The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.

« Funds Rate Probabilities: Locking In 5 | Main | CPI -- Cool... CPI ex Food and Energy -- Cool... Median CPI -- Not So Cool »

March 16, 2006

Thinking About The CPI

Whatever today's CPI report might bring, David Wessel sets us up for the inevitable grumbling about why it does not mean whatever it seems to mean.  Writing in today's Wall Street Journal (page A2 in the print edition), Messr. Wessel covers most of the usual complaints:

If You Think Inflation Is Contained, You Don't Live In My House

Some consumer skepticism is inherent in the way a price index is calculated. It is an average -- and no one is average. It measures the change in prices in a basket of goods and services that matches no one person's shopping list: Medical care accounts for 4.8% of the CPI. But if you are chronically ill, you will notice rising health-care costs more. College tuition accounts for 1.1%. But if you have a child in college -- well, you get the idea. The CPI is a national average: The fine print shows that prices rose 1.7% in 2005 in the San Francisco area, but 4.8% in the Miami area.

Give a Statistician an Inch...

Monthly changes in the CPI, the ones that get the most media attention, are adjusted to remove usual seasonal fluctuations. "In many cases, we may talk about a component declining when it actually rose, but less than it historically did in that particular month," says Patrick Jackman, the Bureau of Labor Statistics price maven. "It doesn't make much sense to the man in the street to tell him that gasoline prices declined, when they just rose less than the seasonal pattern expects."

U.S. government inflation figures also recognize that a $1,500 personal computer available today is more powerful than a $1,500 PC purchased three years ago. If it is twice as powerful but carries the same price tag, then the BLS counts that as a price cut...

And then there is housing. To the consternation of some economists, the CPI doesn't track the rising price of houses directly, but instead relies on rents, which have been rising more slowly than house prices...

If You Think Inflation Is Contained, Try to Live On My Paycheck

Consumers who complain that the government understates inflation may really be saying that their paychecks don't seem to go as far as they once did. For many Americans, that is true. But it isn't that prices are going up faster than the government estimates. It is that their wages aren't keeping pace.

Should we have sympathy with these arguments?  Sure,  They in part reflect the different uses to which any price index is put.  But from my perspective, none of them are really compelling reasons to suggest that policymakers should discount what the official figures have to say.  The blunt tools of monetary policy can at best control the average pace of price-level growth.  Likewise, monetary policy is not the right place to look for a palliative for weakness in real income growth.  And while seasonal adjustments, quality adjustments, and the exclusion of asset prices are all reasonable, and well-vetted, issues for debate, there are good reasons for all of those decisions.  At the very least, not seasonally adjusting the price statistics, ignoring quality improvements, and confounding durable expenditures with the price of consumption flows would be just as problematic.

David Wessel may be right...

... the complaining...never stops.

... but I'm willing to think that the CPI really does tell us something about what monetary policymakers should care about -- the general purchasing power of money.  Heck, I'll even take the "core CPI" seriously.  But that's another argument.

March 16, 2006 in Inflation | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Thinking About The CPI :


David Wessel -- "U.S. government inflation figures also recognize that a $1,500 personal computer available today is more powerful than a $1,500 PC purchased three years ago. If it is twice as powerful but carries the same price tag, then the BLS counts that as a price cut..."

I did not know that.

If the BLS is going to play it that way, then I want them to count as 1/2 all items sold that are not as large as they used to be. The list is long, but here are some examples:

1. candy bars - come on, they are getting smaller unless you go for the monster bars
2. any jar of peanut butter that is no longer filled near the top, say down 1.5 inches
3. cereal boxes - weight is ok (maybe), but that box looks half full
4. minature drink bottles
5. underwear packages - I swear there used to five in a package, now it's three - it should a law that BLS price them individually - now try to tell us they are cheap)
6. Ritz crackers - can anyone really open those plastic wrappings without destroying some crackers? Geez. Where are the scissors?
7. All computers shipped without systems recovery discs - add $100-$500 for the product liability pain and suffering
8. Diet Coke 3 liter bottles - no sane person drinks the bottom 1/4 if over two days old as the fizz is gone; might kill you anyway (it's really a 2 liter bottle, BLS)
9. batteries - I know that they don't last as long as they used to; everyone knows that!
10. chips in a tube - hey, where are the rest of the chips (use a ruler; remember your measurement for six months)
11. All containers of food goods that magically shrink
12. windshield wiper blades - they don't last nearly as long as they did a few years ago, and they're not cheap

BLS doesn't have this type of list, right? Start one...


Posted by: Movie Guy | March 16, 2006 at 07:21 AM

Again, I urge readers to speak to economists and BLS statisticians if you want to know about the Boskin PCE deflator.

But if you want to know how hot inflation is, speak to a small business owner or a housewife as to whats happening with prices.

Or, check out your favorite restaurants -- and look at what is happening to menu prices . . .

Posted by: Barry Ritholtz | March 16, 2006 at 11:09 AM

Movie Guy: Your list is exactly why the BLS does its "hedonic" adjustments to adjust for quality. If Mars sells a smaller candy bar for the same price, that's a price increase to BLS statisticians.
Those who argue against hedonic adjustments would argue that the price of the candy bar hasn't changed at all.
These same guys are whining that the CPI market basket changes over time, so it's not really tracking prices of the same goods. Those are the guys who want BLS to track the price of 8-track tapes, buggy whips and Edsels.

Posted by: fred c. dobbs | March 16, 2006 at 11:37 AM

You're cracking me up Movie Guy, but you forgot another example... movies! How do they hedonically adjust for the fact that you now have to sit through 25 minutes of silly ads even as Hollywood puts the screws on directors to make their films shorter and shorter?

Posted by: Anonymous | March 16, 2006 at 11:49 AM

I'm not sure why you're sitting through 25 minutes of ads. Adjusting arrival time is easy. What's irritating is the exorbitant cost of refreshments. I only take dates that have large purses!

Posted by: cb | March 16, 2006 at 11:59 AM

I didn't think anyone who read this blog could GET a date. I'll have to adjust my assumptions about people who read economics blogs.

Posted by: fred c. dobbs | March 16, 2006 at 12:46 PM

Dave, IF CPI "is an average" then shouldn't it be relatively easy for Economists to build a representative sampling from our population who, at the end of the year, are net "winners" from the lower inflation we've seen since the Boskin recommendations were adopted? Obviously, it would exclude all social security recipients. I would guess it would exclude families with significant ongoing medical or college expenses, as well as those doing major home remodeling & those having more than 2 kids. Has anyone conjured up a sampling representation of a portion of our population whose yearly purchases are likely to be growing at a rate less than the CPI?

Posted by: bailey | March 16, 2006 at 04:30 PM

Anyone purchasing food or energy or education or capital goods or housing or luxury goods or entertainment or anyone who pays taxes is a loser under the current calculation. Anyone who depends on defined benefits investment income, pays for medicine or insurance or healthcare is a loser. All the rest of us are winners.

Posted by: Robert Cote | March 16, 2006 at 05:39 PM

A study done here in Colorado reports that households with income at or below the poverty level spend 16% of their income on energy, compared to 5% by households at the median income level. Clearly, these two types of households saw inflation very differently over the same recent periods. Is there data available to contrast the weights given to different items in the CPI basket for different, say, income quintiles?

Posted by: Michael Cain | March 16, 2006 at 05:58 PM

bailey and Michael -- I think you might find what you have in mind in the annual reports of the Consumer Expenditure Survey: http://www.bls.gov/cex/
If you are looking for raw data, you can find it here: http://www.nber.com/data/ces_cbo.html

MG -- great post. In the future I will try not to read your comments when I'm supposed to be listening to other people, lest they think I am laughing at them.

fred -- ditto.

Posted by: Dave Altig | March 17, 2006 at 04:36 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search

Recent Posts



Powered by TypePad