Nowcasting Migration Trends
Yesterday’s News
Lag is a fundamental challenge with data. Imagine trying to guess today’s weather by only reading about yesterday’s weather. You could probably figure out a range of possible temperatures. Also, if it rained yesterday, it might be more likely to rain today. But these inferences don’t provide concrete evidence of what’s happening outside. And unlike this weather hypothetical, large, difficult-to-gather data sets often lag by well over a year. The obvious solution is to simply open a window and see what’s going on outside, and while that works well for weather, for national demographic trends, such as migration, there isn’t a reliable, publicly-available window to look out of. In short, this is the problem I’m trying to solve: find a good window.
… for national demographic trends, such as migration, there isn’t a reliable, publicly-available window to look out of. In short, this is the problem I’m trying to solve: find a good window.
Digital Exhaust
Similar to Newton’s Third Law where every action has a reaction, any time anyone does anything on the internet, a piece of data is born. And when a bunch of people do the same thing in large numbers the “digital exhaust” they create can be highly indicative of national trends.
…any time anyone does anything on the internet, a piece of data is born… the “digital exhaust” they create can be highly indicative of national trends
For migration, moving companies in particular create some interesting digital exhaust. Some of these companies actually publish a portion of their data in annual reports, but once again, there’s a huge data lag looking at year-end trends. Instead, at any given moment we can look at how much it costs to rent a moving service one-way between cities. When there’s significantly higher demand for truck rentals from one location to another than in reverse, these companies price the two trips differently to compensate for the fact that they’re likely going to have to drive their trucks back to the city that everyone is leaving.
Since there have been so many articles written about people moving from San Francisco to Austin, let’s use that as an example. As of writing, a small truck from San Francisco to Austin costs ~$2,600, while going from Austin to San Francisco only costs ~$1,500. The reason for this difference is that fewer people are moving from Austin to San Francisco than vice versa. Now if we scale up this data and look at all US cities with a metro population of at least one million people (plus a couple cities I find interesting, just for kicks), we can investigate which cities are more “in-demand” and which cities are falling out of favor.
I will caveat that this is an imperfect dataset. I’d like to think that this dataset captures all demand for all people moving. But since there are many ways to move across the country, we can only capture some of this demand. On top of that, this data completely misses any foreign immigration. What the data captures specifically is DIY movers who have enough stuff that they need to rent a truck. And with these disclaimers out of the way, let’s look at some maps.
The Top Ten
First, here are the top ten markets for demand. These are markets with the largest premiums for one-way trips compared to the same trip in the other direction.
- Birmingham, AL
- Oklahoma City, OK
- Charlotte, NC
- Louisville, KY
- Jacksonville, FL
- San Antonio, TX
- Houston, TX
- Austin, TX
- Raleigh/Durham, NC
- Richmond, VA
As someone who follows the national narrative about migration trends, the fact that all ten of these cities are in the greater Southeast isn’t all that surprising. However, I was genuinely surprised about which cities make up that trend and which cities are missing. First, the fact that Birmingham, Oklahoma City and Louisville are in the top four is shocking. No offense to anyone associated with any of these cities, but they’re not typically regarded as high-growth hotbeds. Birmingham actually lost population in 2021, according to the most recent census estimate, but maybe that’s due for a reversal.
In terms of markets missing from the top ten, Dallas is notably absent. Likewise, it’s surprising that Jacksonville is the only Florida city on this list. But I’m only including large cities, and many of Florida’s boomtowns have been smaller, suburban areas.
The Bottom Ten
Now looking at the other side of the equation, which cities have the most people leaving? Again, below are the top ten cities where it is significantly more expensive to rent a one-way moving van out of that city than renting one going to that city.
- San Jose, CA
- Los Angeles, CA
- San Francisco, CA
- San Diego, CA
- Riverside, CA
- Sacramento, CA
- Fresno, CA
- New York, NY
- Seattle, WA
- Hartford, CT
Speaking as a native Californian, I find this list deeply troubling. There’s a pretty clear trend in the top ten – top seven specifically – and it doesn’t look good for California’s population growth. This finding also aligns with recent articles published about California’s population decline.
City Relationships
Next, we can look at the top cities for in-migration and examine specific migration links. Use the layer button in the top right corner of the map below to add or remove cities from the map, and click the connecting lines to check which city the line is pointing to.
To me, one of the most fascinating parts of this exercise has been seeing the different sources of in-migration for each of these cities. Just looking at the top five, Birmingham, Louisville and Oklahoma City are all drawing primarily from California markets, while Charlotte and Jacksonville draw more people from East Coast markets like New York and Boston.
The out-migration segment could similarly be titled Where All the Californians are going. Still there are some surprising differences between California cities.
All these cities have some ties to Birmingham, Louisville and Oklahoma City, which makes sense given where these three cities rank for inflow demand. However, there are some big differences between the Northern California cities and the Southern California cities. San Jose and San Francisco show a heavy out-migration to Reno, while LA, Riverside and San Diego all favor Las Vegas as their preferred Nevada city. Sacramento is also an interesting case, showing up as an outflow city for San Jose and San Francisco, while also ranking as the sixth market for outflow demand. Perhaps most surprising, there doesn’t seem to be a ton of demand between the Southern California cities and Arizona. The Northern California cities are shedding people to Phoenix and Tucson, but Southern Californians seem to prefer Texas.
Going Forward
Of course, the real value in data is seeing how it changes over time. What I’ve established here is a baseline for comparison going forward. Starting in six months or a year, I’ll review how these demand shifts have been changing month-over-month and see how they correlate to other available data.
Also, if there’s a specific city you’re interested in, which wasn’t discussed above, send me a message. I’m happy to share copies of the underlying data.
More Insights
A look at Cloudline’s market selection process. What are the demand drivers we should be focusing on and how do they compar across major markets?
posted 2/18/2022
Supply is often just as important as demand, and even a great market can still be overpriced.
posted 2/18/2022
All about how syndications work, from start to finish.
