Tuesday, October 30, 2012

Money and the Medium of Account


Scott Sumner has defined money to be the medium of account and has argued that the equation of exchange applies to the medium of account.

I define money to be the media of exchange.  I don't have much use for the equation of exchange, but to the degree it does have value, its value relates to the medium of exchange, not the medium of account.

Sumner argues that whether money is the medium of account or the media of exchange can be best explored by considering hypothetical monetary regimes where the medium of account is different from the media of exchange.

I agree that such regimes are instructive.    I began to study this issue nearly thirty years ago after reading Robert Greenfield and Leland Yeager's 1983 paper, "A Laissez-Faire Approach to Monetary Stability "      They were influenced by Fama's 1980 paper, "Banking in the Theory of Finance."   That is why Greenfield and Yeager included Fama's name in their proposed BFH payments system:  the Black-Fama-Hall payments system.

The Greenfield and Yeager system defines the unit of account in terms of a nearly-comprehensive bundle of goods and services.   In their first paper, the media of exchange were a variety of checkable mutual fund accounts.   Jurg Niehans defined the medium of account as the good that defines the unit of account, so Greenfield and Yeager were proposing that the medium of account be a bundle of goods and services.   There is no notion that people use the bundles to purchase items or even that anyone actually gathers up  all of the items in the bundle and trades them all at once.

The Fama paper comes closer to what Sumner discusses.   Fama suggested a steel ingot as the medium of account.   He describes a variety of scenarios regarding the media of exchange, but one interesting one is an Accounting System of Exchange where payments are made with ordinary stocks and bonds.   No one buys anything with steel ingots.   Those producing them sell them for media of exchange and people needing to use steel purchase it with media of exchange.

The price level is the inverse of the relative price of the medium of account.   In Fama's scenario, that would be steel.   The relative price of steel depends on supply and demand, as usual.    If there is an increase in the demand for steel, perhaps because a new use for steel is discovered, its relative price rises.  This implies a decrease in the prices of all other goods and services.   If, on the other hand, there is some technological improvement in steel manufacturing, its supply increases, its relative price falls, which implies an increase in the prices of all other goods and services.

How does that work exactly?   There is an increase in the demand for steel, and a shortage of steel at its current, fixed by definition, price.   Why does everyone else in the economy respond to a shortage of steel by lowering the prices they quote for other goods and services?   Why should they care about conditions in the steel market?

Now, if there were a Walrasian auctioneer, he would note that there is a shortage of steel.   Because steel is the numeraire, he cannot raise the price of steel.   He instead calls out lower prices for all other goods and services.   And so, the reason why other prices fall is simple.  The Walrasian auctioneer lowers them.

But there is no Walrasian auctioneer in reality.   In the real world, why does everyone else lower their prices because there is a shortage of steel?    In truth, the most likely scenario is that those actually quoting steel prices would raise the ingot price of an ingot.   That would mean that the definition of the unit of account in terms of the medium of account would be ineffective, and the "steel ingot" would be an abstract unit of account not tied to anything.

With the Greenfield and Yeager system, this is a much more serious problem.   We can at least imagine that those buying and selling steel would insist that a steel ingot is a steel ingot and so the price is fixed by definition.   When a bundle of goods and services is used as the medium of account, the price of each item in the bundle is free to vary.   It is just that the prices of the items must sum up to the defined total.   If the relative price of the bundle as a whole rises, then the prices of other goods must drop.   But why would any of those actually setting the prices of bundle items make sure that they total up to the defined amount, much less  those producing other goods adjust their prices to make that total consistent with general equilibrium?

If there were a Walrasian Auctioneer, there would be no problem.   While making sure that the prices he calls out for the bundle items add up the the correct total is more complicated than keeping one price fixed, it isn't too difficult.  And if the relative price of the bundle rises, the Walrasian auctioneer just calls out lower prices for the other items.

In my view, Fama never solved the problem.   Like usual, he just assumed market clearing.   And it wasn't like he was proposing the steel ingot as a medium of account in reality.   But, like usual, his equilibrium always perspective blinded him to important aspects of the economy.

Greenfield and Yeager, on the other hand, did solve the problem by examining how the media of exchange must be tied to the medium of account.   A "dollar" claim in the payments system would be settled with some "settlement media" that had an actual market price equal to the sum of the actual market prices of the items in the bundle.   This would create powerful market forces--monetary forces--that would keep the sum of the prices of the items in the bundle at the defined amount and cause all other prices, including wages, to adjust as well.   If the relative price of an item in the bundle rose, then all other prices in the economy would fall enough so that the sum of the prices of bundle items would equal the defined amount.

The solution for Fama's scheme is more or less the same.   But what it amounts to is thatthe  Accounting System of Exchange must set up procedures to make sure that the steel market clears at a price of one ingot of steel per ingot of steel.

Sumner imagines a system where Zimbabwe is on a gold standard.   Gold serves as medium of account.   All prices and wages and contracts are in terms of gold.  However, the people use Zimbabwe currency as the medium of exchange.    The amount of currency needed to make a payment is determined by dividing the amount of the payment in gold units by the gold price of currency.   Sumner supposes that there would be a sign by each cash register showing the gold price of Zimbabwean currency.

Sumner supposes that the Asians demand more gold, and this leads to deflation in Zimbabwe.  The demand for the medium of account has risen, its equilibrium relative price rises, and this implies a lower equilibrium price level.    Sumner just assumes that the price level drops.   Just like there is a Walrasian auctioneer noting that at the old relative price of gold, the old price level, there is a shortage of gold.   And so, new prices must be called out to everyone supplying and demanding the other goods in Zimbabwe.  Sumner supposes that wages, quoted in gold, are sticky, so the lower price level raises real wages and so employment falls.    (Or maybe it is that the lower price level, multiplied by the given level of real output makes up a smaller number.  When that is divided by the given wages, that implies less hours worked.)

What is it exactly that motivates people selling bread, gasoline, diamonds or anything else to lower their gold prices?   Is it because they become more anxious to sell these goods?   Surely the more plausible explanation is that people in Zimbabwe reduce their purchases of bread, gasoline, diamonds and other goods at the currently quoted prices.   Why?

Now, the growing Asian demand for gold could strip the Zimbabwean gold shops of their stock.  There is a shortage of gold at the fixed by definition price of gold.   Zimbabweans going to the shops to get gold for dental work or jewelry find the shelves empty.  

How does that cause Zimbabweans to reduce their purchases of bread, gasoline, diamonds, and the like?   Why would firms be more anxious to sell bread, gasoline, or diamonds?    Why would those setting the prices of those goods lower them?

Sumner seems to think that the experience of the thirties is relevant.   In the thirties, money was redeemable for gold and so there was  obvious process by which an increase in the demand for gold would create a shortage of the media of exchange.   People short of money sell assets and spend less.  This results in reduced demands for goods and services.   This creates an incentive for those setting prices to lower them.

But in Sumner's Zimbabwe example, the currency used as the medium of exchange is not redeemable in gold.    But what exactly determines the gold price of Zimbabwean currency in this imaginary world?   Sumner assumes it is determined on the foreign exchange market.   Perhaps he is thinking back to an international gold standard, so that the foreign exchange market of Zimbabwe determines a gold price of Zimbabwean currency.   But in the real world, the foreign exchange market determines an exchange rate for Zimbabwean currency against U.S. dollars, Euros, and Yen, not gold.

So, what happens?    The purchases of gold in Asia raises the U.S. dollar price of gold.   Given the exchange rate between U.S. dollars and Zimbabwean currency, this raises the price of gold in terms of Zimbabwean currency.    But the Zimbabweans turn this around and say it has determined the gold price of Zimbabwean currency and the price of currency has fallen.  

When this lower price of currency is announced around the country, the amount of currency that must be paid for each and every good is now higher.   Alternatively, the quantity of currency measured in gold units is lower.   Given the demand to hold currency, there is now a shortage of the medium of exchange.  People spend less.   Firms lower their prices--the gold ones that the  quote.   Given the gold price of currency, this is also reducing the amount of currency that must be paid for each item.

Anyway, the reduction of spending on output results in a recession.   Yes, it is the greater demand for gold that gets the process started, but it is only because of the effect on the medium of exchange that it actually causes spending to fall.  And it is that fall in spending that directly causes the recession.

Now, once a market process is established such that market forces--monetary forces--will cause the price level to adjust, speculation will tend to cause appropriate changes in supply and demand.   In a traditional gold standard, where all the media of exchange are redeemable in gold, if people think the demand for gold is going to rise, prices will fall.   There is no need for the demand for gold to rise, money be redeemed for gold, the quantity of money fall, and a shortage of money cause less spending, and the lower spending cause lower output and prices.   The expectation of lower output and prices in the future will cause people to cut spending now.   But the market process must be there if speculation is going to forestall it.  

Anyway, the medium of account is not money.   The bundles of goods and services in the BFH system are not money.   The steel ingots in Fama's story are not money.   The money is the media of exchange.   But that doesn't mean that shifts in the relative price of the medium of account don't cause changes in the price level.

Money is whatever it is that such that changes in its "quantity" lead to inflation or deflation?   When the medium of account is an ordinary good like steel, talking about changes in the quantity of steel is wrongheaded.   It is flow supply and demand that counts.   When the medium of account is a bundle of goods, talking about the quantity of bundles is nonsense.

Further, recession is not caused by a deflation of prices pressing against sticky wages.   Recession is caused by a decrease in money expenditures on output along with a failure of prices and wages to fall enough to make it unnecessary for real output and employment to fall as well.

Money is what is spent.  Money is the medium of exchange.  It is intimately involved in the problem of too little or too much spending.  

But when money is tied to a separate medium of account, then the quantity, price, or yield on money must adjust so that there is a level of spending consistent with clearing the market for the medium of account at the defined price.   This forces the price level to change if necessary, but at the same time, it means that shifts in the supply or demand for media of exchange must be offset if they would otherwise force changes in the price level.

What does this have to do with Market Monetarism?   Well, I don't really understand what is the medium of account with NGDPL targeting, but the nominal anchor is the level of spending on output.   And the quantities of media of exchange must adjust so that they match their demands when nominal GDP is on target.   It is not currency or any type of deposit that serves as medium of account and whose quantity determines spending on output.   It is the level of spending on output that determines the needed quantity of money.

Sunday, October 28, 2012

Coffeeland: The Next Episode

Coffeeland is a small open economy on a gold standard.   It produces and exports coffee and imports all other goods.   In the first episode, foreign investors used equity to fund a railroad to a central plateau perfect for growing coffee.   There were two scenarios, one where coffee prices remained high, and everything worked out well both for Coffeeland and the investors.   The other scenario involved low coffee prices.   Things went bad for both the people of Coffeeland and the investors.

The next version of the story had the "railroad boom" funded by debt.   First by 20 year bonds and then by 1 year CDs, both sold to foreign investors.   In the happy, high coffee price scenario, all the investors receive their money back, principal and interest.   And the people of Coffeeland have higher real and nominal wages, and higher real and nominal income.

In the unhappy, low coffee price scenario, there is a finanical crisis.   The railroad cannot repay its debt.   The foreign investors lose.  And the people of Coffeeland suffer at first mass unemployment and then permanently lower real wages and real income.   The problem, however, was not the debt-financed boom.   It was rather the low coffee prices.

For the final episode, there is only one small change.   Rather than issue one year CDs, the Coffeeland bank funds its loans with checkable deposits.   It pays sufficient interest so that foreign investors are willing to hold enough checkable balances to fund all of the renewable 5 year loans that the Coffeeland Railway needs to fund construction.

Now, the railway boom is funded almost entirely by newly-created money.   The quantity of money in Coffeeland rises.   As before, the Railway company purchases steel rails, picks, shovels, and axes overseas. They also withdraw gold coin from the Coffeeland bank, which imports it from foreign lands.   As the Railway hires workers from the coffee plantations, nominal and real wages rise.   Again, coffee production falls, and due to the higher real and nominal wages, imports of consumer goods rise.   With the steel rails, picks, shovels, and axes all being imports as well, Coffeeland has a significant trade deficit.   However, despite the increase in the quantity of money, both deposits (held by the foreign investors) and the imported gold coin, there is no inflation.   Coffeeland is so small that its purchases of consumer goods has no appreciable impact on their prices.

Once the railroad is complete, most of the railroad workers are laid off, but they are employed in the new, more productive plantations in on the interior plateau.   Coffee productions and exports rise, which allows the Coffeeland Railway to pay back its loans and the Coffeeland Bank to pay off the foreign investors, principal and interest.   And they live happily ever after.

But, there is the unhappy scenario where coffee prices fall.   There is again a terrible depression and mass unemployment.   Only when nominal wages and real wages fall does employment recover, but real output and real income remain low.   Further, the Coffeeland Railway can never repay its loans.  And the Coffeeland bank fails as well.  It is unable to pay off the foreign depositors.

Was the depression causes by the "railroad boom?"   No.   Was the fact that the railroad boom was funded by money creation rather than CDs, 20 year bonds, or equity make any difference?  No.   Did the creation and destruction of money in Coffeeland impact the price level?  No.  

Suppose we develop a large data base looking at the economic performance of many countries like Coffeeland?   Would they really tell us much about the Great Recession?

Cowen on Currencyless Payments

Miles Kimball said something about abolishing currency being the cure for  aggregate demand problems and Tyler Cowen decided to respond.   Cowen quoted Matt Yglesias on the matter.  And also Ryan Avent.

I am not sure what Kimball said (I guess I should start following him on Twitter.)   Still, I think Cowen's discussion is confused.

Would abolishing currency end aggregate demand problems?    Not necessarily.    A shortage of the deposit accounts that allow for electronic or check payments would result in reduced spending on output and so an aggregate demand problem.  

The solution to that problem is to either increase the quantity of the deposits or else reduce the demand to hold them by reducing their yield.   The yield on deposits can be negative.  That is, people can be charged to hold money in the form of these deposits.

Currency is a problem because deposits can be redeemed for currency.   If the yield on deposits becomes too low, then people will be motivated to redeem them for currency.  To avoid those redemptions, those issuing deposits, banks and central banks, cannot lower the interest rate on deposits much below zero.   That only leaves the option of expanding the quantity of deposits.  

At least in the short run, expansions in the quantity of money could be unprofitable.   If those issuing deposits purchase a limited set of safe and short financial assets (like T-bills,) then their yields could be driven as low as, or even lower than, those on deposits.   If, on the other hand, longer term or riskier assets are purchased, then there is a chance of loss.    (Market Monetarists insist that a central bank should take that risk, though we also favor a target--the level of nominal GDP--that makes this problem less severe.)

If there was no currency, then those issuing money could purchase short and safe assets, and as the yield on those assets was reduced, lower the interest rate paid on deposits.   Alternatively, they could purchase longer term and riskier assets, while reducing the interest rate paid on deposits, increasing their interest margin to compensate for the greater risk.

How does this work in practice?   Step one is to cease redeeming deposits with currency.   During the 19th century, this was a common practice.   When did it happen?   During bank runs.   Of course, during that time, interest rates on short and safe assets spiked as banks sold them in a scramble for cash.   And each bank and all banks were seeking to get people to deposit currency back into the banks as soon as possible so that that redeemability could be resumed.

If there is no problem with printing up currency to meet demand, but those issuing money consider it unprofitable because the risk-adjusted yields on the assets they will buy are too low, then there is a second step beyond stopping currency redemptions.   The second stop is to make currency no longer be acceptable for deposit at par in the future.   Deposit the currency now at par or else only at a discount in the future.   This discount compensates the currency issuers for the low yields on their asset portfolio.

Cowen makes the error of assuming that suspending the issue of currency is the same thing as taxing currency.   He then makes various speculations of what would happen if there were a currency tax.

However, the point isn't to suppress the use of currency.   It is rather to make sure that there is sufficient spending on output.    If there is some spending of currency on output, then that is just fine.   The goal is to make sure that total spending, mostly by check or electronic payment, is adequate.  (For market monetarists, that is on the target growth path.)

What could go wrong?   Suppose once banks no longer redeem deposits for currency, people begin to buy currency with deposits on the market.   Currency goes to a premium.   People will only sell a $20 bill for $25.   The buyer writes a check for $25 and receives the $20 bill.   The seller of the currency deposits the $25 check, and can now write checks or make electronic payments for $25.    A $20 bill can be indirectly be exchanged for goods and services whose total price sums to $25.

While avoiding the low (or negative) interest rates on deposits might make holding currency attractive now, the fact that it will never be accepted for deposit, except at a discount, should be kept in mind.   Still, we can imagine that the currency would never be deposited again, and that the currency permanently floats against deposit money.   There is no reason to worry about that.   Surely, even if most currency was thrown away, some of it would continue be held by collectors.

Cowen imagines that some currency substitute might develop even if currency use were suppressed by taxation.  Suppose that people began to use gold or silver coins.   It is entirely reasonable that low or negative interest rates on deposits (and other short and safe assets) would motivate people to invest in gold.   The result would be an increase in the price of gold.  In a certain world, the price of gold would immediately rise to a point such that its expected rate of decrease equals the negative yield on other assets.   More realistically, the price of gold would rise to a point where the risk of capital loss makes additional purchases unattractive.   To some small degree, those who were already holding gold become wealthier and may choose to consume more.   This quasi-Pigou effect would tend to raise the natural interest rate, and so the return that can be earned on the assets that banks and central banks purchase.

Cowen says that there would still be a bond-currency margin, even if currency were taxed.   Consider the scenario above.   Yes, there would be a bond-gold margin, whether the gold was in bars or coined.   If people chose to hold paper currency divorced from the banking system, there would be a currency-bond margin.   But with the price of currency being free to vary with supply and demand, there is no problem.    The price of currency (relative to deposits and short and safe bonds) rises until the risk of capital loss, or the certain future decreases in its price, allows the market to clear.

Finally, what about the possibility that deposits would disappear if there is a currency suspension and no expectation of a return to currency redeemability at par.   I think that this scenario is implausible.   How many people would quit their jobs because the money deposited in their accounts cannot be withdrawn but solely spent on goods and services?   How many retailers would refuse to accept checks or electronic payments by check card because they can only make deposits and then write checks of their own and cannot make currency withdrawals from their bank?

Personally, I favor the complete privatization of hand-to-hand currency.   It should only be issued by private banks, and deposits and currency should be redeemable for deposits at the central bank.   Banks would cease issuing currency if it was not profitable.   And they should have a call option that allows them pull outstanding currency out of circulation or else only accept it at a discount in the future.   What discount?   One based upon the interest paid (or costs of holding) deposits during the interim period.

If issuing deposits at any given interest rate becomes unprofitable, then the interest rate on the deposits simply needs to be decreased.   The quantity of deposits and the yield on them can be adjusted in a way that avoids any shortage of them with spending on output at an appropriate level--on target.

Of course, if people become more optimistic about the future and are inclined to save less and invest more, and if people are willing to use credit for transactions rather than deposits, then this will make issuing money more attractive and holding money less attractive.   The result would be a higher interest rates on deposits.  It would also make issuing hand-to-hand currency more attractive, and so, any conveniences associated with using paper currency for some transactions could again be obtained.

Interestingly, if privately-issued hand-to-hand currency did not receive deposit insurance, and was made junior to deposits, it might be possible for relatively risky currency to be issued and used for small-denomination, face-to-face transactions even while the interest rates on the more senior deposits, and especially insured deposits, are negative.   In fact, allowing retail firms to issue their own unregulated hand-to-hand currency when interest rates are low would be a relatively simple solution to a currency shortage.   Dating the currency would make it a poor store of wealth, and so avoid problems with hoarding, even while allowing for the convenience of using hand-to-hand currency for some transactions.

Saturday, October 27, 2012

Credit Boom in Coffeeland

In the previous post, Coffeeland had a railroad boom funded solely by direct foreign investment.   The people of Coffeeland only use gold coins for money, but still there was a boom, with higher nominal and real output and higher nominal and real wages.   In the first scenario, the improved economic performance was permanent, because coffee prices remained high.   In the second scenario, a shift in tastes away from coffee to tea by foreign consumers led to a deep depression.   The problem was that coffee prices were low.   While there was initially mass unemployment, lower nominal and real wages allowed employment to recover, though nominal and real output remained depressed.

Suppose that the foreign corporation that developed the railroad did not solely use equity finance, but also sold long term bonds.   The firm is quite conservative and plans a sinking fund, so that not only can it pay the interest on the bonds, it can pay off the principle.   As before, the funds raised are used to purchase steel rails, picks, shovels, and axes, as well as bags of gold, and shipped to Coffeeland to build a railroad to the central plateau.

The economics are no different.   In the first scenario, the debt-financed boom results in a permanent improvement in the economy of Coffeeland.   The new coffee plantations on the central plain generate enough business for the railroad to earn profits for the stockholders and pay all the interest and principle on the bonds.   The second scenario remains bad.   There is a depression as before, but just like before, it was caused by the decrease in the demand for coffee.   It is true, however, that the railroad company defaults on its bonds.   There is a "financial crisis."  There is not enough money generated from coffee to pay all of the interest and principle promised on the bonds.   The stock prices crash harder too.  

Now, if Coffeeland insists on closing down the railroad for an extended period of time until its finances are sorted out, then the inability to shift coffee to the exterior will cause tremendous disruption in the economy, especially in the central plateau.   But if they have any sense, and realize that continuing to operate the railroad will reduce the losses of the bondholders, then that should not be a problem.

So, Coffeeland has a debt-financed boom.   If the debt was used to fund a profitable investment, then all is well.   If the investment turns out to be a malinvestment, then there was waste--losses for the investors.  

What about financial intermediation?   Suppose some Coffeeland investors start up a railroad company and other Coffeeland investors open up a bank.    The Coffeeland railway investors put up minimal capital and obtain 5 year renewable loans from the Coffeeland bank.  The Coffeeland bank investors put up minimal capital and fund the loans by selling one year certificates of deposit to foreign investors.  

The economics are the same, but the railway boom is financed bank credit.     A trade deficit develops.   Real and nominal wages increase.   Real and nominal output increase.   And the railroad makes enough money to pay off the interest and principle on the bank loans, and the bank makes enough money to pay the interest and principles on the C.D.s     Opening the fertile plateau allows for permanently higher coffee production.  Coffee exports are enough to pay back the foreign investors.

Of course, if the shift away from coffee to tea leads to low coffee prices, Coffeeland suffers a deep depression as before.   The Coffeeland railway cannot repay its loans to the Coffeeland bank.   It defaults.   And the Coffeeland bank cannot repay the CDs to the foreign investors.   It defaults.

Was the depression caused by the previous boom?   No.   Was the bust caused by excessive bank loans funding the boom?   No.    The problem was the decrease in coffee prices.


Boom and Bust in Coffeeland

Coffeeland is small open economy that grows coffee and imports everything else.   It is on an international gold standard and everyone in coffeeland uses gold coins.   Currently, all coffee is grown on the coastal plain, but beyond the coastal mountain range, there is a plateau perfect for growing coffee.  

Foreign investors sell stock on their home country, purchase steel rails, picks, shovels and axes, and ship them to the small open economy along with some bags of gold coins.   They hire workers to begin constructing a railroad over the mountains.

Competition for workers results in higher nominal and real wages.   The least efficient coffee plantations must curtail production.   Exports fall.    The steel rails, picks, shovels, and axes all count as imports.   And both the new railroad workers and those workers remaining on the coffee plantations increase consumption due to their higher wages, and so, imports of consumer goods rise.   A trade deficit develops.

There is a modest increase in the domestic quantity of money.   Those bags of gold that are being used to pay the workers are an increase in the quantity of money, but as it is actually spent by the rail company to hire the workers, they spend it on imported consumer goods and so it is re-exported.

There is no price inflation because the small open economy provides only a tiny fraction of world demand for the imported consumer goods, or the steel rails, picks and shovels.

Nominal and Real GDP both rise.   While coffee production does fall, the most efficient coffee plantations continue to produce coffee.  The partially finished railroad counts as the rest of output, measured at cost.

There is a great railroad boom, with a small increase in the quantity of gold coin, no price inflation, and higher nominal and real wages.  

The railroad is completed and the plateau is opened to coffee cultivation.   While many of the railway workers are laid off, they are soon employed in the new coffee plantations.    Only the most efficient coastal plantations survive, and the total productivity of labor in coffee production is enhanced.   Real and nominal wages remain high, real and nominal GDP remain enhanced, due to the additional coffee production as well as the railway services.

The railroad company earns profit and gradually repatriates is initial investment.   Coffee exports expand.  This is both because of enhanced productivity due to coffee being grown on more productive land, but also because the railway construction workers return to coffee cultivation.   While consumer good imports are higher due to the higher real wages, there are no longer any imports of steel rails or picks and shovels.   A trade surplus allows the repatriation of profits and the original investment.

In the first scenario....and they live happily ever after.

In the second scenario, the foreigners decide that they prefer tea to coffee.   The demand for coffee falls as does its price.   As a result, our coffee producing land suffers a deep depression.   Wages are sticky, and so as coffee prices fall, the coffee plantations lay off workers, resulting in mass unemployment.  

Eventually, both nominal and real wages fall.  This reduces the cost of producing coffee and so coffee production recovers.  While it might strain credibility just a bit, Coffeeland is so small that it doesn't produce a significant share of world coffee and so as its coffee production recovers, coffee prices remain at the same.   (If, on the other hand, coffee prices fall, then this results in an increase in the quantity of coffee demanded, allowing for added real sales, increased coffee production, and additional employment of coffee workers.)

The sorts of Keynesian stories that emphasize the spending of coffee workers simply do not apply.   While the spending of coffee workers on imported consumer goods would tend to reduce the sales of the foreign consumer good manufacturers, this is insignificant because the Coffeeland is so small.   Further, there is expansion of nominal and real income by those working on tea plantations somewhere else in the world.   Presumably they spend part of that added income on consumer goods.

While coffee production and employment do recover, both nominal and real GDP is much lower because the coffee is worth so much less.  

Unfortunately for the railway company, it also takes a loss.   It turns out that the railroad was a malinvestmentment.   The stockholders will never be able to recover all of their initial investment.   The stockmarket of coffeeland (made up of the shares of a railroad entirely owned by foreigners) crashes.

Because of the nature of railroad construction, particularly through mountains, the loss minimizing strategy for the railroad company could easily involve them maintaining the railroad in perpetuity.   And real wages and real GDP would remain permanently higher than it would have been if the railroad was never built.   While coffee is worth less, the plateau really is more productive.

Did the railroad boom cause the bust?  No, the bust was caused by the shift in tastes away from coffee towards tea.     While building the railroad was not an effective use of resources (perhaps it should have been built in Tealand,) even if it had never been built, there still would have been a decrease in nominal GDP, real GDP, nominal wages, and real wages in Coffeeland.

What about credit and debt?   Watch for the next installment.


Tuesday, October 16, 2012

This Time is Different

I read Reinhart and Rogoff's book some time ago.   It is mostly a warning about making short term investments in other countries.   "This Time it is Different."   Unlike all the other times when investors have been burned by lending short to foreign governments and banks, this time the rapid flow of hot money is sustainable.    We will be able to get our money back.  Not likely, say Reinhart and Rogoff.

Of course, foreigners have been investing in the U.S. for some time now.    The U.S. government has been funding a large and growing national debt with T-bills sold to foreign investors.     And there has been substantial foreign borrowing by U.S. banks, including shadow banks.

Reinhart and Rogoff also show that a country receiving substantial foreign investment ends up in very bad shape when the investment flows are reversed.   When the foreigners stop rolling over their short term government bonds and bank deposits, then the net capital outflow is going to result in a reduction in real income.

Unfortunately, they and others have been misusing their work to provide excuses for the Federal Reserve's poor performance over the last 4 years.   Rather than the true message that financial crises caused by a large sudden withdrawal of foreign investment is bad for both the foreign investors and the country losing the capital, they have turned this into a claim that a financial crisis necessarily results in a deep and persistent recession.

In reality, the U.S. recession and slow recovery was and is due to incompetent Fed policy--the use of an interest rate instrument to target the inflation rate.   The solution is nominal GDP level targeting, and a seamless transition to the use of base money as an instrument any time interest rate targeting becomes ineffective.

Now, if foreigners lose confidence in the willingness of the U.S. government to repay its debts, or they decide that private investment in the U.S. is a bad idea, and a large net capital outflow develops, then real income in the U.S. will suffer.   In my view, the least bad policy would remain keeping nominal GDP on a stable growth path.   Trying to maintain the dollar exchange rate or prevent inflation--immediately due to rising prices for imported goods--would exacerbate the problem.

On the other hand, shifting to a more rapid growth path for nominal GDP would not prevent the inevitable loss in real income.    Reinhart and Rogoff have plenty of examples (it is almost the norm) where nations suffering the net capital outflow have massive growth in nominal GDP and massive inflation as well.   My view is that contracting nominal GDP to defend the exchange rate or the consumer price level is as bad, if not worse, than expanding nominal GDP beyond its existing trajectory in a futile effort to maintain real income.

But the U.S. has not suffered a net capital outflow at all.  Instead, the "crisis" here has been associated with a net capital inflow.

How many of Reinhart and Rogoff's episodes were about a financial crisis "caused" by increased foreign investment?    The U.S. in the Great Depression?    Was that inevitable?    Or was it a Federal Reserve disaster?

Anyway, it is still possible that the U.S. could suffer from a large net capital outflow.   This was the crisis that many economists were expecting.   Surely, it was the crisis that Reinhart and Rogoff were expecting--the one they wrote their book about!  

Running huge budget deficits and building up a massive government debt does not help avoid that problem--it makes it more likely.    The answer is, and has always been, a new monetary regime--a nominal GDP level target.


Wednesday, October 10, 2012

What Trend? 1991 or 1985?

During the Great Moderation, Nominal GDP remained on a remarkably stable growth path.   Generally, I have been calculating a log linear trend from the first quarter of 1985 to the last quarter of 2007.    Why not start at the beginning--1984?    I used to, but the 10 percent growth in the second quarter of 2004 caused me to place that year as part of the Reagan-Volcker nominal reflation.

Others have instead started their trends in the next decade.    As far as nominal GDP is concerned, it makes little difference.   The trend starting in 1985 has a 5.38 percent growth rate and the trend starting in 1991 has a 5.28 percent growth rate.    While the fit is a bit better for 1985 to 2007, the 1991 to 2007 trend does almost as well.


Not much difference to be seen at this scale.  Still, for the second quarter of 2012, the gap between current nominal GDP and trend is 15.2 percent for the 1985 trend and only 14.1 percent for the 1991 trend.    A difference of one tenth of a percent in growth rates adds up over the years.

No doubt it was due to my Market Monetarist priors that I have always fit all the other macroeconomic variables to the same trend--1985 to 2007.    The trend inflation rate using the GDP chain-type deflator is 2.3 percent and in the second quarter of 2012, the price level was 2.42 percent below trend.   

Using 1991, on the other hand, provides a trend inflation rate of 2.05 percent.   And the price level was not below trend in the second quarter of 2012, it was .3 percent above trend!    This is as close to being on trend as can be expected.     As Bullard and also Davies (at all) have pointed out, the Fed is doing a great job in targeting a 2 percent growth path for the price level.  (At least last quarter.)



The 1991 trend fits the first part of the nineties (up until 1997) very well.   It also fits the Great Recession.     The 1985 trend fits just about nothing other than the peak of the "Housing boom," maybe 2006 and 2007.   

Of course, from a Market Monetarist perspective, the price level should vary with "supply shocks."   All that is important is for nominal GDP to remain on trend.   

Here is the price level again with the 1991 trend.


After the 2 percent targeting regime was established in the nineties, everything goes well until 1997 when inflation falls below 2 percent and the price level begins to fall below trend.    The gap between the price level and its trend grows and reaches 1.5 percent in 2002.   It only really starts shrinking in 2004 and has returned to trend by 2005.    However, it overshoots, and reaches 2.8 percent above trend in the second quarter of 2008.   Then it begins to shrink (should I add, "because the Fed took action?")   By the middle of 2009, it was back to one percent above trend, and has continued to gently return to trend, nearly reaching it by the middle of 2012.

Of course, because the trend for nominal GDP using the 1991-2007 period is approximately the same as for the longer 1985 to 2007 period, the deviations from trend are about the same.   During the late nineties, when the price level was below the 1991 trend, nominal GDP was above trend, reaching 2.59 percent in the third quarter of 2000 using both the 1985 and the 1991 trend.    At that point of "high" spending on output, the 1985 trend for the price level is about one percent "too high," while the 1991 trend for the price level was about 1 percent "too low."

And what about the "Housing Boom?"   Using the 1991 trend, nominal GDP is above trend from the third quarter of 2005 to the third quarter of 2007 and it reaches .8 percent above trend during the first and second quarter of 2006.    Looking at the longer, 1985 trend, only the first and second quarters of 2006 are above trend at all, and only about .2 percent.    

So nominal GDP trends and deviations from trend are about the same, but the price level trends and deviations from trend are very different.   Obviously, the difference must be in real GDP.


The 1985 trend tracks real GDP very close from 1984 to 1987, then it rises above trend a bit before falling below trend for a recession in 1989.   There is a very slow return to trend and then a pronounced Dot.Com boom.  The 2000 recession is just a return to trend and then real GDP remains on the trend until the Great Recession.

The 1991 trend starts with real GDP above trend in 1984 and the gap grows reaching nearly 4.2 percent in the first quarter of 1989.   The "recession" was simply a return to trend, which was reached in 1991.   This begins the long period of real GDP remaining very close to trend until the Dot.Com boom begins.    Just like the trend from 1985, the recession of 2000 was just a return to trend.   And then both remain at trend until the Great Recession begins.   Interestingly, real GDP is below the 1991 trend exactly like nominal GDP by the second quarter of 2012.  This, of course, follows from the price level being approximate at trend.

Here is real GDP and the 1991 to 2007  trend for real GDP:


The "Reagan Boom" is clear on this diagram, along with the Dot.Com boom.  There is really no "housing boom."

Consider the CBO estimate of potential output.   How does that compare?

There is no "Reagan Boom" in the late eighties according to the CBO, and the recession in the early nineties involved real output less than potential. While both the 1985 and 1991 trend show real GDP above trend, the 1991 shows a very large gap before the recession, and at least the 1985 trend shows real GDP below trend when real GDP is below the CBO estimate.    

Both the 1991 and the 1985 trends show real GDP above trend during the Dot.Com boom, and the CBO has real GDP above potential.   Interestingly, both the 1991 and 1985 trends have real GDP almost exactly on trend during the "Housing Boom" and the CBO estimate has real GDP equal to potential.   

Looking at real GDP,. the 1991 trend takes the slow recovery of the nineties to define the trend.   It shows real GDP far above trend during the late eighties.  The 1991 trend shows a huge "boom" in the price level during the "Housing boom" when real GDP remains at trend and real GDP remains equal to potential output.

Is the 1991  trend "best?"   I have my doubts.   









  




Sunday, October 7, 2012

More Inflation Counterfactuals

Lars Christensen used the CBO estimate of potential output to show what inflation would have been if nominal GDP had been on a 5 percent growth path.    Marcus Nunes critiqued James Bullard's claim that the Federal Reserve has been successfully targeting the price level, since the Consumer Expenditures Price Index is close to a 2 percent growth path starting in 1995.  

Nunes reproduced Bullard's diagram.



When I read Bullard's presentation, I was horrified.   Woodford endorsed nominal GDP targeting, but only as a  more politically saleable alternative to gap adjusted price level targeting.   Bullard (and the other economists helping him at the St. Louis Fed,)  forgot the "gap adjusted" portion of Woodford's proposal and went to straight price level targeting.   As Nunes points out, the consequence of the run up in oil prices for various consumer energy prices  in 2007 is shown on the chart.   Does Bullard really want to claim that a monetary policy that reverses that is desirable?    The diagram also shows that this measure of the price level is above target now.   Perhaps some further disinflation is in order?

As I have mentioned in my response to Selgin and my response to Christensen, a hypothetical target for some nominal variable does not explain actual excess supplies or demands for money and their disruptive impact on the economy.   Bullard at least says that central banks began thinking about a 2 percent inflation target in 1995.   I am pretty sure that Bernanke only got one adopted by the Fed recently.  The actual trend inflation rate for the Great Moderation was closer to 2.3 percent.   At least that is for the GDP deflator, which is the appropriate measure anyway.

However, my primary concern with Bullard's presentation came later with his sly implication that nominal GDP targeting was responsible for the Great Inflation of the Seventies.   While the actual culprit was probably unemployment rate targeting, this does translate well into real GDP targeting.   Both are a very bad idea.   And nominal GDP targeting is not at all the same.

The concept behind nominal GDP level targeting is to keep spending on output on a steady growth path.   The target for spending on output should grow at a slow, steady rate.  What happens to real GDP and the price level depends on what is happening to the productive capacity of the economy.

Rather than start with the CEP price index in 1995, suppose nominal GDP targeting had started back in 1959.   The output gap was was less than 1/2 percent.   To start with, suppose the target growth path was 5 percent.   First, look at the actual and target paths for nominal GDP.


Nominal GDP is much higher today than it would have been if a 5 percent growth path had started in 1959.   What do the growth rates look like?


The rising trend of nominal GDP growth starting in the mid-sixties is apparent.    And what are the implications for inflation and the price level?   Suppose potential output followed the CBO estimate and real GDP remained at potential.   What would the price level have been?


The actual price level is much higher than what would have happened if a 5 percent nominal GDP target had started in 1959.     It is 136% higher.  (And no, I don't think that targeting a 57 percent decrease in the price level would be wise.)

What about inflation?   

With a 5 percent growth path for nominal GDP starting in 1959, the inflation rate during the seventies would have remained below 2 percent.   It is only in 1980 that a slowdown in productivity growth that began in late 1979 would have raised inflation to nearly 3 percent.    According to the CBO, anyway,the "high" inflation period would be now--from 2008 to 2014, with inflation being only slightly above 3 percent.  We should expect a slight moderation of inflation in the near future.

Actual inflation during the seventies was much higher than what would have happened with a 5 percent nominal GDP growth path.   Aside from the 12 percent spike in 1974, it got worse and worse over the decade, getting as high as 11 percent.

Why?   Was it really because of poor productivity during the seventies?

Well, aside from the 1.5 percent acceleration from 1.5 percent to nearly 3 percent at the very end of the decade (and really in the early eighties,) the problem is clearly shown to be the accelerating growth of nominal GDP.   The Great Inflation was the result of a failure to control excessive growth of nominal GDP.   A nominal GDP target would have avoided the problem.

My own preferred target is for a 3 percent growth path for nominal GDP.  Again, using the CBO measure for potential output and starting in 1959, what would the price level have looked like?


The price level today would be 85 percent lower than its actual value.  (And no, I don't favor a monetary policy aimed at an 85 percent deflation of prices.)

What about inflation?


With a 3 percent nominal GDP level target, nearly all of the seventies would have been characterized by a mild deflation--less than one percent.   Only at the end, and in the early eighties, would there have been a few quarters with inflation nearly one percent.   Only now, from 2008 to 2014, would we suffer over one percent inflation.    And what of the deflation?    How bad would that have been?  Only in 1966 would it have been slightly more than one percent.

Saturday, October 6, 2012

Inflation Counterfactuals

Leading Market Monetarist Lars Christensen calculated the inflation rate for the U.S. if nominal GDP had grown 5 percent during the entire Great Moderation.    The growth rate of real GDP he assumes is from the CB0 estimate of potential output. He produced the following diagram:



He argues that this diagram shows that inflation was too high during the housing boom and too low during the Great Recession.    In my opinion, Christensen   makes the same error as Selgin.   While a 5 percent growth path for nominal GDP might be a good idea, the actual growth path during the Great Moderation was slightly higher.   While 5.4 percent might be only a bit higher, it makes a difference when trying to judge whether a past policy was inflationary.

I calculated P* which is the actual trend of nominal GDP from first quarter 1985 to the last quarter of 2007 divided by the CBO estimate of potential output.    The CBO estimates potential output for the next decade and the trend from the Great Moderation can be extended as well.

During the Great Moderation, the trend of P* is almost the exact same as actual trend for the GDP deflator.   It was about 2.4 percent.   (And the trend for real GDP was equal to the trend for the CBO estimate of potential GDP, both 3 percent.)

According to the CBO, potential potential began to grow for slowly in 2002, but this was from a level that had risen above trend during the Dot.Com boom.    Potential income fell below trend in 2005 and performed poorly since.  In no quarter was the estimate of potential income growth negative.   But during the depths of the Great Recession the growth rate fell to slightly below 1.5 percent, approximately half of the long run trend.    Potential income is 6.9 percent below its trend from the Great Moderation (and much longer.)     By 2022, the CBO projects the shortfall from the long run trend for the U.S. to be 14 percent.  (I think that is a more or less a disaster.)

The persistent reduction in the growth rate of potential income will result in higher inflation.   With the actual nominal GDP trend of 5.4, the level of inflation is higher, though the increase is the same.   The increase in the price level is substantial -- approximately 17 percent higher than with a 2 percent inflation rate target.   Again, if the CBO estimates are correct, then the inflation rate would above 3 percent in 2020.

Here is the estimated price level and the trend from the Great Moderation (which is the same as the actual trend of GDP deflator.)


Here is the estimated inflation rate and the trend from the Great Moderation:


If the CBO is correct, then nominal GDP level targeting, if more or less perfectly successful, would have resulted in inflation rising to nearly 4 percent from the 3rd quarter of 2009 to the first quarter of 2011.   According to their forecasts, inflation would only gradually slow down, though remain higher than the 2.4 percent trend of the Great Moderation for the next decade.

The more I do these calculations, the less faith I have in the CBO estimates.   Further, if there really is a productivity slowdown, then the proper solution isn't a slower growth rate of nominal GDP in order to keep inflation on target.   The solution is to fix the productivity slow down.   

Consider a world where wages and other money incomes continue to grow at trend, but slow productivity growth implies higher inflation.    Policy-oriented politicians claim that the problem is that we are having difficulty in producing goods.   That is why prices are going up.    The solution?   Come up with ways to increase efficiency.   Figure out ways for each worker to produce more.   Get inflation back down.

Consider another world.   The price level always grows at a stable rate.   When productivity slows down, wages and other money incomes must grow more slowly.    To actually get the wages to grow more slowly, the unemployment rate is high.   New entrants to the labor force have trouble finding jobs.   Firms have to adjust their compensation programs so that new workers start at lower pay and existing workers get fewer pay increases.    

Policy-oriented politicians say that the solution is to figure out ways to increase productivity.  Each worker needs to produce more.   How does that message play?     During the adjustment period when few new jobs are being created, getting existing workers to produce more appears to be in direct opposition to creating jobs for new workers.    It will lower inflation and then the central bank will adjust policy so more rapid nominal GDP growth will create more jobs?   Sounds like a tough sell.    

Several times recently I have accepted the RBC notion that if people choose to work less, then stable nominal GDP growth is undesirable.   Nominal wages and prices both rise.     It would seem desirable to instead let nominal GDP grow more slowly or even fall.    The wage rates could continue to grow on trend (or above trend because of an increase in labor-share of income,) while prices remain at trend.   Total wages would fall since people are choosing to work less.

However, there is a policy alternative.   For the U.S. anyway, increased immigration can offset any decrease in the desire of natives to work.   While this might not be terribly important in many circumstances, the problem of retiring baby-boomers makes it very important today.   As baby boomers retire, the growth rate of the labor force will tend to fall.   With nominal GDP level targeting, this will tend to force up money wage rates as well as prices.   

Now, we are seeing shortages of labor, rising wages, and rising prices.   Policy-oriented politicians propose that immigration restrictions be liberalized.    It so happens that taxes on the wages of the new immigrants will also support the medicare and social security expense of the baby boomers.    The immigration occurs during a time when there appears to be plenty of jobs--vacancies created by the retiring baby boomers.

On the other hand, suppose that an inflation targeting regime makes sure that nominal GDP grows more slowly and wages remain stable as baby boomers retire.   Because the bulge of retirees collects social security and medicare, severe budget problems develop.    A possible solution is to bring in more immigrants.   Policy-oriented politicians make that proposal.   But where are the jobs for all of these immigrants?   Is the answer to slow the growth rate of wages so employers will take them on?   Won't lower wages result in less social security and medicare taxes being collected?   Or is it that the slower growth of wages will result in inflation below target, and then the central bank will expand monetary policy enough to get nominal incomes growing.   Sounds like a hard sell.

For the U.S. during the next few decades, nominal GDP level targeting looks like a good idea.   If there is a productivity slow down or slow labor force growth, then those are the problems that need to be fixed.   Let price and wage inflation signal those problems.


Monday, October 1, 2012

Selgin on Intermediate Spending Booms

George Selgin has a post arguing that even modest deviations of nominal GDP from trend can lead to "booms."    He provides the following graph.

He goes on to argue that while the upward deviation beginning in 2002 and ending in 2009 appears small, it could result in significant market distortions.    Immediately, I thought that this graph looked wrong.    Here is the graph I just posted yesterday:


I generated this trend using the natural log of nominal GDP from first quarter 1985 to fourth quarter 2007.    It looks to me like the actual value of nominal GDP just barely kisses the trend in 2005 before its growth rate slowed. 

Of course, that was the whole Great Moderation.   Here is a close up:


It looks about the same.    I then graphed the nominal GDP "gap," which is the difference of nominal GDP from trend.   



The Dot.Com boom was associated with a large upward deviation of nominal GDP from trend that reached slightly more than 3 percent in the second quarter of 2000, and then rapidly collapsed.     The gap had turned negative by the fourth quarter of 2001, and reached a -2.73 percent by the second quarter of 2003.     

And so, what about the housing boom?   Well, it does peek above trend, sure enough.    It was slightly less than 1/4 percent above trend in the first quarter of 2006 and continued above trend, a bit less than 1/5 of a percent in the second quarter of 2006.    Both the quarter before and the quarter after, the shortfall was greater.

At first I assumed that Selgin must have used some different time period to generate the trend, but right there on the chart it says the growth rate is 5 percent.   Well, the trend for nominal GDP during the Great Moderation was 5.4 percent, so that explains the differences in the diagrams.

I realize that Sumner advocates a 5 percent growth path, but I don't try to understand what happened in the Great Moderation by comparing the actual trend to Sumner's preferred target.  I favor a 3 percent growth path, and so, the nominal GDP growth during the Great Moderation was far above what I consider ideal.   Still, I think shifting down to my preferred growth rate would likely create the usual disinflationary disturbance.  More to the point, I don't think that the many quarters with nominal GDP growing more than 3 percent involved an excess supply of money.

As for the theory, Selgin writes:

Such distortions can be significant even when they don’t involve exceptionally rapid growth in nominal income, because measures of nominal income, including nominal GDP, do not measure financial activity or activity at early stages of production. When interest rates are below their natural levels, spending is re-directed toward those earlier stages of production, causing total nominal spending (Fisher’s P x T) to expand more than measured nominal income (P x y). Assets prices, which are (appropriately) excluded from both the GDP deflator and the CPI, will also rise disproportionately. It is the possibility that such asset price distortions may significantly misdirect the allocation of financial and production resources that lies at the heart of the "Austrian" theory of boom and bust.

Selgin's argument would be appropriate if Market Monetarists favored targeting nominal consumption spending.   And, of course, it is on the mark as a criticism of targeting a consumer price index--which is what most central banks actually do.   But Market Monetarists favor targeting nominal GDP, and it includes both the prices and quantities of goods at early stages of production.    Those goods of the "higher order" that last more than one year count as final output as much as do consumer goods and services.   Since capital goods that last a very long time and provide services in the distant future are especially "early" in the production process, they are measured quite well.   (Single family homes, for instance.)

But what about intermediate goods?   These goods are produced and then used up in producing some other good.   Since basic GDP accounting requires that intermediate goods be "netted out," this might suggest that an excess supply of money matched by an excess demand of such intermediate goods would fail to show up in nominal GDP.  

But this is a mistake.  Intermediate goods show up as part of inventories.   When a steel firm produces steel and hasn't sold it yet, the dollar value of the steel is part of nominal GDP.   If a car manufacturer purchases the steel, and has yet to produce any cars with it, then the dollar value of the steel counts as nominal GDP.   (Not twice, of course, but only once.)   If the car company makes engines with it, the steel is included in the value of the engines, even if the engines are left sitting about waiting to be assembled into cars.   

Now, inventories are not included in Final Sales of Domestic Product or Final Sales to Domestic Purchasers.   That is one reason why those are not the proper measures of spending on output.   

None of these measures, of course, measure the prices or quantities of financial assets.    However, if the concern is distortions in the allocation of resources, the production of inappropriate specific capital goods, and the pattern of employment, it is unclear that financial assets (or raw land or other nonreproducible real assets) matter.

Selgin criticizes both Wicksell and Market Monetarists, arguing:
Missing in both perspectives is any attention to the way in which interest rate movements redirect demand from certain markets to others.
Now, I believe that an excess supply of money typically reduces the market interest rate below the natural interest rate in a credit money system.    And I also think that lower interest rates raise the demands for goods with a higher interest elasticity of demand more than other goods.   But that isn't a "redirection" of demand.   The demand for other goods and services, with smaller interest elasticity of demands, just have smaller increases in demand.    Of course, there is a distortion in the composition of demand, but an excess supply of money and a market interest rate below the natural rate doesn't cause demand to fall for some goods as would be the situation with a tax and subsidy scheme. 

Market Monetarists have been especially interested in situations where nominal GDP is below trend (or target) and an expansionary monetary policy is aimed at getting nominal GDP back up to the target growth path.   In those situations, the point of the monetary policy is to create a nominal "income effect."   Market Monetarists insist that nominal and real interest rates can be expected to return to equilibrium once this is accomplished, and so, can immediately begin to rise due to that expectation.

In 2003, nominal GDP was far below trend, and an expansionary monetary policy could return it to trend.  It barely made it back to trend.   Of course, the Fed wasn't targeting nominal GDP at all, but rather inflation and the output gap.   And they were doing this by making periodic adjustments in a policy interest rate.   Such a policy almost seems tailor-made (or is it Taylor-made) to create the problems that concern Selgin.

If instead, we had a central bank that was targeting nominal GDP, it is inevitable that mistakes would be made.   An excess supply of money could develop and the market rate could be pushed below the natural rate.   Only later would spending on output (including capital goods, partially finished goods, and inventories of intermediate goods) begin to rise above target.     Would those firms that faced what turns out to be an sustainable rapid increase in demand commission specific capital goods (say construct buildings) and hire workers?   Certainly that seems possible.   

However, to the degree this is caused by "low interest rates," these interest rates must be expected to persist into the future.   Too much Austrian analysis just assumes that everyone takes  any change in the policy interest rate as permanent and uses it as a reasonable basis for making long range plans.     Isn't the most likely scenario one in which both the monetary authority and entrepreneurs make the same error?  

Perhaps they all believe there has been a permanent decrease in the natural interest rate.   Or, perhaps the problem is that the natural interest rate has increased, and both entrepreneurs and the central bank have failed to see it. 

Market Monetarists, like Sumner and I, who are interested in index futures targeting, have in mind a system where monetary policy is generated by the market.   If those trading the index futures, bond investors, and entrepreneurs all think that the natural interest rate has fallen, then it is quite possible that unsustainable investments will be made before excessive growth in spending on output shows that no, the natural interest rate was not as low as everyone thought.   But would such errors be especially the fault of those trading the index futures?   Wouldn't those holding long term bonds also be a  fault?   And how about entrepreneurs that initiate projects that are only feasible if interest rates remain low?

The central bank lowers the interest rate, entrepreneurs take that to be permanent, and make bad investments.   Even with inflation/output gap targeting, it seems like the entrepreneurs should accept responsibility for figuring out that an excessively low policy rate now will be offset by a higher rate in the future.   And if they think the policy  rate is not excessively low, isn't the problem that they have a mistaken estimate of the natural interest rate now and in the future?   (Just think about scenarios where the natural interest rate falls now and rises again later.  Entrepreneurship is not easy.)

Anyway, I think there was a housing bubble.   The list of things that increased the demand for homes was legion.   The tiny overshoot of nominal GDP in 2006 seems like a pretty small problem in comparison.