CPI: A Busted Metric?

Jason Stipp: I'm Jason Stipp for Morningstar. As Bob Johnson, our director of economic analysis, has reported, inflation has been remarkably mild in recent times, but can you trust the CPI. Here to talk about some of the misconceptions around the consumer price index is Bob Johnson, our Director of Economic Analysis.

Thanks for joining me, Bob.

Bob Johnson: Great to be here today.

Stipp: You say that there are four reasons that people doubt the CPI being a good indicator of inflation, and we're going to talk through some of those and some of the facts behind them.

The first one is, people will say, CPI is not accounting for energy and for food because they are so volatile, so of course it seems lower than it should.

Johnson: I get a lot of questions on that. The Social Security adjustments, the tax bracket adjustments, the TIPS (the inflation-adjusted bonds), and so forth are all adjusted on a broad measure of price inflation, the CPI-U, for urban wage owners, and that includes food and energy.

I think a lot of the confusion comes from the fact that every time you read something on inflation, [it says], the "core" inflation (excluding food and energy) is … And the Fed is a big one to look at that metric to see what the underlying trend is in inflation.

In reality, the number that's used for all the adjustments, and all that anybody cares about, does indeed include food and energy, and they're not insignificant portions of the index. They are probably upward of 20% of the index.

Stipp: The core inflation doesn't include food and energy, but this is one that Fed talks about a lot. Why do they talk about that one, and why is it called "core" if it doesn't include food and energy?

Johnson: That's always been a mystery to me. In most of my pieces I won't even talk about what core is, because I really think that it's such an important part; it's over 20% of the CPI, and it's very important to how people feel about their spending, and you can't avoid it. So I think any index that looks at [inflation] without food and energy is worthless. So I refuse in my weekly pieces to even talk about the core number.

The Fed likes it because food and energy do tend to be volatile, and they tend to say, let's leave that out because that's really volatile stuff and maybe it will be different next month. I tend not to want to do that. I will look at a three-month moving average on some of my [inflation metrics] to smooth out those things. Also, maybe in a month where I know the CPI was very high and it might have been [because of] gasoline, and that comes out on, say, the 16th of the following month. Well, I already know what the prices are at the 16th in the following month. So if a refinery blew up somewhere and temporarily caused a spike in prices, I can look at the data and say, this month gasoline prices are going to exactly reverse what we just saw last month, so let's ignore that. I'll do that, but I won't just throw [energy] out willy-nilly.

Stipp: The important point is most of the adjustments that you'll see out there on wages, on Social Security, on Treasury Inflation-Protected Securities, will account for food and energy. So that's the first point.

<TRANSCRIPT>  

The second point is, some people will say, well, substitutions will happen. So if steak gets really expensive, people will buy more chicken, and so CPI is not capturing that, because it's weighting chicken more and [chicken is] lower priced. What do you say to that?

Johnson: The government does try to capture some substitution effects, and a lot of people think, it's steak, then it's chicken, then it's beans, and you keep going down the pyramid of substitutions and that that really depresses the CPI.

But that's really not true, because the substitutions are fairly narrow substitutable items. If something gets a little cheap within one category, they say, yes, people will switch, and it would be way overstating if we assume people are going to keep buying the most expensive orange juice when the trends are clear they'll buy the one that's on sale this week. People react to prices, and therefore that should be included.

Now, they do that within narrow categories, like orange juice in Chicago; it's geographic and by product. But what they won't do is beef to chicken. Those are two separate categories. So they won't change the weightings on those. They do take a look-back every two years and try to adjust the inflation rate, but [the CPI weightings] don't capture the short-term [consumer adjustments from one category to another category].

And by the way, even in the substitution effect, if the price of chicken starts to go up faster than the price of beef, because people are substituting it, then that has been offset in the CPI, too. So net, I don't think it's as large an impact on the data as people think.

Stipp: Bottom line, if the substitution you're making is a big substitution out of the category, it's really not going to have an effect on CPI, because they're not going to change the weightings for a substitution like that?

Johnson: Right, exactly.

Stipp: One thing, though, is the concept of the chained CPI, which is another way that they measure inflation. What is that and how is that different than the substitution calculations that we just talked about?

Johnson: It's a third category of inflation that's reported in the data. The CPI-U, the CPI-W, and then you have the chain-weighted CPI as well. So it's always available, and that one does allow for immediate substitutions. That's the one that they're proposing if we want to keep Social Security growth under control--let's switch to this more conservative method of calculating the CPI, which allows for this unlimited substitution effect.

That does indeed, and you can see the numbers that are published every month, bring down the rate of inflation and why it's somewhat popular as a backdoor way to reduce Social Security benefits without really raising all the outcry. It is something that does depress the CPI numbers. There is some justification, because people will substitute between categories, but then you have the effect of the steak for chicken that begins to creep through in the numbers.

Stipp: Another issue that can happen when prices increase is that they're increasing not just because you have to pay more for the same thing, but because you're actually getting more for what you're buying, even though maybe it's still a computer.

Johnson: There are two categories where this really stands out--one is autos and one is electronics. Those are items that frequently have a lot of new benefits on the new devices, and old devices are often discontinued. So even though a laptop computer costs $500 now, and it cost $500 a year ago, the one this year has a lot more features, and the government goes in and makes an adjustment and says, the price of that computer probably actually fell.

You're comparing apples to oranges if you don't make some type of adjustment, and many of those [new features] provide real world benefits to you. In autos, putting the airbag in. The price of the car goes up, but [CPI] says, well, there wasn't a price increase because that was a benefit. Well, obviously if you're in an accident, it saves you hospital costs and a whole bunch of other things. It's the same way with the computer; there are benefits by speeding up a computer. Now you shop online, you don't have to drive, so it's really benefitting your cost of living, too.

You can see the point. If I have to own a car, and I have to buy it with these features, it's less money that I can spend on other things. But on the other hand, you are getting a lot of benefits that are accruing to you over time, maybe not just this month.

Stipp: Your last point, Bob, is that people in their personal lives might have a different experience of inflation, depending on their consumption pattern. You might say, inflation has been very tame over the last year, but someone can come back to you and say, well, I'm paying a lot more for X, Y, and Z. How can that be?

Johnson: Absolutely true, and I get so many questions. They are realistic in many cases. For example, one of the big things holding the CPI back this year has been drug costs. Granted a lot of men over 40 are on Lipitor, and it's a large part of the population, but if you weren't on Lipitor, you probably think drug prices went up last year. It very much depends on what you're using, what your mix of products is. Maybe you're somebody that doesn't own a car. Maybe for you, gasoline shouldn't be in the index. Or maybe you're using a lot of medical goods, and that's generally, until recently, been a very high-inflation category. So [CPI] really didn't reflect your personal rate of inflation, which is a big deal.

As an economist, I have to see [inflation] as a general thing: How do I adjust retail sales overall? How do I adjust wage growth? But I can't talk to everybody's individual experiences, because those are very different. Maybe nobody gets on an airline, and I do a lot, so I notice that one a lot more. So there are a lot of reasons why personal inflation may indeed be different.

Then there's also the whole effect of every day, the food and gasoline [prices are] in your face. You drive by a sign every day, and you've got that metric right there. Nobody says, this happened with home prices because the mortgage rates went down, or the car cost the same as it did last year, [because they] didn't really notice. I buy a car every 10 years.

Stipp: Bottom line then, Bob, as an economist, do you trust the CPI as a good metric for gauging inflation?

Johnson: I think it's an excellent vehicle to measure the overall economy and compare across economies to detect whether policies are working. They've got a terribly hard job. How do you compare something from 20 years ago when there were things like phonograph records and transistor radios in the mix? We don't have those anymore. There are cell phones now; there weren't then. How do you even think of blending them together? And I think they do an outstanding job of capturing that. I think they try to stay apolitical; obviously, with the chain weighted CPI and Social Security, that gets a little bit more political.

But nevertheless, I do trust the data. I think it gives us a good overall impact, but you have to know your own situation. If somebody is going to retire and say, the CPI is X … no, you have to say, am I going to spend a lot more than the average guy on drugs? Am I going to own a car? Am I going to live in a low-cost state? And you just can't take the numbers as a whole to calculate your financial situation.

Stipp: Bob, some excellent insights on inflation and the CPI. Thanks for joining me today.

Johnson: Thank you.

Stipp: For Morningstar, I'm Jason Stipp. Thanks for watching.