MIT’s Billion Prices Project

Copyright 2011 by John T. Reed

Alberto Cavallo is an MIT economist. He created a price search engine that gathers daily prices of 5 million products sold on the Internet. He calls it his Billion Prices Project.

He is from Argentina. They know hyperinflation.

Cavallo is from Argentina where his father worked for the Argentinian government on taming that country's hyperinflation. He understands that governments deliberately change the calculations to make the inflation rate look lower than it really is.

Yet he says he is trying to work with the U.S. government so they can use his BPP to make better decisions than they would using the Consumer Price Index (CPI). Hey, Alberto, the U.S. government, like the Argentinian government, does not want to make better decisions. They want to make their elected bosses happy which is done by helping them get re-elected which requires falsely low figures.

As I related in How to Protect Your Life Savings from Hyperinflation & Depression, 2nd edition

private companies and citizens have to create their own price indexes—the Me CPI—during hyperinflation like they have often had in Latin America, in order to make decisions—like what prices to ask for the goods and services you sell. Cavallo's BPP is essentially the sort of thing Argentinian companies were doing in that country during their hyperinflation.

Untrustworthy and too slow

In my book How to Protect Your Life Savings from Hyperinflation & Depression, 2nd edition, I have a chapter on price indexes. I said that they were run by the government and therefore could not be trusted. I also said that they were too slow. Many assets or payment streams that are supposedly adjusted for inflation are only adjusted annually (Social Security) or semi-annually (TIPs bonds) and have lags between data collection and publication of the CPI like the 90-day lag on TIPs adjustments. These delays are totally unacceptable in hyperinflation and render the the CPI useless for all but very low inflation.

In my book, I recommended using a commodity price index instead. It is adjusted by the minute from sales of commodities on commodity exchanges. True, it leaves out manufactured goods, raw materials that are not traded on exchanges, and services, but at least it is somewhat diversified and calculated in real time continuously.

Diversification is important because relying on one commodity like gold results in your being behind or ahead of true inflation rates almost all the time.

Cavallo's BPP is more diversified than commodity indexes. Indeed, his BPP may include commodity prices, but I have not seen any mention that it does. It could.

Daily may be too slow

And, believe it or not, even a daily calculation like Cavallo's is too infrequent during hyperinflation. In German hyperinflation in 1922, workers would run from their job the moment they were paid to spend their marks before they fell in value. Employers had to pay the workers multiple times per day and shut down after each pay hour.

Commodity price indexes are well stablished, which is nice to have before you embed them in long-term contracts. The BPP may not be sufficiently well-established yet to be putting it into long-term leases and such. But it is arguably a more diversified and therefore more reliable measure of true loss of dollar purchasing power, and it is fast. I recommend we keep an eye on it and use it more as it becomes more reliable.

I expect the U.S. government will try to suppress it. In February 2011, the Chinese government announced it would no longer release many stats on prices—at a time when their inflation was increasing.

So much for putting a cost-of-living clause in a long-term contract in China.

John T. Reed