Watch Out for Surge in CPI
Just to be clear, I'm also worried about a surge in inflation, but that's not what I'm talking about here. I don't know the future, so I try to stay away from predictions. But you don't need to know the future to "predict" a surge in the Consumer Price Index. All you need is to know the recent past.
I distinguish between inflation and the CPI. Inflation is the money becoming less valuable. The Consumer Price Index is an indicator of recent prices based on the prices of a standard basket of consumer goods.
The CPI is just a number. The number for June (reported July 15th) was 215.693 (versus a base of 100 in 1983).
Most people, though, don't really care about the number--they care about changes. And that makes sense--rising CPI values are an indication of rising prices, and rising prices are an indication of inflation.
(As an aside, rising prices don't necessarily mean inflation. Lots of things, such as changes in styles, tastes, and consumer preferences can change relative prices. Some things, such as taxes, technological change, and resource depletion, can change absolute prices. But none of those things are inflation, even if they result in rising prices.)
Almost all the attention to the CPI focuses on two things:
- The change from last month
- The change from last year
And it's that second one that prompts me to warn of a surging CPI.
After peaking in July last year, energy prices fell for months. From July 2008 to March of 2009, energy prices fell 37%. (Data from Bureau of Labor Statistics.) Energy prices amount to about 7.5% of the index, so that fall, all by itself, has had the effect of subtracting something like 3 percentage points from the overall CPI change from a year ago.
The thing is, most of the recent fall occurred in the last three months of last year. In the CPI index values published in the coming November, December and January, those big declines will "drop off" the year-ago comparisons--which will produce the surge I'm talking about. The "change from a year ago" values reported for the CPI for those months is going to be much higher--pretty much regardless whether there's any inflation.
Does it matter? Well, it matters to some people--people who invest in TIPS, for example, care a lot about the reported value of the CPI. (People on Social Security also care a lot, but as it happens the base year for figuring Social Security cost of living adjustments ends in September, so the effect of this spike won't show up until 2011.)
In fact, though, what really matters to real people is how their own cost of living changes.
I just wrote a post making fun of another economist's prediction of inflation, so I want to be really clear about the distinction here: What I'm talking about is a statistical artifact of the way the CPI gets calculated. For the past year, falling energy prices have made the inflation numbers look artificially low. For the next few months, CPI calculations of "changes from a year ago" is going to show inflation numbers that are artificially high--even if the value of the money were stable, we'd still see a spike in the CPI simply because the base month for comparison is going to have lower and lower fuel prices.
Last October I wrote a post called Inflation is going away for a while. Despite my general reticence about making predictions, things seemed clear enough to risk that one. This post is my official announcement that "a while" is just about over.
With that in mind, I'll direct you to a post of mine from earlier last year: How to live with inflation. Everything old is new again.