Understanding Inflation – Part 4 of 6

If you haven’t read parts one, two and three of this series on inflation, you might want to read those first.

The 1999 Switch to a New Formula

In 1999 the weight given to items in the CPI index was changed from straight arithmetic to geometric weighting. Instead of straight math, the new formula puts more weight on prices that have increased less and less weight on prices that have increased the most.

Example:

Let’s take an example of the geometric mean in action from the BLS’ own website :

If you buy one pound of carrots and one pound of peas, both at a dollar a pound, you have spent $2.00. If the price of peas goes up to $1.50 per pound, it will take $2.50 to buy that same amount of goods. Using the old way of calculating inflation — which was straight math — we have inflation of 50 cents in this example. But when the Bureau of Labor Statistics (BLS) switched from the arithmetic mean to the geometric mean, they moved away from the math that you and I know. If you like numbers, you can see the formula used in this example on the BLS’ website, but for those of you who like to keep it simple, I’ll tell you the end result. Using the new formula that was put in place in 1999, the CPI calculation implicitly assumes that the consumer will buy more carrots and less peas, resulting in an actual expenditure of only $2.45 (when the consumer really wanted to spend $2.50 for equal amounts of peas and carrots). Using the old formula, inflation would have been 50 cents in this example. Using the new formula, inflation is only 45 cents.

How do they do it? The new formula implicitly assumes that consumers will buy 0.816 pounds of peas and 1.225 pounds of carrots instead of one pound of each. The new formula takes care of the item substitution automatically. The long and short of the matter is that the new formula results in a lower inflation number than what we would have gotten if we used the old calculation that relied on straight math.

The theory is that consumers are no worse off by substituting an item in the same category to compensate for price increases. Advocates say that this statistical way of doing things is widely used and accepted and results in a more accurate CPI than using straight math. I have to disagree. If I eat one Snickers bar to cap off my lunch everyday, the new formula implicitly assumes that I will switch to 3 Musketeers if the price of Snickers goes up. The problem is: I don’t like 3 Musketeers! Snickers has always been my preferred choice of candy bar. For me, it’s Snickers or nothing. I don’t like a formula that implicitly assumes that I’ll be just as well off with an Oh Henry! or a Butterfinger. The BLS and I disagree on this issue. They maintain that a consumer’s standard of living does not decline as long as they can substitute an item from the same category when prices rise.

I want to take the time to address one popular urban legend. The formula has been wrongly accused of substituting hamburger for steak. Fortunately, the CPI calculation does not assume that you are just as well off if you have to buy hamburger because of increasing filet mignon prices. The reason is that steak and hamburger are in different item categories, and the formula only allows substitution within categories, not across categories. However, the formula does assume that you’ll be just as well off if you have to switch from 90% lean ground beef to 80% lean beef! I completely disagree with the BLS on this issue — and so do my arteries.

The CPI calculation assumes that consumers shift their purchases toward products whose prices are rising less (or falling more). However, supporters maintain that even though high prices caused you to buy something that you wouldn’t have bought otherwise, the goods you bought are not necessarily less desirable. I think the BLS is correct that consumers will shift their purchasing patterns toward items that have fallen in relative price. But just because consumers will buy different items doesn’t mean that they want to make different purchases. If you are forced to by something that you wouldn’t buy otherwise, you are adjusting your standard of living because of price increases. That’s inflation.

Mathematically speaking, a geometric mean will always be less than the arithmetic. The end result is that the use of the geometric mean results in a lower CPI, which means a prettier inflation number.

Stay tuned for part 5 of Understanding Inflation …

by Wade Young