# Median Household Income Trends: Wisconsin vs. Minnesota

No surprise, Minnesota beats Wisconsin, again.

Some individuals have touted trends in Wisconsin’s median household income relative to Minnesota and the Nation. The Census Bureau warns researchers from doing intertemporal comparisons using the standard series (e.g., on FRED), given data breaks due to different methods of interpolation. Here I plot the American Community Survey (ACS) series for Minnesota, Wisconsin and US, which are not subject to the same problem of comparability, and are estimated with greater precision due to larger sample sizes.

Figure 1: Median household income for Minnesota (blue), Wisconsin (red), and US (black). A reading of 0.045 means that the value is 4.5% higher than it was in 2010. Source: American Community Survey/Census Bureau, and author’s calculations.

Use of this series reverses the impression from the standard series that the gap between Wisconsin and Minnesota closed going from 2012 to 2013 (see Figure 6 in this post). In fact, the gap has been widening since 2010.

Once again, it is useful to read the footnotes.

## 19 thoughts on “Median Household Income Trends: Wisconsin vs. Minnesota”

1. Phil

It is WORKING! But only for the fortunate 1%

That is OK with Walker and his contributors.

2. Patrick R. Sullivan

So, how does Walker keep winning elections in Wisconsin if he only helps the top 1%?

1. Menzie Chinn Post author

Patrick R. Sullivan: Helps to have billionaires bankrolling you.

I am still waiting to hear you admit you were in error regarding depth of the downturn in Canada vs. US during the Great Depression. As you recall, you stated unequivocally:

And this statement is wrong.

3. AS

Professor Chinn,
What is the proper analysis if the median income for workers in Wisconsin has increased more than the median income for workers in Minnesota, while the median household income is the reverse as you show above? If I calculated the numbers properly it looks like the increase for workers has been about 3.6% in Wisconsin compared to 2.9% in Minnesota using logs normalized to 2010 as used above.

1. Menzie Chinn Post author

AS: Not sure. No reason why the two should match up exactly, although you’d think that over time the two should correlate. I think both series contain important information (I don’t know the median worker income series — all I know of is the average).

1. AS

Professor Chinn,
Thanks for the comment. I notice that the “median earnings for workers” is about 6 lines down from the median household income on the American Community Survey/Census Bureau link that you provided. I wonder if the difference in percentage change between the household income and [individual]worker income is due to more or (different income bracket) family members leaving the workforce compared to individual workers when comparing Minnesota and Wisconsin.

4. macroduck

Walker still has a shot at the nomination, but he is a smaller part of the GOP conversation now that Trump is giving the press something fun to write about. Maybe the lack of scrutiny is good for Walker. Give the world time to think about his economic record isn’t likely to help his chances.

Wisconsin has been lagging Minnesota for decades.

There may be a greater drain of higher income and higher educated people, both old and young, in Wisconsin compared to Minnesota.

Minnesota and Wisconsin: How did two peas in a pod grow apart?
March 7, 2015

“From the 1940s to the 1960s, Minnesota and Wisconsin were as alike as “two peas in a pod.”

But then something happened: Minnesota rose while Wisconsin fell.”

http://www.startribune.com/minnesota-and-wisconsin-how-did-two-peas-in-a-pod-grow-apart/295426901/

6. Mike V

I’m working on a masters in statistics, so I am quite math-literate; but I cannot comprehend why you insist on using logs for absolutely every graph. It seems so unnecessary to complicate what should be a simple time series graph by using a log transform. Why not just index the value to 0 or 100 at the start like all other economists do?

1. Menzie Chinn Post author

MikeV: I will merely refer you to a post by this fella, who wrote a book you might recognize called “Time Series Analysis”. The text has a mere 16,200 citations in Google Scholar. (If you are not familiar with it, then you should ask your econometrics professor about it.)

If you plan to do serious work in economics, and particularly econometrics, you might seek to overcome your aversion to logarithms.

1. Mike V

I’m not at all averse to logs, but they have a time an a place. You are trying to point out relatively small changes in income over a short time-series – not the nearly exponential changes in the S&P 500 over the last 100+ years.

I just think you lose a lot of people by using logs for every. graph. Here is your data (excluding the U.S. median) visualized similarly, but with using cumulative percentage change instead of logs : http://i.imgur.com/9AHEszF.png

I bet most would find the second one more intuitive and easier to understand at a glance.

1. Menzie Chinn Post author

Mike V: Thanks for the advice on how to present data.

My view — if you don’t want to bother to understand a log, then how can I get you to understand why a error correction model might be more appropriate than a simple OLS regression. But if that bothers you, by all means, go ahead and skip my Econbrowser posts. (I am thankful you did not express yourself as vehemently as this like minded fellow).

1. Mike V

Alright, thanks for the condescension and snark. Your graphs are hideous, by the way. Take a poll on your site about which graph is more attractive and easier to understand and I’ll shut up about it.

And based on my statistics background, I still see logs as being completely unnecessary for this particular data.

2. Menzie Chinn Post author

Mike V: Thanks again for your input. I will put up a post soon for your benefit. In the meantime, I’ll just place you and James Sexton in the group of like minded individuals.

I might note that while the aesthetics of my graphs might leave something to be desired, it has not seemingly had a negative impact on my ability to convey information, insofar as referees at peer reviewed journals are concerned. Nor do I recall any complaints when writing memos as CEA staff economist.

Mike, one of my econ instructors at CU-Boulder, who also happened to get a Ph.D from Berkeley, said to the class one day he was most interested in statistics, but went into econ, because statistics is a “dead field.”

I guess, he meant there was nothing more to learn in statistics and there remains lots to learn in econ (in the 19th century, economists of the day were known as moral and social philosophers).

The econ department actually preferred math majors over econ majors, and there’s a lot of math in upper division undergrad and grad econ classes.

In undergrad econ, you learn a little about a lot, and in grad econ, you learn a lot about a little (i.e. mathematical and empirical models, methodology, terminology, etc.).

7. Samuel

Dear Dr. Chinn.

Just something I read today which I thought might interest you.

Madison WI has one of the highest concentrations of employment in software publishing in the nation.
http://www.bls.gov/opub/ted/2015/area-employment-concentration-in-software-publishing-industry-december-2014.htm

In addition, Dane County (i.e., Madison and surrounding communities) consistently has the lowest unemployment rate in the state of Wisconsin.