The scary difference between the GDP deflator and CPI – the case of Japan

Most inflation targeting central banks in the world are targeting inflation measured by the Consumer Price Index (CPI). However, if you want to target inflation CPI is probably the worst possible measure to focus on. Why? Because CPI includes both indirect taxes and import prices – something the central bank can certainly not control.

If the central bank targets CPI it would in fact have to tighten monetary policy in response to negative supply shocks such as rising oil prices. Similarly the CPI targeting central bank would effectively be “forced” to tighten monetary policy in response increases in indirect taxes. Do you think this is foolish? Well, the ECB is doing it all the time…just think of the catastrophic rate hikes in 2011 in response to higher oil prices and austerity induced indirect tax increases across the euro zone.

A much better measure to target – if you want to maintain an inflation targeting (I don’t…) – would be to target the so-called GDP deflator as this measure of prices by definition excludes import prices and indirect taxes. Targeting the GDP deflator therefore would reduce the problem of monetary policy reacting to positive and negative supply shocks.

You might think that the difference between CPI and the GDP deflator is small and frankly speaking that used my view. However, the difference is far from trivial, which the case of Japan’s deflationary experience over the past 15-17 years clearly illustrates. The graph below shows the development in the Japanese price level measured by both CPI and the GDP deflator.

While CPI indicates that the Japanese price level today is around 2% lower than in 1995 the GDP deflator is telling us that prices have dropped nearly 20% in the last 17 years. The difference is stunning and is certainly not something that should be ignored, but unfortunately I doubt that most central bankers are aware about just how great these differences are.

It should of course be stressed that it is not normally so that CPI will be upward biased compared to the GDP deflator, but if tight monetary policy is leading to long periods of low or no growth and that forces the government to increase indirect taxes to improve public finances – as it has been the case in Japan – then there very likely will be an upward biased in the CPI compared to the GDP deflator.

This conclusion obviously is highly relevant for the conduct of monetary policy in the present situation – particularly in the euro zone, where governments around Europe are increasing indirect taxes in a more or less desperate attempt to improve public finances. With the ECB’s focus on consumer prices (the HICP in the euro zone) rather than on the GDP deflator higher indirect taxes implicitly leads to tighter monetary policy – something which is hardly warranted in the present situation.

Therefore if central banks want to continue targeting inflation they should at least change from CPI targeting to GDP deflator targeting – that would be a small, but important step away from repeating the Japanese scenario.

PS This discussion is less relevant for the Federal Reserve as the Fed is targeting a the PCE core inflation measure, which is much closer to the GDP deflator than to CPI.

Related posts:

The dangers of targeting CPI rather than the GDP deflator – the case of the Czech Republic
Failed monetary policy – the one graph version

Leave a comment


  1. Great graph Lars. Do you ever adjust for the difference between CPI and PCE when looking at Breakevens/TIPS? I usually think of breakevens as a forward inflation indicator (in the monetary policy sense), but I don’t think that is quite correct because TIPS adjust off of CPI, not PCE. The mismatch makes TIPS an interesting asset, since the payoff increases on both the right and left tails (higher growth or higher prices from supply side shocks).

  2. Aha! Didn’t realize the gap was so large in Japan… that explains a lot. And if there is such a big divergence between the headline target rate and the rate that actually has an impact, wouldn’t wage-setting proceed along the headline rate, leading to labor-market tightness for over a decade…

  3. Chris

     /  November 8, 2012

    Is there a comparison to other OECD countries available to evaluate this theory?

  1. The scary difference between the GDP deflator and CPI – the case of Japan « Economics Info
  2. Malaysia should peg the renggit to the price of rubber and natural gas « The Market Monetarist
  3. Quick Digression on Japan’s Upcoming Election « The Law School Tuition Bubble
  4. Ambrose on Abe « The Market Monetarist
  5. Answering Tyler’s question on Japan with old blog post « The Market Monetarist
  6. Why the Yen Is Now Fairly Valued, USD back as Preferred Funding Currency - SNBCHF.COM
  7. Transmission Channels of Central Banking and QE | worldofinterest
  8. profitfunction
  9. Phillips + MMT + Zeitgeist = Živi zid | Tko je John Galt?

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 )

Facebook photo

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

Connecting to %s

%d bloggers like this: