Scott Sumner  

How interest rates matter, and how they don't

PRINT
Krugman's Bait and Switch... Consider the Substance...

There are certain topics that are extremely hard to explain to the average person. Interest rates are one of those topics. I've often argued that interest rates don't really matter that much; they are a sort of epiphenomenon of monetary policy. I've also argued that the ECB adopted a tight money policy in 2011, and that this triggered a double-dip recession. And what's the easiest way for the average person to visualize the ECB's tight money policy? The fact that in 2011 they raised their interest rate target from 1.0% to 1.25%, and then to 1.50%. But if interest rates don't matter very much, how could mere 50 basis point increase cause a recession, especially given that they soon started cutting rates again?

One problem is that most people think we are always in the short run. No matter how many times you teach students that tight money raises rates in the short run (liquidity effect) and lowers them in the long run (income and Fisher effects), when the long run actually comes around they will still see the fall in interest rates as ECB policy "easing". And this is because most people think the term "short run" is roughly synonymous with "right now." It's not. Actually "right now" we see the long run effects of policies done much earlier. We are not in an eternal short run. That's the real problem with Keynes's famous "in the long run we are all dead."

I'm going to try to explain why even a 1 basis point increase in the interest rate target could cause a Great Depression, and then afterwards explain why interest rates actually aren't very important at all. Bear with me.

In most macro models, interest rate pegging is a knife-edge equilibrium. If the rate is pegged slightly below the natural rate, the economy will spiral up into hyperinflation. If slightly above the natural rate we'll spiral into hyperdeflation. That's because when you peg the interest rate above or below the natural rate, the resulting movement in the economy starts pushing the natural rate further and further away from where you are pegging interest rates. You gradually get further and further off course.

Perhaps an auto steering analogy will help. Suppose you set off west across the Bonneville Salt Flats. You try to lock the steering at a position where you will go straight west, but the wheel is accidentally set 1 millimeter off center. After 100 yards the car has drifted an inch to the right. That's not too bad, it seems like you won't end up too far off course. But here's the problem, the error will gradually get worse and worse. After 200 yards you'll be 3 or 4 inches off course, and after that the car will rapidly deviate further and further to the right. The path will be a gradual curve to the right, and before very long you'll be headed straight north. (Don't ask for details, I took calculus 40 years ago.)

Why don't we see real world central banks make even more errors than they do? Because they usually adjust the steering as they see the economy go off course. They adjust their interest rate peg up or down (and do QE at near zero interest rates.) That's like a driver nudging the wheel a bit to the left or right as he observes the car drift slightly off course. It's easy for drivers to do, and it's easy for central banks to do. When they produce bad policy (as the ECB did in 2011) it's because they were trying to produce the policy we now judge as bad---they were trying to reduce eurozone AD to bring down inflation.

Over at MoneyIllusion a commenter thought I'd disagree with this analysis by Brad DeLong:

Taylor says the Federal Reserve kept interest rates two percentage points to [sic, 'too'] low for three years. If you are buying a long-duration asset like housing, such an interest rate break leaves you willing to pay 6% more than you would otherwise have been willing to pay. Are we really supposed to build a 50% nationwide housing bubble on top of the 6% impetus? Only a market already fully infected with the bubble disease could see such a small impetus have such a large impact.
I actually agree with DeLong. Although a tiny change in interest rates might cause the money supply to gradually change in such a way as to create rapid NGDP growth, the actual policy done by the Fed did not do that in the early 2000s, at least no more so than in the 1950s, 60s, 70s, 80s, or 90s. So while in retrospect policy may have been a bit too expansionary, it didn't miss by all that much. And yet the house price boom was much more dramatic than during more inflationary decades.

Now perhaps interest rates worked on house prices through some channel other than NGDP growth. In that case I'd point to DeLong's argument, the interest rates were not off base enough to explain such a dramatic price increase. If you are looking at interest rates and home prices from a microeconomic perspective, then you throw out the knife-edge equilibria models, and use common sense. Oddly, while a one basis point interest rate increase can cause a Great Depression in a macro model, it's very hard to see how even a 200 basis point rate cut lasting three years can boost home prices by 50% in a microeconomic model.

Here's another way of making the same point. Monetary policy matters because money is the medium of account, the driver of NGDP. If the government were able to reduce interest rates via a non-monetary mechanism (say a smaller budget deficit) then you'd expect a far less dramatic impact on the economy, as compared to Fed policy moving rates around. With Fed policy it's the change in the monetary base (relative to base demand) that really matters, interest rates are simply the signaling device that people focus on.

The 2011 ECB tightening mattered because:

1. The ECB has a monopoly on the euro base, and hence controls the path of AD.

2. The two rate increases were a signal they wanted to slow the growth in AD.

3. AD growth did in fact slow sharply.

Any two of those would not be enough, but combining all three we can assume that the ECB was the cause of the sharp slowdown in AD growth after 2011. The size of the rate increase is totally irrelevant. The eurozone didn't have a double dip recession because slightly higher rates discouraged people from borrowing, but rather because the ECB caused eurozone M*V growth to slow sharply.


Comments and Sharing






COMMENTS (25 to date)
foosion writes:

Many members of the Fed want to raise rates within the next few months to head off inflation (sounds like the ECB) or because higher rates are more normal or because they don't like workers getting raises.

Will this increase cause AD growth to slow?

If AD growth slows following an increase by the Fed, can we assume the Fed was the cause?

Ray Lopez writes:

I don't buy this logic. If country A never raises interest rates, and country B raises interest rates by 50 basis points, then lowers them again (as the ECB did for a period in 2011), and both country A and B end up, in 2015, with the same interest level, Sumner's logic would mean that country A has tight money and country B has easy money? Incredible, as in not credible.

In other news today: "The European Central Bank’s imminent bond-buying plan has left $1.9 trillion of the euro region’s government securities with negative yields." Tight money?

Kevin Erdmann writes:

I don't understand why everyone is so obsessed with short term rates and their effect on home prices. It's long term real rates that matter, and the Fed doesn't have much influence on those, short of creating demand shocks that create a sense of hopelessness.

My latest housing post addresses this. There is no mystery to solve.

Andrew_FL writes:

It hasn't been so long since I've taken calculus, but since you specified the distance the steering wheel is off-center by, rather than the degrees, I'd need to know the radius of the steering wheel to solve the problem.

"So while in retrospect policy may have been a bit too expansionary, it didn't miss by all that much. And yet the house price boom was much more dramatic than during more inflationary decades."

The problem with thinking about it like this-if loose monetary policy was responsible for the housing bubble, why doesn't loose money always cause a housing bubble?-Is that the "transmission mechanism" isn't literally a machine that produces the same result every time. If the kind of investments that turn out to be malinvestments were the same every time during a monetary expansion, it'd be trivial for entrepreneurs to short circuit the business cycle.

But in reality, the specific effects of monetary expansion depend upon a variety of factors. Probably it'd be fairly typical to see a "bubble" form in an area that would be expanding anyway for real (ie non monetary) reasons. That's precisely why it fools people, and why afterwards it's possible to say some of the growth looks justified for non-monetary reasons.

So for example in the 19th Century, you might get booms and busts concentrated in railroads, or land. In the 1990s, a dot com boom and bust. You wouldn't see a railroad boom and bust today (well, unless the government really got on a high speed rail kick).

This also gets to a common problem. The usual definition of a bubble involves a process which is irrational and self reinforcing. The actual boom and bust, while policy induced and completely unnecessary, is rational and self reversing.

James writes:

Scott Sumner:

If the overnight market is in equilibrium, then interest rates reflect the marginal rate of substitution between nominal dollars now and nominal dollars later. I'm sure you know this. Do you think there are any other markets, financial or otherwise, which have this property, that marginal rates of substitution are not very important?

Kevin Erdmann:

Short term rates matter because they are economically linked to long term rates. If they are not linked, then an arbitrage exists. If you have a macroeconomic model which implies an arbitrage for anyone who is aware of the model, you are either sitting on a goldmine or you are wrong.

Kevin Erdmann writes:

James, real long term rates were similarly low in the late 1970s and the 2000s & 2010s during very different fed funds rate regimes. How does your arbitrage work?

Scott Sumner writes:

foosion, It's hard to say, interest rates are not a good indicator of the stance of monetary policy.

Andrew, In my view expansionary monetary policy always has the same effect, rising NGDP, mostly because I define easy money in terms of NGDP growth.

Regarding the railroad booms, it's worth thinking about the fact that those occured under the gold standard.

James, I'm not quite sure what you are asking, but let me say that interest rates are important in some respects but not others. My point is that they are not very important for aggregate demand. As an analogy, the price of zinc is very important for the zinc market, but not for AD.

Don Geddis writes:

@Ray Lopez: "country A ... country B ... interest rates ... end up ... with the same interest level ... country A has tight money and country B has easy money? ... bond-buying plan ... negative yields ... Tight money?"

Since interest rates and bond yields can be associated with either tight or loose money, your hypothetical has no single answer. You haven't provided enough information (namely, what happened to aggregate demand) to answer the question about whether money was tight or loose in your scenario. It could be either.

You really need to read Sumner's OP again. See in particular the critical points #2 and #3 under "the 2011 ECB tightening mattered because", which aren't mentioned in your example.

Ray Lopez writes:

@Don Geddis - your points don't refute mine. I am questioning the framework as nonsense. Sumner writes: "2. The two rate increases were a signal they wanted to slow the growth in AD. 3. AD growth did in fact slow sharply."

So, taken to the extreme, you can simply say: if country B raises interest rates by 50 basis points, and suffers from a double-dip recession, unlike country A which left rates alone, a monetarist would say that this resulted from country B responding to the "signal" from the central bank that it intended to tighten money. Fair enough. But if country B fails to suffer a double-dip recession, then the "signal" was not credible, and not responded to, hence no recession (would say the monetarist). Can you see how illogical this thinking is? And how unscientific and untestable it is? Akin to the scholastic metaphysics of the medieval ages.

Andrew_FL writes:

@Scott Sumner-Scott, I'm really not sure how to interpret a statement like that. Is NGDP (high) growth the effect of loose monetary policy or is it itself loose monetary policy? If the latter, you've either denied that loose money does anything, other than just sort of...exist; or plead ignorance as to the specifics.

I'm not surprised, otherwise, that you'd think the effect should be the same every time. I just don't agree. But better people than me have tried for way too long to settle that argument so there's really no point in rehashing it.

foosion writes:

Scott, I postulated essentially the three factors that caused you to conclude the ECB was the cause of the slowdown in 2011. What's the difference between the Fed case and the ECB case?

foosion writes:

Scott, did you see this paper http://econweb.ucsd.edu/~jhamilto/USMPF_2015.pdf?

We examine the behavior, determinants, and implications of the equilibrium level of the real federal funds rate, defined as the rate consistent with full employment and stable inflation in the medium term. We draw three main conclusions. First, the uncertainty around the equilibrium rate is large, and its relationship with trend GDP growth much more tenuous than widely believed. Our narrative and econometric analysis using cross-country data and going back to the 19th Century supports a wide range of plausible central estimates for the current level of the equilibrium rate, from a little over 0% to the pre-crisis consensus of 2%. Second, despite this uncertainty, we are skeptical of the “secular stagnation” view that the equilibrium rate will remain near zero for many years to come. The evidence for secular stagnation before the 2008 crisis is weak, and the disappointing post-2008 recovery
is better explained by protracted but ultimately temporary headwinds from the housing supply overhang, household and bank deleveraging, and fiscal retrenchment. Once these headwinds had abated by early 2014, US growth did in fact accelerate to a pace well above potential. Third, the uncertainty around the equilibrium rate argues for more “inertial” monetary policy than implied by standard versions of the Taylor rule. Our simulations using the Fed staff’s FRB/US model show that
explicit recognition of this uncertainty results in a later but steeper normalization path for the funds rate compared with the median “dot” in the FOMC’s Summary of Economic Projections.

Bob Murphy writes:

Scott, as usual I think I get the general gist of your argument, but I have problems. Specifically this part:

=========
The 2011 ECB tightening mattered because:

1. The ECB has a monopoly on the euro base, and hence controls the path of AD.

2. The two rate increases were a signal they wanted to slow the growth in AD.

3. AD growth did in fact slow sharply.

Any two of those would not be enough, but combining all three we can assume that the ECB was the cause of the sharp slowdown in AD growth after 2011.
============


You say any two of those would not be enough. OK so if the following two conditions held:

1. The Fed has a monopoly on the dollar base, and thus controls the path of AD.

2. Growth in AD from 2008 onward did in fact slow sharply,

...but we didn't see rate increases from the Fed, then you wouldn't be able to conclude that the Fed had caused the slowdown in AD growth?

TomM writes:

@Ray Lopez

Don is right. I think where your example is incomplete is the fact that we have no information about NGDP growth in either country.

If Country A had NGDP consistant NGDP growth- lets assume 5% and thats the objective of the central bank, then there would be no need to increase the MS.

If Country B had NGDP growth below there target- say 4% then pursued a contractionary policy, then yes I would assume markets would react by assuming the Central bank wanted lower NGDP growth. Once the central bank realized there error, and pursued more expansionary policy- and growth returned to 5%.

However, if country B had NGDP growth ABOVE there target- say 6% and then pursued a contractionary policy, then I would assume they would simply be adjusting to hit there 5% goal- no recession, just slower NGDP growth.

In both scenarios country B has pursued contractionary policy and there are two different outcomes.

Ray Lopez writes:

@TomM - it's 'their' not 'there'. But there's a bigger problem with your example; it fails to account for when markets do NOT perform as the central bank assumes. Surely you don't believe that NGDP is always and everywhere a target that can be set by a central bank (cb), do you? Do you? Only a real diehard MM believes that. So the question remains: if a small interest rate rise by a cb does not lead to a double-dip recession, given Sumner's scenario re the EU and the ECB, then what can we say? We say that the rise in rates was 'not credible' in the eyes of the market. But this is nonsense I posit, akin to metaphysics.

@ Ray Lopez writes:

1) It absolutely is a target that can be set by a central bank. Although shocks can disrupt the path in the short run, the central bank absolutely has the ability to return NGDP growth to trend.

If the trend growth is 5% even if rGDP is -4%, a CB could raise inflation by 9% to hit that target in the short run. Whether or not it would, or that would be the optimal policy is irrelevant. They absolutely could do it.

2) You are simply talking past the point because you have not established the scenario. If that small rate rise did not lead to a double dip recession it could be for a million reasons if you have not outlined the scenario...

The fact that the rate hike DID cause a double dip recession indicates a scenario similar to:
1) NGDP in Europe is below trend
2) Markets expect the ECB to provide more expansionary policy to get back to trend
3) ECB provides more contractionary policy
4) ECB has signaled to markets that it is comfortable with below trend growth or that growth is not a concern to them.

Scott Sumner writes:

Ray, You are confusing moves in NGDP (monetary policy) with moves in RGDP (effects of monetary policy.)

Andrew, Ideally we'd have a NGDP futures market, and we'd use that price as the monetary policy indicator. Unfortunately we lack that market, so I often look at actual NGDP as an indicator of the stance of monetary policy. So do other economists like Ben Bernanke, so it's certainly not just my idea.

foosion, Sorry I misread your post. Yes, if AD slows we can assume the Fed was the cause. I originally thought you were asking if we could assume AD growth would slow.

Thanks for the link, I'll take a look, but the growth of the past year is not at all impressive for a recovery year.

Bob, Good point, those two would be enough. With the three you'd also know the effect was intentional.

bill writes:

Slightly off topic. In a post today called "Slippery Slope of Disinflation", Paul Krugman says:

.....zero lower bound (which isn’t as binding as everyone thought, but there are still limits to rate cuts)

I know quite a few people that Krugman has just assumed out of existence. I always thought we were somebody.

Ray Lopez writes:

@ the anonymous poster 'at Raylopez': "If the trend growth is 5% even if rGDP is -4%, a CB could raise inflation by 9% to hit that target in the short run. Whether or not it would, or that would be the optimal policy is irrelevant. They absolutely could do it. "

That seems to be the heart of our disagreement. What evidence (cite please) leads you to think a central bank (CB) could "absolutely" raise NGDP? If money is superneutral, how could a CB do this? After all, in today's zero lower bound environment, a lot of economist were fooled that printing more money and taking on more paper by CBs would not result in more inflation, as it has not. In short, CBs seem to be pushing on a string, contrary to your 'absolutely' qualifier.

Jack Davis writes:

Kevin Erdmann is correct when he says short term interest rates have little impact on long-term rates.It's somewhat Disappointing that so few supposed expert commentators on the business channels mention this.
See this paper for more detail:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2031729

As it points out, the FFR went up from 2004-2006 but the long-term rate went down.

A writes:

Does this article imply that the Fed could raise the target rate today, and simply communicate that they are doing so because of a satisfactory inflation rate? The NGDP effect should be neutral, assuming steady RGDP, unless the liquidity effect is high enough to "test" the Fed's true zone of acceptable inflation.

A writes:

Does this article imply that the Fed could raise the target rate today, and simply communicate that they are doing so because of a satisfactory inflation rate? The NGDP effect should be neutral, assuming steady RGDP, unless the liquidity effect is high enough to "test" the Fed's true zone of acceptable inflation.

Bill Woolsey writes:

Ray:

Superneutrality means that changes in the growth rate of the quantity of money impact nominal variables only--the inflation rate, the growth rate of money wages, the nominal interest rate, and so on. It doesn't impact real wages, real output or real income, real interest rates or any other real variable.

Real variables are variables corrected for inflation. For example, real wages is the amount of goods and services workers can by for every hour they work. The real interest rate is the amount of real purchasing power, goods and services, borrowers must pay lenders.

Since nominal GDP is a nominal variable, superneturality implies changes in nominal GDP. The growth rate of nominal GDP would be in proportion to the growth rate of the money supply. For example, if the growth rate of the money supply when up from 2% to 3%, then nominal GDP growth might go up from 3% to 4%.

If money is neutral, then if the quantity of money rises once and for all by 10%, then all nominal variables will rise 10%, including nominal GDP. No real variables would change.

There are reasons why persistent expected changes in the growth rate of the money supply will impact real variables while a one time change in the money supply will not.

By the way, neutrality and superneutrality would usually be considered a good thing.

You seem to be using "superneutrality" to mean that the quantity of money has no effect on anything. That isn't what it means at all.

You really should lean some economics.

Major.Freedom writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring your comment privileges. This is your final notice. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Comments for this entry have been closed
Return to top