.   The result would be somewhat similar to the sorts of suspensions of convertibility that occurred periodically in the 19th century, except that under this institutional framework, there is no obligation to issue hand-to-hand currency.

What would be happening is that the issue of hand-to-hand currency would simply be too costly.    As interest rates on earning assets (such as T-bills) fall, they will eventually be too low to cover the cost of maintaining the currency in circulation.    When nominal interest rates are negative, clearly, issuing zero-nominal yield currency and using it to fund a portfolio of T-bills priced above par is a losing proposition.

If the central bank dropped out of the hand-to-hand currency business, and commercial banks issued hand-to-hand currency under normal conditions, then there would be no "policy" decision regarding currency.   If nominal interest rates fall, and even turned negative, then the profitability of issuing hand-to-hand currency would fall as well.   When nominal interest rates became too low, then banks would quit issuing hand-to-hand currency.   Deposits, on the other hand, perhaps with negative nominal interest rates, would continue to be profitable.   The greater part of the economy, would be able to operate with minimal disruption.

Finally, consider the alternative.     In the privatized scenario, banks could continue to issue hand-to-hand currency by using it to fund riskier, higher yield assets.    Suppose a bank created a poorly capitalized subsidiary and funded junk bonds with hand-to-hand currency.  While the interest rate on safe, short term to maturity securities, like T-bills, might be negative, and safe (perhaps even FDIC insured) deposits might have negative interest rates, currency could still be issued.    Such currency would be a poor store of wealth.   In fact, those receiving it would probably deposit it just about as fast as they received it in payment.

Now, suppose that instead of this privatized alternative, the central bank wants to continue to provide hand-to-hand currency that is perfectly safe from credit risk.    What can it do?    Just like the private banks, it could shift from holding a portfolio of safe, low risk T-bills, to holding a portfolio of riskier assets--say longer term government bonds.    Of course, if the central bank chooses to provide a perfectly safe asset with a zero-nominal yield, then everyone will hold that rather than lend at negative nominal interest rates.   And so, the central bank has created the zero nominal bound.

Reflecting on that situation, maintaining macroeconomic equilibrium could require the central bank to issue large amounts of zero-interest currency and bear substantial risk.   Should it?   Is it the role of government to bear risk for people?

I don't think so.

Sunday, July 10, 2011

New Blog: Uneasy Money

David Glasner has started blogging at Uneasy Money.    David has advocated free banking combined with index futures convertibility.   Rather than proposing to use index futures on the prices of goods and services or else nominal GDP, he proposes using a wage index.

While I am not very familiar with wage indexes, the proposal has some advantages over GDP targeting.    For example, suppose wages are targeted to grow with labor productivity.   If labor's share of income should fall relative to capital, then the trend of money wages would remain unchanged.   Nominal income would grow more rapidly, nominal capital incomes would grow more rapidly, and the price level would rise to a higher growth path.   The inflation in final goods prices would slow the growth in real wages, and reduce them to a lower growth path.

The disadvantage is that because wages are sticky, monetary disequilibrium could impact output and even employment while having little impact on the growth path of wages.   With index futures convertibility, those speculating in futures in a way that corrects the disruption would profit little, while those who were taking positions on the futures in a way that exacerbates the problem would suffer little loss.

With nominal GDP targeting, when monetary disequilibrium impacts both real output and flexible prices, the errors show up in what is being targeted.   Those trading the futures that drive the quantity of money will obtain profits or suffer losses more consistent with the benefits or costs their actions are creating for the economy.

In a recent post, Glasner discussed the Pigou effect, Keynes, Sumner, and Krugman.    With the Pigou effect, a lower price level raises real wealth, reduces the supply of saving, and raises the natural interest rate.   If the natural rate is negative, and the market rate is at zero, then equilibrium returns when the price level is low enough for real wealth to be high enough that saving is low enough that the natural interest rate rises back to the market rate.    (This only works with outside money.  With inside money, a lower price level just transfers wealth from debtor to creditor.   Of course, in a  gold standard world, there is outside money.)

Glasner wonders how different economics would have been if Pigou had instead suggested that the inflation rate rate increase enough so that the real market rate fall enough to match the perhaps negative natural interest rate.   I am not sure whose "effect" this is, but if the price level falls below its expected value in the long run, then the expected inflation rate will reduce the real market rate, perhaps making it sufficiently negative to equal the natural rate.

If the monetary system has no anchor for the price level, for example, due to inflation targeting, this doesn't work.   To the degree people project past deflation into the future, moving to a lower price level ends up having perverse and perhaps catastrophic consequences.    With a gold standard, the price level is anchored, more or less.

The flavor of Glasner's comment, however, is more consistent with a policy aimed at causing more rapid inflation from the current price level--shifting the price level to a steeper growth path.    The price level doesn't need to fall.   Prices begin to rise more rapidly, and the real market rate falls to the perhaps negative natural interest rate.    (In the context of an economy that already has suffered a large deflation, this could just be reflating the price level  to some past level.   He is discussing Keynes and Pigou.)

For those of us favoring a rules-based monetary regime, whether targeting a growth path for wages, prices, or nominal GDP, the anchor for the price level means that deflation can be brought to a halt as long as people have confidence that the regime will eventually bring the price level back to the explicit or implicit target.    This is especially true if the problem is that experience of deflation causes people to expect more deflation so that a zero nominal interest rate becomes a positive real interest rate greater than the natural interest rate.

But suppose the problem isn't expected deflation.   If the price level falls to a lower level, expectations are that it will stay there or perhaps rise back to a normal level.   No, suppose the problem is that the natural interest rate is negative.

If the problem is just self-fulfilling expectations--high saving and low investment because of poor economic performance is projected into the future--then the same process should work.   As the real market interest rate falls to the negative natural interest rate, the economy recovers, and so, there is little reason to expect poor economic performance in the future.   The natural interest rate rises again.

But suppose the problem is more persistent.   Suppose the natural interest rate is negative, even when economic performance is good.   People want to save a lot, and the investment opportunities, even in the context of full employment, appear somewhat poor on the margin.

With a rule-based monetary order, having the experts adjust the inflation rate to clear markets seems counterproductive.   It would be possible to make the rule inflationary, so that a zero nominal interest rate will imply a sufficiently negative real rate that it could never be below the natural rate.     I favor price level stability, and so would like to avoid a rule mandating a rising price level because something might happen.

In that situation, if the of the nominal market interest rate needed to equal the natural interest rate is negative, either because of expectations of deflation or a negative real natural interest rate, then the "problem" is that the issue of hand-to-hand currency becomes unprofitable.

What does it mean that the issue of hand-to-hand currency is unprofitable?   The nominal interest rates on the assets purchased by the issuer of the hand-to-hand currency is not sufficiently greater than zero to cover the cost of intermediation.

In the context of free banking, the answer is simple.   Don't make zero-nominal-interest hand-to-hand currency central to the monetary order.  Base everything on interest bearing deposits, and make the issue of currency a minor convenience.   And, if no bank wants to issue it, then that convenience will not be available.

Do we set a permanently higher inflation rate so that we can be assured that hand-to-hand currency can always be issued at a profit?   Or, do we deflate all prices and incomes so that the real value of hand-to-hand currency adjusts to a given quantity?   Perhaps it is just better to avoid having an entire economy held hostage to hand-to-hand currency.     If the issue of hand-to-hand currency becomes unprofitable, then let those who would use it find alternatives.