Guest post: Why “Integral” is wrong about Price Level Targeting (by J. Pedersen)

I have always said that my blog should be open to debate and I am happy to have guest posts from clever and inlighted economists (and non-economists) about monetary matters. I am therefore delighted that my good friend and colleague Jens Pedersen (I used to be his boss…) has offered to write a reply to “Integral’s” post on price level targeting versus NGDP level targeting. Jens who recently graduated from University of Copenhagen. His master thesis was about Price Level Targeting.

Jens, take it away…

Lars Christensen

Guest post: Why “Integral” is wrong about Price Level Targeting

by Jens Pedersen

The purpose of this comment is two-fold. First, I argue that ”Integral” in his guest post ”Measuring the stance of monetary policy through NGDP and prices” is wrong when he concludes that the Federal Reserve has done a fine job in achieving price level path stability and by this measure does maintain a tight stance on monetary policy. Second, I present a way of evaluating the Fed’s monetary policy stance based on the theory of optimal monetary policy.

“Integral” assumes that the Federal Reserve has targeted an implicit linear path for the price level since the beginning of the Great Moderation. Following Pedersen (2011) using the deviation in the price level from a linear trend (or the deviation in nominal GDP) to evaluate the stance of monetary policy needs to take into account the potential breaks and shifts in the trend following changes in the monetary policy regime. Changes in the monetary policy committee, changes in the mandate, targets etc. may lead to a shift or break in the targeted trend. Hence, the current implicit targeted trend for the price level (or nominal GDP) should correctly be estimated from February 2006 to take into account the change in president of the FOMC, or alternatively take account of this possible shift or break.

Changing the estimation period changes the conclusions of “Integral’s” analysis. Below, I illustrate the deviation in the log core PCE index from the estimated linear trend over the period 2006:2-2006:12. As the figure show Fed has significantly undershoot its implicit price level target and has not achieved price level path stability during the Great Recession. Currently, the price level gap is around 3% and increasing. Hence, looking at the deviation in the price level from the implicit price level trend does indeed suggest that monetary policy should be eased.

Following, Clarida et. al. (1999), Woodford (2003) and Vestin (2006) optimal monetary policy is a dual mandate which requires the central bank to be concerned with the deviation in output from its efficient level and the deviation in the price level from its targeted level. The first-best way of evaluating Fed’s monetary policy stance should be relative to the optimal solution to monetary policy.

However, this method requires a clear reference to the output gap. Common practice has been to calculate the output gap as the deviation in real output from its HP-filtered trend. This practice is by all means a poor consequence of the RBC view of economic fluctuations. Theoretically it fails to take account of the short run fluctuations in the efficient level of output. Empirically it does a poor job at estimating the potential output near the end points of the sample.

Fortunately, Jordi Galí in Galí (2011) shows how to circumvent these problems and derive a theoretical consistent output gap defined as the deviation in real output from its efficient counterpart. The efficient level of output corresponds to the first-best allocation in the economy, i.e. the output achieved when there are no nominal rigidities or imperfections present. Galí further shows how this output gap can be derived using only the observed variables of the unemployment rate and the labour income share.

The chart below depicts the efficiency gap in the US economy. Note, that this definition does not allow positive values. It is clear from the figure that at present there is significant economic slack in the US economy of historic dimension. The US output gap is currently almost 6.5% and undershoots its natural historical mean by more than 1.5%-points.

Hence, the present price level gap and output gap reveal that the Federal Reserve has not conducted optimal monetary policy during the Great Recession. Furthermore, the analysis suggest that Fed can easily increase inflation expectations by committing to closing the price level gap. This should give the desired boost to demand and spending and further close the output gap.


Clarida, et al. (1999), “The Science of Monetary Policy: A New Keynesian Perspective”

Galí (2011), “Unemployment Fluctuations and Stabilization Policies: A New Keynesian Perspective”

Pedersen (2011), “Price Level Targeting: Optimal Anchoring of Expectations in a New Keynesian Model”

Vestin (2006), “Price Level Targeting versus Inflation Targeting

Woodford (2003), “Interest and Prices”

Leave a comment


  1. jpirving

     /  June 21, 2012

    Nice post Jens. Though I wonder if it makes sense to consider the Greenspan -> Bernanke handover as a break. It doesn’t seem to me that expectations should have shifted to a slightly higher trend in that switch over, I vaguely recall Bernanking being view as less of a maverick. Anyway, point stands, I have calculated similar trends before on headline CPI starting from 2000 to 2006q4 and get a similar result. They are clearly failing by on both ends of the mandate. You could show the same thing with the GDP deflator too.

  2. Here is my latest post “FOMC: ditch the sandbags, grab the buckets.”

  3. Jens Pedersen

     /  June 21, 2012

    jpirving, my point is that, if you are assuming an implicit targeted trend in the price level or nominal GDP, you need to consider possible breaks and shifts following changes in the policy setup. Fed lacks transparency, so things may change when the FOMC or the president changes. Remember, we do not know the true trend.

    More importantly you need to check the stationarity properties of the variable of interest. If bygones are bygones, then it is a clear sign that this variable has not been of concern to the central bank.

    As you will find in my thesis, Fed has kept the price level stationary, which shows that Fed has been concerned with price level path stability. I am not sure you will find the same result for nominal GDP. Hence, the observed nominal GDP path stability is really a result of Fed stabilizing the price level and real GDP separately, and not an explicit focus on nominal GDP.

    To me that is the strongest argument for Fed to announce a price level path target and returning to its pre 2007 policy, and not a target path for nominal GDP. And as my post shows, doing this will leave plenty of room to ease monetary policy.

  4. Integral

     /  June 22, 2012


    Thank you for the response. I’ll try to have a longer reply ready later this week. For now, a few notes:

    1. There are certainly arguments for using core CPI instead of headline: the Fed ostensibly responds to core rather than headline, core leaves out particularly volatile prices, et cetera. I use overall CPI because it seems to tell the story best: by eyeballing it, you can “see” the “bubble” in 2006-2008 and the “crash” back to the old trend line. Furthermore it was headline inflation that the Fed cited as suggesting “equal risk of negative growth and inflation risk” in its September 2008 decision to keep interest rates steady. Reading the minutes, the Fed clearly keeps an eye on both core and headline inflation when making decisions. I could, however, be convinced to use core if I found the arguments strong enough.

    2. I also agree that we can draw any line on a graph we want, but it doesn’t necessarily mean anything of importance. I’ve never been a huge fan of trying to glean the stance of monetary policy by simply looking at atheoretical trends. Despite these reservations, I use the full Great Moderation trend, rather than just Bernanke’s tenure, because the GM was generally considered a period of “good monetary policy” and many Market Monetarists argue that we should go back to the GM trendline. I wished to show that in terms of prices, we already were back on that trendline. That conclusion is true for headline measures though not for core, as you have pointed out.

    At best the price level evidence is ambiguous: while by one measure the Fed is failing its mandate (core CPI) it is not at all clear that the Fed is failing if you look at a broad swath of price indicies. It is immediately clear, however, when one looks at the NGDP evidence, which is what I wished to show.

    3. Jordi’s paper is interesting: I’m always skeptical of deriving output gap measures from models without capital, but it’s a good first step. I’d like to see more discussion of the stance of monetary policy using model-consistent metrics rather than mere trendlines and welcome your contribution. i certainly agree that the output gap is likely negative and likely large relative to past recessions, and I certainly agree that the output gap evidence suggests room for monetary policy action. I derived similar conclusions using the nominal GDP evidence.

  5. Jens Pedersen

     /  June 22, 2012


    I very much appreciate your response and the opportunity to have a discussion of these matters. And of course a thanks to Lars for facilitating an excellent forum for discussion of monetary policy.

    Since Fed has never been very explicit about its target for monetary policy, using an implicit statistical derived trend as a reference point for current stance of monetary policy is attached with many problems and uncertainties. First of all, one needs to make sure that there at least exists an implicit targeted trend. On this point your use of nominal GDP fails as a valid reference point for the current monetary policy stance. In a proper-scaled graph nominal GDP does look to have been stable around a trend during the random chosen sample period – Great Moderation. However, I would argue that nominal GDP has in fact not been stationary; hence Fed has allowed shocks to nominal GDP to persist. Thus, no implicit targeted trend exists and your graph therefore does not tell much about the stance of monetary policy.

    This is of course not to say that announcing an objective to increase nominal GDP by X% would not work. I do agree, that looking at the development in nominal GDP does paint a rather gloomy picture of the US economy, and I certainly hope that it is part of all central banks’ research palette. In the case of Fed, the foundation for doing so, in my view, is just very weak.

    Regarding price level targeting, the opposite holds. It is possible to show that Fed has kept the price level stationary before 2007. This feature would greatly help communicate an explicit commitment to raise the price level. Looking at the recent decline in US inflation expectations this may be exactly the injection needed for the US economy to move out of the slump. And of course one needs to assess the level of real slump as well, which is really the most correct way to evaluate whether monetary policy has done its job. Galí’s output gap is of course based on strict assumptions, but is, nonetheless, a simple and theoretical consistent way of assessing the economic slack, which does not suffer from all the problems the random use of statistical trends do. (The great use of random statistical trends was actually what motivated me to write the response to your post in the first place) I agree that a more detailed model including capital is warranted. On this subject, I can recommend Galí, Smets and Wouters (2011): “Unemployment in an Estimated New Keynesian Model”.

  1. R. Glenn Hubbard – Romney’s economic adviser and the confidence fairy « dajeeps

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: