Introduction

In late November, 2007, the Commerce Department’s Bureau of Economic Analysis (BEA) announced the United States had achieved a third quarter real Gross Domestic Product (GDP) gain of 4.9 percent. The price index for gross domestic purchases, which (theoretically) measures prices paid by U.S. residents, increased 1.6 percent in the third quarter.

At approximately the same time, the Department of Labor’s Bureau of Labor Statistics (BLS) reported the Consumer Price Index (CPI-U) increased by 2.36 percent from Q3 2006 to Q3 2007 (calculated using actual BLS data, not seasonally adjusted).

These numbers appear to be overly optimistic. With an accelerating rate of inflation and declining economic activity, how could the economy of the United States manage to achieve a high growth, low inflation, performance?

In this essay, we examine how the CPI is calculated. In a following essay, we will look at the computation of GDP. 

Why Should We Care About The CPI?

Because it has a key role in determining and measuring America’s economic performance: 

• Approximately 30 percent of Federal spending is tied to movements in the CPI. 

• The interest rates paid by Treasury Inflation Protected Securities (TIPS) are tied to movements in the Consumer Price Index. 

• Thousands of civilian contract obligations are tied to the CPI, and 

• Deducting the rate of inflation from current dollar GDP is a way to measure “Real” GDP.

For individual consumers, the CPI is especially important:

• If you are getting Social Security (or expect to be on SS in the near future), the amount you receive each month is tied to the CPI.

• If you are getting Supplemental Security Income (SSI), a veterans pension, or military retirement, your benefit is tied to the CPI.

• If you are getting an inflation adjusted civilian pension, the size of your check is probably tied to the CPI.

• If you are working under a labor contract that includes a CPI adjustment clause, the size of your pay check is tied to the CPI.

• Tax receipts, including individual income tax brackets, personal exemptions, and the standard deduction are tied to movements in the CPI.

And the biggest reason of all. 

If the BLS is deliberately underestimating the rate of inflation,
you are receiving a smaller paycheck than you would get
if the CPI were calculated using a balanced methodology.

Now the truth. A reality check. It is in the selfish-best-interest of the federal government to keep the CPI calculation as low as possible. Doing so saves the government money.

Deception?

One would think the methodology used by the BLS to calculate the CPI would be an important issue in the upcoming elections for Social Security recipients, veterans, pensioners, and anyone working under a CPI adjusted contract. In addition, it should be a hot topic for millions of “Baby Boomers” who will have to decide when to take Social Security during the next administration’s term of office. Since most of these people depend on their benefits check to pay for essential items such as food, clothing, shelter and transportation, one would think they would be very sensitive to the amount they receive each month. To them, the CPI represents a measure of what is costs them to pay their bills. But there is a problem:

If we examine how the CPI is calculated,
we find multiple examples
of questionable data manipulation
and institutional deception entrenched in public policy.

First of all, although the CPI is called a consumer price index, it is NOT a price index. To quote the BLS: “… the CPI focuses on approximating a cost-of-living index not a general price index.” Although originally introduced in 1978 to measure price changes, the CPI was changed to a “buying habits” index during the Clinton Administration. It no longer measures price change. It is not even a good measure of the cost of living. 

And this raises a question: if the CPI is not a price index,
then why is it called a price index?
Isn’t this a bit deceiving?

Media references to the CPI frequently refer to it as a way to measure price change over time. It is widely referenced in discussions about the cost of living. Yet both references are incorrect.

So. What does the CPI measure? The key words are “buying habits” and the introduction of this concept has a long history.

Before the Clinton Administration, the American Consumer Price Index actually tried to measure the prices consumers paid for an identical “basket” of goods each month. If steak cost $1.25 last year, and then went up in price to $1.30 this year, the annual rate of inflation for a pound of steak was calculated at 4 percent. If the average consumer price for a refrigerator was $300 in the first quarter of last year, and went up in price to $325 during the first quarter of this year, the rate of inflation for that refrigerator from Q1 last year to Q1 this year was calculated at 8 percent. And so on, for a long list of items. Price points for each period were determined by conducting a survey of retail industry outlets and associations. The change in total cost from period to period determined the rate of inflation for maintaining a constant standard of living.

During the first Bush Republican Administration, Chief economist Michael Boskin and Federal Reserve chair Alan Greenspan lobbied for a change in this methodology. They believed that when consumers could no longer afford a particular item, they would purchase a cheaper substitute. If steak, for example, became unaffordable, the consumer would switch to hamburger. If cars with V8 engines became too expensive, the consumer could purchase a car with a smaller engine. And so on.

The CPI, they argued, should reflect actual purchase decisions, rather than a fixed basket of goods that would gradually become irrelevant as consumers continued to substitute cheaper products for those on a fixed list of goods. In effect, they wanted the BLS to find ways to decrease the reported rate of inflation by tracking consumer buying habits as they struggled to find cheaper goods and services.

They got their wish. During the Clinton administration the BLS initiated a long and complex process to measure the rate of inflation based on “value” rather than “price”. It works this way. If the BLS believes the value of an item has increased from one period to the next, it decreases the item’s new price point by the value of the improvement. Thus if the car you buy this year has more features than the one you could have purchased last year, the price point is deflated to account for the added value of the new features. If this year’s health care is presumed to be superior to last year’s available health care, the added value is deducted from the CPI health care price point. Since this year’s personal computer has more power and features than last year’s PC, the added “value” is deducted from its new price point. And so on. Product after product. The adjusted cost of an item, as measured by the CPI, may go down even though the actual cash you pay for the item is going up. The technical term for this highly subjective data manipulation is called hedonic regression. It guarantees the actual cash you pay for goods and services is more than the phony price the federal government claims you paid for these goods and services.

The BLS further manipulates price data by tracking consumer substitution. Thus, if we can no longer afford steak, we purchase a cheaper pound of meat. If we can no longer afford to buy a mid-sized car, we purchase a smaller vehicle. If the price of cereals, eggs, poultry, and milk become unaffordable, the consumer is expected to substitute cheaper foods. By this process, the BLS uses a heavily manipulated CPI to track the prices consumers pay for the goods and services they actually buy. It does not, however, provide comparative pricing.

To quote the BLS: 

“Method evaluation. …. Before 1999, CPI used only Laspeyres indices, measures of the price changes in a fixed market basket of consumption goods and services of constant quantity and quality bought on average by urban consumers, … . The Laspeyres index, however, systematically overstates inflation because it does not take into account changes in the quantities consumed that may occur as a response to price changes. ….

Chained CPI for All Urban Consumers (C-CPI-U). This index applies to the same target population as the CPI-U. The same raw data are used, but a different formula is employed to calculate average prices. The chained CPI was developed to overcome a shortcoming of the CPI-U series, which does not account for the changes that people make in the composition of goods that they purchase over time, often in response to price changes. The alternative method of the C-CPI-U is intended to capture consumers' behavior as they respond to relative price changes.”

The BLS CPI is no longer a relevant measure of what it costs
to maintain a standard of living, dollar for dollar, period to period.
Rather, it has become a way to measure the cost of human survival.

There are (at least) three fundamental philosophical problems with the BLS methodology:

One: Taken to its logical conclusion, the BLS assumes consumers will continue to make quantity and product substitutions until they no longer have any options. One presumes this means hamburger will be substituted for steak, dog meat will be substituted for hamburger, a diet of grass will be substituted for dog meat, and starvation will be substituted for eating. 

It’s all very logical. Very sophisticated. And very theoretical.

But is it moral? This methodology ignores a key reason for measuring CPI in the first place – what does it cost the consumer to maintain a fixed level of well being? Or, to put it another way:

Two: Hedonic regression breeds phony money analysis. Economic worth becomes a complex game of pedantic simulation. The connection between simulated worth and actual cash value must inevitably become increasingly obscure. For consumers, hedonic regression means the aggregate rate of inflation will always be less than the actual prices buyers are charged for goods and services.

Is this institutionalized deception?

Three: For consumers, chained dollar values are meaningless economic drivel. Today’s consumers spend today’s dollars and pay today’s prices. Try buying a loaf of bread with chained “inflation adjusted” dollars. Or a refrigerator. A pair of shoes. A gallon of gasoline. It can not be done. For the consumer, “real” value is cash in hand. A pay check. Charges on a credit or debit card. We need to know how much stuff our cash will buy.

Isn’t it time the CPI became a bona fide Consumer Price Index?
Or perhaps the BLS should change the name.
Call it the Consumer Moribund Lifestyle Until Absolute Poverty Index.

Inflation Factoids

Medical Costs

Employee medical care costs were up over 5 percent in 2006 and over 10 percent in 2007. Individual health care coverage can cost $400 to $700 per month, and family coverage can cost $800 to $1,500 per month. The cost of health care has risen far faster than the published rate of inflation for the last 10 years. Rapidly rising and very expensive health care receives constant media attention.

Yet. The BLS claims these costs are up less than 5 percent.

Should we re-examine how health care inflation is calculated?

Food Prices

It would sure be nice to know where the BLS does its food shopping. They must get one heck of a discount. There appears to be a total disconnect between the food costs claimed by the BLS versus what consumers are really paying for food. From September 2006 through September 2007, commodity prices for corn went up ~43%, soybeans went up ~46%, wheat went up ~62%, milk went up ~64%, oats went up ~12%, and so on. But the BLS claims food and beverage prices only went up 4.4% during this same timeframe.

Huh? Does the BLS data make any sense? How long does it take for higher commodity prices to work their way through the food chain?

Look at the following food price inflation analysis. We do not, of course, have access to the raw data used by the BLS in computing food costs. But we can develop our own from industry reports and supermarket data. The “BLS” column shows the food inflation data reported by the BLS. The “TCE” column shows my results. Data is for September, 2007 versus September, 2006. Unadjusted indexes. BLS data is weighted by item. TCE data is weighted by volume and then cumulated. 

The BLS claims food prices increased by 4.4% during this period. Fudging as best I could to reduce my estimates, it would appear consumer cash outlays for food actually increased by at least 11%.

Should we challenge the BLS CPI data?

It would appear real world food prices accelerated at a rate that is (at least) 2.5 times faster than the rate of inflation reported by the BLS.

WHY?

Fuel Prices

The BLS reported motor fuel prices went up 8.6% from September 2006 to September 2007. Household energy prices (natural gas, fuel oil, propane, electricity, etc.) went up 1.8% during this same timeframe. How can this be true? According to the Energy Information Administration (EIA), oil prices went up 10.7%, propane went up 5.5%, gasoline went up 9.2%, diesel fuel went up 6.1%, heating oil went up 10.4%, natural gas went up .2%, and electricity went up 1% during this period. Although the BLS motor fuel price increases appear reasonable, average home fuel costs actually went up ~ 3.3%, not 1.8%. 

Should we challenge the BLS CPI data?

Weighting

The BLS weights consumer spending by category. Price increases are calculated by category and then multiplied by this weighting to determine the CPI. The following Table shows the importance of these weights (Price Indexes: Percent of all items, CPI-U, U.S. city average, December 2006 Base). The challenge, of course, is that these weights must change from period to period if they are to properly reflect how consumers are actually spending their money. Based on a careful analysis of BLS data tables, my estimates for Q3 2007 are shown in the TCE column.

The importance of this chart is not our disagreement with the BLS weighting (although that discrepancy should be resolved). The “real” importance is the projected changes in consumer spending. For several months pundits have been asking if higher food and fuel prices will force the consumer to spend less on housing, apparel, medical care, recreation, and so on. As shown by this chart, the answer is yes. Either consumers shift their budgeted spending to food and fuel, or go into debt to sustain an existing lifestyle. If they chose to accumulate more debt, most of it will accumulate on their credit cards.

By the way, creditcard.com reports two out of three Americans say they'll cut back on spending for other things as a result of higher energy costs in 2008, with nearly a quarter saying they'll cut back significantly on other spending.

If consumers chose to borrow against their credit cards in order to sustain existing spending patterns plus the added costs of food and fuel, it would appear uncollectible credit card debt could exceed 2% of consumer spending.

Compounding

The price deviation of the CPI from the prices consumers actually pay for goods and services compounds year by year. In order to make an estimate of the effect compounding has on the CPI, we can compare a sample of prices from 2000 (the last year of the Clinton administration) versus the prices consumers paid in 2007. We use Q3 data, the latest available for 2007.

The data in this Table shows that only one item increased in price at a rate that is below the BLS index. The rate of inflation for most of these items is substantially higher than the index.

Does this mean the CPI understates the rate of inflation? Did the BLS food index for Q3 2007 versus Q3 2000 understate the rate of inflation by 56%? Did the BLS total price index for Q3 2007 versus Q3 2000 understate the rate of inflation by 46%? 

So OK. There is no weighting to this comparison, the sample is much too small to be representative of the whole, and the car you buy today is superior to the iron we could purchase in 2000. But to a consumer, a banana is a banana. These are real world prices, and our purchases must be made in current, unadjusted dollars. If the index doesn’t mirror these prices, then it is NOT a price index.

If the CPI is understated, compounding also reduces your wage or benefit check. Let’s assume you get a check every month for $1,000 in year one. Then for the next 9 years, you get a CPI linked increase of 3%. Your total benefits of wages over this 10 year period would be $137,566.55. Now suppose the CPI should have been 4% each year. Your benefits or wages would be $144,073.29. If the CPI has been underestimated by 1 percent each year, you lost $6,506.74 in benefits or wages. 

How the CPI is calculated does make a big difference in the size of your paycheck. 

And the economic viability of the Social Security “trust” fund.

Alternative Calculation of CPI

If we use the weighting and data points from the above factoids to calculate an alternative estimate of CPI, we get a very different picture of American inflation from Q3 2006 to Q3 2007. There is a dramatic increase in food and housing costs. Note I accumulate my index calculations. 

Granted. Accuracy would require the acquisition and analysis of a lot more data than assembled for this effort. But the large discrepancy suggests something is wrong with either the survey methodology or the process of analysis. 

Whereas the BLS reported a CPI increase of 2.36% for this period,
the actual rate of inflation was more like 4.02%.

Blame

We can not blame the BLS for understating the CPI. These people are just doing what they were told to do, and they are doing their job with great enthusiasm.

But this begs a question: 

Who is responsible for determining the objectives
of the CPI calculation methodology?

Congress. Bill Clinton. George Bush. You know. Our elected officials. 

Since the current policy and methodology was established during Bill Clinton’s term in office, it is perfectly logical to ask Hillary (who also wants to be President) what steps she is willing to take to ensure the CPI mirrors the actual rate of inflation. As discussed below, this is a simple management problem. A United States Senator, Hillary Clinton already has the necessary procedural power. 

Hillary merely needs to demonstrate she has the leadership skills to make it happen.

Two Recommendations

In the spirit of being constructive, I offer the following recommendations. 

It is time to step back from the minutiae to consider the objectives of public policy. Let us ask our political leaders (including our candidates for President) what they want to accomplish when the Federal Government reports the Consumer Price Index.

• Do we want the CPI to actually track the rate of inflation? 

• Should it measure the rate of inflation based on what consumers actually pay for goods and services? 

• Do we want a fair and balanced calculation methodology? 

• Is consumer well-being an important consideration in calculating the rate of inflation?

• Should the BLS incorporate a consumer’s quality of life in its CPI calculations?

Once there is a consensus on these objectives, then the Government Accountability Office should be asked to review the BLS CPI methodology for compliance and procedure. The results will tell us what needs to be done to ensure American wage earners and beneficiaries are being treated fairly.

Make sense?

2. If the BLS wants to use value pricing in its calculation of the rate of inflation, then let us view this price change from the consumer’s viewpoint. We can do that by tabulating data against cost-of-ownership and cost of acquisition pricing. For vehicles, the cost-of-ownership can be compared, period to period, on a cost per mile basis. For homes or rentals, the price can be tracked on a cost per square foot basis. Non-durable goods and services can be tracked on a cost of acquisition basis.

This methodology provides a basis for comparison that can be expressed in terms of actual dollars spent.

And that has meaning for the consumer.

Conclusion

The BLS November 2007 Consumer Price Index (unadjusted)for All Urban Consumers (CPI-U) was 4.3 percent higher than in November 2006. In my opinion, the actual number is closer to 6 percent. Although the BLS index for energy looks more realistic with a 12 month increase of 21.4, the food index is still much too low.

One can challenge my methodology. But it should be obvious to any reasonable person: we have a management problem. One our political establishment needs to address. The objectives and methodology of the Federal Government’s Consumer Price Index need a thorough review.

"Oil, Jihad and Destiny"	"Detensive Nation"