Wednesday, May 04, 2016

Bathtub Model of Migration

The United States reached peak millennial in 2015. Demographer Dowell Myers takes this simple fact a step further and predicts a stronger outflow of young adults from the urban core:

The simple metaphor of a bathtub is widely used to capture these dynamic relations within human migration.  The level of population in a city is like the level of water in a bathtub, held steady by an inflow from the faucet and an equal outflow from the drain. A rising water level could be evidence of faster inflow, or it might be that the drain has been clogged. Similarly, a falling water level might be attributed to an inflow made slower, signifying a weakened preference by human movers to arrive, but it just as well could be due to a more open drain. Demographer Ken Johnson is the most prominent to have voiced the view that postrecession population gains in cities are due to a clogged outflow rather than a stronger inflow (Johnson, Winkler, & Rogers, 2013).

In urban affairs, far more attention is paid to the inflows because they are much more visible and researchable. Those newcomers are present to be interviewed, whereas the departed outflow has disappeared to other, unknown destinations. Thus, for these reasons, many times urban analysis is subject to inflow bias: our interpretations are blind to the equal effects of the unobserved outflow. However, if we survey today’s residents about their future intentions, we might learn about their intended outflow, and yet we cannot ask future incoming residents about their intentions, about their intended inflow. Accordingly, surveys of intentions have an opposite interpretative error, one of outflow bias.

I like this bathtub model of migration, which forces the analyst to disaggregate the flows and consider how bias might be distorting the picture. In the Rust Belt, most assume brain drain is driving the population loss (outflow bias) when the data often reveal weak inflow. Residents don't think about newcomers failing to show up in their neighborhood. They notice the people who leave. As for city revival narrative, stronger inflows receive all the ink. That too makes sense. Something not happening hides behind something that is happening. Why would anyone consider the possibility of a clogged outflow?

The specter of inflow bias and outflow bias puts coverage of population dynamics in a different light. Now, all I see are bathtubs. Demographic hysteria in the Bay Area:

More than one-third of Bay Area residents say they are ready to leave in the next few years, citing high housing costs and traffic as the region's biggest problems, according to a poll released Monday.

Hey Dowell, what does the bathtub say about San Francisco dying? "[S]urveys of intentions have an opposite interpretative error, one of outflow bias."

The headlines say nothing about a collapsing inflow. How could we conduct a survey of future intentions to move to the Bay Area? Furthermore, we don't know if those who intend to leave will actually move to another region. This story suffers from outflow bias.


JRoth said...

I'm unclear on how a "clogged drain" could lead to growth. I understand that a failing city like Flint, where home sales have collapsed, could be a place where people wish to leave but can't, but surely those places usually have drastically reduced inflows as well.

Is it a failure of imagination on my part, or is clogged drain growth the one quadrant that rarely, if ever, happens?

And is birth/death rate relevant in this metaphor?

OwenW said...

@JRoth: I think that the 'clogged drain' is meant to stand in for any decrease in outflow from the recent historical norm, not just one that is in some sense "caused" by the city in question itself, as in your Flint example (people not being able to move because of problems in Flint).

I think you are probably correct that the Flint-type situation is unusual, but outflow also depends on the receiving city on the other end - that is, if, say, mortgages suddently get harder to get everywhere for those not already relatively well off, then increasing rents in a place like San Francisco will fail to drive as many as people as previously to places like Phoenix and Vegas. (This would be the Kevin Erdmann hypothesis, which I find pretty convincing.)

Jim Russell said...

is clogged drain growth the one quadrant that rarely, if ever, happens?

And is birth/death rate relevant in this metaphor?

Last question first, birth/death rate indirectly impacts the bathtub model of migration concerning the size of the age cohort and the life cycle state of that cohort. The elder millennial age cohorts should be flowing out of the city at this life stage, but the drain appears clogged with a fire hose of the peak millennial age cohort going through their city-slicker life stage. The bathtub is backing up, giving the appearance of an urban boom signaling a changing of preferences. Typically, a recession will cause a clogged drain and long-term trends suggest declining geographic mobility.

These days, a clogged drain wouldn't be unusual.