Right-wing and anti-Social Security forces in the U.S. have been successful in suppressing retiree benefits, workers’ wages/salaries, and programs indexed to the consumer price index (CPI).  Because of many techniques for suppressing the official measure of the CPI –advanced by conservative economists over the past few decades – middle and low income wage/salary earners and retirees have lost ground to the cost of living.

One of the gimmicks now offered by conservative economists, politicians, bureaucrats, and pundits, is known as the “chained CPI.”  What is the “chained CPI?”  This bulletin will “boil it down” and provide a catalyst for a wider discussion of this misguided notion.  In general, it must be said that in the history of big pseudo-scientific ideas, the chained CPI ranks right up there with the homunculus, Piltdown Man, the free market, the bleeding of patients with leeches, eugenics, and Lysenko genetics.

The chained CPI and its underlying justification by the Bureau of Labor Statistics (BLS) through a subset of assumptions about human behavior are little more than cynical attempts at manipulation of the CPI to the detriment of workers and retirees and to the advantage of corporations and the super-rich.  This bulletin will cover three concepts essential to the chained CPI construct: (1) “substitution bias,” (2) the “hedonic adjustment,” and (3) the geometric mean versus the arithmetic mean.

Any lay person with no more than a basic knowledge of high school math can understand these concepts and no one should be intimidated by them.  Trust me; the chained CPI is the invalid reduction of the inordinately complex dynamics of consumer behavior and price fluctuations in the world’s largest economic system to a few simple mathematical models (the U.S. officially has a GDP of nearly $16 trillion or $16,000 billion).  In the world of science, this fallacy is known as reductionism.  You will see that fallacy clarified as you read on.

Let’s look at the three concepts underlying the chained CPI construct:

Substitution Bias

Substitution bias simply means this:  when prices rise on a staple or service regularly purchased or needed by an individual or group of individuals, the individual or group of individuals will substitute a cheaper product or service for what was regularly purchased or would be purchased; ipso facto, when prices rise, the cost of living actually declines.  For instance, if you are accustomed to buying steak and the price of steak increases, you are compelled to purchase ground round; therefore, your cost of living has actually gone down.

Believe it or not, that is the theory.  That’s pretty much all there is to it.  Even though it doesn’t pass the laugh test, economists have been all too eager to come forward with complex “econometric models” to overlay this nonsense with a patina of science and mathematical validity.  As a statistics professor, I can tell you that these models are of no higher quality, i.e. statistical validity, than the nonsensical assumptions on which they are based.

Hedonic Adjustment

Do you think that higher prices for goods and services imply that you have a much better product than you had at a lower price, and therefore your cost of living actually decreases because of all of the bells and whistles on the ever expanding array of the gewgaws, electrical devices, apparel, etc. that are offered up to consumers these days?  Have you looked at the charges on your phone bill lately?  Were they there in 1950, 1960, 1970, 1980, or even 1990?

It is this notion that has led the BLS to lower the CPI in recent years in spite of so many increases in the prices of necessities without a correlative increase in quality.  What are students and families paying for education these days versus the 1960s or even much later than that?  What does your health insurance cover these days for the price increases you have seen?  OK.  Enough of this – we could go on and on about the absurdity of this notion.

Geometric Mean Versus the Arithmetic Mean

Considering that the CPI involves a survey of thousands of products in a “market basket” of goods and services, calculating averages is a major mathematical technique for arriving at changes in prices.  There is more than one method for calculating an average.  For instance, by middle school, practically all numerically literate people have a good handle on the arithmetic average.  Unfortunately, only mathematicians/statisticians have exposure to the other types of “means.”  For instance, the “geometric mean” – applied now by the BLS in calculation of the CPI – is calculated by taking the product of all values in a dataset rather than summing all of the values.

Here is an example of an arithmetic mean versus a geometric mean:  three people ages 19, 20, and 21 would have an average age of exactly 20, i.e. (19 + 20 + 21)/3 = 20.  A geometric mean would be the cube root of 19 * 20 * 21 = 19.98331943.  The geometric mean in this case is ever so slightly lower than the arithmetic mean.  Three issues are important here: (1) very small reductions in the CPI due to mathematic gimmicks are compounded over time and result in major losses of benefits and wages, (2) as the number of values and variance of the distributions of values entered into the calculation of the geometric mean, divergences of the geometric mean and the arithmetic mean increases, and (3) the CPI has been suppressed by about ½ of 1 percent each year by this technique throughout the past decade (thinks to the Boskin Commission in 1996, which is a subject for a later bulletin).


No valid, credible, empirical/scientific evidence exists to support substitution bias and quality improvement, which have resulted in the hedonic adjustments, lower CPIs, and, consequently, a call for aneven more conservative CPI known as the chained CPI. Nevertheless, the BLS, an agency responsible for fairness and economic justice involved with the determination of price increases and charged with looking out for the well-being of all U.S. citizens in adjustments in the CPI, has decided to apply a mathematical gimmick in repressing benefits, wages, and salaries.  They are, in fact, making life more difficult for the broad mass of working people and retirees.

No doubt, tracking changes in in prices of goods and services within the U.S. economic system is a massive and complex undertaking.  Nevertheless, we expect more than overly-simple models, the origins of which can be traced to conservative think tanks and the results of which advantage the wealthy and penalize the masses.

It is important that we call out the stewards of our economic well-being and demand that we have an open and honest explication of reasons for suppressed wages and benefits along with discussion of poor and middle class quality of life and how it has changed due to the disappearance of masses of good paying employment and major increases in necessities of life – including education, health care, and housing.

Officially, the CPI has increased at an average of 2.1% from 2000 through 2011.  Even considering the negative -.34 CPI change in 2009, it would seem as though 2.1% is an unbelievably low average increase in the CPI over the first decade of the current millennium.  Enough is enough.  We expect fairness and will demand it.


  1. It’s really very complicated in this full of activity life to listen news on TV, therefore I simply use the web for that purpose, and take the most up-to-date news.

Comments are closed.