The Cleveland Fed is getting close

The previous post linked to Mark Thoma, and here is another interesting post by Mark.  It contains a very good article discussing the advantages of price level targeting over inflation targeting.  That is something I have been emphasizing all year.  But there is just one problem:

One drawback of a price-level target is that it necessitates stimulating the economy whenever prices fall””no matter what the cause. For example, an expansion driven by a positive supply shock would naturally put downward pressure on prices and upward pressure on the real rate, but few economists believe that monetary policy accommodation is helpful in such a situation. An inflation target can potentially be changed, to respond to unusual economic conditions,but a price-level target has the advantage of responding according to a very simple and easy-to-understand rule.

Hmmm, supply-shocks.  Anyone want to guess how Bill Woolsey or George Selgin or I would respond to this “drawback” if the Fed bothered to ask us for advice?

PS.  If you don’t have time to read the whole thing, skim down to the graphs.  It is a good way of visualizing the difference between inflation and price level targeting.

HT.  Leigh Caldwell



20 Responses to “The Cleveland Fed is getting close”

  1. Gravatar of DW DW
    22. December 2009 at 20:47

    Target NGDP?

    [nervously waiting to see if I’m right]

  2. Gravatar of Doc Merlin Doc Merlin
    22. December 2009 at 22:18


    Also, I would say that, the fed can be wrong. plus it adds a level of uncertainty to the market. A nominal target, like price level or NGDP or the Taylor rule would be way better, precisely because expectations would then be set.

    I am not convinced the fed could actually successfully maintain a target like this, but I would agree that trying to maintain a nominal target with no manipulation power by the Fed would be far better than what we have now.

  3. Gravatar of Jon Jon
    23. December 2009 at 00:11

    Perhaps his most interesting remark is the one about level-targeting allowing monetary policy to stabilize with a lower trend-rate of inflation. Now that is fascinating.

    Ymm. Zero-percent inflation trend. Why now, we can go back on to gold. Amazing.

  4. Gravatar of bill woolsey bill woolsey
    23. December 2009 at 04:33


    Money and credit still confused! Was it “simplification” or do they really focus on a credit transmission mechanism?

  5. Gravatar of Scott Sumner Scott Sumner
    23. December 2009 at 05:22

    DW, Yes!

    Doc Merlin, I agree.

    Jon, I have frequently argued that Bill Woolsey’s proposal for a 3% NGDP target makes more sense that my 5% target if the Fed had a sensible monetary policy to prevent liquidity traps.
    The same logic applies to George Selgin’s proposal.

    Bill, I just skimmed the piece, and missed that. It is a sad comment on modern macro.

  6. Gravatar of David Pearson David Pearson
    23. December 2009 at 07:19


    Sorry to be off-topic, but the rise in longer-term TIPS inflation spreads begs comment. We are now in the range of where we were pre-recession on the 5 and 10yr spreads (roughly 2.1% 5yr, 2.35% 10yr).

    With a credibly “symmetric” Fed, NGDP rate or level targeting does not produce divergences between short and long term inflation expectations. That’s because any increase in short term inflation rate will be “taken back” in the future.

    Is our Fed credibly “symmetric”? Of course not — it has had a policy of “asymmetry” (quick to cut, slow to raise) at least since 2001, and it has been reinforced in a big way during the past two recessions. That is why long term inflation expectations are rising. The markets expect that structural high fiscal deficits will at least partly be financed by the Fed. This is the case of a fiscal horse pulling the monetary cart.

    Your proposal (NGDP targeting) would make the long-term expectations problem disappear, if the Fed were credible. If the Fed is not credible, then an attempt to juice short-term NGDP growth will result in higher long-term expectations. Let’s say the Fed adopted a 5% 2-yr target (or the level equivalent) tomorrow. What would 10yr TIPS inflation do? In your mind, it would fall (to closer to 2%). In the real world, it would spike up. The markets would believe the Fed’s intention to create inflation, and they would dis-believe the Fed’s willingness to tighten when necessary.

    Targeting needs to get us from the point “A” of low Fed credibility to what we would agree is a great deal of credibility in post-adoption point “B”. The problem is, you can’t get to point “B” without the credibility, because long-term expectations would climb, and high nominal rates have a disproportionate impact on the economy right now given leverage and overcapacity. Think about the impact on housing: 8% nominal mortgages rates would not be offset by rapid house price inflation, as happened in the 70’s. The same could be said of consumer debt service vs. wage growth (the latter would lag given unemployment). As happens in many emerging markets post-devaluation, rapid inflation would bankrupt levered consumers and firms, and after that “contract-clearing” event, you do get the benefits from higher inflation. A case in point is Argentine real estate, which plunged after the abrogation of convertibility in 2002, and thereafter performed marvelously.

  7. Gravatar of StatsGuy StatsGuy
    23. December 2009 at 07:22

    Bill W:

    “Was it “simplification” or do they really focus on a credit transmission mechanism?”

    Oh no, they really do… Remember Mr Fisher of Dallas Fed “Credit Blob” Fame?

  8. Gravatar of Scott Sumner Scott Sumner
    23. December 2009 at 19:26

    David, I have trouble following your scenario. You seem to assume one group of people expect high inflation and another do not. Are you assuming that lenders would expect high inflation, and thus charge high nominal rates, and yet home buyers would not expect high inflation, and thus would not be willing to pay high nominal rates? I don’t see why that would occur.

    In any case, I am not concerned about high interest rates anytime soon. Interest rates do not merely depend on the rate of growth of NGDP. NGDP growth was very high in the 10 years after 1933, and yet interest rates remained extremely low. The level of NGDP (relative to trend) also matters.

    I should add that the Fed is not going to target NGDP anytime soon, I am just trying to talk up the idea for the next business cycle. No central bank would even consider such a step until it had been debated for years. I do understand that NGDP targeting would work much better if adopted before a recession, and at this point it would be a bit trickier to shape expectations properly. But I really believe the Fed needs to first decide what it wants to do, and then do it consistently. Right now they give the impression that they don’t want both inflation and output to fall sharply, as these are their dual goals. And yet they allow it, which makes the markets lose trust in the Fed. Even worse, after it happens they indicate no desire to catch up to the old trend. How can that be justified? The only way they will gradually be able to rebuild trust is to start doing what they promise.

    If they just target one year forward, there will be less of a credibility problem. I agree that the Fed may currently not be able to convince the public about its 10 year inflation goals, as by that time there will be a different FOMC in place.

  9. Gravatar of Doc Merlin Doc Merlin
    23. December 2009 at 20:02

    “NGDP growth was very high in the 10 years after 1933, and yet interest rates remained extremely low. The level of NGDP (relative to trend) also matters.”

    Also the demand for loans as opposed to other types of financing affects interest rates a lot, too. In the 90’s before Sarbane’s-Oxley, lots of financing was equity based, which reduced the need for debt in business financing for example.

    Also, the demand for risk matters even more. After 33, people were very wary to get loans, and people were very wary to give loans to the low end regardless of rate. So, you could get loans but only if you didn’t need them.

    Also a note: For the record, it seems equity based financing instead of debt based financing is less damaging to the market when a crash happens. (Compare the tech crash to the recent crash).

  10. Gravatar of rob rob
    24. December 2009 at 01:19

    Scott, if I could bet money against it, I’d bet NGDP targeting won’t work. Why? The wisdom of crowds seems to be against you on this. You are just a monetary crank. You have Utopic visions.

    On the plus side, I think I have found my internet trolling home at the Flat Earth Society. So I won’t be bothering this site as much, hopefully.

  11. Gravatar of van van
    24. December 2009 at 05:42

    this question is specifically for scott and prof woolsey, but anyone should jump in:

    as a layman, i am always confused a bit by economists analysing “real” versus nominal variables, in particular, interest rates. and my confusion centers around causality, versus coincidence. if i were to explain real rates to wife or my mother, how should i explain the concept? that is, until a formalized TIPs market came about, we were left to infer real rates – an abstract concept. therefore its very difficult to explain how a movement in real rates CAUSE behavioral change. yet, economists’ analysis are very much concerned with teh movement in real rates, and how those changes affect economic activity. yet 99% of teh world’s population only think in nominal terms, and seem to act in nominal terms. so how do economists make the leap that real factors are determining behavior? does this make any sense?

  12. Gravatar of van van
    24. December 2009 at 05:43

    i should add that, given what i just said, the concept of NGDP targeting is VERY clear and tangible

  13. Gravatar of David Pearson David Pearson
    24. December 2009 at 07:42


    My example is a very practical one. Inflation expectations are embedded in interest rates, but take time to show up in house price increases. Take an increase in long term inflation expectations from 2% to 5%. This would, all else equal, cause the mortgage rate to rise from 5% to 8%. In turn, this would result in an immediate jump in mortgage payments for anyone considering buying a house. Now, that person might think that house prices would rise in the long term, but it won’t make a difference to his ability to qualify for a loan or make the current payments. In essence, this is an underwriting and cash flow issue, and not a present value issue. So there is one group — the highly levered consumer — for which the issue is not expected house price increases, but their immediate ability to access credit and service debt payments.

    BTW, the same dynamic occurs more broadly for consumers. You want real wages to fall. The practical consequence of ZIRP is speculation in commodities via the carry trade, which leads to higher oil and food prices. Consumers feel the pain of higher prices, but get little in the way of wage increases. This is what you say you want. And yet you still expect this same pool of consumers — one with record-high leverage — to maintain real consumption levels such that aggregate demand increases. Krugman and others have been arguing that this may not happen, and I would argue that in the special case consumer spending depressed by leverage and debt service, combined with an depleted stock of private savings, AD will have trouble rising in the face of an inflation-induced fall in real wages.

  14. Gravatar of Scott Sumner Scott Sumner
    24. December 2009 at 07:45

    Doc Merlin, Yes, the main problem is that the banking problems caused monetary policy to go off course in a way that the tech crash did not.

    rob, Thanks. Happy Holidays!

    “yet 99% of the world’s population only think in nominal terms, and seem to act in nominal terms. so how do economists make the leap that real factors are determining behavior? does this make any sense?”

    Actually, the experience of the 1970s showed that people respond to real variables, not nominal variables. Workers understood they were losing purchasing power, so they demanded double-digit pay increases.

    In the 1970s the public understood that housing prices were rising rapidly, so they gladly paid double digit interest rates on mortgages. Now that housing prices are not rising, you’d have a hard time finding anyone willing to take out a mortgage at a 12% or 15% interest rate.

    So I think people do respond to real variables, not nominal. BTW, this is also true in other countries where inflation has been much higher than in the US. Conversely in Japan there is little demand for credit despite very low interest rates. Why? Because they have deflation.

    But I agree that the average person does not think about these things using the language of economists, rather it is instinctive.

  15. Gravatar of ssumner ssumner
    24. December 2009 at 08:00

    David Pearson, Here is what I find confusing about your example. When monetary policy causes higher inflation, it doesn’t happen in a vacuum. The way it works is that monetary policy increase AD, and this causes higher inflation and higher real growth (unless at capacity.) You seem to be arguing that higher inflation would not in fact boost AD. But without the higher AD, there would be no higher inflation in the first place.

    Consider the market for new houses. Suppose the monetary stimulus fails to boost house prices. Then it will also fail to boost inflation expectations, as house prices are strongly correlated with future expected house prices. Alternatively if house prices do rise, this will spur more construction, as it will increase the profit of building houses (assuming wages are sticky). So it’s either one or the other. Either inflation doesn’t rise at all. Or both inflation and AD both rise. You seem to be arguing that inflation will rise, making things more expensive so that AD won’t rise. But without higher AD, inflation can’t rise.

    [BTW, I understand that adverse supply shocks can raise inflation too, but we are talking about monetary policy, which only affects the demand side of the economy.]

  16. Gravatar of van van
    24. December 2009 at 08:36

    Its a Christmas Miracle! I understand it clearly now,Scott, thank you. My sense then is that money illusion is very much human nature. when i posed the standard question to my coworker, he chose the former (assuming 100% likelihood either choice happening): would u choose a 2% pay raise with 2% CPI rise; or no pay raise, and -2%CPI fall? His reasoning, at least with a pay raise, i can cut my costs here and there, and end up ahead. I am sure many would choose this…

  17. Gravatar of van van
    24. December 2009 at 08:38

    sorry the first choice was supposed to be4% pay raise, not 2%

  18. Gravatar of David Pearson David Pearson
    24. December 2009 at 09:45


    Inflation is lumpier than you imply. Sure, house prices are correlated with inflation over the long term. Short term, a rise in spike in nominal rates can have an adverse impact on house prices at the same time that headline inflation is rising. In fact, we are seeing this happen now. What I tried to do is introduce a mechanism by which this happens: large segments of the house buying market are constrained by leverage, debt service, and credit access from capitalizing on an expected rise in house prices. This can happen at the same time that long-term inflation expectations are climbing.

    In fact, the above process is self-reinforcing: the Fed’s ZIRP policy raises long-term inflation expectations but not short term (aside from an increase in the price of food and energy). Short term AD is sluggish, and the market expects the Fed (and fiscal policy) to DO MORE in the future, which raises future inflation expectations further. The more AD fails to respond beyond commodity prices, the more the market perceives that the Fed may lose control of inflation expectations in the future.

    You may think the above scenario is unlikely, and yet it is par for the course for emerging markets with a politicized central bank, large fiscal deficits, sluggish real growth, and very high term premiums caused by long term inflation expectations. In such markets, more sluggish real growth is taken as a sign to AVOID long term bonds, and it results in yet higher term premiums, until there is very little financing for long term investment.

  19. Gravatar of ssumner ssumner
    26. December 2009 at 16:16

    You’re welcome Van.

    David, Even if we assume the Latin American scenario is a real risk, I would argue that monetary expansion reduces that risk. The long term fear of inflation is coming from large fiscal deficits, and the fear that those deficits will be monetized. Of course it is tight money in 2008 that produced those huge fiscal deficits in the first place.

    But I would also emphasize that these risks are hypothetical, long term T-bond yields are relatively low, so people obviously are not expecting much inflation.

    I do agree that we need much more inflation over the next two years, not 10 or 20 years out. And that requires a much more expansionary monetary policy.

    I agree that inflation is “lumpy” and indeed have emphasized that an expansionary monetary policy will raise commodity prices, stock prices and real estate prices before other goods and services. Again, however, monetary policy can only raise commodity prices by raising AD (or lowering the value of the dollar.) So if monetary policy pushes up commodity prices, the economy will expand more rapidly. High commodity prices are only contractionary if they result from less supply.

  20. Gravatar of Mark A. Sadowski Mark A. Sadowski
    26. December 2009 at 18:08

    I spent the weekend (other than enjoying the holidays) trying to decompose the effects of policy and shocks from Joe Gagnon’s recent policy proposal and my own rules of thumb. The bottom line to me seems to be the following:

    1) The Wealth Effect
    Compared to expected nominal growth net wealth has declined by about 160% of GDP since peak. If the wealth effect is approximately 4% per year of level then the economy is facing a headwind of about 6.4% of baseline GDP.

    2) Interest Rate Policy
    According to the FRB/US model each point deviation from neutral policy increases real GDP growth vs baseline by about 1.7% 8 quarters hence. Year on year GDP deflator change is predicted to be about 0.9% this quarter and the same a year from now. The effective nominal FFR is about 0.1% right now. Let us suppose that the natural real rate is about 2%. Then the effective real FFR is about -0.8% right now and since this is 2.8% below natural, this should generate 4.8% increase in GDP relative to baseline eight quarters hence.

    3) Quantitative Easing
    The Federal Reserves’ expansion of balance sheet in late 2008 lowered the 10 year bond rate by about 67 basis points. Since each 75 basis point change results in a 3% increase in real GDP 8 quarters hence relative to baseline, and it dissipates by about a sixth over the next 6 years on average this means that this expansion should still add 1.5% to real GDP relative to baseline 8 quarters from now.

    4) Discretionary Fiscal Stimulus
    DB analysis indicates little change from the present baseline in level effect on GDP. My own estimates based on CEA and CBO analysis indicate approximately a 0.4% reduction compared to present baseline due to stimulus withdrawl.

    Sum it up. Using my estimates it comes to 0.0% effect. In other words I expect that the output gap, currently about 7.7% by my own personal estimates will still be 7.7% two full years from now.

    The major models (IMF, OECD, FRB, CBO etc.) suggest that the output gap closes by 1.8-2.3% two years from now. Am I missing something?

    In other words, is there some factor that they are weighting more heavily that I am not (such as an expected increase in net asset values)? Please, someone who is familiar with these models, inform me of my missconception.

Leave a Reply