What do U-haul prices tell us about America?

Read what Chris Lightfoot has to say about this question here.

The origin of the idea for the analysis appears to be in this marginalrevolution post.

The idea is that large differences in prices for one-way trips from Detroit to Las Vegas compared to one-way trips the other direction reflects differential migration. The answers aren’t so surprising: the flow tends to be South and West, and especially towards Las Vegas.

In one sense, this analysis is very Freakonomics-y. He answers a question by cleverly using data that were created for a completely different purpose, but contain useful information nonetheless. In another very important sense, however, there is something unFreakonomics-y about it. If we want to know about American migration patterns, there are easy ways to use census data to do it. In general, I’m not in favor of doing things the hard way, just for the sake of it being hard. So, at one level, using U-haul data for this question is kind of pointless. When I take roundabout approaches to answering questions, it is only because there is no easier path I can see to the answer. This is a key insight that people often overlook.

Still there are at four reasons to applaud the U-haul analysis:

1) It is clever and insightful. Seeing it, others might be inspired to find another
application in the same vein.

2) It was a lot of hard work, and I am generally in favor of rewarding hard work.

3) It uses prices. Prices are an incredibly valuable tool for understanding what is going on. Indeed, the Center I run at University of Chicago is called the Initiative on Chicago Price Theory for just that reason.

4) It provides a good test of indirect strategies to identifying patterns of interest. Because we have more direct measures, we can compare the results. In this case, what Chris Lightfoot found matches well what we expected, validating the methods.

(Thanks to Tom Steinberg for forwarding the link.)


Anonymous

http://www.irs.gov/taxstats/indtaxstats/article/0,,id=96816,00.html

IRS sells County to County migration data based on the tax returns ... while this will be a year or so later than some other data, I'd think it would be rather definite.

Anonymous

One rather large caveat is that not everybody uses U-Haul, and you can probably think of good reasons for why the bias goes towards a very specific socio-economic class. Managers relocating from Las Vegas to New York are not captured here. Disgrundled waiters going the other way are.

Ken

The IRS data is very good, but it only counts people who file in consecutive years, so it misses out on some of the very poor and unemployed, people who are dependents one year but not the next (e.g. students graduating who move), etc.

There is a great report, but now a decade old, that compares different data sources here. At least back then, the poor and unemployed were much more likely to be net emigrants, there was a much smaller difference at the high end, althought net migration was slightly negative (away from CA) for high incomes, too.

See also William Frey's site, which is mostly census-based research.

Foreign migration seems to be a much bigger factor than taxes. Foreign immigrants seem to be, to some extent, displacing native-born citizens.

Read more...

Acad Ronin

Students also make some (much?) use of U-Hauls.

The great advantage of U-Haul data is that it is available on a more frequent basis than the census or tax data, and can show how markets adjust to events such as Katrina pretty much as it happens, seasonal events such as the school year, and the like. By combining U-Haul data with census data one could, perhaps, look at short-run versus long-run responses.

doc

Along with three students, I've started to look at the U-Haul pricing data, and, so far, it appears that the patters is: High rates from larger to smaller cities, low rates from smaller to larger cities, independent of population growth rates. Maybe, just maybe, the issue is the cost of relocating the trucks.

Xmas

Depending on how quickly U-Haul changes their prices, this sort of data can determine seasonal and other forms of migration on a short-term basis. Census data is only a once-a-year process, along with IRS and most other government data.

So, if U-Haul has short-term pricing adjustment for one-way hauling, the data could pick-up short term movement trends.

Anonymous

doc, the original marginalrevolution post showed data from LA-Las Vegas and vice versa, which would seem to control for the cost of relocating trucks, as both are large cities (in fact, LA is larger).

Barry Ritholtz

I am compelled to comment on the errors in reasoning in the original.

Las Vegas is the most rapidly growing region in the U.S.; most of its explosive growth is relatively recent. California is a mature economy, still suffering from the dot com and tech bubble.

People, this is not rocket science: The cost of sending a U-Haul from a mature slow growth area to a less expensive high growth area has to include the cost of THE RETURN TRIP.

Duh

Christopher Davis

Then there was Jartran (founded by James A. Ryder, yes, the same guy who founded Ryder); when they started out, they advertised some really cheap one-way rental rates.

When you asked, however, they turned out to be one-way rates from Detroit to anywhere. That way, you could pay for the privilege of being their delivery driver....

StCheryl

I did two cross-country moves (NYC - SF Bay Area, SF Bay Area - Washington, DC) in the 1980s by UPS and US Postal Service. It is possible that neither the managers whose moves are paid for by their employers, nor the disgruntled waiters (cited by anonymous above) are captured by the UHaul approach.

Anonymous

While there are other, direct indicators of this trend, this might be a good predictor which could let someone with this advance knowledge take advantage of a little known early indicator.

Many years ago, an executive at a state insurance fund told me that he could accurately predict the local unemployment rate by the change in the ratio of disability to workmen's compensation claims. That sort of predictive indicator could provide a finacial benefit.

Anonymous

http://www.irs.gov/taxstats/indtaxstats/article/0,,id=96816,00.html

IRS sells County to County migration data based on the tax returns ... while this will be a year or so later than some other data, I'd think it would be rather definite.

Anonymous

One rather large caveat is that not everybody uses U-Haul, and you can probably think of good reasons for why the bias goes towards a very specific socio-economic class. Managers relocating from Las Vegas to New York are not captured here. Disgrundled waiters going the other way are.

Ken

The IRS data is very good, but it only counts people who file in consecutive years, so it misses out on some of the very poor and unemployed, people who are dependents one year but not the next (e.g. students graduating who move), etc.

There is a great report, but now a decade old, that compares different data sources here. At least back then, the poor and unemployed were much more likely to be net emigrants, there was a much smaller difference at the high end, althought net migration was slightly negative (away from CA) for high incomes, too.

See also William Frey's site, which is mostly census-based research.

Foreign migration seems to be a much bigger factor than taxes. Foreign immigrants seem to be, to some extent, displacing native-born citizens.

Read more...

Acad Ronin

Students also make some (much?) use of U-Hauls.

The great advantage of U-Haul data is that it is available on a more frequent basis than the census or tax data, and can show how markets adjust to events such as Katrina pretty much as it happens, seasonal events such as the school year, and the like. By combining U-Haul data with census data one could, perhaps, look at short-run versus long-run responses.

doc

Along with three students, I've started to look at the U-Haul pricing data, and, so far, it appears that the patters is: High rates from larger to smaller cities, low rates from smaller to larger cities, independent of population growth rates. Maybe, just maybe, the issue is the cost of relocating the trucks.

Xmas

Depending on how quickly U-Haul changes their prices, this sort of data can determine seasonal and other forms of migration on a short-term basis. Census data is only a once-a-year process, along with IRS and most other government data.

So, if U-Haul has short-term pricing adjustment for one-way hauling, the data could pick-up short term movement trends.

Anonymous

doc, the original marginalrevolution post showed data from LA-Las Vegas and vice versa, which would seem to control for the cost of relocating trucks, as both are large cities (in fact, LA is larger).

Barry Ritholtz

I am compelled to comment on the errors in reasoning in the original.

Las Vegas is the most rapidly growing region in the U.S.; most of its explosive growth is relatively recent. California is a mature economy, still suffering from the dot com and tech bubble.

People, this is not rocket science: The cost of sending a U-Haul from a mature slow growth area to a less expensive high growth area has to include the cost of THE RETURN TRIP.

Duh

Christopher Davis

Then there was Jartran (founded by James A. Ryder, yes, the same guy who founded Ryder); when they started out, they advertised some really cheap one-way rental rates.

When you asked, however, they turned out to be one-way rates from Detroit to anywhere. That way, you could pay for the privilege of being their delivery driver....