CPI Null Set (via The Economist)

This is a classic example of working the refs: If enough long-only fund managers/sell side pundits spew the same misleading nonsense long enough, it will eventually find resonance somewhere in the mainstream media.

Today’s case in point is from The Economist, a publication that ought to know better. They not only focus on the Core CPI, but they take it a step further: Core CPI ex Housing OER.

First off, I need someone to explain why food and energy is considered volatile — if the volatility has been exclusively in one direction. That’s not merely volatility, that’s a trend

Second, let’s credit The Economist with creating a whole new category of measuring increasing prices: Inflation ex inflation (ex inflation). That’s right, we take the basic measure of inflation, remove the vital components of life consumed by every living human being (food and energy). Then, just in case that is still too inflationary, remove Housing (see chart below). 

That’s Inflation ex inflation (ex inflation). 

We should take it a step even further. I propose a whole new category to measure inflation: CPI Null Set. Start with the basic CPI, then go to the Core, removing food  and energy. Then take out housing. Then remove everything else. Ta-da! No inflation!

Here’s an excerpt:

"But much of that jump is thanks to a sharp rise in the cost of housing (which makes up almost 40% of core CPI), particularly the category of “owners’ equivalent rent” which estimates the cost of living in a house by looking at rents charged on similar properties. Although this measure makes sense in theory (by living in your house you forgo rental income), it may now be overstating inflationary pressure.

As the housing market has slowed, fewer people are buying property, choosing to rent instead. That has pushed up rents. In turn, owners’ equivalent rent has risen too, even though homeowners have seen no change in the actual costs of owning their house. Because owners’ equivalent rent is estimated net of utility prices, recent falls in gas and electricity bills have paradoxically made matters worse.

Statistical quirks, in short, are distorting the picture. But what should central bankers do about it? Some suggest that owners’ equivalent rent should simply be dropped from the inflation index. That is what European statisticians have done. But credible central bankers cannot suddenly ignore an inflation component when it starts behaving in ways they do not like. That was the mistake made in the 1970s, when officials deluded themselves that inflation was under control by excluding ever more prices from their indices."

Funny, I do not recall The Ecomomist discussing how OER understated inflation over the past 5 years. Where were you guys? Oh, that’s right, no one was working the refs then. I guess quirks in reporting standards only matter if it impacts someone’s portfolios/bonuses.   

As poorly considered as that was, the Brits do manage to redeem themselves with a paragraph towards the end of column:

"The bigger point is that even if you take out housing costs the recent acceleration in core consumer prices does not disappear (see chart). And a variety of other gauges suggest that underlying inflation is on the high side and rising. The deflator for core personal-consumption expenditure (PCE), Fed officials’ favoured index, was up 2.1% in the year to April. The “trimmed-mean PCE deflator”, calculated by the Dallas Fed, which excludes those prices that have risen and fallen the most before taking a weighted average of the rest, is up 2.4%. The “median consumer-price index”, calculated by the Cleveland Fed, is up 3%. Look at these figures and the surprise is less that the central bankers are now so jumpy about inflation than that they sounded so sanguine earlier this year." (emphasis added)

Glad to see someone across the pond hasn’t drank the kool-aid.

Hey guys: You were on the verge of being punk’d. Get with it.


American inflation: Feeling the heat
Jun 22nd 2006 | WASHINGTON, DC
From The Economist print edition

What's been said:

Discussions found on the web:
  1. rick commented on Jun 26

    A great idea would be now to substitute
    the avg price of housing for rent component, since it seems
    that house prices could fall now while rents could
    rise,,,, that way , inflation could be well under
    control once again in the core.

  2. erikpupo commented on Jun 26


    I dont remember the Economist talking about the understated effects of rent when OER was low the last 4-5 years, due to the housing bubble. Everyone knows if OER had not been used and if the full real estate appreciation measure had been, the inflation over the last 4-5 years would have been higher and the Fed might not have cut to 1%.

    Everyone also knows OER is used now because when the Feds used real estate appreciation in the 1970’s, the inflation index got skewed even higher, so they switched to a rent measure to try to make the CPI go down.

    But heck, if the Feds are allowed to do it, may I remove the price of housing, food and energy from my budget calculations? I can just tell the bank “dont have to pay my house, doesnt count” and the grocer “food is free, doesnt count” and the gas station “fill er up!”

  3. Trent commented on Jun 26

    Hey, it’s about time you strategists got a taste. We stock pickers have been dealing with EBITDA, EEBS (earnings excluding bad stuff) and EWE (earnings without expenses, which is used to value Baa securities) for some time now.

  4. FliteTime commented on Jun 26


    Now, now. We still have to pay the mortgage, gas, & groceries. We just don’t have to pay “more” for them than we did a few years ago. Got an ARM? No problem! You still shouldn’t be paying any more than 4%! Gas is still a ‘meager’ $2 a gal, and all that Kool Aide that I get from Publix is still under a buck! Yay, Kool Aide!

  5. KL2005 commented on Jun 26

    excluding housing as an exercise provides another view of what is going on. I think we all agree there is inflation. I think we all agree the FED should tighten.

    The debate begins about how bad is inflation and how far to tighten. If you think housing will fall, you do not want them to tighten to far.

    They report both core and non core inflation. The FED should not ignore either (and I do not think they do).

    The Fed is not ignoring inflation they have just adopted a more gradual .25% approach. They have been tightening for quite some time now.

    It is only prudent they consider what happens if certain items stop rising (oil, housing). It does not mean they stop tightening, but I also do not want them to tighten to far.

    Greenspan screwed up and created a nasdaq bubble. Creating more imbalances is not the solution. Bernanke has been dealt a tough hand to play.

  6. GRL commented on Jun 26

    As the housing market has slowed, fewer people are buying property, choosing to rent instead.

    And exactly what statistics support this assertion? How many of those statistics are “seasonally adjusted” or otherwise artificially constructed?

  7. Si commented on Jun 26

    Nice one Barry. The potential and ability of these policy makers to mess around with numbers is unnerving.
    The Brits are certainly not immune from the this either.
    I was over in the uk during their election time, I remember their finance minister trotting out a nice, big, bright line graph of the kind you would show nine year olds, revealing how interest rates had fallen over time.
    It brought a smile to may face, I wonder how many Brits knew that the relationship in inflation calculations between the start and end of the graph was heavily compromised. Not many it would seem.

  8. Big Al commented on Jun 26

    As long as the US government can report “No inflation”, they don’t have to increase SS payements,etc. what would the deficit be like if the true inflation had been used?

  9. DavidB commented on Jun 26

    True enough Big Al but because they are fiddling with the numbers to keep payments low to senior citizens what they are doing is bordering on either theft, fraud or embezzelment. Why doesn’t the AARP get ahold of a keen economist and put a class action suit toghether against the government? Sure, it may drive up the deficit but better the money be put in serniors hands than in the hands of the five fingered crooks behind the till.

  10. JoshK commented on Jun 26

    Or, you could just make an index based on computer component costs and then show deflation. You could keep the same acronym, CPI – computer price index.

  11. Tim commented on Jun 26

    The Economist is stealing my charts:

    How Not to Fight Inflation

    To get the full effect though you really have to see OER along with the Core CPI and Core CPI ex-OER, then go on to suggest that the Fed should lower rates to fight inflation.

  12. Blissex commented on Jun 26

    «That is what European statisticians have done.»

    That is because index-linking is even more pervasive than in the USA and removing rising prices from an ”inflation” index saves a lot of money to companies the the government.

    «But credible central bankers cannot suddenly ignore an inflation component when it starts behaving in ways they do not like.»

    No? They cannot? They have and they will continue to do so. Are they still «credible»? Well, being a central banker is a big confidence game. Credibility has nothing to do with it, it is all about confidence, hitting the right buttons, not making persuasive, credible arguments.

    «That was the mistake made in the 1970s, when officials deluded themselves that inflation was under control by excluding ever more prices from their indices.”»

    Nice little ominous reminder :-)

  13. Blissex commented on Jun 26

    «As poorly considered as that was, the Brits do manage to redeem themselves with a paragraph towards the end of column:»

    Barry, I suspect that you are not used to dealing with the enigmatic English… Most people at the Economist are not just enigmatic Englishmen, they also are mostly impenetrable Oxbridge mandarins.

    These people are trained to express themselves subtly and hypocritically according to a a set of conventional dissembling techniques. Here is the decoder ring:

    * Always argue both sides of an argument.
    * Make sure the side you prefer subtly looks better.
    * Make sure that the side you don’t prefer is first.
    * Ostensibly, always endorse the official line.
    * You can sneak in a bit of subversion too.

    So the article reads like ”we gotta say that inflation ex-inflation ex-inflation is all fine and good, because that’s the official line, but then we suavely add what we actually think, that things are not looking good”.

    My understanding of the yankee way is instead:

    * Only argue your side of the argument.
    * Exaggerate wildly that side of the argument that benefits your wallet most.
    * Put the wildest exaggeration first, and the weakest last, unlike Oxbridge dons no real yankees read to the end.

    Now you can see a great potential for misunderstanding… :-)

  14. cm commented on Jun 26

    Blissex: Excellent summary!

  15. Blissex commented on Jun 26

    «First off, I need someone to explain why food and energy is considered volatile — if the volatility has been exclusively in one direction. That’s not merely volatility, that’s a trend. »

    Well, for energy the explanation is easy: retail energy prices do go down even in nominal terms.

    Note that the CPI is really not an inflation gauge for policy purposes, whatever ”inflation” is, it is first and foremost an indexing tool for COLAs.

    Now, if retail energy prices go down, and they have done so at several points in the past, COLAs might end up negative, but that is considered politically impossible, so the direct influence of energy prices is excluded.

    For those countries where most energy is imported energy prices are excluded from COLAs, quite deliberately and explicitly, because workers should then not be protected from energy price increases (and equivalently should not be penalized when energy prices go down).

    Let me explain here: COLAs were invented to forestall constant renegotiation of wages, because that can easily get into a positive feedback loop as categories of workers leapfrog each other (a bit unrealistic today that workers’ leverage is low, but it happened in the 70s).

    The idea of COLAs is that they protect all workers from the inter-sector distributional effects of price increases, so they can just enjoy their COLAs without renegotiating their wages all the time.

    But if energy is mostly imported, COLAs including energy in effect end up protecting workers from redistribution to/from foreign entities, not workers in other sectors, and this is just not fair; if a country’s GDP is redistributed away to the oil sheiks, the country becomes poorer, and the COLAs should not attempt to shield a category from a country-wide effect.

    Also, and importantly, retail energy prices correlate very highly with wholesale internationally traded energy prices, and very quickly too.

    Excluding food prices instead has essentially no justification, because retail food prices bear essentially no relationship to wholesale food prices, as the cost-of-goods in retail food prices is really small.

    All this said, some deindexing of COLAs actually has some merit: because the problem is that if COLAs reflect, and quickly, 100% of price increases but 0% of price decreases (if any) then they may end up triggering the inflationary spiral that they are designed to forestall.

    So dampening a bit the index in which COLAs are built helps avoid the possibility of a positive feedback loop.

    Ideally COLAs should be based on a good and accurate (a representative basket, not one ”adjusted”) index, the index should exclude prices of things that are mostly imported, the COLA should be say 95% and not 100% of the movement in the index, and should negative COLAs should be possible; that is they should protect most of the relative wage differentials among categories of workers and retirees.

    But this is politically infeasible, so the indices on which COLAs are based are tricked, first with the best intentions (to dampen them a little, a little hysteresis helps), and then immediately thereafter when the politicians realize that they can get away with it, mercilessly for expedient reasons.

  16. cm commented on Jun 26

    There are fundamental philosophical difficulties, and resulting controversies, how to construct indicators (and which) that have, or can be construed to have, a “social goal function” aspect. For example, CPI can be viewed as defining a particular lifestyle to be targeted or accomodated.

    One can then argue what are “necessities”, “staples”, “luxuries”, etc. For example, should the Fed target the price of alcoholic beverages? Or of house prices (as opposed to rent)? Jewellery vs. bread?

    At any rate, that’s not at issue here. The problem is that there is an apparent preference for removing items that are largely non-discretionary, because they for that very reason have price leverage, and show inflation trends reliably, if not accurately.

  17. cm commented on Jun 26

    Blissex: My latest comment was not meant in response to yours (I hadn’t seen that), but it worked out nicely, didn’t it.

  18. Estragon commented on Jun 26

    My biggest problem with the Economist article is “owners’ equivalent rent has risen too, even though homeowners have seen no change in the actual costs of owning their house”. The arithmetic suggest otherwise.

    The actual costs of owning a home are surely affected by the discounted terminal value of your investment. Until recently, discount rates were dropping and expected terminal values were increasing. Assuming a fixed initial price and stable carrying costs, overall cost was dropping. Now the discount rate is increasing, and expected terminal value growth has levelled off, so the current cost of housing is increasing.

    Few homeowners may be explicitly aware of this calculation, but people are good at absorbing changes implicitly and tend to act accordingly.

  19. VJ commented on Jun 26

    Oh come on. It’s not as if Americans spend money on food and energy every day. What are you tryin’ to do, put a crimp in brokerage commissions ? Have you seen how much BMWs have gone up in price ?

  20. rick commented on Jun 26

    LCD’s are really dropping in price, lets get that
    into the core, take out rent, and put LCD’s,
    that should create deflation in the CPI !!

  21. knzn commented on Jun 26

    Before 1971, the critical target for monetary policy was inflation-ex-everything-but-gold. That actually seemed to work pretty well, so I don’t think it’s such a problem to exclude food, energy, housing, and a few other things from a broad index.

    Even if all the volatility in energy is upward, it should still be excluded, because the objective is merely to anchor monetary policy while keeping employment high. If the Fed had to constantly counteract a rapidly rising price of energy, wages would have to fall in order to maintain employment in the face of an increasing resource scarcity. Cutting wages is tough, and what usually happens instead is falling employment. Better to have the aggregate price level trend upward (in a controlled manner) than to have protracted economic stagnation. (Unless, of course, your opinion is that, what is good for long-term bondholders is good for America and vice versa.)

    Housing is another issue. The reason to exclude it is that it reacts perversely to monetary policy. By tightening, the Fed increases the cost of rental housing, which leads to more tightening. The result is a policy that is more volatile than necessary. We saw this in reverse a few years ago, when the excess policy volatility was on the easy side. It is a general problem with products that are capital-intensive, but especially with housing services, which are essentially pure capital.

  22. Big Al commented on Jun 26

    knzn. Do I understand you? Inflation is a monetary thing, Cost Of Living is a population thing and there is little or no connection?

  23. Estragon commented on Jun 26

    knzn – I’m not clear on why something that “reacts perversely to monetary policy” should be excluded from CPI. I understand your point that (all else held equal) it sets up a positive feedback loop through monetary policy, but it is what it is, isn’t it? If housing costs are increasing, shouldn’t that show in the data whatever the source? Surely the past to more stability in monetary policy is through a deeper understanding of the data, not willful blindness.

  24. eric commented on Jun 26

    Ah! ….. so what about HGX jan. 07 puts not below 200?

  25. jopo commented on Jun 27

    i would like to point out that the economist actually did run an article about cpi understating inflation in its economics focus about a year, maybe a year and a half ago where they quoted a study that said if a truer measure of housing costs (i.e. not oer) was used headline inflation would have been running at closer to 6% over the previous 3-4 years. conspiracy anyone? press me and i could find it, as a bond vigilante myself, i have saved it somewhere.

  26. knzn commented on Jun 27

    I’m not saying we should stop looking at the headline CPI altogether (or God forbid, stop even reporting it). I’m just saying that, to guide monetary policy, and to judge monetary policy, we shouldn’t use the headline CPI as the primary objective. Of course bondholders do have to concern themselves with whatever purchasing power is relevant from their personal point of view, and in most cases that will be closer to the headline CPI than to the core (any core). To some extent it’s just a value judgment, but I don’t think that it should be the Fed’s job to protect bondholders from all forms of rising prices.

  27. knzn commented on Jun 27

    To modify my last comment, where I said the headline CPI is more relevant for most bondholders; this is only true in the sense that the headline CPI is (closer to) the thing for which they would want a long-term forecast. Historically, I think the conventional core CPI is better at forecasting the headline CPI than is the headline CPI itself. (That may have changed now, but I doubt it, for reasons that I won’t go into.) In making such a long-term forecast, it may also be helpful to exclude temporary business cycle effects, of which housing may be one. But for a bondholder, all this is only justified if one can show that it produces a better forecast for a comprehensive price index such as the headline CPI.

Read this next.

Posted Under