Poring over inflation with the Consumer Price Index in hand

Paul sez,
I just spent some quality time poring over the latest Consumer Price Index (CPI) tables, which are reported each month by the Bureau of Labor Statistics and determine the U.S. inflation rate. According to the release, prices for urban residents have risen 4.0% over the past year, and it's fun to look at the different numbers and see how they contribute to the overall result.

For example, in the Transportation category (up 8.2%), a dramatic increase in Gasoline prices (26%) is balanced in part by a more heavily-weighted decline in the cost of New vehicles (-1.1%). In the Food and beverages category (up 4.4%), relatively modest increases in Meats, poultry, fish, and eggs (3.8%) and in Fruits and vegetables (1.7%) are counteracted by sharper increases in Cereals and baking products (8.1%) and in Dairy and related products (11%).

Meanwhile, the entire index is pulled lower by something called "Owners' equivalent rent of primary residence," in the Housing category, which accounts for 23.942% of all expenditures and rose 2.6%. I'm a layperson in economics and statistics, but I'm hooked-- I'm looking forward to seeing April's numbers, which come out May 14.



  1. Check out the May 2008 issue of Harpers

    The contention made by Kevin Phillips is that over the past 25 years, the calculations for the leading economic indicators, specifically CPI (consumer price index), the GDP, and the unemployment rate have been tweaked in order “to create a false sense of economic achievement and rectitude, allowing us to maintain artificially low interest rates, massive government borrowing, adn a dangerous reliance on mortgages and financial debt even as real economic growth has been slower than claimed.” Phllips does not lay the blame at any particular political faction, stating that this process was abetted by Republicans and Democrats alike in an effort not to rock the boat.

    Consumer Price Index
    – Nixon: Separated “core” inflation rate from the headline inflation rate. Core inflation excludes food and energy prices.
    – Reagan: Substituted housing prices with “owner equivalent rent” in calculation of CPI, masking high housing prices.
    – Bush I: Proposed new calculations for inflation to lower inflation numbers
    – Clinton: Implemented Bush’s proposals.
    – Bush II: Created experimental CPI calculations that resulted in even lower figures. Stopped publishing M-3 money supply numbers, which captured rising inflationary impetus from bank credit activity.

    – Johnson: Created the “unified budget” which combined Social Security with other federal outlays to mask deficits.

    Unemployment Rates
    – Kennedy: Re-categorized people who have given up on looking for jobs as “discouraged workers” rather than “unemployed”.
    – Nixon: Unsuccessfully requested that the Labor Dept publish only the lowest of the seasonally adjusted unemployment rate and the un-adjusted rate.
    – Reagan: Categorized military personnel as “employed” instead of “outside the labor force”.
    – Clinton: Excluded discouraged workers from calculation of the overall workforce. Decreased monthly household economic sampling from 60,000 to 50,000. a disproportionate of those excluded were from the inner cities.

    If the economic indicators were to be calculated based upon the criteria from 25 years ago, they would be as follows:
    Unemployment Rate: 9 to 12 percent
    Inflation: 7-10
    Economic Growth: little economic growth since 2001

    Some very disturbing and sobering figures.

  2. Zuzu, actually those changes make sense to me. How can you categorize an Army Sargent drawing a paycheck for 30 years as not employed? That is just crazy.

    Also, they made the correct decision on the “discouraged worker” thing. How can you be considered “unemployed” is you aren’t looking for work? The Bureau of Labor Statistics says a discouraged worker is someone who has not been actively seeking employment. How can they be “unemployed” if the aren’t even looking!! Anyway, if you want to see how many “discouraged workers” are out there, the BLS keeps track of that too, so its not like it is being hidden.

  3. Cory, this leads me to believe that you are a numbers geek. I’m one too. I have an interest in any sort of more or less raw data that I can examine and play with. The CPI is a great example, but I have also been known to create rather complex spreadsheets that calculate fuel economy, average price per gallon of fuel for the month/year, price per mile for each tank, each month and each year, and other assorted (mostly useless) pieces of information from the gas receipts and odo readings of my car.

  4. I came to the realization that canned goods are going to increase in price. They’re heavy, and the food will last a very long time. This is exactly what can’t be cheap for very much longer. Fill your bomb shelters now!

    Seriously, I love anchovies and canned fruit. I’d hate for them to be prohibitively expensive, like mediocre beef.

    Soylent Green is People!

  5. I don’t know how much longer it will be up, but the inflation for the past year is presented in a neat style chart here:

    The basket of goods in CPI does not measure inflation; it’s merely an indicator of inflation.

    The price of gold (which has relatively flat supply and demand) is also an indicator of inflation. (When the price of gold goes up, that’s actually the value of the dollar going down. The price of gold recently hit $1000/oz.)

    The money supply data (e.g. the M3) also provides indicators of inflation in the form of how the source of inflation — increase in the supply of money and credit — trickles down through the economy (aka the “non-neutrality of money”). Oh, except that the Federal Reserve stopped publishing that indicator in 2006.

    Another way that they under-report inflation is called geometric adjustment.

  6. Keep in mind the failure of CPI and pretty much all measures of inflation (as well as GDP etc) to capture the full effects of improvements in quality and product mix. This manifests itself in two ways.

    First, quality: an iPod today still costs about the same in nominal terms as it did when new, but is far from the same product. Cars today in real terms are not only quite a bit cheaper than 10, 20, 30 years ago (etc), but you get more in a sub-$20k hatchback than a Cadillac which a baby boomer may have lusted after, so the actual change in price is massively underestimated.

    Second, this especially being pertinent to CPI which measures prices of baskets of goods which are chosen somewhat arbitrarily: as new goods enter the product mix, they aren’t initially considered part of any basket; they have to be entered into the records. This generally doesn’t happen until they have fallen to a fraction of the price they originally commanded, when a new and novel technology. The archetypal example is the VCR, which debuted at something like $30,000.

  7. @RRSAFETY #2:

    I think in anyone’s book, someone who is not working is not contributing to the economy while someone who is working, is contributing to the economy (broken window fallacy notwithstanding). Therefore, if someone was working, and no longer is working for whatever reason, the economy takes a hit.

    In that sense, it doesn’t matter if that person is “unemployed”, “discouraged”, “disabled”, or “not part of the labor pool”.

    This is why some people choose to look at the Employment data rather than the Unemployment numbers.

    I also disagree with your contention that separating out food and energy from the core inflation rate was a good idea. I agree that it’s nice to see the inflation rates of individual items, but when you get right down to it, you’re looking at an average across the entire populace, and while different people spend predominantly on different things, the *overall* economy takes a hit regardless of what item shows inflation.

  8. It’s also important to keep in mind that the CPI is only a statistic, and that there are wide variations in people’s actual experience of price changes.

    As an example, here in NYC, the price of take-out food has skyrocketed in the past year. Almost every restaurant has upped their prices at least 10-20%. So this doesn’t matter to people who cook their own food in Peoria, but for a culinarily-challenged person like myself, this definitely makes a lifestyle impact.

    Regarding the manipulation of statistics for political purposes, I would generally take that with a grain of salt. As previously noted, the numbers are still out there, and you are not required to concentrate on those the politicians choose to trumpet. The people who produce these numbers are generally apolitical professionals.

    The more interesting question is not so much how much prices are going up, but why. We seem to be entering a low growth, high inflation period of the economic cycle, the dreaded “stagflation” of the 1970’s. The fact that high petroleum prices happened then give no solace to our current prospects.

    A great source for the statistically inclined:

  9. How can they be “unemployed” if the aren’t even looking!!

    That’s mental Twister. Unemployed means not employed. Zuzu’s point is that we’ve used semantics to manipulate statistics to manipulate people to manipulate policy. Plain language gives a bleak picture, but an accurate one. Too bad nobody notice it a decade ago.

  10. Rent is the only thing guaranteed to rise more than the cost of anything else you buy on a monthly basis. Which have you experienced more often? A yearly rent hike of $50 a month or a yearly pay raise of $50 a month?

    And yeah, I’d like to see the “real” employment statistics: meaning “full-time, 40 hours a week” vs “languishing in a temp job where my company considers 30 hours to be full time”.

  11. I couldn’t find a consolidated historical table of what w-grl wanted (hours worked), but the Bureau of Labor Statistics does make available “Total unemployed, plus all marginally attached workers plus total employed part time for economic reasons, as a percent of all civilian labor force plus all marginally attached workers” (Series ID LNS13327709). Of course, there is no direct link to LNS13327709, you have to construct it via http://www.bls.gov/webapps/legacy/cpsatab12.htm (chalk the inability to directly cite to bad software design, rather than deliberate government obfuscation)

    That’s a pretty inclusive definition of people who in a lousy work situation and the data is for the last ten years.

    Here’s a picture of that inclusive definiton over the last ten years. And yes, it has worsened over the last year, in addition to being far above the 10-year low of October 2000. Then again the plain old unemployment figures look pretty much the same as the all-inclusive figure, though the vanilla figures start at a lower base. Both give a similar picture of national unemployment.

    Going back to the CPI and its components, at least computers are cheaper : )

Comments are closed.