Sunday, July 24th, 2011
The Internal Revenue Service just released its benchmark place to place migration data for 2009. This tracks moves from county to county and state to state for people who filed tax returns between calendar year 2008 and 2009. My initial look at the data confirms what other sources such that Current Population Survey have shown, namely that migration has slowed during the Great Recession. I’m going to cheat though and not actually show much of that despite my enticing title, and instead illustrate a few other points the come out from this data.
As I’ve said before, this data is a gold mine of information. Few people seem to use it though, probably because it is so cumbersome to work with for non-specialists. One of the biggest reasons I built my Telestrian system was to create a platform that would actually make this data usable for me. It is by far the best system I’ve ever seen for analyzing this data. And not only is it better, it’s cheaper. Just to get the IRS to ship you the full data set – in the form of over 3,000 Excel spreadsheet I might add – is $650 dollars, but you can get access to to it via Telestrian for only $49/year.
(If you are a researcher or other person who would be interested in getting a complete copy of the data in a file format that is ready to import into a database, email me. In addition to the base IRS files in a usable form, I’ve created a full matrix of in, out, gross, and net migration for people (exemptions), households (returns), income (AGI), household size (exemptions/return), and household income (AGI/return) for not just county-county and state-state, but also MSA-MSA and MSA-state for years 1996-2010. This is about 2 million rows per migration type, plus there are dozens of other summary statistics available from the files).
With that out of the way, let’s take a look at some interesting things. I’d previously showed a chart of gross migration between Chicago and other major Midwest metros. Gross migration is one of the better measures of true human capital circulation between cities, and indicates which metros have the tightest talent linkages with Chicago. Here’s the update for 2000-2009 (note that year 2001 is people who moved between 2000-2001, so that’s why that’s the base year. This applies to all the charts and data).
You can see here the generally slowing migration I alluded to earlier in the 2009 data series though in some cases it’s still higher than the dot com recession earlier in the decade. The gap between St. Louis in Columbus is quite large as you can see. Clearly, certain nearby cities, as expected, have greater flows and thus are in a way better connected to Chicago.
You might be wondering what the net flows are. I’m glad you asked. Here’s the totals:
Unsurprisingly, struggling Detroit and Cleveland lead the list for inbound migrants. Looking at this chart, I also can’t help but wonder if the high net flows from Chicago to Indianapolis is what enables Indy to maintain a slight population growth lead over places like Columbus that are also rapidly growing in general.
Of course, people is one thing, but you might suggest that it is the high talent individuals that matter more for a city’s economic future. The IRS doesn’t measure that directly, but it does give data on income. You could potentially use this as a proxy. Here is a chart showing total income (AGI) flows:
Note that this data is in thousands, so in this case Detroit actually sent over the last decade households that earn $400 million per year. The total income lost over the decade actually is far higher, because as I note, that’s per year income, so we should really look at it on a total accumulated loss basis perhaps, but I didn’t calculate that for now.
That’s total dollars, but what about the income per person? We can’t quite get there, but we can look at the average per household (i.e., tax return). If we do that, here’s what we see:
Now there are some limits to this data. The annual values are arithmetic means. Probably median would be better, but that data is not available. Also, to save time I used a straight average of the annual values, when probably some type of weighted average would have been better. But as a quick look, this is fine. What it shows is that with the exception of Indianapolis, Chicago is losing higher income people to other cities. Again, I’d caution to do a more rigorous calc before reading too much into this.
The Indianapolis case is interesting. It is a high net importer, but is importing lower income households than it is exporting, which is contrary to the other cities. I took at look at the disaggregated inbound and outbound flows quickly, and it looks like the households going back and forth between Indy and Chicago are lower income than those of any other city. Plenty to explore here.
Again, you start looking at this and the possibilities for serious research are endless, but alas, we’ve only got time for a blog post.
What Type of City Are You?
Another interesting thing jumped out at me when looking at cities. It is the migration pattern differences between the large tier one cities and smaller regional cities. Let’s illustrate this by showing the top 10 sources of gross migration (circulation) for Chicago and Indianapolis. Here’s Chicago:
|1||New York-Northern New Jersey-Long Island, NY-NJ-PA||8,819||8,603||8,409||8,260||9,117||9,478||8,991||9,572||9,685||80,934|
|2||Los Angeles-Long Beach-Santa Ana, CA||9,041||8,540||8,439||8,510||8,052||8,327||8,146||8,358||7,902||75,315|
|5||Milwaukee-Waukesha-West Allis, WI||7,001||6,586||5,955||5,901||6,241||6,634||6,528||6,562||6,180||57,588|
|6||Dallas-Fort Worth-Arlington, TX||6,929||5,750||5,283||5,272||5,616||6,159||6,163||6,206||5,955||53,333|
|7||Atlanta-Sandy Springs-Marietta, GA||5,931||6,101||5,237||5,309||5,574||5,665||5,781||5,697||5,050||50,345|
|10||Minneapolis-St. Paul-Bloomington, MN-WI||5,721||5,429||5,199||4,878||5,019||5,251||5,216||5,327||5,091||47,131|
And here is Indianapolis:
|6||Terre Haute, IN||1,791||1,620||1,807||1,606||1,871||1,611||1,546||1,756||1,694||15,302|
|7||Fort Wayne, IN||1,644||1,552||1,415||1,423||1,641||1,737||1,633||1,756||1,664||14,465|
Interesting, isn’t it? While Chicago does exchange people with regional cities, it also has a lot of circulation with other national and tier one cities, including it’s top two of New York and LA. Indianapolis, by contrast, apart from its huge circulation with Chicago seems to primarily recirculate people within Indiana. Its human capital networks are much more parochial.
New York City
I thought I would also take a quick look at New York metro. Here’s a map of net migration from 2000-2009 using a five bucket sort with separate positive and negative gradients:
It looks like New York is drawing in from the Heartland while exporting everywhere else, though the magnitude of the imports is far less than the exports. This is not surprising given that New York is well known for very large net out migration. (This is not necessarily nearly as bad a thing as it is frequently made out to be, as I explained here).
I’ve also read a lot about income migration from New York, so thought I’d throw that up. Here are the top ten places New York imports income from. Again, this data is in thousands and the cavet that this is not a total accumulated gain/loss applies:
|4||St. Louis, MO-IL||14,706||4,064||18,140||5,696||5,728||-14,400||9,811||16,568||26,365||86,678|
|7||Ann Arbor, MI||7,171||-1,867||-1,721||-4,718||6,062||2,824||12,527||20,105||-1,356||39,027|
|10||Kansas City, MO-KS||32,966||-2,018||-6,633||-2,593||-2,244||-2,827||8,343||8,852||1,645||35,491|
Here’s the bottom ten:
|1||Miami-Fort Lauderdale-Pompano Beach, FL||-489,537||-637,986||-488,454||-656,464||-921,299||-749,266||-656,371||-578,875||-319,368||-5,497,620|
|6||Tampa-St. Petersburg-Clearwater, FL||-141,225||-150,747||-158,583||-172,940||-252,097||-244,928||-157,668||-139,623||-53,674||-1,471,485|
|8||Atlanta-Sandy Springs-Marietta, GA||-65,596||-86,237||-80,997||-118,411||-171,883||-221,298||-246,739||-165,034||-83,497||-1,239,692|
|9||Los Angeles-Long Beach-Santa Ana, CA||-123,833||-153,255||-117,068||-149,971||-150,245||-123,540||-125,267||-131,519||-3,222||-1,077,920|
|10||Charlotte-Gastonia-Rock Hill, NC-SC||-18,899||-44,940||-46,497||-83,099||-118,675||-126,562||-179,474||-146,129||-69,665||-833,940|
Miami says thanks. No two ways about it, New York metro is exporting a lot of income. For those who are interested, you can download the full matrix of net AGI flows for New York.
Like I said, there’s some seriously interesting stuff in here, and I plan to keep digging into the new data for quite a while, so stay tuned. For more of what you can do with this, click here.