The Minneapolis Miracle

Part 3 of a Cantankerous 3 part series.

A claim I keep seeing popping up is the idea that Minneapolis is a notable example of effective supply side reforms in housing.

I am afraid that, currently, it is not.

It could be. It might be on its way. It may be heading in the right direction. But, the data doesn’t yet confirm a meaningful success there. And, we would be wise not to act like it is, because if we jump the gun on this and it turns out that Minneapolis hasn’t really changed the trend in local housing costs, then we’ve just created a big false negative narrative that can be cited as a place that proved supply-side reforms don’t work.

This is an example of how the things I have discovered about the Great Recession and its aftermath are important details for understanding what is happening today.

Figure 1 contains 3 charts from a Financial Times piece on it.

Figure 1

There definitely was a sharp downshift in rents in Minneapolis in 2021 and 2022. And there is a clear correlation between housing construction and rent trends. Those FT charts look compelling. Is this a YIMBY win?

The FT article also mentions Auckland, New Zealand. Figure 2 shows construction in Auckland. Auckland has implemented effective supply-side reforms. The results of those reforms appear to be notable. Rents are down noticeably in Auckland. This is the sort of thing we should be looking for in an American city. Auckland appears to be a clear YIMBY win.

Figure 2

Figure 3 shows total housing permits for the Minneapolis metro area, Nashville, and Charlotte. Can you pick out the supply-side miracle here?

Figure 3

Taking a broader look, Figure 4 compares trends in permits in the Minneapolis area to the US and to several other cities. This highlights two points.

1. Minneapolis (along with Charlotte and Nashville shown in Figure 3) construction has run remarkably parallel with the national average. (Unrelated note: I wonder what caused permits in Minneapolis to peak in 2004, a year earlier than the rest of the country.)

2. There is a lot of variation around the average, from Pittsburgh to Austin, for example, in Figure 4. If there has been a supply breakthrough large enough to flow through to current rents, we should be able to see something. Even then, it would be hard to know if the difference was from changes in demand or changes in supply. But, the thing about Minneapolis is that, compared to the national average, there wasn’t much change to explain.

Figure 4

In Figure 3, I tried to pick other interior metro areas with relatively healthy local incomes and economic trends. If you compare Minneapolis to a rust belt city, its permitting since 2008 tends to look better. I think that is what is happening in the FT story.

I suppose this is another way where a comprehensive understanding of the 2000s boom and bust is necessary to understand new data in an unbiased way. The mortgage suppression after 2008 created outcomes that were correlated with income. It became harder for households in poorer cities to get mortgage funding, and this shows up clearly in post-2008 trends. Poorer cities have built less since 2008. But this has been due to federal lending standards, not changes to local land use rules.

Figure 5 shows the relative change in the number of permits issued per capita after the Great Recession compared to before the Great Recession. Poor cities permitted fewer homes after the recession because poor families (and middle class families) can’t get mortgages very easily. Before the Great Recession this sort of thing didn’t happen. Generally, through boom and bust cycles, the pace of construction in cities tended to rise or fall proportionately. The post-Great Recession correlation of relative construction trends with incomes is really odd.

Figure 5

Since the effects of mortgage suppression have been largely unnoticed, those trends are attributed to other things. In this case, Minneapolis compares favorably to nearby metro areas, which tend to be struggling with rust belt economic challenges. But, I suspect this is more due to effects of mortgage suppression than supply-side reforms.

Figure 6 compares the incomes of the cities that were included in the FT charts. I think what we have here is a spurious correlation. FT can be forgiven for missing it. Nobody seems to have noticed the most devastating policy change of a generation. Nobody except EHT readers knows about this.

In fact, I wonder how many countless spurious correlations there are in the post-Great Recession academic literature, in any number of topics, because it is rare to see this controlled for.

In the FT charts, the cities with the most new dwellings had the lowest rent inflation. From top to bottom, the order of cities on their left panel (showing housing production) is the opposite of the order of cities in the central panel (showing rent inflation). So, new home construction is negatively correlated with rent inflation. That is very true.

But, Figure 6 shows incomes in those cities. The order is the same in Figure 6 as it is in their left panel (except that Columbus and Indianapolis have changed places). Minneapolis is an outlier among the cities in the FT chart because Minneapolis incomes are higher.

Figure 6

Figure 7 shows the log change in rents (from Zillow) among 49 major metro areas, arranged by income. Poor cities have built less since 2008 and rents in poor cities have gone up more since 2008. Income (and mortgage access) is driving these patterns. Not differences in local land use regulations. (It is true that the causation goes in the other direction, over the long-term. Strict land use regulation leads to higher home prices and outmigration of poor residents. That causes average incomes in cities with strict land use to rise over time. But that’s not what is going on in Minnesota and the rest of the country since 2008.)

Figure 7

Keep in mind. This correlation between income, new home construction, and rent inflation is really weird. And, to make it even weirder, rent inflation after 2015 is highly correlated with price deflation before 2015. The more home prices went down, the more rents subsequently went up.

Now, of course, if you have the right model, it’s not weird at all. Most of what happened after 2007 was the result of overzealous mortgage suppression. Where it bit the most, it lowered prices and construction the most. Rents had to rise the most to bring home prices back up to equilibrium with the new construction market. It actually makes perfect sense. But, you have to have noticed the most devastating policy change of a generation for it to make sense. And the most devastating policy change of a generation doesn’t appear much in the history books or the academic literature.

So, make no mistake, the point the Financial Times is making is correct. Irrefutably, more new homes push rents down. The question remains: Is Minneapolis a good example of this and are Minneapolis supply-side reforms important?

Well, even my Figure 7 shows Minneapolis as a bit of an outlier. Rent inflation has been about 10% lower than the average metro area of its income level. In the FT graphic, Minneapolis rent inflation was about 20% to 40% lower than the other cities. So, much of the difference between Minneapolis and the other cities in the FT charts is from incomes (and, I would assert, mortgage access). But, even after accounting for incomes, Minneapolis still appears to be an outlier.

So, was it a change in supply or a change in demand that pulled Minneapolis rents down 10% or more?

Well, as I mentioned in the previous post, a round estimate of demand elasticity for housing is about 0.5. In other words, each 1% increase in the housing stock should lower rents by about 2%. In fact, in a basic regression of rent inflation among the 30 largest metro areas against housing permits and population growth since 2015, either 1% more population growth or 1% lower housing growth are both associated with about 2% more rent inflation.

In that regression, Minneapolis is an outlier. Rents are almost 10% below where we would expect a city with Minneapolis’ income and housing production to be. That suggests that something other than new supply has lowered rents.

Furthermore, for supply to have made that much of a difference in rents, Minneapolis would have needed to increase the housing stock by roughly 4% or 5%. (5% more homes means 10% lower rents.) In any given year, Minneapolis increases its stock of homes and its population by about 1% to 1.5%. So, 3 years of extra supply should be associated with a 10% drop in rents. In Figure 4, Minneapolis should look something like Austin. Nothing like this has happened.

In apples-to-apples comparisons, Auckland seems to be a good example of effective supply-side reforms. Minneapolis doesn’t. It is possible that some quality of life issues or lower population growth related to the local economy or the George Floyd aftermath temporarily reduced local rents. Did supply-side reforms help construction rates remain stable for a few years when a collapse in local demand would have caused it to dry up? Possibly. But, it would be an amazing coincidence if reform-boosted construction followed a level path that perfectly counteracted highly unusual temporary local demand shocks.

Also, a while back, when I tentatively hoped we were seeing a start to a California YIMBY turn, it wasn’t because construction had increased. It was because prices in the EHT looked like they were starting to reflect more affordability, and I thought those prices might reflect expectations of moderating future rents.

In the case of Minneapolis, the drop in rent has not been accompanied by a drop in prices. The EHT supply measure for Minneapolis has improved slightly since 2020, but at about the same scale as the average major metro area. So, home buyers, sellers, and builders don’t seem to have changed their expectations about Minneapolis housing costs. And, in the most recent data, Minneapolis permits are declining (see Figure 3 and 4) and rent inflation is reconverging with the US average. Zillow estimates that the median rent in both Minneapolis and nationwide are up about 2% since July 2023. This is what we would expect to see if Minneapolis had a one-time downward shock in demand rather than effective supply-side reform.

I hope Minneapolis has implemented and will continue implementing reforms that are effective. But, I don’t think we can say that yet.

If we don’t heed my “Part 3” complaint here, we’ll find ourselves buried even deeper in the nonsense of my “Part 1” complaint, and Minneapolis will be used as rhetorical ammunition. Anti-YIMBYs will try to claim Minneapolis as an example where even if you make housing legal, oligopolistic builders will refrain from building in order to keep prices high.

Don’t create false negatives! That doesn’t help.