Archive for February 16, 2010

Oh, (Yeah!) Canada!

February 16, 2010 5 comments


Who knew that all the world needed to heal from the global economic crisis was the Olympic games?

Stocks rallied 1.8%; commodities soared 3% as the dollar declined; and 10y note futures rose 7.5 ticks. The passage of the weekend without further incident was probably more to credit for the sigh of relief as the global sense of camaraderie that accompanies the Games, but as long as we have a reason to thank Canada by golly let’s do so!

The economic data was modestly supportive, but only modestly. The Empire Manufacturing Index┬árose to 24.91, but New Orders (8.78 vs 20.48) and Shipments (15.14 vs 21.07) – both important components – both fell and “# of Employees” (5.56 vs 4.00) barely rose. So “General Economic Conditions” (which is a separate question) are much better, but components of the business weren’t quite as good this month versus last. This falls into the “probably doesn’t change our null hypothesis” category.

Why CPI Is Not Bogus

On Friday the BLS will release the Consumer Price Index, which is a very important release and one of the most maligned even though CPI is one of the more carefully-designed and researched numbers in the entire list of government releases. This is partly because so much is riding on it, between securities linked to non-seasonally-adjusted CPI (such as TIPS) and contracts such as Social Security and others. To some people, this just opens the door for more government shenanigans, since arguably the government stands to gain the most from monkeying with CPI.

But if people with lots of money on CPI didn’t fundamentally believe in the veracity of the number, then the $500billion-plus TIPS market would be in real trouble. Indeed, in some countries where there is better reason to doubt the government’s accountability on such matters, there are private as well as public inflation indices and there are securities are linked to each index. (Brazil is one example.)

That doesn’t mean that these investors are right, of course, but you can believe that they’ve looked pretty hard at the number. Sure, investors have also looked very hard at equities and concluded that they are entitled to a very lofty multiple right now, so we can’t just rely on that. I will tell you, though, why inflation cannot possibly be as high as some folks believe it is, and why you very likely feel that inflation is higher than is being reported by the government.

What follows is an enumeration of some of the cognitive errors that people make with respect to CPI. The list isn’t complete, but I think I’ve hit on the biggest of them.

As a first point: CPI does what it is supposed to do very well, but that might not be what you want it to do. CPI is a cost-of-living index, which means that if your standard of living improves then the price you pay for that standard of living should also increase (that is, your outlays should increase faster than general inflation); if your standard of living is static then your outlays should increase with inflation. CPI measures, in other words, the cost of an unchanged standard of living.

This important point gives rise to the concept of hedonic adjustments, which adjust the price recorded by the BLS for a particular good to account for changes in the quality of those goods. This is a crucial adjustment for certain goods that change significantly in quality, such as cars, computers, and medical care. But this is one source of complaints of people who don’t bother to understand the CPI: people don’t mentally record hedonic adjustments; people measure cash out of their pockets. So when you buy a new computer and it’s lots better than the last one but costs the same, you experienced deflation in the sense that the cost of your old lifestyle…which you no longer have…costs less. Since you spent the same amount, it doesn’t feel like deflation to you, but since your standard of living improved while the costs were unchanged, that’s deflation in a cost-of-living-index sense.

Second, your consumption basket may vary. For most people, the broad CPI index is a reasonable measure, but each person’s consumption is different. Some people spend more on Apparel and less on Recreation; others are the opposite. CPI is supposed to measure the average experience, and no one is exactly average.

Third, CPI excludes taxes that don’t have anything to do with consumption, since CPI is only supposed to measure changes in the costs of things you consume. But taxes definitely affect our standard of living, so if income taxes rise and that causes a decline in your standard of living, that sucks but it’s not inflation. Don’t blame CPI – blame the dudes who are taxing you!

Fourth, people tend to remember price changes of small, frequently-purchased goods rather than large, infrequently-purchased goods even though the latter are more important to your cost of living. For example, your house is typically no more than a once-a-year negotiation (if you rent) or even less frequent if you own. But it is a huge part of your cost of living. When milk doubles in price, or gasoline spikes, you notice it a lot but it’s a much smaller portion of your consumption and matters less.

Fifth, you may be the victim of classic attribution bias. When you go to the store and you come home with a bunch of stuff at higher prices, you say it’s inflation; when you come back with a bunch of stuff at lower prices, it’s “good shopping.” Combined with the prior effect, the rapid oscillations in food and energy prices seem to us to be lots of inflation interspersed with lots of good shopping.

These are the predominant cognitive and comprehension errors that most people make when they think about inflation. But some complaints about CPI go way beyond these innocent and entirely normal perceptual biases.

Some complaints of CPI are just silly. Shadow Government Statistics, which has made quite a great business catering bad data to conspiracy theorists – and I won’t link to the site; if you need to find it for some reason I am sure you can – has a chart of what inflation would be if the BLS used 1980 methods, with the implication being that the guvmint is trying to hide the 9% rate of inflation (the site says inflation is understated by about 7%). The choice of 1980 is very adept, since it was in 1982 that the BLS changed the method of computing the cost of housing to remove investment-value-of-the-home considerations (such as the mortgage rate) and focus on the consumption-value-of-the-home (which is best represented by what it would cost to rent). This was done after much research, many public papers and debate, and is absolutely the right way to measure inflation in the cost of housing consumption as distinct from changes in the value of the home as an asset. There are lots of other improvements that have been made to CPI, and they really are improvements. Not everything from 1980 is better than the 2010 version. Computers, cars, medicine. I’ll concede music.

But we can rely on a very simple argument to prove that true inflation cannot be at 9% (but it involves math). I presume that most readers can recall what they were paid ten years ago, or at least can very easily figure out what their income was back then. Government statistics say that the average increase in wages and salaries (in the Employment Cost Index) has been about 36% over 1998-2008 (for some reason I am having trouble finding 2009 data, perhaps because it is subject to revision). I assume that we don’t think the government is exaggerating that number on the low side for some sinister reason. Now, if inflation is really running at the 9% or so that Shadow Government Statistics says it has for the last decade, then while your wages have grown 36%, cost of living has risen 136% (the government says inflation has been more like 29%).

More concretely: suppose you made $60,000 in 1998, took home $40,000 after tax and were just breaking even with your cost of living at $40,000. According to the government, you ought to be making about $81,500, and if taxes were the same your take home pay of $54,333 would leave you with about $3,000 better with your cost of living about $51,500. If, however, inflation was really at 9%, then your cost of living is now $94,700, and you declared bankruptcy several years ago. You don’t have to be able to track your receipts to see that 9% is not the right rate of inflation – you just need to look at the compounded outcome.

Consider a longer period of time for a more-poignant comparison. The person making $30,000 and taking home $20,000 in 1980 is now making $89,000 and the $59,300 take-home pay (2/3, assuming improbably that taxes were unchanged) has improved your standard of living somewhat as the old standard of living now costs $52,800 using CPI. Using a 7% higher rate of inflation, the same standard of living you enjoyed in 1980 for $20,000 now costs $353,000.

That is nonsense. And by the way, it also means that housing over the last decade not only wasn’t in a bubble, it didn’t even come close to keeping up with inflation, and neither did any other asset in the world. That’s worse than nonsense; it is an offensive ignorance of mathematics.

CPI is not a perfect number, and moreover it may not be a perfect number for what you want it to do. But it does what it is supposed to do, and it does it very well. I am certainly no apologist for the government and the way it is run, but on the occasions that the bureaucrats get something basically right, I think it’s okay to say so.


Housing Starts is the first release tomorrow. The consensus expects a rise to a 580k pace of Housing Starts, an improvement from 557k last month. This figure hasn’t been over 600k since the crash down from 2273, so a rise to 580k isn’t anything to write home about as the chart below amply illustrates. But, since it would be less bad (and because the stock market today seemed to be feeling rather ebullient), the market may well react well to any print near 600k.

Even a great number on Wednesday isn't cause for celebration!

Similarly, Industrial Production and Capacity Utilization (Consensus: +0.8%/72.6%) should show decent improvement (although be sure to look at ex-utilities numbers as January was pretty chilly) with Capacity Utilization up from 72.0%. But considering that 73.5% was the lowest reading of the last recession, it is again merely less-bad. Economists who rely on Keynesian analysis of inflation will say this is the best of all worlds: a rapidly improving industrial sector with tremendous slack to be taken up before inflation becomes a possibility. Monetarists will stifle giggles.

Also out tomorrow are the FOMC minutes from the recent meeting. I doubt there will be anything in there we didn’t learn from the text of the Chairman’s speech last week, but be alert for headlines at the 2:00 hour.

Categories: CPI, Good One
%d bloggers like this: