Does the CPI understate inflation?

You’re hearing this a lot these days, most of all from Kevin Phillips.  David Leonhardt sets the record straight.  Here is one excerpt:

During the 1980s and 1990s, though, did you ever stop and marvel at what a small share of your paycheck you were spending at the supermarket? I didn’t. I also didn’t really notice that gas cost less in the late 1990s than it had in the 1980s. Yet lately, every time my wife or I pass a new benchmark for filling up our tank – $40, $50 and now $60 – we have a conversation about it.

Price increases are simply more noticeable – more salient, as psychologists would say – than price decreases. Part of this comes from the notion of loss aversion: human beings dislike a loss more than they like a gain of equivalent size. If you have to sell your house for less than you bought it for, you’re really unhappy. You hate that ground chuck now costs $2.83 a pound, but you didn’t notice that oranges are 31 percent cheaper than they were a year ago.

…The price of major appliances has been flat over the last year. Furniture is 1 percent less expensive. A decade ago, a basic four-door Toyota Corolla LE cost $16,018, according to the company. The 2009 basic model costs $16,650, and it’s a safer, more powerful, more fuel-efficient car than its predecessor.

To top it all off, most people don’t buy any of these items very often. “People tend to remember things they do frequently,” says Stephen Cecchetti, an economist at Brandeis University who studies inflation. “And what do you buy more frequently than gas and food?”


Does anyone really think that inflation is overstated? How ignorant is that? Stop and think for one minute about how much the computer you are using to read this would have costed 15 years ago. To what extent is that taken into consideration when caclulating CPI?

Now do the same for your car, your TV, your MP3 player, your DVD player, your ..., well, you get the picture.

Yes Joe, but for the things that I buy most often prices have gone up a lot lately. I rarely buy T.V.'s, MP3 players and DVD players.

For the people whose consumption as a percentage of income is mostly spent on food and transportation (poor people in the U.S.), inflation is almost surely understated. And let's not forget the food riots in other countries...

After I read the article this morning, I was left thinking "Give me a break." Maybe Leonhardt went to lunch with Ben Stein and decided everything was dandy, no reason to worry.

Two things I find curious in the CPI calcs:

1. Imputed rent. The largest single component in the calc. This would suggest to me that as house prices fall, a huge chunk of the CPI would also fall--suggesting better times when in fact the result is people experiencing quite the opposite.

2. In the NYT infographic recently, health insurance seemed to constitute a tiny chunk of the index. (No percentage given.) What's with that?

I understand that if you are in an accident with a Corolla, it automatically turns into a casket. Very efficient.
Consumer electronics are obsolete at least every 3 years. No wonder they are cheap. Look at VHS tapes. Useless. My kid's MP3 player didn't last 6 months. And does the CPI measure taxes? Property taxes never go down.

Yeah, VHS tapes are useless, unless you have a VCR, in which case they still work just fine. I watch my old VHS tapes all the time.

And great example, by the way, since VHS was introduced almost 30 years ago and were the market standard for about 15 years (1985-2000).

"technology advancement isn't deflation. At least not in my book."

I don't understand this view. If I bought a computer last year for $2,000, and now if I went to buy a computer for $2,000 it would be twice as powerful, then wouldn't I expect the store to have a computer equivalent to mine at something like $1,000? How is that different than deflation?

"Despite the housing bubble and the food price rise, in general the portion of money we spend on food/transport and housing has been going down. This means the overestimation of inflation has been going up (since new technologies and quality improvements in all other areas are understated)."

The 'rule of thumb' for purchasing a primary home used to be 2-3x household income. What is it now - twice that?

Does the fact that savings rates are at an all-time low imply that everyone is spending discretionary income on baubles and trinkets, or is it possible that more of people's incomes are being allocated for increasingly expensive necessities? Real income has been falling for many Americans.

Just because there are goods whose prices actually decreased is still not a valid counterargument. It's not pure speculation to point out that prices of these goods would have even decreased more without a general increase in the price level. Inflation is not about some shifts in relative prices (it's a fallacy to believe that high petrol prices drive inflation), but about general price increases. Obviously, the CPI does understate inflation only due to the fact that they adjust their methodology in order to keep rates lower as they would be otherwise.

I don't know how credible the source is, but it is nevertheless astonishing to which extent CPI rates differ according to the methodology employed (compare pre-Clinton and today).

Following my plan, I will concoct a theory of loss aversion, then do the background reading. I have reasons for purposefully being backward.

My theory, people fear price volatility and have a high preference for stability. Loss aversion would be derived from our fear of volatility.

The cause would be simple, we live our lives with promises of future production, promises which are estimated under the assumption of some stability. A loss of stability requires a near term renegotiation of the future, an increase in stability allows us to hold off the expensive renegotiation.

Now, I go back to the material, looking for a specific relationship between loss aversion and volatility.

OK, as far as the efficacy of the CPI, it can be manipulated for propaganda purposes. And we can argue about it as an accurate indicater of inflation.

For the sake of argument, when shortages develop, profits increase, and more producers enter the field. At least in an open market.

Let's assume that manufactured goods are as good or better than goods of the past and productivity has negated inflation. But what about the long-term price of energy? We are left with the fact that demand for commodities such as oil has outstripped supply. We are left with this reality until supply increases or demand declines, or, more likely, a combination of both. It takes time to increase the supply of commodities.

What are the implications for this? Smaller cars, smaller houses, more energy efficient appliances until a new "equilibrium" is reached, that is budgets of households are adjusted for new price levels. Or will Congress allow oil companies to drill for more oil? How long will the adjustment period last? If more cars in the world mean more demand for oil, which should government policies be?

Has the enrgy crisis been engineered by political interests to push voters into a certain political or "crisis" frame of mind? Or for servcies, is there a shortage of medical services or has government restricted the supply through medical licensing and restrictions on building new hospitals? Is this truly an open market? Who is doing the policy analysis?

The rubber meets the road in buying patterns. If we can continue equal buying year after year, on a similar wage, without hardship, there probably isn't much price inflation. If the same shopping, the same lifestyle, no longer works on that same wage, there probably is price inflation.

I think what happens with food and fuel is that people too close to the edge, suddenly find their budgets blown.

(FWIW, my two largest fixed costs, my condo association dues and my health insurance, both went up 10% last year.)

Cliff: How anyone can have any confidence in the accuracy of the inflation measures is beyond me. Hedonic adjustments, substitution effects, these are very complicated and difficult in practice to turn into an equation.

This hits the nail on the head. The real question is why so many economists take CPI so seriously when it is just a subjective formula, subject to manipulation depending on the interests of the politicians and bureaucrats who set it. Furthermore, for inflation CPI is a trailing indicator, lagging far behind price changes in commodities. It is utterly stupid to use CPI as an economic decision-making tool.

good reference: Time Well Spent: The Declining Real Cost of Living in America ( check out the pdf at )

it's interesting that from 1920 to 1995 the average number of hours needed to work per square foot of average housing for the average worker hovers around 5-6 hrs/sqft for the entire perioud. that makes makes 2005 about 13-14 hrs/sqft? amazing...

Isn't this rather confusing? I thought a traditional definition of inflation generally meant something along the lines of 'reduced buying power' not 'certain items cost more in real terms than in times past'. I mean isn't it inevitable that certain sectors are volatile whereas other are relatively stable? Or other commodities are rather prone to shocks of shortages and surpluses? If everyone used a fixed currency of a, say, 10 million gold coins wouldn't prices still fluctuate? Or alternatively if we're investing in real estate we're happy to see prices rise even though would-be home buyers aren't? You know?

Okay, I'll go there. Inflation is the increase in money supply chasing goods.

It's also an indicator for how good the government is doing. The government compiles this indicator themselves. During periods of cost push supply shock inflation, they take out "volatile" food and energy, even though the high costs of food and energy take money away from consumers who could have driven up other prices. Do they ever put food and energy inflation back into the calculation after the volatile periods are over?

Motive, means, and opportunity.

I don't understand this view. If I bought a computer last year for $2,000, and now if I went to buy a computer for $2,000 it would be twice as powerful, then wouldn't I expect the store to have a computer equivalent to mine at something like $1,000? How is that different than deflation?

Ah, but there's the trick. I could buy something as powerful as my last $2000 PC for about $200 these days, but I wouldn't, because I wouldn't be able to buy any software that would run tolerably on it, and I wouldn't be able to do all the normal things that people do with modern PCs.

Expectations have risen. About 15 years ago, I wanted a portable music-playing device, and bought a Sony Walkman. Today, for a similar amount of money I could get an mp3 player. What I probably would do is spend rather more money and get a portable device with a big hard disc that plays video as well.

So have prices fallen, because a Walkman is now "free" from the dumpster, remained constant because a portable music playing device is about the same price, or risen because the portable device that I would like to buy is more expensive than the portable device I wanted 15 years ago?

It's easy enough to pick a basket of food that represents a reasonable varied, nutritious family diet in keeping with the current taste (people today are unlikely to want to buy tripe, for example) and track its cost. It's easy enough to look at the price of a family home, and although family homes have tended to grow (sharing bedrooms is now a rarity) that growth has been fairly slow.

For consumer toys, entertainment and so on, tastes and technology change so fast that it's almost impossible to define an "inflation" in a meaningful way.

So, I just checked what they are charging for rent, at the apartment complex I moved to after I graduated from college, in 1998. It was 1100 then it is 1150 now.

I would imagine that those most impacted by the rise in gas and food prices are lower income people who tend to rent. So, yes, food and gas prices are up but rent is most definitly not.

I second randy's comment from 1038 am on May 9.

One important point to make is that probably at least 95% of economists (conservative, liberal, Democrat, Republican) think that the BLS's measurement of the CPI has improved over the years. Of course, you have to be interested in what the CPI is trying to measure: the cost of reaching a given standard of living through purchases of consumer goods and services. It is not trying to measure asset price inflation, and it increasingly is moving away from measuring the price increased in a fixed market basket of goods and services. Now maybe you think all the economists of various political persuasions for some reason are interested in reducing Social Security payments, or whatever. Still, I think people might want to ask why most economists think the BLS measurement of the CPI has improved.

After reading Kevin Phillips's article, I looked at the Shadow Stats website he cites, and I also relooked at various article by BLS on changes in CPI measurement over the years. The Shadow Stats website does a terrible job of explaining its methodology. I am skeptical that the website does an accurate job of measuring how inflation would look in the CPI under old methodologies. From looking at various BLS articles, the difference between the old and new methodologies appears to be far less than the numbers produced by Shadow Stats. Why people trust this website when they don't trust BLS is puzzling to me.

Finally, for what it's worth, at the present time, if we reverted to measuring owner occupied housing prices by asset prices, measured inflation would probably be lower than it is now, given the decline in owner-occupied housing asset prices.

Randy -

There's no need for conspiracy theories when it's all out in the open. Check out the Boskin report:
An explicit goal of the Boskin report was to curb the growth in social security spending. I think they were right to do so, but it is obvious that they are changing it for political reasons.

Most economists get paid by the state. Professors rely on federal grants. I am continually amazed that people who would never trust the climate science of Exxon-Mobile, somehow completely trust the economic "science" of the federal government. The Fed is both responsible for preventing or creating inflation, and hires the economists who measure it. Why would you trust the Fed to accurately measure inflation, any more than you would trust Exxon-Mobile to accurately measure the impact of CO2?

Being pro-science means anti-cargo cult science. When the number of variables greatly exceeds the amount of data, the result becomes extremely susceptible to picking the variables that give you the results you want.

Comments for this post are closed