April 07, 2009
Some useful resources
Hal Varian and Hyunyoung Choi (paper here) document the usefulness of Google Trends and Google Insights for Search for purposes of updating assessments of current economic magnitudes. I see that Mark Thoma also calls attention to this intriguing data source.
Posted by James Hamilton at April 7, 2009 06:32 PMdigg this | reddit
Posted by: Vicente at April 7, 2009 07:09 PM
JDH: It's interesting to think of Google Trends in the context of this 1947 quote from Hayek:
"The "data" from which the economic calculus starts are never for the whole society "given" to a single mind which could work out the implications, and can never be so given. The peculiar character of the problem of rational economic order is determined precisely by the fact that the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form, ..."
Unless, of course, some very clever people are incrementally doing precisely that kind of concentrating and integrating, particularly with the prioritizing effect of page rank.
Vicente: agreed - it's cool
Posted by: Robert Bell at April 8, 2009 06:16 AM
Fantastic Professor. Thanks.
Posted by: DickF at April 8, 2009 06:37 AM
On the topic of commodities... in your view, do oil prices move in a fashion consistent with a mean reverting random walk? What about real estate?
Posted by: MikeR at April 8, 2009 07:35 AM
MikeR: A random walk can not exhibit mean reversion. My forthcoming Energy Journal article concludes that a random walk is a fair approximation to the series for the real price of oil.
On the other hand, real estate prices are unambiguously not a random walk-- there's significant predictability there. Standard theory predicts that the land component of real estate should appreciate in price at the rate of growth of GDP. What I see in the data are long deviations from that prediction.
Posted by: JDH at April 8, 2009 08:36 AM
I think I confused the statistics terminology. I agree that a random walk is a good approximation for oil. But when the price gets too high relative to other economic variables, it becomes more "predictiable" in that it is likely to revert to some long term trend.
Statistics asside, why should real estate be different than oil? They are both limited in supply and provide value (one provides energy, the other a perpetual stream of rents or farm income).
I will look at the paper, but I think I agree with your explanations based on elasticities.
Posted by: MikeR at April 8, 2009 08:59 AM
I am sending you a paper that I wrote with other colleagues on the usefulness of search query data in predicting economic behavior.
Posted by: Thomas at April 8, 2009 12:27 PM
I read the paper and enjoyed it. You used oil prices from 1970 - 2008. If you play with the start or end points, the ADF test result will change. (ie, cut the 30 year period into three sections so on page 35 from 6/98 - 6/08... clearly this looks like a trending series). The variance is just too high and heteroskadastic.
On page 16 "Both the supply and demand in any given year t are responding to any of a number of factors besides the current price. Imporant amonth these other factors are income (a key determinant of demand) and previous years's price" This statement seems to suggest a non-random behavior, which could feed back into future prices.
This sentence gets back to my main point: Why should oil behave differently than gold or real estate (aside from different elasticities. I also think real estate is an exhaustible resource)?
Posted by: MikeR at April 9, 2009 07:50 AM
MikeR: There are frictions in the real estate market that aren't present with commodities. Because there is only one seller of a particular property, there is no market to arbitrage away bad decisions by the seller. If you've priced your house too high, there is no way for somebody else to sell your house for a lower price; instead, the market just has to wait until you've accepted the current reality. This in my mind explains why real estate prices adjust as slowly as they do to fundamentals.
Posted by: JDH at April 9, 2009 09:34 AM