Scott Sumner  

The ECB is steering the economy (plus a survey of nautical metaphors)

Ask Me Anything Tonight... Diasporas, Swamping, and Open ...

Here's Paul Krugman suggesting that the ECB has been in a liquidity trap for years:

And yes, Europe is very much in a trap. Inflation is falling because the economy is weak, and the economy is being weakened in part by falling inflation. That's the Japan syndrome. It leads eventually to actual deflation, but to the extent that there's a red line (or more accurately, an event horizon), it's crossed when monetary policy starts being limited by the zero lower bound, which happened years ago.
On one level this makes no sense, as the ECB has been almost continually raising and lowering interest rates ever since it began in 1999. Indeed the ECB has cut rates many times in the past few years, and raised them several times in 2011.

I'd guess that Krugman is thinking of something else, the fact that risk free market interest rates in the eurozone are quite low, and hence the ECB might not be able to conduct a highly expansionary monetary policy even if it wanted to. I don't agree, but that's a defensible argument.

But that argument does not mean that fiscal austerity in the eurozone has reduced output. The reason is very subtle, so I'll start with a nautical metaphor. Let's assume a ship is going west, and something goes wrong with the steering wheel. It cannot be turned in a counterclockwise direction, and hence the captain cannot turn toward the southwest. But the wheel can move clockwise, and hence the captain can turn toward the northwest. Now assume that in 2011 the captain turns the wheel toward the northwest, and then occasionally nudges the boat this way or that. In that case the inability of the wheel to move in the counterclockwise fashion would not restrict the captain from going in the direction he chose, and hence would not be constraining the ships direction. Any wind that buffeted the ship would not throw it off course, as the captain would simply adjust the steering.

The ECB starting raising rates in 2011, from 1.0% to 1.5%. Those rates are still quite low, and it's possible that a rate cut to zero in 2011 would not have helped all that much. But that doesn't matter. The ECB was steering the ship, and hence the new Keynesian model applies. The fiscal multiplier was zero, as the ECB would have simply offset the impact of more fiscal stimulus with tighter monetary policy. It's revealed preference 101.

[Technically my metaphor requires a weird steering wheel that determines absolute direction, not direction relative to current course. But you get the idea.]

Nautical metaphors are an excellent way to learn monetary economics. Here are 7 more:

1. The Fed should steer the nominal economy between the Scylla of high inflation and the Charybdis of high unemployment (or better yet between 4% and 6% NGDP growth.)

2. Because of policy lags, many people believe steering the economy is like piloting a huge tanker. It responds slowly to a shift in the rudder. In fact, policy lags are much shorter than many people assume.

3. Monetary offset prevents fiscal stimulus from having an expansionary effect. If I am steering a ship and my daughter starts to push on the steering wheel, I push back with equal force. This keeps the boat on that path that I choose. My daughter (fiscal policy) doesn't get to choose the path for NGDP.

4. Fiscal and monetary policy are not equivalent ways of impacting nominal spending. Fiscal stimulus is costly, analogous to revving up an engine with more costly fuel being added (future distortionary tax increases.) Monetary policy is costless. Therefore it makes more sense to think of monetary policy as a setting of the steering wheel. It doesn't cost more to set the steering wheel at one setting compared to another. It's not "more" it's "different." Setting a NGDP target at 5% doesn't require any more costly fuel than 4%, for two reasons. You probably would not have to do any more open market purchases, as a higher NGDP target raises base velocity. And even if you did OMPs are essentially costless.

5. Nick Rose likes to point out that if you want higher nominal interest rates and higher NGDP growth, you have to cut interest rates in the short run. This would be like a ship with a steering wheel that had to be turned right to move the ship left.

6. The central bank should target the forecast, but very few actually do so. They should set policy so that expected NGDP growth equals target NGDP growth. Failing to do so is equivalent to a captain setting the steering wheel at a position where he expects to end up in Boston, even though the ship's destination is New York. So adjust the steering Captain Ben! You should expect to arrive at the place that you wish to arrive. BTW, an NGDP futures compass might help the captain.

7. The liquidity trap is a tough one. Interest rate targeting could be compared to a steering mechanism that works fine except when you need it most; it locks up in rough weather. Alternative mechanisms include QE, which would be like side jets that help steer the ship, or forward guidance. Forward guidance would be like shooting a long cable to the spot you want to arrive at, and then using a winch to reel in the ship.

OK, the last one is a bit far fetched. But hey, I grew up in Wisconsin and never saw salt water until I was 20 years old (at 3am while driving in Tampico, Mexico.) So I'm a bit rusty with some of the nautical terms.

Bonus analogies: ECB = Costa Concordia.

Pre-Abe BOJ = Sargasso Sea.

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (23 to date)
Nathan Smith writes:

Maybe sometime you could post on why to target NGDP rather than asset prices.

It seems like much of the success you claim for market monetarism over the past few years would be equally well achieved by targeting asset prices.

A Federal Reserve targeting asset prices would have tightened in the late 1990s to curtail (what turned out to be) the stockmarket bubble, and in the mid-2000s to curtail (what turned out to be) the housing bubble. That might have saved us a lot of trouble.

But from late 2007 or early 2008 it would have been highly expansionary, just what you want.

QE succeeded in large part because it reflated the stockmarket, no? Or at least, a Federal Reserve that wanted to get stockmarket and housing wealth back up to its long-run trend would have pursued QE aggressively?

Asset prices are more observable than future NGDP, and they can be managed directly. Politically, the dangers of speculative asset bubbles could be as good a cover as any for the Fed when it needs to tighten to rein in an unsustainable boom.

If you prefer NGDP targeting, why? And would monetary policy have been better over the past 25 years if "sustainable asset price growth" had been its top priority?

Effem writes:

@Nathan: Targeting stock prices makes no sense to me. If you look up the Kalecki profit equation you will see that consumers and the government essentially need to dis-save to generate ever-higher profits and therefore stock prices. Doesn't seem sustainable to me. To some degree that's what the Fed has done by constantly driving down rates and allowing people (and government) to increase leverage.

Matt writes:

On #3, maybe instead of using the daughter metaphor, you could say that fiscal policy is more like ocean currents. You simply assume that the captain would correct for them even though it's possible that they won't, and you can't really control them once you set sail. In other words, you can't really go back in time and elect someone else to president or congress but you can choose them in the long run (push off in summer (Democrats) or winter (Republicans)).

I'm sure you could refine that, but I think it would sound like less of a dig on Keynesians then calling them little girls.

Kenneth Duda writes:

Nathan: which assets do you propose for price targeting, and how do you propose picking the target prices for those assets?

NGDP is a marvelous thing to target because it is just one thing (no need to pick a bunch of weights for a bunch of different asset classes), it's easy to pick a suitable target (simply last year's target plus 5%), and it is so closely linked to actual economic activity. There are new types of assets being dreamed up all the time and their prices go up and down with the whims of the market. I don't see how targeting their price level could be helpful.


Kenneth Duda
Menlo Park, CA

Capt. J Parker writes:
Monetary policy is costless
I'm doubtful that that statement is always true. When the time comes for the Fed to raise rates and shrink its balance sheet it's quite possible that the Fed will ask Treasury to pony up to either increase IOR payments or cover losses on asset sales by the Fed.
Scott Sumner writes:

Nathan, A NGDP target does a much better job of stabilizing the labor market and the financial system as compared to a asset price target. Ideally the Fed should create a NGDP futures market.

Matt, Actually I didn't think of the little girl analogy as a dig at Keynesians, but I suppose it could be read that way.

J Parker, QE is a reflection of a failed monetary policy. If you use monetary policy to target NGDP you should never have to do large quantities of QE.

But I agree that massive QE could involve some costs. Even so they'd be trivial compared to the cost of fiscal policy.

Michael Byrnes writes:


There is no such thing as "not steering". The rudder isn't removed from the water any time the boat is on course.

There are aspects of the ship's voyage that are out of the captain's control, such as prevailing winds and currents. It is not the captain's responsibility to try to adjust these or even care about them, only to respond appropriately to them.

As a passenger on a ship, you don't care at all about the specifics of how the captain turns the wheel or the specific course the ship follows through the water. You care that the ship arrives at the correct destination at roughly the time it was supposed to arrive. If you arrive at the wrong place, you won't be mollified by the captain's explanation that his various turns of the wheel were consistent with past trips in whuch he did arrive in the right place, so what else could he possibly have done?

On cutting rates in the short run if higher long-term rates are desired, maybe a better analogy is a bus swinging wide left to take a right turn?

Yancey Ward writes:
NGDP is a marvelous thing to target because it is just one thing (no need to pick a bunch of weights for a bunch of different asset classes)

But it isn't just one thing- it is the aggregate of billions of things. The entire edifice rests on the assumption that providing, let's say NGDP futures markets, then allows the market to distribute that NGDP infusion into the billions of transactions that make up RGDP. I still have a problem seeing this as superior in principle to the Fed doing the infusion in a different way, like buying up Treasuries willy nilly, or MBSs.

Scott Sumner writes:

Michael, Excellent metaphors.

Yancey, I do favor having the Fed inject money by purchasing Treasuries, even if they use an NGDP futures market.

Don Geddis writes:

#5: you can get intuition for counter-steering, by thinking of a bicycle or motorcycle. You need to first change the distribution of the center of mass (by briefly turning the "wrong" direction), before getting on a consistent turning path.

Unfortunately, that's not a nautical example, so it doesn't fit in with the theme of your post. But I don't think boats use counter-steering, so the analogy kind of breaks down with #5.

Don Geddis writes:

@Nathan Smith: To choose a monetary policy, you need to start by figuring out what problem you are trying to solve. Market Monetarists would say that "the dangers of speculative asset bubbles" should not be a goal for monetary policy.

The real danger that they're worried about, is unanticipated fluctuations in the value of the Medium of Account. In other words, you agreed to a mortgage ten years ago, you're still paying it off, what is the current real burden of making those debt payments? If the real value of the MOA changes suddenly, then observation has shown that it causes unnecessary changes in real output and unemployment (probably because of sticky wages and prices).

The theoretical ideal would probably be to stabilize nominal (hourly?, average?) wages, but NGDP is very close to that, and perhaps easier to work with. Your idea, asset prices, is also correlated with NGDP, so it would be a lot better than most real central bank monetary policies. But in the ways that it is different, it is probably worse.

The point is, that stabilizing asset prices is not a direct goal of monetary policy (because it has few direct real consequences). The goal is to stabilize the value of the MOA, and/or national income (because not doing so causes real recessions).

Nathan Smith writes:

Scott writes:

"Nathan, A NGDP target does a much better job of stabilizing the labor market and the financial system as compared to a asset price target. Ideally the Fed should create a NGDP futures market."

This is a quick brush-off-a-commenter answer, which is fine. But I'll persist...

Why do we want to "stabilize" the labor market? If unemployment fluctuated a bit more, how bad would that be?

As for the financial system, it seems obvious that huge fluctuations in asset prices can destabilize it. The housing bubble inexorably lured the financial sector into speculative investment in ever-increasing housing prices. When the bubble popped, the banks were thrown into crisis. It's less obvious how NGDP fluctuations directly cause financial turmoil.

Don writes that:

"The point is, that stabilizing asset prices is not a direct goal of monetary policy (because it has few direct real consequences)."

Why would you think that asset prices have few real consequences?

High stock prices spur corporate investment via Tobin's Q. They also spur consumption as consumers get wealthier.

High house prices also spur consumption, see the refi boom in the middle of the last decade. And they trigger a boom in home construction.

And surely there's at least some truth in the conventional wisdom that our two great depressions, 1929-33 and post-2008, were a result of asset price bubbles, in stocks and houses respectively?

In the long run, NGDP targeting and asset price targeting don't differ much, since asset prices can't permanently grow faster or slower than GDP (at least, any trend in the K/Y ratio must be very slight on an annualized basis). But in the short run, the AD effects of people's wealth roller-coasting up and down seem at least as serious as the AD effects of sticky wages.

And targeting asset prices (maybe in the form of total net worth of all Americans) seems like the most direct way to block the speculative asset price bubbles that seem to cause most of the Really Big Crisis. (See Japan too.)

Nathan Smith writes:

A couple more points:

1. Let's assume that we'd typically prefer for any variable in the economy to be more stable, i.e., level or growing steadily, to dramatic fluctuations. Wages are pretty stable anyway. Asset prices fluctuate wildly. Does that suggest that the marginal value of stabilizing asset prices a little bit is greater than the marginal value of stabilizing wages a little bit?

2. Some have noted the danger of "fine-tuning" in monetary policy, i.e., it's hubris to think you can get the target variables to perform exactly the way you want them to. Because asset price fluctuations are bigger, there's more room to manage them without fine-tuning. It was obvious in the late 1990s that stock prices were well above trend, so the Fed could have reined them in without pretending to be able to fine-tune them to an exact target. It was obvious in the mid-2000s that housing prices were above trend, so the Fed could have reined them in without attempting to fine-tune.

3. Asset prices are a crude proxy for future NGDP anyway, since stocks reflect long-run dividend expectations and house prices reflect long-run expectations about buyers' purchasing power. An NGDP futures market would be small and could be vulnerable to bubbles and/or manipulation, which would be disastrous if Federal Reserve policy were automatically determined by it. Asset prices in the economy as a whole are too big a variable to be manipulated, and their sheer size is some protection against bubble dynamics.

Michael Byrnes writes:

Nathan Smith wrote:

"Why do we want to "stabilize" the labor market? If unemployment fluctuated a bit more, how bad would that be?"

I think "stabilize the labor market" is sort of an unfortunate shorthand. What proponents of NGDP targeting really want is to avoid (or minimize) nominal shocks that disrupt the labor market. Problems that startwith money and end with millions of people involuntarily out of work.

Scott has a great example of a "stable" labor market: the labor market in the immediate aftermath of the housing construction bust (2006-2007). Home construction crashed from record levels to its lowest level in decades, many construction workers who had been employed building homes were no longer employed building homes... and unemployment ticked up from 4.7% to 4.9%.

That's a stable labor market, in the sense that Scott Sumner wants it to be stable. A labor market in which jobs that are lost are offset by jobs that are gained, and there is plenty of turnover (turnover falls precipitously during recessions).

Don Geddis writes:

@Nathan: I agree with Michael Byrnes (above). More details: "It's less obvious how NGDP fluctuations directly cause financial turmoil." The public's demand for money rises, the Fed doesn't accommodate this additional demand, thus money becomes more valuable, thus the value of the MOA rises, thus the real burden of paying wages and paying debts rises, but unfortunately the labor market (and some others) is not very liquid, and so sticky wages prevent the labor market prices from adjusting rapidly to the new value of the MOA. So instead, as the labor market prices are out of equilibrium, there is an excess supply of labor (unemployment), and a lack of demand for labor (no jobs). Which results in an unnecessary drop in real output, aka a self-inflicted recession.

"And surely there's at least some truth in the conventional wisdom that our two great depressions, 1929-33 and post-2008, were a result of asset price bubbles, in stocks and houses respectively?" No, there's essentially no truth in that "conventional wisdom". At best, your "bubbles" were the initial spark for the subsequent economic turmoil. But the economic damage itself, in both cases, was because the Fed did not maintain a stable value for the MOA (or NGDP).

"speculative asset price bubbles that seem to cause most of the Really Big Crisis" Asset price "bubbles", by themselves, cause no economy-wide crisis. The stock market crash in 1987 was on the same order as in 1929. But the overall economy hardly noticed in 1987, because the Fed kept NGDP on target (unlike in 1929).

Scott Sumner writes:

Nathan, There is not a shred of evidence that the Great Depression was caused by asset price bubbles. In fact, it was caused by tight money (a 50% fall in NGDP.) A very similar stock crash occurred in 1987 and there wasn't even a tiny recession, because the Fed kept NGDP growing. There are good theoretical reasons why NGDP instability causes business cycles (and financial crises), and no good theory explaining why asset price instability would cause a business cycle if NGDP was stabilized. The correlation between NGDP instability and cycles is much better than the correlation between asset bubbles and recessions. Central banks have no way of knowing the "proper" level of asset prices. Last time the Fed tried to pop a bubble was 1929, and it didn't go well.

Regarding NGDP futures, I have a paper at Mercatus that addresses many of the criticisms. But if no market exists, I have no objection to using market indicators like stocks and TIPS as tools for predicting NGDP. To the extent that asset price fluctuations reflect expectations of unstable NGDP, by all means stabilize asset prices. That's why we are called "market" monetarists. The Fed cut rates after the 1987 crash, which seems appropriate. Perhaps they should have raised them earlier.

I believe labor market instability is the biggest macro problem, far larger than other cyclical problems. Of course there are also real problems, and I am a moderate supply-sider in terms of addressing those real problems.

Michael Byrnes writes:

If the US government started issuing "trills" (as suggested by Shiller), would that serve the same informational function as a futures market?

Nathan Smith writes:

re: "Nathan, There is not a shred of evidence that the Great Depression was caused by asset price bubbles. In fact, it was caused by tight money (a 50% fall in NGDP.)"

Yes, that's one view. The question has been much debated. Obviously causation is complicated, and the popping of an asset price bubble could cause a fall in NGDP via wealth effects, Tobin's Q, etc.-- all the stuff I said. Indeed, you seem to suggest the same thing yourself a couple of sentences later.

"A very similar stock crash occurred in 1987 and there wasn't even a tiny recession, because the Fed kept NGDP growing."

Which is exactly what the Fed would have done if it were targeting asset prices. Asset prices fall, Fed loosens monetary policy.

"No good theory explaining why asset price instability would cause a business cycle if NGDP was stabilized."

Tobin's Q and wealth effects explain why asset price instability could cause business cycles. The "if NGDP was stabilized" clause is unfair. Part of the point is that that may be very hard to do, and sustainable asset price growth might get you the same good effects while being easier to implement.

1929 is obviously not a data point against "sustainable asset price growth" as a monetary target, because asset prices were not growing sustainably, but plunging. Obviously, if the Fed had been targeting sustainable asset price growth in 1929-33, it would have pursued a highly expansionary monetary policy in order to reflate stock and housing prices. Which is pretty similar to the market monetarist recommendation.

Off the top of my head, I can't think of a plausible example where NGDP targeting and sustainable asset price growth targeting diverge dramatically. For the record, though, you're a much more accomplished macroeconomist than I am, and I'm willing to place a relatively low subjective probability on sustainable asset price growth being the best approach, in mere deference to your intellectual authority. For now.

Don Geddis writes:

@Nathan: "Part of the point is that [stabilizing NGDP] may be very hard to do" You haven't presented any evidence that it is difficult, and MMs certainly believe that it would be relatively straightforward.

Your only argument for choosing to stabilize asset prices instead of NGDP, seems to be that you imagine it would be "easier to implement". But this isn't an important problem to solve, because NGDPLT is already easy to implement.

Meanwhile, you haven't suggested any additional benefit to targeting assets (assuming that NGDP is stabilized). And we know (from theory) why NGDP volatility leads to real economic recessions. Why not stabilize the variable that's actually of interest? What's your attraction to this second best, "almost the same" alternative target?

Nathan Smith writes:

Thanks Don. As for evidence that sustainable asset price targeting is easier to implement, in a crude sense this is very obvious. What's the first thing that moves when the Fed announces a policy? Interest rates and the stock market, i.e., asset prices. The effect is immediately seen. NGDP is slower to move and is observed with a much longer lag. There's more to the story than that but that's a first pass.

When you say that we know from theory why NGDP volatility leads to real economic recessions, I think you're referring to downward nominal wage stickiness. If NGDP falls below trend, that causes a recession because wages don't adjust properly, leading to unemployment. That's theory. But theory also suggests that when asset prices are high, consumers' intertemporal budget constraints are shifted out, so they should spend more. And firms should invest more, since (Tobin's Q) the market is placing a higher value on existing capital. It's more expensive to buy competitors, so build your own capacity. It's a good time for IPOs. Etc.

You say "why not stabilize the variable that's actually of interest?" But asset price volatility does have welfare costs, real effects. House prices rise, you buy a house on spec, house prices fall, you're underwater. You can't sell. You want to move but can't. Mobility falls. Or you walk away from the house, and the bank has to foreclose. Major real damage happens as you don't mow the lawns, bills go unpaid, the neighborhood starts looking junky. Again, suppose you planned to retire, then your portfolio crashes in value. You have to keep working. If you'd known what would happen, you'd have planned differently. How do these costs compare to the downsides of a little more unemployment? Remember, there are "efficiency wage" stories about the *upside* of unemployment.

It would be very odd to deny that the housing bubble (as it turned out) caused a lot of real hardship for people, and was the major cause of the financial crisis of 2008. Monetary policy could have prevented that. Could NGDP targeting have prevented it? It's not obvious how, though maybe it would have made it less damaging. You can say "Bubbles are hard to identify," but that misses the point.

Suppose the Fed thinks there's an asset price bubble, but it's wrong. Stocks are soaring, not because there's a bubble, but because there really has been a fundamental revolution in productivity growth that justifies the higher prices. The Fed tightens monetary policy to stop the asset price bubble. Consumer price inflation goes to zero or negative. Maybe NGDP growth even slows down a bit. But, *ex hypothesi,* asset prices are still high. That can only mean a surge of growth in the future. Downward nominal wage stickiness raises real wages and causes job cuts in stagnant industries. That's great! Labor gets reallocated to fast-growing sectors. Workers still in stagnant sectors get higher wages, they get to share in the boom too. The Fed guessed wrong about the asset price bubble, but that doesn't lead to disaster, it's just what the doctor ordered.

Nathan Smith writes:

By the way, I've become more heterodox out of disillusionment with the "Great Moderation." For twenty years we were quite successful in keeping unemployment and inflation low. What did it get us? Ever-increasing indebtedness, unsustainable growth of leverage, a bloated financial sector, asset price bubbles that lured people into making a lot of bad life decisions, growing inequality, stagnant median incomes, the Great Stagnation. Of course, it wasn't a disaster like the 1930s, but deference to the conventional wisdom of ten years ago doesn't seem to be the most appropriate attitude to take at this point.

Michael Byrnes writes:


That's a good case for why we should consider abandoning flexible inflation targeting for NGDP level targeting. :)

TallDave writes:

Forward guidance would be like shooting a long cable to the spot you want to arrive at, and then using a winch to reel in the ship. OK, the last one is a bit far fetched.

Actually found this the most apt. Expectations uber alles!

Comments for this entry have been closed
Return to top