Proof of rampant U.S. deflation

From San Diego Tribune via the Big Picture, irrefutable proof of steady deflation due to enlightened government policy!

Fodder for thought for those curious about U.S. economic statistics.  While conventional wisdom (or publicized rationale) often claims that the recent adjustments to the core methodology remove components that are "more volatile", the evidence here suggests that the new method basically removes 3% from the previously computed rates.  Remarkable how stable this difference is over time.

This is a fantastic chart that makes its point clear and loud.  The secret is in picking the right comparables.  To vet the data analyst with shifting metrics, it is not necessary to prove that the new metric is or is not more accurate than the old.  Oftentimes, by tracking both the old and the new, the effect of the change is revealed.  Here, the adjustments just keep taking the rates down.

Reference: "The Fed's inflation guage isn't realistic, critics say", San Diego Union Tribune, April 17, 2008.

Rant of the day: Typepad's new editor continues to impress -- changing font size wipes out all hyperlinks in the text being formatted!

Comments

[Bob]

"Deflation" means negative inflation, i.e. rate of inflation less than zero??? The heading should be something like "Proof of rampant decline in U.S. inflation"?

[Christian vd Berge]

interesting to see though that allthough the trend looks the same, the delta between starting point and the last point is very different;
pre 1983 method starts at 10% and ends at 11,6% while the newest method seems to be almost steady; a little below 4% and 4% in the end....

[AdamV]

A very interesting chart. I find it particularly ironic that these changes were to remove volatile elements from the model, yet each iteration appears more "jagged" than the last. The oldest model has the smoothest line, implying that it actually has the least short-term volatility overall.

Christian makes an intriguing point - a trendline would be good to see on these charts, although of course the careful selection of exact start and end dates could be manipulated to show one model growing and another declining over the same period.

It would be nice to see the data starting from 1998 when they switched models - how subtle was this at the time?

I read an interesting post somewhere recently about a similar change in the UK government economic model for inflation - basically property values were stripped out of the model as prices rose, to make inflation rates reflect "true" economic growth more accurately. No distortion and propaganda involved at all, honest.
Now that property prices are starting to fall, maybe they should bring them back in to drive inflation rates down?

[derek]

I like the graph, well done to the San Diego Union-Tribune. But I would have liked to see the residuals, to show how much less volatile the components taken out are than the components that were left in.

[Joe Harris]

It seems clear to me that the rate is being manipulated to minimise the cost of inflation linked payment policies for Social Security and other programs.

Imagine what 11% payment increases would do to the deficit.

Joe

[RichmondTom]

Not everyone is convinced that the CPI number is understated. See this NYT article. The economists at the Bureau of Labor really seem to make a good-faith effort to model a complex issue. Their modifications are driven by legitimate concerns. If you want to see some really manipulated economic numbers, however, check out the unemployment figures or Federal debt. Pure fiction.

[RichmondTom]

Not everyone is convinced that the CPI number is understated. See this NYT article. The economists at the Bureau of Labor really seem to make a good-faith effort to model a complex issue. Their modifications are driven by legitimate concerns. If you want to see some really manipulated economic numbers, however, check out the unemployment figures or Federal debt. Pure fiction.

[Mike]

How can anything be measured if they keep changing the measuring device? Thank God that a yard-stick is still 36 inches long or else we'd all be living in homes that fall on our heads..... which would be much like our government: BROKEN.

[Kaiser]

RichmondTom: thanks for referencing the Leonhardt article. while I like much of his writing, this article I find to be sloppy.

He somehow concluded that "So when the new inflation numbers come out next week, they will indeed be misleading. They will be artificially high." His basic argument was: anyone who had doubts about the number suffers from loss aversion so inflation was not understated; the rent component might be overstated, therefore inflation was overstated.

Even if the effects cited are directionally correct, no numbers were used so how would we know that those factors contributed enough to affect the inflation rate?

In any case, as I pointed out, the point is to explain the variance between the old way and the new way, not to prove that the new way is "accurate" because that is not provable.

[DQKennard]

I seem to recall that the change in 1983 was pretty openly admitted to be explicitly to lower the COLA for Social Security and Federal employees. Reagan and his handlers "felt" the inflation rate was too high, and directed a recalculation be done, which (lo and behold!) came up with a lower rate.

The Clinton Administration denied it had anything to do with COLA-tied contracts, but was instead only to "improve accuracy". Right.

The inflation rate is what the Administration says it is, and has to be considered closely tied to the larger context of political and economic factors. This is also true of most (all?) other economic measures.

[John S.]

Sometime around 1972, I spent $69 for a four function calculator. The other day at an equipment show I was attending, several booths were passing them out for free. But let's say they really cost $1 each. That's equivalent to an annual rate of deflation of 11%.

[Michael]

My golden rule of media skepticism is to think about what isn't said in the article.

What isn't shown is the inflation rates prior to 2001. Is this because they could only find data going back to 2001? Or is it because data prior to 2001 doesn't fit their story?

[Mike]

@DQKennard: I spent 58 cents per gallon of gasoline in 1972. This morning I spent $4.05. In 1972 I also purchased:

Green Grapes: 49 cents per pound. Today: $2.99 (when on sale).
Chicken: 89 cents per pound. Today: 4.xx per pound (on sale).

Too bad we dont eat and power our vehicles (that allow us to get to work in order to buy the food to eat) with calculators or we'd all be doing great, eh?

Inflation is rampant. It has ZERO to do with what the government tells people. Watch CNBC and their own correspondents cant even agree with inflation or deflation or recession or not.

No one knows a darn thing. Except..... we do. We know whats going on. The people who whip out the wallet and pay for the items we need to buy in order to survive each day.

Those items have skyrocketed and will continue to. Enjoy ground beef today at 5.xx a pound. Next year this time: at least 7.50/lb.

Yea, thats rampant deflation for ya.

[Sean Carmody]

The San Diego Tribune piece takes their information from John William's site Shadow Government Statistics. Although I certainly haven't made a close study of the different methods used in each CPI series, there is something about this web-site that suggests it's somewhere out of the conspiracy theory fringe. Has anyone else looked into the details?

[John S.]

"Too bad we dont eat and power our vehicles (that allow us to get to work in order to buy the food to eat) with calculators or we'd all be doing great, eh?"

No, but we run our businesses on computers. What do you think a machine as powerful as the one you're using would have cost in 1972?

Look, I'm not saying the price of everything has gone down. I'm just saying there are methodological isues when you try to compare today's prices with those of thirty years ago.

http://www.slate.com/id/2142241/

[Jon Peltier]

John S -

Look at an employers' costs. Computers are way down. Health insurance is way up. Other items are in between, mostly heading up.

Any calculated inflation index is an attempt to look at overall cost of business or cost of living.

[R Simmon]

Shouldn't the difference between the different indices be plotted, rather than stacking the three methods on top of each other? Our eyes automatically calculate the shortest distance between two lines, not the vertical distance at a specific point.

[Ric]

It's all cool to measure things. At the end of the day, the same dollars I got last year and the year before, since no wage increases have happened, the same dollars buy 20-30% less. Gasoline is an artificial price, else why doesn't the price of diesel drop with the price of 87 regular? The price of computers is an artificial price, since the cost of HAVING a computer has risen anyways. The cost of saving money has gone up astronomically. Up.

I frankly think inflation is more like 16%