Thursday, July 27, 2023

On the 2% inflation target

Project Syndicate asked Mike Boskin, Brigitte Granville,  Ken Rogoff and me whether 2% is the right inflation target. See the link for the other views. I pretty much agree with them in the short run -- don't mess with it -- but took a different long run view. Apparently Volcker and Greenspan were fans of price level targeting and hoped to get there eventually, which is the sort of long run approach I took here. 

I also emphasize that any inflation target is (of course) a joint target of fiscal as well as monetary policy. Fiscal policy needs to commit to repay debt at the inflation target. 

My view: 

No, 2% is not the right target. Central banks and governments should target the price level. That means not just pursuing 0% inflation, but also, when inflation or deflation unexpectedly raise or lower the price level, gently bringing the price level back to its target. (I say “and governments” because inflation control depends on fiscal policy, too.)

The price level measures the value of money. We don’t shorten the meter 2% every year. Confidence in the long-run price level streamlines much economic, financial, and monetary activity. The corresponding low interest rates allow companies and banks to stay awash in liquidity at low cost. A commitment to repay debt without inflation also makes government borrowing easier in times of war, recession, or crisis.

Central banks and governments missed a golden opportunity in the zero-bound era. They should have embraced declining inflation, moved slowly to a zero-inflation target, and then moved gently to a price-level target.

Why not? Some focus on the short run and say that central banks should raise the inflation target, because getting inflation to 2% will require too much pain in the form of unemployment. But inflation is falling alongside very low unemployment, proving this argument wrong again. And shifting the goal posts undermines the stable expectations that allow relatively painless disinflation.

The other argument says that a higher inflation target creates more room to use rate cuts to stimulate the economy in times of recession. But that is like wearing shoes that are too tight all day, because it feels so good to take them off at night. This argument presumes that expected inflation is set mechanically by previous experience. Moreover, the evidence that slightly lower overnight rates provide much stimulus is weak. The potential benefit is not worth permanently abandoning a stable value of money.


Many comments here and on twitter ask about a nominal GDP target. I'm not a fan, for three reasons. 

First, just what does the Fed do to hit a nominal GDP target? That objection is common to the price level target, but it's an important point. Nominal GDP targeting advocates seem to think it solves the whole conundrum of just how do interest rates affect inflation. No, it's just a different centering point of the Taylor rule. Raise interest rates if nominal GDP growth is high, lower it if low. 

Second and more to the point, it assumes that "potential," "supply" or "neutral" real GDP growth is constant or at least slow moving and known. If potential grows 2% real and you want 2% inflation, then the nominal GDP growth target is 4%, and the idea is that you let the economy take care of the split between real and nominal in the short run. I'm of the view that there is a lot more high frequency movement in "potential" than commonly thought, so even if the Fed achieved steady nominal GDP growth, there would be needless inflation volatility or needless deviation from neutral real growth. Real GDP grew 4% 1950-2000 and 2% since then. How long does it take the target to adapt to this sort of thing? 

The idea is that the Fed isn't smart enough to separate nominal GDP growth to real growth and inflation, so let that be endogenous. Of all the problems of monetary policy, this doesn't seem like the worst to me. 

Third, I like the clarity of a price level target. Even if you make it level, not growth, of nominal GDP, it's muddy just how much inflation you should expect when borrowing money, financing a project, etc. Keep the units pure. Moreover, if you don't like price index measurement issues, wait until you look in to GDP measurement issues. GDP is not consumer surplus. 

Others ask what about mismeasurement. Answer: fix the measurement. 

Back to, just what does the Fed do differently? Presumably, the Fed raises interest rates when nominal GDP is higher or growing faster than target, and vice versa. That's awfully close to a Taylor rule. 

Take a nominal GDP growth target. Then the Fed would follow interest rate = a + (coefficient) * (inflation + real GDP growth - target). That's the same as interest rate = a+ (coefficient) * (inflation-inflation target) + (coefficient) *(real GDP growth - its target). So, we're just arguing about whether the inflation and GDP coefficients should be the same -- the Taylor rule is 1.5 and 0.5 -- and whether to use growth or gap. Other versions are just other modifications of the Taylor rule.  

It's not magic. There is nothing about a nominal GDP target that makes the Fed any more able to hit it than it does an inflation target. 

The main argument for targeting the level of NGDP is that it enforces forward guidance. In the 2010s, the Fed was searching for was to promise it would keep interest rates low for a long time, and allow inflation to surge in the future. NGDP, as to a lesser extent a price level target does that. However, so did the Fed's new strategy. The Fed waited an unprecedentedly long time to raise rates, and in the view that raises inflation, they did their job. NGDP is right back on the pre 2007 track. Would the NGDP target have been more effective a promise than the flexible average inflation targeting? Maybe, but it's just a different flavor of the same thing. 

Meanwhile, the big missing question is just whether low interest rates raise future inflation at all. We're sitting on the roadside in Iowa, with a blown tire, and discussing whether we want to go to Disneyland or Universal studios first. 


  1. Why not a Nominal GDP target over a price level target? Wouldn't that handle supply shocks better?

    1. Why not target both inflation (low) and real GDP (high) independently of each other? Handles both supply AND demand shocks.

    2. Nope.

      1. You can target anything or things you want. Whether you hit your target(s) is a separate matter altogether.

      2. See two interest rate Taylor Rule for an approach to targeting both inflation and Real GDP growth rate. This:

    3. As it turns out, the US already adopted several targets:–Hawkins_Full_Employment_Act

      Explicitly states that the federal government will rely primarily on private enterprise to achieve the goals.

      1. Instructs the government to take reasonable means to balance the budget.

      2. Instructs the government to establish a balance of trade, i.e., to avoid trade surpluses or deficits.

      3. Mandates the Board of Governors of the Federal Reserve to establish a monetary policy that maintains long-run growth, minimizes inflation, and promotes price stability.

      Let's see that is five targets established by an Act of Congress.

  2. It is a commonplace that in a governmental monopoly fiat money system, inflation represents a tax on existing savings. I have therefore argued for many years that a non zero inflation target is a tax that has not been authorized by Congress in violation of Article I Section 8. Further it is a direct tax that does not comply with the apportionment requirement of Article I Section 2, so even Congressional authorization would be fraught with difficulties.

  3. Why not broaden your view and adopt an Aggregate Nominl Spending (NGDP) level target? It really is a much better & effective target!

    1. Why not adopt both an inflation (low) and real GDP (high) target?

  4. Inflation is a form of constructive default.

  5. is this the FED that has printed $24,000 per American since 2008 to ASSURE wall street bonuses? The Debt is going to destroy America. You can delay economic and biology...but not stop it!

  6. What about the common argument that nominal wages are sticky downwards and that a zero inflation target would therefore raise unemployment?

  7. The stronger counter-argument for me is monetary illusion. I believe it's quite empirically overwhelming that workers do not accept easily nominal wage cuts, while they can accept reale wage cuts (just look at what is happening in the labor-protected Europe in the past two years). Even the government many times uses the same trick. The only way in which public expenditure can be cut is by freezing the nominal expenditure

  8. I hope you'll respond to Scott at . I think Scott is onto something very important with NGDP targeting. We can have better things if smart economists can unify behind better monetary policy.

  9. Did you see the Inflation numbers this morning? Aaahhhhhh!!!!!!!! Zimbabweeee!!!!!!!!, Weimer Republiccccccc!!!!!!!!, Venezuelaaaaa!!!!!!!!!! Wheelbarrows full of cash for a loaf of breaddddddd!!!!!!!!! Aaaaahhhhhh!!!!!

  10. Good luck getting the government to do its side of the "equation" leaving the Fed to have to react. And too many seem to overestimate the power of the Fed it seems to me.

  11. "I'm of the view that there is a lot more high frequency movement in "potential" than commonly thought, so even if the Fed achieved steady nominal GDP growth, there would be needless inflation volatility or needless deviation from neutral real growth." -- JHC.

    "Potential" (adj.) "Capable of coming into being or action, latent; [...]." -- The Concise Oxford Dictionary (1951); Oxford: The Univ. Press.

    Potential [economic] capacity is reserve, latent or underutilized capacity of the economy to produce goods and services. Macroeconomics' new Keynesian DSGE model theorizes that the potential capacity of the [modeled] economy is the economic output in goods and services that would be produced if the economy was not subject to 'friction' and competition was not imperfect. In other words, 'potential capacity' is a normative concept that is (a) unobservable, and (b) an ideal condition unattainable in fact.

    If 'potential capacity' has the characteristics commonly attributed to it, as in the previous para., then (i) it is unmeasurable (because it is unobservable), and (ii) "there is a lot more high frequency movement in "potential" than commonly thought" is a statement which is non-falsifiable because the underlying state variable is unobservable.

    The terms 'potential capacity', and 'potential GDP', as used in a macroeconomic sense, are synonyms. The growth rate of 'potential GDP', or 'potential capacity' is unobservable because the underlying state variable is unobservable. In a similar vein, the 'trend in potential GDP' is unobservable. These terms are normative concepts. Basing policy decisions on unobservable state variables is not likely to yield the results claimed for the policy in question.

    It is difficult to soften this conclusion, if effective monetary policy is a fundamental requirement demanded of the monetary authority.

    1. "Basing policy decisions on unobservable state variables is not likely to yield the results claimed for the policy in question."

      1. News flash - all forward looking economic measures (potential GDP, inflation expectations, etc.) are best guesses.

      2. "Continuous effort — not strength or intelligence — is the key to unlocking and using our potential." - Liane Cordes

    2. Yeah? One might be inclined to remark that 'best guesses' are inherently no better at predicting the present or the future course of events than is licking your finger and holding it up in the air to see which way the wind is blowing at that instant in time.

      It is simply a charade. Performing a "rain dance" would be as effective as exerting the effort to compute "potential GDP". There is no connection between a "rain dance" performance and the arrival of rain on the parched landscape. In a like vein, the "output gap", x(t), is an index without a basis. The output gap of new-Keynesian modelling theory is defined as x(t) = ln(Y(t)) - ln(Y⁺(t)), where Y⁺(t) is the "potential output" in period t, and Y(t) is the actual output in period t. Y⁺(t) is defined as that level of output in period t that would be obtained but for the frictions in the economy and the imperfect competition that prevails between producers in the economy. Since removal of economic frictions is all but impossible, and imperfect competition prevails because consumers value different characteristics differently thus giving rise to the ability to chose different flavors and textures in the commodities they consume, Y⁺(t) is both unattainable and unobservable. But, if Y⁺(t) is unobservable, then, as a consequence x(t) is unobservable. And, if x(t) is unobservable, then the new-Keynesian DSGE model is merely a theoretical construct and presents no practical capacity for predicting the course of the economy in real time.

      For the theoretical academic this poses no difficulty -- the model is 'real', in the sense that a Hollywood movie is 'real'.

      But, when the academic states that "potential GDP" will be used as the basis for determining a price index that will determine monetary policy, then that is when academic practices (modelling) begin to go wrong and wrong in the catastrophic sense of Long-Term Capital Management L.P. (LTCM) which is the exemplar of academic hubris that affected the real economy in ways that the academics didn't anticipate and couldn't control.

      The shortest distance between two points is a straight-line, except when it isn't. In Euclidian space, a straight-line is always and everywhere the shortest path between two non-coincident points. But, space isn't Euclidian except as an approximation in the near-neighborhood of a point.

      "Potential GDP" and "NAIRU" are simply academic constructions which have no analogues in the real economy, and as a consequence are of less value than Euclidian geometry. With Euclidian geometry as your frame of reference, you will eventually get where you're going albeit not in the least amount of travel time.

      When unobservable state variables, e.g., NAIRU, are used as a basis for monetary policy what you get are episodic periods of recession resulting from inappropriate monetary control responses triggered by URATE dropping below NAIRU. Those responses have had real uncompensated negative economic consequences that destabilized employment and business investment, often permanently.

      To paraphrase the warning that the ancient and now extinct Trojans failed to heed, ‘beware Academics bearing gifts.’

    3. The time series GDP, adjusted for CPI, i.e, GDP(t)/CPIAUCSL(t), exhibits behavior usu. associated with a process having unit-root. See, e.g., which plots log[GDP(t)/CPIAUCSL(t)] vs. calendar time, t. The financial crisis of 2007-Q4 thru 2009-Q3 put the American economy on a different trajectory from the trajectory that connects 2000-Q2 and 2006-Q4.

      The post-pandemic period appeared to be returning to the trajectory that prevailed from 2009-Q3 to 2019-Q3, but in 2021-Q4 a smaller unit-root break appears in the time-series and persists thru 2023-Q2.

      If the GDP/CPIAUCSL series does present a unit-root process, then it will not be possible to achieve a target real GDP trend growth rate through time.

      The nominal GDP time series also exhibits what appears to be a unit-root process behavior when the data is plotted on semi-log paper. In the nominal GDP time series after 2020-Q3 the unit-root process behavior is not evident, but the time series does not return to the pre-financial crisis trend line after 2009-Q2 at any point up to 2023-Q2.

      This apparent unit-root process characteristic in the two time series renders nominal-GDP targeting difficult, if not impossible, to achieve in practice.

    4. In order to compare "price index" targeting with "real GDP" targeting, a plot of the two time-series on semi-log chart paper for the period from 2000-Q4 thru 2023-Q2 is created – see, e.g.,

      The series are GDP(t)/CPIAUCSL(t) and CPIAUCSL(t) versus calendar time, t, plotted on semi-log chart paper. The data series GDP(t)/CPIAUCSL(t) values appear on the left-hand vertical axis, and the data series CPIAUCSL(t) values appear on the right-hand vertical axis. Straight-lines drawn on the chart represent curves log(y(t)) = b + m ( t – t₀ ) , where y(t) is the value of the index or real-GDP, and b and m are Y-axis intercept and the slope of the straight-line, respectively, and t – t₀ is the time scale gauged on time horizon starting date given by t₀ . The natural logarithm is chosen by the FRED charting program.

      Straight-lines on the semi-log chart paper represent uniform growth rate curves on linear x linear chart paper: Y(t) = exp ( b + m (t – t₀)) which can be simplified to Y(t) = Y(0) exp( m t ) by rescaling the axes such that Y(0) = exp(b), and t₀ = 0.

      Semi-log chart paper allows the analyst to more easily pick out the changes in the trend line regimes. It also shows the difficulties to be anticipated in 'GDP targeting' and 'Price-level targeting'. Those difficulties are evident in the FRED chart (see the URL above).

      And that may be why neither P. Volcker nor A. Greenspan ever managed to change the monetary policy control target during their respective terms as Chairman of the Board of Governors of the Federal Reserve System.

      If it were easy to make the conversion, it would have been made by now, if it could be done.

    5. The question posed by Project Syndicate to the four experts was direct to the appropriateness of the annualized rate of inflation target of 2% ± 1% as a monetary policy control point.

      Michael Boskin’s comment frames the question, but doesn’t answer it. He raises the interesting point that the current measure of the rate of inflation overstates the ‘actual’ rate of inflation by 100 basis points.
      Bridgette Granville assets that the target of 2% ± 1% per annum is arbitrary. She states that pursuing tighter monetary policy conditions in the economy would “..mean further compression of real household incomes, which have been stagnating or declining, on average, for the past decade, and heightening the risk of social and political disruptions.” Her preferred approach would be translated into a higher target range for the monetary policy control point.
      Kenneth Rogoff discusses the pros and cons of the current target (2% ± 1% p. a.) and then opts to stay with the status quo monetary policy control point because it is a settled thing and change would be disruptive.
      John Cochran offers the view that the current monetary control point of 2% ± 1% p. a. is akin to reducing the length of the standard metre stick by 2% each and every year. The analogy is stark, but how appropriate is it? He suggests changing the control range to 0% ± 1% p. a., arguing that the reasons put forward for retaining the status quo monetary policy control point are of dubious merit.

      On the narrow question of the appropriate level for the monetary policy control point, PS received one punt, and three answers: 0% ± 1% p. a., 2% ± 1% p. a., and somewhere north of 2% ± 1% p. a. Taking median, the resolution is to stay with the status quo, i.e., the policy control point of 2% ± 1% p. a.

    6. "But, when the academic states that potential GDP will be used as the basis for determining a price index that will determine monetary policy, then that is when academic practices (modelling) begin to go wrong and wrong in the catastrophic sense of Long-Term Capital Management L.P. (LTCM) which is the exemplar of academic hubris that affected the real economy in ways that the academics didn't anticipate and couldn't control."

      The issue isn't what happens if the academic gets things wrong, the issue is who bears the risk when the academic gets things wrong.

      See equity financing.

  12. "Others ask what about mismeasurement. Answer: fix the measurement."

    Can you elaborate on this, or point to posts in the past that have? Quite interested in more context on the problems and proposed fixes.

    1. You can't fix it; inflation even as a theoretical idea becomes incoherent over long enough time periods because any basket of goods is going to change in ways that can't be quantified.

  13. “aiming for 2 percent inflation every year means that after a decade prices are more than 25 percent higher, and the price level doubles every generation. That is not price stability, yet they call it price stability. I just do not understand central banks wanting a little inflation.” - Paul Volcker, former Fed chairman

  14. I once told Prof. George Selgin, an advocate for nominal GDP target, that a major problem I find is that, to achieve a nominal GDP target, the Fed will have to two variables: inflation--which has proven to be difficult even to forecast--and real GDP, which is not affected by monetary policy, at least in the medium term. Thus, instead of one, the Fed might be trying to achieve two goals, a more difficult job.

    Selgin answered that nominal spending is only one variable, which I quite not see clearly. For example, what will happen if real GDP is slowing down? Will the Fed push for more inflation and face the risk of stagflation?

    1. Total nominal spending (NGDP) is one variable in the sense that it's (relatively) easy to define and measure. Inflation is incredibly complex to define and measure (it requires measuring "quality differences" across years in the items that make up the CPI) and RGDP is of course just NGDP adjusted for inflation, so it's downstream of both the NGDP measurement and the inflation guesstimate.

      Advocates of NGDP targeting (including me) believe that social welfare is maximized when total nominal spending is stabilized. "Stabilized" typically means kept on a publicly prescribed level-path increasing say 4% per year. If real growth slows, then inflation will be higher, but assuming real growth is at least zero, inflation won't go above 4%. I'm not sure what you mean by "stagflation" --- the 1970's are not possible under NGDP targeting --- but yes, if we had many years of 0% population growth and 0% productivity growth, then under NGDP targeting, there would be 4% inflation. If we believed that population and productivity growth would be zero for a long time, we might lower the NGDP growth target closer to 2%. But, in worrying about 2% vs 4% inflation, I think you need to consider the alternative, which is when the Fed allows NGDP to collapse like in 2008-9, throwing tens of millions of people needlessly out of work and causing them to needlessly lose their homes. That can't happen under NGDP targeting either, which is a hugh improvement.

  15. I suppose I'm confused how the Fed would target the price level or NGDP, especially if you believe Hayek that prices communicate relative scarcity. I don't see how targeting those metrics would work well if prices fluctuate based on scarcity and/or cost to produce, as factors of production fluctuate, too.

    Bazookas may solve temporary price pain but at a certain point, the money has nowhere to go, so the only thing left to do is bid up prices with larger wads of cash, and meanwhile, producers want a chunk of that surplus liquidity, too. There's an equilibrium sure, but demand-pull and cost-push inflation are not going away anytime soon.

    I remember in my econ classes long ago Pigouvian taxes effectively reduced externalities (tax roads to reduce number of cars on the road). Well, if inflation is an externality, reduce consumption via demand and turn on the crimp on the spigot in the form of higher rates. Yes, durable goods suffer, financing operations gets more expensive, and hurts the poor disproportionately.

    Subsidies for production to lower costs? The government can probably borrow at cheaper rates compared to anyone. Or do a reduced temporary sliding scale on taxes to mitigate the pain. But these have problems, too. No medicine is perfect. But I don't know...

    Inflation does seem to be going away, but the Fed has been hawkish after their self-inflicted wounds with the whole transitory mess.
    2% is a reasonable inflation target. It prevents deflation to an extent, but that specter is always lurking. A little inflation gives people time to adjust. Beyond 4% that’s when it feels like getting punched in the face by Iron Mike. It triggers a scramble to adjust.

  16. How exactly would we define or measure "price level" ?

  17. re: "First, just what does the Fed do to hit a nominal GDP target?"

    That's a piece of cake. You target legal reserves.

  18. I'm going to nitpick you here, fiscal policy cooperation would not be necessary for price level targeting. It would be necessary for some sort of 'real return' policy regime in which you were targeting the real value of the currency net of interest. This should be intuitive for you given you have a firm understanding of FTPL.

    The bigger problem is that a constant price level target would likely require the use of negative interest rates to achieve effectively.

    Overall, it's not clear to me that the price level target would be any better than an inflation target. You run into most of the same issues.

    1. Fiscal cooperation is absolutely required for price level targeting. The Treasury must commit to repaying its debts at the price level target, and not count on inflation. A price level target is prime FTPL. As was the gold standard, which is similar. The gold standard needs fiscal backing, enough taxes or future taxes to get gold if needed. FTPL book goes on about this at length. Negative interest rates are not needed.

    2. This comment has been removed by the author.

    3. You’re confusing a price level target and a return target. Say a company wants to target the price of its shares. It can reverse split the stock in response to a negative shock or split the stock in response to a positive shock.

      In the same manner the monetary authority can control the price level without explicit fiscal cooperation. But it would likely have to reverse- split the debt stock vis a vis negative rates.

      The monetary authority could not guarantee a specific price level each period, but it will be able to achieve a longer term target.

    4. You're thinking about returns here not prices. Say a company wants to target the price of its stock. In response to a negative shock it can reverse-split its shares, in response to a positive shock it can split its shares. The monetary authority can do the same to achieve a price-level target over the long-term.

      Note that a reverse split and negative rates are the same thing.

    5. "This stipulation makes it all but impossible for the U.S. Treasury to manage the government's budget, and the Federal Reserve Board of Governors to maintain the central bank's independence (such as it is). Congress will not bind itself to cooperation to maintain a price level target (entailing a loss of independence of action)."


    6. Ahh Blogspot said there was an error so I posted my response twice.

      Re: Eagle Eye
      The main thing here is that an exchange rate peg and a target are subtly different things. Under a true peg the returns on money must always be equal to the returns on whatever the asset it's being pegged to is.

      A true peg requires a monetary (interest rate) commitment because covered interest parity/impossible trinity is a thing, and it requires a fiscal commitment because the government is guaranteeing a return.

      A target does not require a fiscal commitment because you aren't guaranteeing a return, you're just trying to keep the price level constant over time. You can do that by contracting/expanding the value of debt vis a vis interest rates.

      Long-term debt makes things a bit more complex, but we can imagine fiscal regimes where the government issues only short-term or floating rate liabilities.

  19. I think the 2% inflation target is popular because that's the point at which most people will either not notice or just not bother to include inflation into their consuption decisions.

    Decades of data reveal that expected inflation does nothing to the economy. 2% is identical, in terms of consumption decisions, as -2%.

  20. Wouldn't restating the inflation target higher simply reprice rates higher as well? Accordingly, bringing about the very same pain that folks are saying can be avoided by the abandonment of the 2% target.
    It's unclear to me how this corrects anything.

  21. Apparently, you know nothing about legal reserves. You can't run a regression test on the time series. And Dr. Richard Anderson screwed up the reconstruction after the DIDMCA.

  22. I was thinking for a while why everyone is focusing on IT. IR is very transitory when price level dynamics are bad. So the path to IT can be false guide. Therefore I am defending price level adjustments or targeting more crucial to prevent cyclical movements. Glad to see others are on the same page


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.