Here is a riddle for you. Follow along closely.
Two products sell for $100 apiece. A year from now, one sells for $200 and the other for $50. An index constructed of these two products would have increased 25%, from $200 ($100 + $100) initially to $250 ($200 + $50) a year later.
Now let’s look backwards. The products sell today for $100 apiece and, like before, the price of one is declining 50% annually and the price of the other is increasing 100% annually. That puts the price of the first item at $200 a year ago and the price of the other at $50. An index constructed of the two products would have been $250 a year ago ($50 + $200) and $200 today ($100 + $100). Looking backward, the index decreased 20% during the year.
Reread the last two paragraphs. The prices of two products are changing at the exact same rate, yet an index of the two products is declining by 20% when we look backwards and increasing by 25% when we look forward. We assure you that the math is correct.
The deception in the riddle is a common trick played by those wanting to manipulate indices to prove a particular point. The sleight of hand has to do with the weightings of the items in the index.
When we looked forward, the beginning weight of each item in the index was 50% (each is $100 of the $200). When we looked backward, the beginning weight for the item whose price was declining was 80% ($200 out of $250) and the beginning weight for the item that was increasing 100% a year was just 20% ($50 out of $250). The beginning weight of each item was different, which caused the information conveyed by the index to be quite different despite the prices of each item in the index increasing and declining by the exact same rate.
The initial weightings of the items in the index, in other words, have an enormous effect on the result of the index. It seems almost obvious, yet we accept the results of an index without really examining or questioning the weightings of the various items in the index.
It might surprise you, for instance, that while the United States spends about 18% of its GDP on healthcare, the weighting of healthcare in the consumer price index is just 7%. Since the growth rate of the cost of healthcare is about 40% higher than the CPI in general, underweighting the cost of healthcare lowers the CPI. But it clearly doesn’t lower the actual cost for anyone that spends a high percentage of their income on healthcare.
Many of the costs of owning a home are shockingly also excluded from the index. The index uses the cost of renting a home as a proxy for “the cost of shelter” and the cost of a mortgage is only indirectly included as a component of rent. The cost of a 30-year fixed-rate mortgage has increased about 130% since 2021. Much of that increase is not yet included in the CPI.
The Federal Reserve claims that omitting food and energy from the index (so-called core inflation) is its preferred measure of CPI. Really? Once again, the experience for anyone that eats is quite different than the CPI.
Indices are unreliable conveyors of information, especially when the people that control the weightings of the items in the index also have a vested interest in its outcome. The status of the people citing these indices implies a gravitas that simply doesn’t exist. An index can be manipulated to show whatever the keeper of the index wants it to show. The Consumer Price Index is one big deception. We ignore it.
