Archive for October 2012


Kocherlakota is almost there—just needs to drop the word “too”

Johnleemk and Nic Johnson sent me this great quotation from Minneapolis Fed president Narayana Kocherlakota.

[The Fed] has also been targeting a fed funds rate of under a quarter percent for nearly four years””and anticipates continuing to do so through mid-2015. In the language of central banking, the Fed’s policy stance is considerably more accommodative than it was five years ago. . . .

Some observers argue that the Fed has done too much, has been too accommodative. I strongly disagree. … In light of the unusually large macroeconomic shock, I believe that it is misleading to assess the FOMC’s actions by comparing its current choices to policy steps taken over the past 30 years. Instead, we have to assess monetary policy by comparing the economy’s performance relative to the FOMC’s goals of price stability and maximum employment. In particular, if the FOMC’s policy is too accommodative, that should manifest itself in inflation above the Fed’s target of 2 percent. This has not been true over the past year: Personal consumption expenditure inflation””including food and energy””is running closer to 1.5 percent than the Fed’s target of 2 percent.

But this comparison using inflation over the past year is at best incomplete. Current monetary policy is typically thought to affect inflation with a one- to two-year lag. This means that we should always judge the appropriateness of current monetary policy using our best possible forecast of inflation, not current inflation. Along those lines, most FOMC participants expect that inflation will remain at or below 2 percent over the next one to two years. Given how high unemployment is expected to remain over the next few years, these inflation forecasts suggest that monetary policy is, if anything, too tight, not too easy.  (Bold print added by Johnleemk.)

It’s really gratifying to see Fed people like Dudley and Kocherlakota getting closer and closer to the truth, that Fed policy has been ultra-tight since mid-2008, just as Fed policy in the 1930s was ultra-tight, despite near-zero interest rates.  Over time, people will stop talking about policy as being “too tight” and just start calling it “tight,” as Friedman did in 1997 when Japanese policy was too tight to hit their inflation target, and interest rates were near-zero.  After all, you rarely see people say; “Those 1000% interest rates show that monetary policy was tight during the German hyperinflation, albeit too expansionary relative to the needs of the economy.”  They simply cut to the chase. Policy was too easy!

When I started out in late 2008 and early 2009 my claims that tight money was the problem were met with incredulous stares.  Now we’ll see who gets the last laugh.  Indeed (here’s a teaser) we may find that a month from now I’ll be able to get the last laugh on some of my early critics.  I have a long memory.

I’ll get back to the MOE/MOA debate tonight.

PS.  Do you see a bit of sarcasm when an academic says “in the language of central banking.”

When we have internecine battles, you know that market monetarism has arrived

Now we’re talking!  Nick Rowe has a new post up responding to my claim that the medium of account (MOA) is the essence of money, not the medium of exchange (MOE.)  Nick disagrees, as do all the other market monetarists who have weighed in.  But that’s never stopped me before.  Let’s think of a basic monetary model of the price level:

P = Ms/(Md/P),  Where Ms = Md.

Oops, that’s a tautology, not a model.  Let’s try again:

Md/P = f(i, Y)  Where real (base) money demand is negatively related to i and positively related to Y.

Ms is set by the Fed.

Now consider what we mean by “M.”  If M is the medium of account, but not the medium of exchange (as in my Zimbabwe example) then the model works fine.  If M is the MOE but not the MOA, then it doesn’t work at all.

Here’s Nick:

Scott Sumner argues that it is the medium of account function that matters. My view is different. Here is my view:

Demand and supply of the medium of account determine the equilibrium price level.

Demand and supply of the medium of exchange determine whether the economy is in a boom or a recession.

If the MOA determines the equilibrium price level, then ipso facto it determines NGDP.  So unless I’m mistaken Nick agrees that the market for the MOA determines NGDP.  I claim that business cycles are caused by NGDP shocks in the presence of nominal wage stickiness, or perhaps both wage and price stickiness.  Once one accepts that MOA shocks cause NGDP shocks, it seems to me that the battle is over.  In my Zimbabwe example the MOA got more valuable due to higher gold demand, and the MOE got almost worthless from excessive money printing, and yet Zimbabwe experienced deflation as the Z$ price of gold soared.  But maybe I am missing something.

Nick’s a very smart guy who knows monetary economics better than I do, so let’s try to figure out where the disagreement comes from:

1.  I look at recessions as big drops in output, associated with involuntary unemployment.  I am pretty sure that Nick views them as big drops in spending, where monopolistically competitive firms are not able to find buyers.  Perhaps that’s why he focuses on the MOE.  But in the end, gross domestic expenditure equals gross domestic production.  So that really shouldn’t be the decisive difference.

2.  Nick seems to assume a disequilibrium model, whereas I assume the MOA and MOE markets are continually in equilibrium.  Here’s Nick:

But what determines Y when we are out of equilibrium, because the Emperor Diocletian has issued an edict forbidding any price changes?

Thought-experiment 1. Start in equilibrium, hold all prices (update: both Ph and Ps) fixed, then halve the stock of gold (medium of account). What happens?

There is an excess demand for gold in the gold market, but nothing else happens. The market for haircuts continues as before. People want to sell some of their silver for gold, but they can’t, because nobody wants to take the other side of the trade. The production and sale of haircuts for silver continues just as before, because there is no change in the relative demands for silver and haircuts. There is no change in the marginal utility of silver, the marginal utility of getting a haircut, or the relative price of haircuts and silver Ph/Ps. So trade of silver for haircuts continues just as before.

An excess demand for the medium of account does not cause a recession.

But why would the gold (or money) market ever be in disequilibrium?  In the early 1930s when a big increase in the demand for the medium of account (gold) was causing worldwide deflation and depression (remember sticky wages), anyone could freely get gold any time they wanted it. There was no shortage at all.  Indeed the same is true for currency, which is easily available via ATMs, even when a reduction in the growth rate in the currency stock is triggering a recession (as in late 2007 and early 2008.)  A shortage is very different from a reduction in supply.  A shortage means people are unable to hold the currency or gold balances that they wish to hold.

But what if prices are sticky in the short run?  In that case wouldn’t a big reduction in gold, or currency, cause a shortage of that MOA?  No, interest rates are flexible, and the liquidity effect of interest rates is what equilibrates the MOA market until prices have had time to adjust.  Then over time NGDP begins to fall, so people don’t need as much MOA, and thus interest rates begin to fall.  In the very long run the NGDP will adjust by enough to fully offset the MOA shock, and interest rates will return to their original level.  At no time is the MOA market out of equilibrium, it’s just that different variables are equilibrating the market; first interest rates (the opportunity cost of holding MOA), and then NGDP, and in the long run only prices change—money is neutral in the long run.

I can’t make sense of macro unless I think in terms of one set of markets always being in equilibrium (stocks, bonds, forex, currency, gold, etc) and another set of markets being out of equilibrium during business cycles (labor, and perhaps some monopolistically competitive goods.)  My opponents talk about scenarios where it seems like almost everything is out of equilibrium.  I’m just not smart enough to figure out what’s going on in that case—it’s like trying to keep track of all the leaves swirling outside my window last night.

One other point.  Assume there is a negative gold market shock, and gold is the MOA but not the MOE.  Also assume a dual MOA, with paper currency being the MOE.  I do realize that in order to maintain convertibility the central bank may have to reduce the currency stock, and that this could be viewed as causing a recession.  Indeed this is precisely what happened in Canada in the early 1930s.  But it’s very strange to attribute that early 1930s deflation/depression to a falling Canadian currency stock.  That response was endogenous.  Rather it makes more sense to view the global hoarding of gold as the cause of the Canadian deflation.  After all, why would the Canadian central bank decide to run a highly contractionary monetary policy in the early 1930s?

When there is a negative monetary shock from the MOA market, lots of other things will happen.  The nominal interest rate might shoot up in the short run.  And the Keynesians will say monetary policy is really all about interest rates.  The central bank will be forced to contract the currency stock, and the non-Sumnerian (non-insane?) market monetarists will say monetary policy is all about adjustments in the stock of MOE, relative to shifts in demand for MOE.  But isn’t the heart of the matter that the MOA got more valuable, forcing global deflation on any country using gold as the MOA?

There is a Canadian Nobel laureate in economics who would understand my argument.  His Nobel lecture claimed that the Great Depression was caused by too much demand for gold.  Maybe he’ll leave a comment.

Paging Siena, Italy . . .

The effect on consumer welfare is the sum of the change in consumer and producer surplus

Tyler Cowen linked to a post (Jeffrey Ely?) that tried to refute the standard economic argument that price gouging is beneficial:

Suppose that an unexpected shock has occurred which has two effects. First, it increases demand for, say bottled water. Second, it cuts off supply lines so that in the short-run the quantity of bottled water in the relevant location is fixed at Q. A basic principle of economics is that if you wish to maximize total surplus then you should allow the price to adjust to its market-clearing level. This ensures that those Q consumers with the highest value for water get it. The total surplus will then be the sum of all their values.

This simply assumes away one of the two major rationales for price gouging.  The whole point of allowing higher prices for water is too encourage producers to truck in water from outside the disaster zone.  So yes, if you assume away any supply response, then the case for price gouging becomes somewhat weaker, but it doesn’t entirely go away.  There’s still the advantage of allocating water to those with the highest need, which is likely to be an extremely important advantage during a natural disaster.

[And don’t say it’s still an allocation argument, as the water trucked in will reduce consumption elsewhere.  In fact, much will come from warehouses.  The high prices will induce warehouses to run a bit closer to the edge, while frantically calling for Poland Spring to ramp up production.  People will drink more water in aggregate, if some is trucked in.]  Jeff continues:

But in fact it is quite typical for the consumer surplus maximizing solution to be a rationing system with a price below market clearing. I devoted a series of posts to this point last year. The basic idea is that the efficiency gains you get from separating the high-values from the low-values can be more than offset by the high prices necessary to achieve that and the corresponding loss of consumer surplus.

Why would we only care about consumers’ surplus and not also the surplus that goes to producers? We normally we care about producer’s surplus because that’s what gives producers an incentive to produce in the first place.  But remember that a natural disaster has occurred. It wasn’t expected. Production already happened. Whatever we decide to do when that unexpected event occurs will have no effect on production decisions. We get a freebie chance to maximize consumer’s surplus without negative incentive effects on producers. And just at the time when we really care about the surplus of bottled water consumers!

This is wrong for two reasons.  We care about producer surplus because producers are people.   Except for the unemployed, people have dual roles; they are both producers and consumers.  So 100% of the loss of “producer surplus” is actually borne by consumers!  It hurts producers who happen to be consumers.  And since all producers are consumers, the net effect on consumers in aggregate is negative, contrary to his argument.  He made the common mistake of (implicitly) equating the effect on “consumer surplus” with the effect on “people who happen to be consumers,” which is everyone.

Now I expect some commenters to jump in with some absurd income distribution argument.  One common mistake is to assume consumers who are non-water producers (like Bill Gates) are poor and consumers who are water producers (like the guy who delivers water to my college, or the immigrant with the 7/11 store) are rich. But producers and consumers are simply two sides of the same coin.  If a society has a policy of “no price gouging,” then competitive industries like water supply have to figure that into their pricing system.  They must charge more during non-emergency periods, knowing they won’t be able to price gouge during emergencies.  Contrary to Jeff’s assertion, in a world of rational expectations there is no free lunch to redistribute to consumers.  If there were, we could simply have the government expropriate any unexpectedly large crop harvest, or an unexpected mineral discovery, and share the proceeds with “consumers.”  Does anyone think that if a country behaved that way producers wouldn’t begin to anticipate the future expropriation, and act accordingly?  Believe me, producers know price gouging is currently illegal, and that already factors into their pricing decisions during good times.

There’s no free lunch, so efficiency is the only relevant criterion to use here, unless you believe the government should also be setting water prices in non-emergency periods, due to some other factor like monopoly power.

And remember, producer surplus goes to consumers—100% of it.

Update:  Suppose Massachusetts decides to discontinue the state lottery.  It will do one last $100 million jackpot.  After the winning number is announced, should the state go “nah nah nah” and take back 99% of the winnings in a special unexpected tax, and spend the money on poverty programs?  After all, the poor need the $99 million more than the guy who would still be left with a million.  And there’s no effect on the future demand, since the lottery is being discontinued.  Or would governments that behaved that way eventually pay a price?  (Hint: consider the history of Argentina.)

What is money? What is inflation?

The answers to these questions seem obvious to me, but I can tell from the recent comments that most people don’t agree.  Saturos and Bill Woolsey have argued (in the comments) that money is the medium of exchange.  I argue that money is the medium of account.  What makes this issue so tricky is that money is almost always both, and the textbooks define it as having both characteristics.  But which criterion is the “essence” of money?

Money is also that thing we put in monetary models of the price level and the business cycle.  That begs raises the question of whether the price level is determined by shocks to the medium of exchange, or shocks to the medium of account.  Once we answer that question, the business cycle problem will also be solved, as we all agree that unanticipated price level shocks can trigger business cycles.

So now we have to figure out what we mean by the term ‘inflation’.  I’d like to propose 5 criteria:

1.  Inflation is a general rise in the sticker price of goods.

2.  Unanticipated inflation helps borrowers and hurts lenders.

3.  When wages are sticky, deflation causes unemployment.

4.  Inflation reduces the value of the medium of account.

5.  Inflation reduces the value of the medium of exchange.

When the medium of exchange and the medium of account are identical, then all five of these statements are true.  If they are split, statement 5 is true for the medium of exchange, and statements 1 through 4 are true for the medium of account.  Let’s take an example to illustrate this confusing issue:

Imagine Zimbabwe uses gold as the medium of account.  Then they have budget problems because their economy crashes when the government tries to take too much wealth from the top 1%.  So they decide to print money.  But the president (who is a madman) tells his treasury minister that he wants to stay on the gold standard, and will not tolerate any inflation.  If the treasury minister violates this demand, he and his family will be executed.  What’s a treasury minister to do?

Easy. Tell the public that they must continue to set all wages and prices and debt contracts in terms of grams of gold.  At the cash register there will be a sheet of paper listing today’s exchange rate between paper Zimbabwe dollars and grams of gold.  People will pay their bills with Zimbabwe dollars.  (I’ve seen something similar in Canadian border towns, with US dollars as the MOE.)

In this example, gold is the medium off account and Zimbabwe dollar is the medium of exchange.  What is the “true” rate of inflation?  In terms of the definitions above, the criteria 1 through 4 apply to inflation in gold terms, and criterion 5 applies to inflation in terms of Zimbabwe dollars.  I would add that the average person would view inflation in gold terms.  Imagine you are a Zimbabwe diamond miner who works on a wage contract promising you 10 grams of gold per month.  You could care less what the prices are in Zimbabwe dollars, you have a fixed wage in gold terms, and when you go shopping you will see price stickers on the goods in gold terms, not Z$.  That’s your “cost of living.” When prices in terms of gold rise then you are worse off.   Zimbabwe dollars are just a medium of exchange.

In any model of the price level you are essentially modeling the real value of the medium of account, and not (necessarily) the medium of exchange.  And I’d add that the same is true of business cycles. Saturos and Bill Woolsey (and I think Nick Rowe) believe that recessions occur when there is a shortage of the medium of exchange. I disagree. I never have any problem getting cash, even during recession years.  A lack of cash never causes me to spend less. A recession is caused by a decrease in supply or an increase in demand for the medium of account. Thus suppose that Zimbabwe was on the gold standard at the same time that soaring gold demand in Asia was pushing its value up, relative to other goods and services. Zimbabwe would experience deflation. If the wages of Zimbabwe gold miners were sticky in gold terms, then the diamond mine would have to lay some of them off. They would be unemployed, and would start buying fewer goods and services. Indeed in some cases market signals (such as higher interest rates) will encourage people to buy less even before the workers are laid off (the effect before the cause). But the fundamental problem is that there is too little NGDP, given the current (sticky) level of nominal hourly wages. The medium of account is becoming too valuable.

In the comments section I get the impression that people believe recessions occur because there is less media of exchange to spend. That’s not a useful way of thinking about the problem, for several reasons.  Start with a gold standard, and think about the effect of worldwide deflation on a small economy like Canada.  If Canada wants to maintain a fixed peg between gold and the Can$, they might have to reduce their currency stock.  Is the lower currency stock “causing” the drop in spending?  Obviously not.  The global increase in gold demand that caused the global deflation in gold terms is the fundamental problem, and Canada’s declining currency stock is a symptom, just as the decline in lots of other nominal variables in Canada are symptoms.

In many cases the stock of currency does not decline during deflation.  For instance, it actually rose in the US during the early 1930s.  That’s because the negative effects of lower prices are more than offset by the increased demand for currency caused by the lower nominal interest rates (which are the opportunity cost of holding currency.)  When this occurs it is especially easy to see why a lack of media of exchange is not a problem.  During 1932 Americans held much more currency than in 1929.  The lack of spending wasn’t because they had less currency in their wallets and piggy banks; they had plenty.  Rather they had a higher demand for money, and simply chose not to spend it as often.

What makes this confusing is that today currency is the medium of account.  Thus an increase in the demand for currency that causes deflation and depression would be the fundamental problem, not merely a symptom of a deeper problem in the ultimate medium of account (as in the early 1930s).  And because these two roles (medium of account and medium of exchange) are now unified, both sides can look at the same set of facts and reach similar conclusions.  It’s only when we separate out the roles of medium of account and medium of exchange that we can clearly see the real essence of “money.”

What’s the role of the financial system in all this?  Financial systems are recent inventions; money and monetary policy and inflation have been around for millenia.  Yes, financial shocks can impact currency demand, but at a fundamental level the financial system is nothing special, just one more factor (like drug dealing) that impacts the demand for currency.  I don’t care if currency is only 1% of all financial assets.  Give me control of the stock of currency, and can drive the nominal economy and also impact the business cycle.  Cut the value of currency by 90% via OMOs, and you will raise the nominal GDP 10-fold, and prices 10-fold, and wages 10-fold, and the nominal stock of all assets by 10 fold.

PS.  The equation of exchange (MV=PY) applies to the medium of account, but of course one could substitute any dollar-denominated asset for M, and it would still be an identity.  After all, the definition of V is NGDP/M.

PPS.  It’s true that the broader aggregates fell in the early 1930s, but you can find plenty of recessions where they didn’t, due to higher demand for liquidity.

PPPS.  Totally off topic, but Matt Yglesias’s critique of the electoral college is a rare note on sanity in an otherwise horribly confused debate.

The monetary base is special

Tyler Cowen has a new post criticizing the idea that we should eliminate hand-to-hand currency so that the Fed could cut interest rates below zero.  He’s right that it’s a bad idea, but not all of his objections are sound:

Furthermore, under some views, this proposal would in essence put monetary policy in the hands of the drug trade.  Cracking down on drug lords, or easing up on them, would become major monetary policy instruments, at least if you take the Fama-Sumner view that currency has special potency over the price level.

4. I do not myself believe that currency per se has such extreme power over aggregate demand, at least not in such a credit-intensive economy as ours.  That means this proposal doesn’t get at the heart of the AD problem, which is closely linked to credit creation.

This is a very common mistake, so it’s important to explain what’s wrong with this reasoning.  Money is not special because it is a big part of wealth, or a big part of credit.  Indeed it’s not even special because it’s the medium of exchange.  It’s special because it’s the medium of account.  All prices (including the price of credit) is measured in money terms, not credit terms.  If everything was measured in apple terms, then the apple market would drive the nominal economy (even though apples would continue to have only trivial impact on long term RGDP growth.)

Even if there is no change in the real demand for credit, a doubling of the money supply will lead to a doubling of the nominal credit stock in the long run.   That’s true even if the stock of credit is 100 times larger than the stock of currency. The reverse is not true.  Base money is special because we price things in terms of base money.  If someone could show me that the previous sentence was wrong, I would disavow everything I’ve written on this blog from day one.  It’s the rock on which all of monetary history theory is built.

Also note that drug dealers would not cause any problem for monetary policy, for the simple reason that they did not do so before 2008, when 95% of the base was already currency.  The Fed can easily accommodate changes in the demand for currency, and does so.  That’s why the big increase in currency demand in the 1990s and 2000s did not put us in a depression.  On the other hand if we had been following a 4% constant MB growth rule, we probably would have fallen into depression as foreign demand for our currency soared.

6. I don’t see how this proposal could work unless it is applied globally, which seems implausible.  If your dollars are being taxed some extra amount, just put them in a foreign bank to earn zero or do some kind of funny quasi-repurchase agreement, with a foreign bank, to avoid having formal ownership of the dollars on the days of the tax.

If I’m not mistaken the electronic money proposal would eliminate hand-to-hand currency, and all other base money would be electronic accounts at the Fed or money embedded in debit cards representing electronic accounts at the Fed.  Continual positive or negative interest would occur via adjustments in those money balances up or down.  Other foreign “dollar bank accounts” could pay positive or negative interest.  So I think it would be feasible, but I’m far less confident on this point than on my previous point.

Currency will obviously be eliminated at some point, but I agree with Tyler that it would be a mistake to rush this solution into effect anytime soon.  For now a much easier solution to our problems is monetary stimulus.  Remember, Bernanke does not say the Fed can’t give us 3% inflation; he says it would be a bad idea.

PS.  The “Sumner-Fama view” is not all that special.  During the Great Moderation the standard view was that the Fed could target NGDP or inflation via adjustments in the fed funds rate.  And those adjustments occurred only because the Fed was able to adjust the base (not via a magic wand.)  And the base was 95% currency.  Furthermore, the vast majority of monetary economists believe the policy could have been implemented without reserve requirements, which is simply a tax.  In that case the base would have been over 99% currency.  Currency seems unimportant because actual operating procedures make the base change first, and then currency endogenously adjusts to a change in the base.  But prior to 1914 the base was 100% currency, and we could easily return to that system and still run monetary policy essentially the same way.  The only difference would be that the Fed would give banks ten $100,000 bills for a $1 million T-bond, not credit their Fed account for $1,000,000.  That’s a trivial difference.

Monetary policy is all about adjusting the stock of currency.