Nick Rowe on Interest Rate Control

Nick must have come back from his American adventure even smarter than when he left, as he has just posted an excellent essay on the problem with relying on interest rate targets.  First I’ll point out a few passages that I particularly liked, and then I’ll provide my thoughts on a few questions that he raised at the end.  Let’s start with how he concludes his discussion of the classical dichotomy.

If you take this homogeneity insight, and add the assumption that the supply of money is exogenous, you get the Quantity Theory of Money (a change in the supply of money will cause an equi-proportionate change in all nominal variables), and the Neutrality of Money (a change in the supply of money will affect no real variable).

Post-Keynesian horizontalists (and we are all horizontalists now, unfortunately, because that’s the underlying problem) reject the Quantity Theory because they reject the assumption that the supply of money is exogenous. But that misses the point. A revised Quantity Theory can be re-formulated taking any nominal variable as exogenous — the price of gold, or nominal GDP futures, for example. The homogeneity insight does not depend on any definition of money supply being exogenous.

The term ‘horizontalists’ refers to the horizontal line for an interest rate peg in a money supply and demand diagram.  I think that Nick’s posts complement mine very well.  Sometimes I lack the background in various theoretical perspectives, and thus rely on making my points through long thought experiments and historical examples.  Nick very quickly gets to the heart of one problem with Post-Keynesianism, which I was implicitly making in my long post on George Warren and FDR’s dollar depreciation program.  But I don’t always have the vocabulary to make the point in the concise theoretical language that economists like to use.

[BTW, Robin Hanson gently chided me (correctly) for the length of my posts.  Nick did the same so subtly that I was probably the only one who noticed.  Read his piece, glance at the two links to my blog, and you will see what I mean.]

Then Nick has a very nice analogy for Taylor Rule-type policies:

Interest rate control is like riding a bicycle. You can’t keep the steering fixed and expect to stay upright. You need to keep moving the steering, and move it faster than your tendency to fall over, if you want to stay upright. And also like a bicycle, you need to steer left if you want to turn right. If you want higher nominal interest rates you first need to lower interest rates, so that inflation starts to rise, and expected inflation starts to rise, at which point you can raise interest rates, and raise them higher than originally, so that inflation and nominal interest rates eventually settle down at some new higher level.

It took me a moment to remember how I used to ride a bike.  And I’ll bet even many economists might have to think twice to understand that in order to raise rates you must first lower them.  BTW, this is more than just a vague hypothesis—it is confirmed every time a dramatic Fed announcement makes short term rates move in the opposite direction from long term rates.  Then Nick raises the following question, and provides one answer:

But there have been many times in the past when central banks needed to lower interest rates, because they saw disinflationary pressures, indicating that the actual rate must have been above the natural rate. Why did it not always end in tears?

The financial crisis must have been the final straw that exposed the inherent instability of interest rate control. When the financial system is fragile, and bits of it break, the natural rate moves more quickly than normal, and the financial system also becomes more fragile when the actual rate is above the natural rate. Falling expected inflation, and falling expected real activity, weaken an already fragile financial system, and this lowers the natural rate by more than it would with a robust financial system.

You may recall that I once argued that the Wicksellian natural rate was quietly falling during the long period from April to October, 2008, when the Fed held its target rate at 2%.  I agree with Nick’s interpretation, but in addition I’d like to put in one more plug for my NGDP rule.  During the first nine months of 2008 we saw some really high CPI inflation from the commodity price boom (especially oil.)  This naturally led inflation-phobic central banks to hold back, despite the worsening financial situation.  But suppose inflation is the “wrong” nominal variable.  Suppose macroeconomic stability requires NGDP stability?  NGDP was only growing about 3% during that 9 month period, well below the norm of 5%.  In that case the Fed would have had ample justification for moving much more aggressively in the 3rd quarter.  So I agree that the financial crisis was part of the problem, but the concurrent oil shock helped create a perfect storm.

You have heard me argue that the stock market crash of early October 2008 (23% in 10 days) was a sign of investors losing confidence in the Fed’s ability to prevent a sharp fall in NGDP.  Here is what Nick has to say about policy credibility:

We knew that implementing monetary policy via an interest rate instrument was an unstable control-mechanism in theory, but it seemed to work well enough in practice, so we shrugged our shoulders and kept doing it. But it only worked because we trusted it to work. There were always two equilibria: one with validated trust in which prices were determinate and fairly stable; a second with validated mistrust in which prices were indeterminate. We were lucky to be living in the first equilibrium. The financial crisis eliminated that first equilibrium. It will be hard to go back to it, even when the financial crisis is over.

We need to switch to some other instrument for monetary policy, one which does not cause that inherent instability. The homogeneity insight tells us that any nominal variable (anything with $ in the units) will work in principle, and make the price level determinate. But some will work better than others. I am not advocating a return to base control, or the gold standard. Though each of these would make the price level determinate, the equilibrium real money base, and real price of gold, are too variable. I’ve written a long enough post, so I am going to leave that second question unanswered.

BTW, what do you mean “long enough,” I’d call that a short post.  Seriously, the question Nick raises here naturally leads right into futures targeting.  You want something as controllable as the price of gold, but with a much closer connection to your macroeconomic objectives.  Why not NGDP futures contracts?


1. Some of you have seen Mankiw’s recent NYT piece discussing a Harvard grad student’s proposal for interest penalties on currency.  I commented on this earlier.  For those of us in the monetary “underground” there is nothing new about this sort of idea.  My fear was that turning Federal Reserve notes into a giant Keno game wouldn’t even come close to political viability, and hence might lead many to also discard the idea of interest penalties on reserves only.  Looking at the bright side, perhaps this will indirectly create interest in more politically achievable policy options, such as the excess reserve penalty rate idea we have been discussing in this blog.  And if his idea is politically feasible, then obviously futures targeting (another foolproof escape from liquidity traps) would be even more politically feasible.

While I don’t know exactly why Mankiw didn’t opt for my reserve-only idea, I believe it might have something to do with the interest rate focus of new Keynesians.  My proposal would not reduce free market nominal rates significantly below zero, as cash in safety deposits boxes would still be an option.  However the proposal would almost certainly reduce cash hoarding by commercial banks, and thus make QE much easier.  Many Keynesians overlook this (monetarist) excess cash balance mechanism.  Repeating what I said in my first Nick Rowe post (regarding Laidler), right now we need fewer bright young grad student theoreticians and more experienced pragmatists.

2.  Those who read my earlier post on utilitarianism might find my frequent use of thought experiments in monetary economics to be inconsistent with my criticism of using such experiments as a tool for evaluating moral issues.  But my actual argument was slightly different.  I argued that if there were no real world examples corresponding to a particular philosophical thought experiment, it might have little or no pragmatic value.  There are many real world examples of extreme deflation, zero interest rates, and hyperinflation.  My thought experiments are pedagogical devices to illustrate insights that I have already developed by integrating monetary theory and history.



14 Responses to “Nick Rowe on Interest Rate Control”

  1. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    19. April 2009 at 09:50

    Not on topic, but in the Responsible Liberals comments, Scott opined that Jim Glass should have his own blog. Well, he does:

    And, it is always interesting, and often amusing.

  2. Gravatar of Nick Rowe Nick Rowe
    20. April 2009 at 04:33

    Thanks very much Scott!

    Here’s a simpler analogy to my bicycle analogy (since most of us ride a bicycle without ever thinking about how we do it):

    Here’s a simpler analogy. Targeting (say) inflation by controlling an interest rate instrument is like balancing a pole upright in the palm of your hand. The position of the top of the pole represents inflation, and the position of the bottom of the pole (the palm of your hand) represents the interest rate.

    First it’s an unstable equilibrium. If you keep the palm of your hand fixed, the pole will fall over. You have to keep moving your hand to keep the pole upright.

    Second, if you want the top of the pole (inflation) to move north, you have to first move your hand south, so the pole starts to lean north. You then need to move your hand north, to stop the pole falling over.

    If it’s a long heavy pole, so only leans slowly (inflation inertia), and if you can move quickly, and anticipate gusts of wind, or observe quickly which way the pole is leaning, you can keep it upright, and keep the top of the pole roughly where you want it to be.

    But there’s a wall to your south which you can’t go beyond (the zero interest floor). And your hand is up against that wall, and yet the pole is leaning to the south. You want to move your hand south, to start the top of the pole moving north, but you can’t.

    Solution: turn the damned pole upside down so it becomes a stable equilibrium, and the target end of the pole always moves in the same direction as your hand.

    I’m still thinking about your NGDP futures proposal. You may well be right. But I don’t want to go there yet, because I find it hard thinking about this stuff, and NGDP futures are a more complicated good than gold, plywood, stocks, or similar assets.

    But there’s NGDP futures as a control instrument, and NGDP futures as a target. The target and control instrument don’t have to be the same thing, though it does make monetary policy much simpler when they are the same thing. Like under the gold standard, when the price of gold was both the target and instrument. “Buy and sell gold to hold the price fixed, both today and next year”.

    You may also be right in saying that the (implicit) inflation target made interest rate control less stable than an NGDP target, because of the oil price shock. In fact, I think you are right, for the US. It wouldn’t work the same way for Canada, because we export oil.

  3. Gravatar of Nick Rowe Nick Rowe
    20. April 2009 at 04:39

    And my post, by the way, was mostly just my trying to say what I thought you might be trying to say.

  4. Gravatar of Bill Woolsey Bill Woolsey
    20. April 2009 at 07:03


    The pole on the palm of the hand is a great analogy.

    I do think that the single interest rate simplication
    that is so useful in general breaks down when only some
    interest rates are at the zero bound.

  5. Gravatar of Vera L. te Velde Vera L. te Velde
    20. April 2009 at 16:17

    In contrast to some complaints you say you are getting w.r.t. the length of your blog posts, I wanted to thank you for filling a niche of in-depth, insightful, extremely well-written economic blogging. Please don’t stop!

  6. Gravatar of ssumner ssumner
    20. April 2009 at 17:36

    Patrick, Thanks for the tip.

    Nick, I like the pole example even better. One reason I used to like read interwar macroeconomists like Irving Fisher is that they had these great intuitive analogies. I use some of these (such as the shrinking yardstick) in my macro classes. But the pole example is much more subtle, sort of a grad level intuitive analogy.

    I’m just thinking out loud here, but let’s start with your view that the movement of your hand is like adjusting the nominal interest rate (and with sticky prices the real rate in the short run.) Then if you are against the wall and you grab the top of the pole with a long hook and pull it off the wall, is that like depressing the real rate through higher inflation expectations?

    Regarding NGDP futures, I have a paper in the (online) Berkeley Electronic Journal Contributions to Macroeconomics (2006). I seem to recall the editor forced me to do the paper in terms of CPI futures, but the basic idea is the same either way. Here’s one way I walked people through the idea.

    1. Start with a monetary dictator.
    2. Go to a committee, and let the median vote set policy

    That’s roughly where we are now, a quasi market–but with no monetary rewards.

    3. Now have salaries partially determined (ex post) based on votes. Everyone agrees on a goal (say 5% NGDP growth.) If the ex post result comes in too high, those with the hawkish votes are rewarded more than those with the dovish votes. Now you are even more like a market. But nothing fundamental has changed, right? If number 2 works, obviously number 3 also works.

    4. Now expand the FOMC to anyone who wants to vote on monetary policy, one man one vote. Keep the reward structure in place. If 3 is workable, I don’t see why 4 wouldn’t work. Now it’s even more like a market.

    5. Same as 4, except replace one man one vote with one dollar one vote. Now you have essentially arrived at my futures targeting idea.

    It’s like putting a frog in water and gradually heating up the pot–he doesn’t even notice anything has happened. I don’t see any sharp break point where the policy would become unworkable. And if it’s workable for CPI futures, it’s obviously workable for NGDP futures

    The weird thing about this idea is that right now, in year 2009, it is 10 times more appealing than in 2006 when I last published it. Then I merely argued that the market might better predict the future than the Fed. I also pointed out that it would solve the problem of “liquidity traps”, indeed the market would essentially determine what monetary policy indicator it thought was best. Little did I know that Fed policy would soon lose credibility. Futures targeting cannot lose credibility, except in the unlikely case investors actually expect the government to default on its futures contracts. So the U.S. would have serenely sailed through last September and October with steady 5% NGDP expectations. Wouldn’t that mean high inflation and falling real GDP if the banking system was tanking? No, because with 5% NGDP growth the banking system wouldn’t have been tanking anywhere near as badly. And even if it was tanking, if there was that much nominal revenue growth out there, firms would have found plenty of sound banks to finance the orders flowing in.

    Nick #2: I did see some similarity in the topics covered, but I like the way you covered them. I was amused by your second link to me, which I think you said was “especially” interesting. Then I noticed that I had made my point better in the brief comment section link, than in my long rambling “official” post on IS-LM. I’m like an actor who does his best performance in rehearsal.

    Bill, I agree with the comment about many interest rates being important. It is one reason why money can still matter when T-bill rates are at zero. I suppose any analogy can only cover so much of a subject like monetary policy before it gets strained, but I was already impressed before he hit the wall, which I thought was a very nice way of describing the current confusion of central bankers. They are so used to moving that hand, that they can’t imagine any other ways of getting the top of the pole to move. Maybe we should send a troupe of magicians and acrobats to the Fed, to give a seminar on unconventional solutions to seemingly insolvable problems. When I was 5 years old I recall a friend of the family trying to solve that connect the grid of nine dots with four straight lines without taking your pen off the paper problem. Maybe future Fed chairmen should be given similar tests, to find out if they can think outside the box. I’ll bet the FOMC from 1929-32 would not have done well.

    BTW, can you believe all the attention Mankiw’s wacky idea is getting? I like those sorts of off the wall ideas, but I wonder whether people realize there are far more politically feasible ways of reducing the real interest rate (such as CPI futures targeting.) It just seems to me that his plan would have zero chance of being adopted, for all sorts of reasons. On the other hand in 30 years we might replace cash with smart cards, and by then his idea might be the ONLY way the Fed conducts monetary policy.

    Vera, Thanks, I appreciate peoples’ willingness to read these long posts. You may want to look at today’s post, which is probably my longest.

  7. Gravatar of Nick Rowe Nick Rowe
    21. April 2009 at 02:32


    Yes, my pole metaphor is better than my original bicycle metaphor. And I only came up with the pole metaphor in the comments section of my post, in replying to critics, just as you made your best statement in a comment, rather than in the original ISLM post. Your comment is what got me started on my post.

    That’s how blogging works, I think. It’s more like a workshop than a journal article. The comments, especially critical ones, and your replies to them, are an important part of why it works.

    It wasn’t that your post was too long, just that it turned out to be as much of a prelude as a conclusion, for one thread of the argument. Which is not a problem.

    “I’m just thinking out loud here, but let’s start with your view that the movement of your hand is like adjusting the nominal interest rate (and with sticky prices the real rate in the short run.) Then if you are against the wall and you grab the top of the pole with a long hook and pull it off the wall, is that like depressing the real rate through higher inflation expectations?”

    Yes. But I would also add two additional channels (and again, I got to thinking about these two additional channels from reading your post, though I’m not 100% sure whether this is what you had in mind):

    1. If the Fed can raise expected future real AD, consumption and investment demand will increase today, which raises the real natural rate today.

    2. If the Fed can raise expected future real AD, financial markets and banks will recover (their assets become worth more, because default risks will fall), so spreads will fall, and the natural real rate on safe assets will rise today.

    (This is where I would bring in Bill’s point, which I totally agree with. But yes, the pole analogy cannot handle everything.)

    Still wrapping my mind around the NGDP futures. Mostly it just needs time, so I can find the right way of thinking about it. I get the “wisdom of the market” bit. Key point is whether and how a deviation of NGDP futures from target causes an increase (if lower) or decrease (if higher) in base money.

  8. Gravatar of Scott Sumner Scott Sumner
    21. April 2009 at 15:43

    Nick, The interest rate point you make is interesting. I have this idea that people focus on the role of expected inflation too much when thinking about liquidity traps. It seems to me they are missing half the picture. What you call AD, I call NGDP. I’ve argued that if we get 5% NGDP growth, we don’t need 2% inflation to get the proper real interest rate. The idea is that the real part of NGDP growth affects the Wicksellian natural real rate. So if you get 5% NGDP growth expectations, it doesn’t matter how much is inflation and how much is real growth. If it’s mostly inflation you depress the real real and exit the liquidity trap. If it’s very little inflation it must be a high level of expected real growth. But that raises the Wicksellian natural real rate up to zero–and again you exit the liquidity trap. I read someone claimed we need 6% inflation to get the correct real interest rate now, which seems absurd because if we had even 5% expected NGDP growth we would have robust real growth (given the slack right now), and obviously much lower inflation.

    I’d like to do an entire macro book without mentioning inflation once. Just do all the nominal shocks as unexpected changes in NGDP, and do the real shocks with real GDP. Do an AS/AD diagram with NGDP on the vertical axis and have the AD curve be horizontal. Does any of this make any sense?

    Inflation is just an arbitrary construct of economists using all sorts of fancy hedonic techniques–how can it actually be an important causal factor in the business cycle?

  9. Gravatar of Nick Rowe Nick Rowe
    22. April 2009 at 02:21

    Scott: thinking about expected future NGDP growth would be a nice simplification, if it worked, but I don’t think it does quite work. The composition between expected inflation and expected real GDP growth would matter. It would matter quantitatively, (or probably would), because there is no guarantee that a 1% increase in expected inflation would have the same magnitude of an effect as a 1% increase in expected real growth. Plus, the qualitative theoretical explanation of the effect of expected inflation is very different from the qualitative explanation of the effect of expected real growth.

    I always like to keep real and nominal variables very distinct.

    But I nevertheless agree with your main point. People look only at expected inflation’s effect on current real demand, and tend to forget about expected real growth. Talking about expected nominal GDP growth would make that impossible, but it is an (over)simplification, I think.

  10. Gravatar of ssumner ssumner
    22. April 2009 at 16:53

    Nick, You may be right. I also try to keep real and nominal variables separate for some purposes. But real and nominal GDP are in fact separate real and nominal variables. NGDP is a 100% nominal variable, just as nominal as inflation. And the way money works is by impacting nominal spending, so I have always thought NGDP is the correct variable for monetary analysis. I think Friedman also thought that way, but I am not certain.

    I do agree that the effect on the Wicksellian nominal natural rate would depend on how much of the 5% NGDP growth is inflation and how much is real. But if you are in a NGDP targeting world then the Wicksellian model plays little role in the transmission of monetary policy. I have always thought that the apparent role of short term nominal rates in the transmission mechanism was mostly a sort of optical illusion, but that becomes near certain when you are pegging NGDP futures contracts (or even the price of gold as you point out in your nice piece on that approach.) When the U.S. tried the gold price strategy in 1933 it had an explosive effect on AD, but nominal rates barely budged. The emptiness of Wicksellian/Keynesian monetary theory was exposed for the whole world to see.

  11. Gravatar of Nick Rowe Nick Rowe
    23. April 2009 at 05:14

    “Changes in monetary policy cause changes in nominal GDP, and the AS curve determines the division of those changes into real growth and inflation”. A nice and sometimes useful heuristic, but it depends on how you define “monetary policy”. What are you holding constant when you hold monetary policy constant, or draw the AD curve.

    I am coming more and more towards your conclusion that interest rates are not at all an essential part of the channel in the monetary transmission mechanism. It all depends on the control variable. (Or, as they used to say in the olden days, going back I think at least to Cantillon, “how the new money enters the economy”). Maybe the inter-war monetary economists (even the Austrians) had a point after all.

    I am being forced to this conclusion via thinking through the transmission mechanism under the gold price control instrument, in the comments section of my recent post.

    Multiple potential channels is actually clearer when you think about alternative control instruments than when you think of a helicopter drop. Or it seems clearer to me.

  12. Gravatar of ssumner ssumner
    23. April 2009 at 06:18

    Nick, I think we see things in a fairly similar way. A few points.

    1. For me the most useful way to think about the AD curve is as a simply rectangular hyperbola, a given level of M*V (or P*Y). I know that is not the standard Keynesian view, and I think Krugman even speculated that the AD curve could be upward sloping. But I don’t find that kind of analysis useful at all. The optimal central bank policy is not to control M, but to control M*V, so that’s what I view as a baseline policy. Thus any sudden changes in M*V are assumed to be monetary shocks in my way of looking at the world–either central bank errors of omission or commission.

    2. My transmission mechanism is as follows. An increase in M that is expected to be permanent causes a proportional increase in expected future NGDP. By “future,” I mean just far enough out that price stickiness is no longer an issue. So if you want to say prices rise proportionately, that’s also fine. Because price stickiness is no longer an issue at long time horizons, the monetarist excess cash balance mechanism fully explains the long run process. The short run effects all flow from the expected long run effects. So if a 10% increase in M makes expected NGDP two years out rise by 10%, that will immediately affect all sorts of asset prices that have flexible prices. Thus the price of stocks will rise relative to dividends. So there is a sense in which you could say that falling interest rates are part of the short run story, after all the dividend/stock price ratio is a sort of interest rate. I just don’t see the ff-rate as being all that important to the story. Lots of things like real estate prices also rise quickly, but construction worker wages are sticky–so home construction increases. These are just examples, there are a million and one transmission mechanisms. But I think they all flow from the fact that next year’s spending is expected to rise (which I think is part of what Keynes meant by “confidence.”)

  13. Gravatar of TheMoneyIllusion » Nick Rowe’s wall and the Great Recession TheMoneyIllusion » Nick Rowe’s wall and the Great Recession
    24. August 2010 at 19:43

    […] Nick Rowe uses the analogy of balancing a long pole in your hand.  If you want the top to go left, you move your hand right.  By analogy, if the Fed wants inflation/growth (and long term rates) to go up, they lower the fed funds rate.  But if you bump up against a tall wall, then you may not be able to move your hand in the direction required to move the pole in the other direction.  You are stuck.  The only solution is to rely on some other method–such as directly grabbing the top of the pole. […]

  14. Gravatar of Nick Rowe’s wall and the Great Recession Nick Rowe’s wall and the Great Recession
    24. August 2010 at 23:35

    […] Nick Rowe uses the analogy of balancing a long pole in your hand.  If you want the top to go left, you move your hand right.  By analogy, if the Fed wants inflation/growth (and long term rates) to go up, they lower the fed funds rate.  But if you bump up against a tall wall, then you may not be able to move your hand in the direction required to move the pole in the other direction.  You are stuck.  The only solution is to rely on some other method-such as directly grabbing the top of the pole. […]

Leave a Reply