Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Wednesday, May 2, 2012

Is higher inflation really the answer?

A lot of people, including those who favor NGDP targeting, want the Fed to raise the rate of inflation; at least, temporarily. Three questions immediately come to mind: [1] What is the theoretical mechanism linking economic prosperity to the rate at which nominal prices rise; [2] Exactly how is the Fed, given the tools at its disposal, supposed to generate higher inflation under current economic circumstances; and [3] What is the evidence to support the belief that more inflation will reduce unemployment (or increase real GDP)?

There are so many different views out there that it's hard for me to keep track of them all. My last couple of posts dealt with the idea of a NGDP target, and it's close cousin, a price-level target. I'm no expert in the area, but if I understand the logic correctly, the idea is for the Fed to reverse what was a sharp and unanticipated decline in the price-level that occurred in late 2008. The presumption is that because debt is denominated in nominal terms, an unexpected permanent decline in the price-level path increase the real value of the stock of outstanding nominal debt. In turn, this imposes a real burden on all debtors, including households with mortgages and the government sector.

There seem to be two aspects to the "price level" surprise shock. First, there is a "fairness" issue. The shock evidently resulted in a redistribution of wealth from debtors to creditors, and it is only fair that this wealth transfer be reversed. (And since the Fed was the agency responsible for letting the price level drop, it should do the undoing -- even if the same might be accomplished by the fiscal authority). Second, there is an "efficiency" issue. Somehow, this wealth transfer has manifested itself as "deficient aggregate demand." I am not exactly sure how this last part works--maybe somebody can enlighten me (in a language that I can understand--a mathematical model!).

In any case, I am not entirely sure I can believe in the quantitative importance of this mechanism. The prescription presumes a sharp and persistent decline in the price-level path, something that I have trouble seeing in the data. In particular, the follow diagram plots the (log) PCE price-level for the U.S. since 1990; the red line is a (log) linear trend. According this data, we are essentially back on the original price-level path (I think the same roughly holds true when the price-level is measured by the CPI or the GDP deflator).

Of course, the "wealth channel" I described above is not the only way in which higher inflation might stimulate economic activity. Here is Paul Krugman for example: Krugman: Fed Should Tolerate More Inflation to Reduce Unemployment.  
"The main thing the Fed can do is promise that they will be very slow to step on the brakes, that as the economy recovers that they will let inflation rise, not to high levels, but to 3 or 4 percent from two percent," Krugman suggested. "That would move the markets quite a lot. It would lead people who are making plans to think that sitting on cash is not a good idea.
I have no doubt that people would think that sitting on cash is not a good idea. The question is: how would people seek to transform their cash holdings? Krugman seems to think that people will want to go out and spend the cash on goods and services. But what if they instead decide to buy gold or Caribbean real estate? There is also the possibility that nominal rates might rise (perhaps not one for one) with higher expected inflation via a Fisher effect, leaving the real return on "safe haven" assets relatively unchanged. Who really knows what might happen?

At the same time, one has to ask how the move to a higher rate of inflation might affect different members of society. Those on fixed nominal incomes are likely to suffer; at least, in the short run (or however long it takes to index those incomes to the higher inflation rate). What about those who have no bank accounts--those people who rely on cash transactions--the poorest segment of the population? This could, in principle, be rectified by cash disbursements to those deemed to be in need, but...well, good luck with that.

And, in any case, just how is the Fed supposed to engineer this smooth ride up from 2% to 4% inflation? Jim Hamilton has a nice post today explaining why it might not be as easy as people generally think it might be; see here: Should the Fed Do More?

I haven't even touched on my third question here, the one dealing with the inflation-unemployment relationship (for long-run evidence, see the data in here). So many questions, so little time...


  1. (1) The talk about Gold/RE in Caraibes is off-target. The nominal interest rate is 0 up to the 2 year maturity at least. Given this, higher inflation expectations would decrease the real rate and encourage higher consumption spending today, esp. on durables. This is just the Euler equation.

    (2) I agree that the debt overhang issue is hard to quantify. However, the current situation is unprecedented and hence historical precedents are difficult to use. Your picture also doesn't allow for the fact that debt was high and the current shock to income is big, so many people are 'underwater', and hence having inflation (even above the trend!) may help. (In many models, borrowers are higher productivity and hence redistributing wealth towards them increases output.)

    (3) As to how to implement it, a first step would be to say that you want to do it. Which the Fed hasn't even done. Second, instead of long-term bond purchases, I would simply send checks to households, telling them that the Fed will not increase taxes to pay for them.

    1. (1) Yep, I understand Euler equations. Typically though, these models are closed economy models. In reality, there are competing stores of value around the world. Moreover, the 0 nominal interest rate at 2 years is endogenous. How do you know it wouldn't rise in response to a higher expected inflation? Central bankers have to worry about risk factors like this.

      (2) I agree that the picture does not capture those things. But then, this is an argument against price-level targeting, right? (Not necessarily NGDP targeting.). And that's a good point you make: the effect could be felt mainly be debt-constrained entrepreneurs, causing a socially-inefficient decline in investment spending. It would be useful to have this likely effect quantified, however, don't you think (I mean, before the Fed embarks on a potentially dangerous policy experiment?)

      (3) You should read the Jim Hamilton piece carefully. Also, the Fed is legally prohibited from writing checks or taxing agents. The Fed is essentially only permitted to buy and sell Treasury (or agency) debt. That is, the Fed is only allowed to swap one form of debt (cash) for another (government bonds). It is not immediately clear how rearranging the portfolio of government debt by asset swaps will be very helpful under present circumstances.

  2. David, I cannot understand your story. Fed keeps short rates at zero, tell me how real short rates don't go down with an increase in inflation expectations. It just doesn't add up to me. Please explain.

    1. I agree with Rollo. By making the proposition that the Fed can no longer prevent nominal rates from raising it seems like you are assuming QE or some other form of "bond support" if you like loses it's potency, yet if inflation rises whilt initially at the ZLB as a result of Fed action, how can this be the case?

    2. Rollo, no, I think you are right. Real short rates would almost surely decline. The issue, I think, is what happens to real and nominal rates at longer horizons (interest rates more relevant, I think, for evaluating NPV of capital spending).

      Consider two bonds, a 30-year nominal treasury and a 30-year TIPS. There is a no-arbitrage condition linking the real return on these two assets (the Fisher equation). If expected inflation rises, what do you expect to happen to the real return and why (and whether the effect would be potent enough to stimulate spending). If you can answer this for me, we can go from there.

    3. David

      You're in effect questioning the entire interest rate channel of monetary policy. Your challenge holds equivalently even in the case when the central bank cuts its short term rate from, say, 5% to 4%, holding inflation expectations constant. If you think that works, you probably have a good theory of why this works.

      In your TIPS/ T-bond example, people would sell the T-bonds and buy the TIPS bonds until such time as the new price of the TIPS securities reflects the new reduced yield. But what does this particular arbitrage have to do with anything? Are you saying that the presence of a 30 year TIPS security pins down the real long rate expectations? That would be a much more difficult thing to prove.

    4. Zen:

      If you think that works, you probably have a good theory of why this works.

      I guess I don't think this "works" in any significant manner.

      As for the second part of your statement, the no-arbitrage condition states i = r + p.

      The experiment assumes that the Fed can increase p (expected inflation rate 30 years out). The question is: what happens to i and r at the 30 year horizon? If i rises in proportion, or roughly so, why should that have any impact on real capital spending today? On the other hand, if you believe i would remain invariant, then r must fall. This, to me, would seem to be a much more difficult thing to prove.

      Of course, there are many complicating factors in the current environment. I think that central bankers are aware of this and are afraid to experiment for fear of losing their greatest accomplishment: a credible, steady long run inflation rate. Krugman may think the risks are worth the experimentation, but I'm sure he is in a financial position that ensures if things go wrong, he won't personally suffer that much.

    5. OK, sorry didn't check this. I don't see any real rate rising according to Fisher, while *some* real rates would decline, right? Actually: *all* real rates would decline:

      Long rates are the combination of short rates. Short rates from here to the future could take two forms: "Fisher rates" (nominal rates depending on the combination of real rates and inflation) and "Liquidity trap" rates (real rates depending on the combination of zero nominal rates and the chosen inflation). Real "Fisher rates" wouldn't be affected (only nominal will); "Liquidity trap" rates would be lower. All long real rates (a combo of Liquidity trap and Fisher rates) would go down.

  3. Nice article, thanks for the information.

  4. Good article. The proponents of higher inflation act as though inflation is like your home's thermostat. Set it at 4% and walk away. They also dont ask themselves how, once inflation has hit its new target for the specified time period, do we get inflation back down to 2%.

  5. Speaking for myself, a) i do not agree the premise that ngdp targeting means a backdoor increase in the inflation rate; b) i do not think further stimulus would be inflationary anyway (to points 2 and 3).

    empirically, wages are sticky.

    if wages are so sticky, further stimulus would not be inflationary.

    I realize PK is on record as tepidly promoting ngdp as a way to increase the inflation rate, which i think is completely wrong.

    We actually do not know (for certain) how an increase in nominal spending will split into inflation or output. In the long run we can probably say that a 4% ngdp target splits into 2.5% real growth and 1.5% inflation. but we do not know over the short run.

    lastly, I place significant weight on the signalling aspect QE (I cite the following mechanism: lots of bank economists sit on capital committees and planning boards and feed their assumptions into CFOs business plans. those economists also are deeply familiar with fed policy and their forecasts, therefore when the Fed says they are committed to lowering the UE rate more quickly they update their assumptions, which feed into lending, capital, and other business plans. some economists also work at investment banks and work their assumptions into presentations which influence CFOs).

    merely signalling its committment to increasing AD and putting the economy back on its ngdp growth path would ameliorate some of the effects of high UE on the housing and construction market (to start with) which is a very large segment of the economy and foster healing and a return to growth in the market (including the construction/manufacturing jobs that go along with it.

    in sum, i totally disagree with the premise that ngdp targeting means higher inflation. I think some like PK have said so, without really deeply diving into it.

    1. dwb:

      I have written elsewhere why data like what you take to be evidence of sticky wages needs to be interpreted more carefully (see: )

      In any case, is your statement "if wages are so sticky, further stimulus would not be inflationary" just your opinion, or is it a proposition that must hold under specified conditions?

      You may be right about the signalling aspect of QE. I am not sure about the mechanism you propose. It would be interesting to interview those people to see if their thinking squares up with your hypothesis. I'd be interested to hear what you find.

    2. Yes, i agree that evidence of sticky wages has to be interpreted carefully. personally, most salaried employees like myself also get variable compensation (which can be reduced if revenues are sub-par), say 10-40% of salary. I also know of many employees at large companies who (recently, last month) took across-the-board 5-10% paycuts to keep their jobs, and i know of many manufacturing employees whose bonus is a function of productivity. Salary is not the only component of costs (theres the 401k and health for example), so yes, i would 100% agree that wages are a lot less sticky in than people assume and the evidence for aggregate wage stickiness needs to be carefully examined. Every CFO will tell you employee costs cannot rise faster than revenues, and jobs cuts are often balanced with rewards (retention bonus) for those who stay to cut unproductive employees and make the ones who stay happy to stay and not jump ship. employee compensation is an extremely complex mix of economics and psychology.

      Anyway, my 30,000 foot interpretation of the FRBSF study is that (for whatever reason) there is a population p of sticky wages stuck at 0% and 1-p of wages (employees with bargaining power, maybe) whose rate of increase is geared to expected inflation rate(and productivity). Those p population wages could be stuck at 0% for a variety of reasons (long term unemployment increases the reservation wage, and so on).

      over-time inflation (or an increase in AD) would reduce the population p to zero, eventually.

      So the answer to your first question, is its a proposition that I think only holds true under certain circumstances, for certain sectors, and the "degree of stickiness" in aggregate depends on the relative wages between the population p and 1-p - as real wages in the p population decline, the stickiness disappears -so not in the long term.

      As to your second question, It's an empirical personal observation having done many of those ppt decks & presentations and knowledge of how these presentations circulate and influence thinking and decisions at MegaBanks and MegaCompanies... its more of suggestion/observation of how information flows around. it was sort of tounge in cheek because i am not sure one really needs a "mechanism" to justify how expectations are formed. when the CFO approves the annual merit increase for the 1-p population he's seen many economist presentations, the business plan, and ample stories on Bloomberg. so there are i am sure many factors...

    3. let me add: I also put weight on the idea that additional nominal gdp will likely be channeled into "deleveraging" i.e. paying down principal in the case of the 25% or so of mortgage holders who are underwater. they might buy gold or Carribean real estate, but not at the median. 70% of GDP goes to wages. you can call it inflation or output, but it still reduces mortgage and debt delinquencies.

      This has the beneficial knock-on effect of fostering reduction in nominal debt contracts which leads to an improved housing and construction market.

  6. Forgive me for what is probably an ignorant question, but isn't inflation targeting just a proxy for what is really desired, which is to equate money supply with money demand? If not, then I'm more confused than I thought.

  7. Prof J: Don't markets always equate supply to demand at some point?

    Besides, the empirical evidence suggesta that money supply and money demand can move around without necessarily impacting the observed price level or inflation rate. That was the observed disconnect that sunk Monetarism if I am not mistaken.

    I thought inflation targeting was designed to stabilize prices and make shifts in relative valuations more transparent. Transparency and accountability tend to interfere with private rent-seeking, and the dead-weight losses that result.

    Great thread. I almost spit my tea all over the monitor upon reading the introduction to the first post on NGDP..... but hey, I want to know what the arguments in favour of NGDP targeting are too.

    1. Westslope,

      The problem is if people increase money demand but money supply doesn't increase to accomodate that the equilibrium has to be set in other markets where money is traded.

      There is an argument that inflation targeting is supposed to make prices more stable, but I think that argument is false. Targeting 2% (or whatever) inflation per year will create problems if the natural tendency of price changes is, say, -2%/year. Kevin Dowd and George Selgin have written a lot about this, and it's impossible to do the discussion justice in a reply to a blog post.

  8. David,

    A little financial repression will take care of the pesky Fisher equation: use international banking regulations to get banks to buy more Treasuries, for example and make it increasingly difficult for US citizens to invest abroad. Oh, wait, the government is doing those already...

    (Half-)joking aside, it appears investors are quite content to accept negative real yields in much of the developed world. e.g. US swaps less inflation swaps are negative out to at least 10y. Of course, a higher inflation target could cause them to change their mind.


  9. David:

    0. Well, we really want temporarily higher NGDP growth, not temporarily higher inflation. We want a temporarily higher sum of inflation plus real growth. We hope it's mostly real growth, but the SRAS curve will determine the actual mix of the two.

    1. Well, because we have left out half of the mechanism (the effect of higher expected real growth on consumption plus investment) this inflation mechanism will be weaker than the full mechanism with real growth included. Nevertheless: higher expected inflation means lower real interest rates on paper assets like money and bonds which increases the demand for real assets. Some of those real assets can be produced, so their production will be increased in response to that increased demand to hold them (investment, Tobin's Q, etc.). The higher demand for non-reproducible real assets (land) will raise their prices. This causes a substitution effect (plus wealth effect in an OLG model) into increased consumption demand. In an open economy the real exchange rate will depreciate, increasing net-exports, starting a currency "war" and "retaliation" by other countries loosening their monetary policies too, which is exactly what the world needs.

    2. Unless you believe the US is stuck in a liquidity trap *that will last forever*, the Fed can always simply promise to create a higher price level at that future date when the US is not in a liquidity trap, then work by backwards induction. Announce a *permanent* increase in the stock of money, conditional on NGDP being below the target level path. Or it could simply switch from targeting nominal bond prices to targeting some real index like the S&P500 today. Right now there is no nominal anchor for expectations of the future. The Fed needs to create one.

    3. Canada 1982, for example, where a tighter monetary policy that lowered inflation temporarily increased unemployment. Because David Hume noticed this empirical relationship too, for example. Etc.

    1. Nick:

      0. So you want higher RGDP. You want more paper to create more RGDP. Why not just say it this way?

      1. You are essentially arguing a Tobin effect. One does not need sticky prices to generate this effect. Why are you highlighting it here now?

      2. This is mostly wishful thinking. It may, or may not work. You want to risk the Fed's considerable reputation for an airy fairy policy experiment? And what do you mean that there is currently no nominal anchor? We have a 2% inflation target. Together with the current price level, this gives the anchor. Take a look at the plot of PCE above...what do you see?

      3. What about Canada in the 1970s? High inflation, high unemployment. Why do you cherry pick the data like this.

      Totally unpersuasive. Sorry.

  10. The effect of higher implied inflation targets on savers is complicated by demographics. The more-numerous older cohorts need to reduce the their allocation to risky assets and increase current income. This is a secular, not cyclical phenomenon. Reducing their expectations of future real returns from income investments -- the effect of NGDP targeting -- might cause them, paradoxically, to have to reduce current real spending to save more for retirement. Granted, social security benefits are inflation-indexed; this just means that younger cohorts will expect their tax rates to go up, forcing them to reduce current income.

  11. Sorry, I meant "reduce current spending in the last sentence".

  12. David,

    There are some who want the higher inflation that goes along with targeting the price level because they believe that inflation (or higher expected inflation has real effects. According to the New Keynesian model, if the Fed can raise inflation expectations, this will help close the output gap. There is a problem with this in that we have to assume that the nominal interest will not adjust to rising inflation expectations. If it does, then the real rate doesn't change and there is no effect on output.

    So here is your question: Why is a price level target beneficial and can anyone show me a mathematical model that says that it is?

    There is, in fact, a model that shows the importance of a price level target without resorting to New Keynesian logic. The model is outlined in the paper "Price Level Targeting and Stabilization Policy" by Berentsen and Waller. I know that you are at least tangentially familiar with the model, which is a based on Lagos and Wright or more aptly on Berentsen, Camera, and Waller. In the model, when the central bank targets the price level, they are able to anchor inflation expectations and therefore use monetary policy to stabilize short-run fluctuations (and this stabilization is welfare improving).

    1. Berentsen and Waller? Figures!

      Yes, as I suggested earlier, I can see how a price-level target might be beneficial. To me, the essence is to provide a nominal anchor. My question was more: What makes people think that adopting a PL or NGDP target right now is going to propel us on a path of greater prosperity?

  13. "[1] What is the theoretical mechanism linking economic prosperity to the rate at which nominal prices rise"

    Lars Svenssons Foolproof Way is a great paper.

    "[2] Exactly how is the Fed, given the tools at its disposal, supposed to generate higher inflation under current economic circumstances."

    Besides the one above, one interesting idea is to increase VAT (sales tax or whatever you call it in US) to created inflation expectations.

    "[3] What is the evidence to support the belief that more inflation will reduce unemployment (or increase real GDP?"
    I know George Akerlof have concluded somewhere in his research that increase in 2% inflation units will permanently lower the unemployment rate by 0,5% percentage units.

    - "But what if they instead decide to buy gold or Caribbean real estate? There is also the possibility that nominal rates might rise (perhaps not one for one) with higher expected inflation via a Fisher effect, leaving the real return on "safe haven" assets relatively unchanged. Who really knows what might happen?"

    It's not very likely people will buy gold for a couple of reasons, first, gold price is strongly related to low bond yields. When yields rise investors might find bonds a more attractive investment due lower inflation expectations in the future (hence higher real return).

    Also, I don't think inflation of about 4% is such a big deal. The average inflation during the entire 19-century was somewhere around 5% still we had the entire middle class appearing after the Post-war period.

  14. DA, wrote, "But what if they instead decide to buy gold or Caribbean real estate?"

    DA, when you wrote this I knew you were just making it up as you went along, for this counterfactual is so wrong it is not even worthy of Ron Paul.

    The sellers of the gold or Caribbean real estate would have to do something with the cash.

    And, the middle men would earn an nice fat commission check, which hasn't happened in a while in real estate.

    IOW, at the end of the day, the money would be spent, not hid in the ground in a Mason Jar.

    1. Anonymous,

      I am not sure what you mean by your last sentence. What do you call the vast USD reserves held in Asian countries (China, in particular)?

      In any case, the point is, I do not see a priori why additional spending has to take the form of expenditure on domestically produced goods and services. It may instead be spent on imports.

      But you seem like you know a lot of things, some of which may even be correct. Why don't you tell me?

  15. Higher inflation will just mean the collapse of the U.S. dollar. The answer is to go back to a gold standard and make gold and silver coins legal tender again. Inflation skyrocketed as soon as we went off the gold standard.

  16. Wow.....What a nice post!!....Your post is an excellent example of why I keep coming back to read your excellent quality content that is forever updated.I would like to read newer posts and to share my thoughts with you.Thanks a lot....

    how to get instagram followers fast