Household Formation: Divorces, Births Correlated with Unemployment Across States

From the Simpsons episode where Milhouse’s parents get divorced:

Kirk van Houten:  “Singles life is great, Homer; I can do whatever I want.  Today I drank a beer in the bathroom.”
Homer Simpson: “The one down the hall.”
Kirk van Houten: “Yeah, and another great thing, you get your own bed.  I sleep in a racing car, do you?”
Homer Simpson:  “I sleep in a big bed with my wife.”

Kirk is not only employing someone who makes race car beds, he’s also taking up another unit of housing. Kirk van Houten is engaged in the messy business economists refer to as household formation, creating another unit of demand for housing (as well as boosting aggregate demand for things like race car beds in weak economic times).  If you haven’t seen it already, there’s several arguments that a wave of household formations will start, taking pressure off the housing market and boosting overall aggregate demand.  See Yglesias here, Krugman here and Karl Smith here.

As you can see here at Sober Look, “Family households have completely decoupled from the US population growth since 2008.”

You’ve also seen this in relation to divorce rates, child-birth rates, households doubling up, etc. – a big drop-off starting in 2008.  But, usually due to data limitations, this analysis is almost always at the national level.  I wanted to find rates at the state level, to use the range of unemployment rate growth across states to see if household-related changes tracked them, boosting the argument that the weak economy is causing these changes. Weirdly, I can’t currently find household growth at the state level for beyond 2008.

But I did find other things.  From the U.S. Census’ Births, Deaths, Marriages, & Divorces data page, here’s changes in divorce rate (2009 subtracting 2005 averages) against the change in the unemployment rate by state (2009 subtracting 2005):

Significant relationship.  Let me clarify this chart.  As unemployment goes up (vertical axis), the divorce rate goes down (horizontal axis).  We knew that the divorce rate across the country declined with the recession; the fact that it correlates with unemployment rates by state gives us more evidence that this is related to the weak economy.

Getting a divorce is an expensive and economically risky move.  Equally expensive and risky is having a child.  Here’s changes in birth rate from 2009 to 2005 against unemployment.  It’s an even stronger relationship:

Same dynamic.  (It is also true if we use “fertility rate” instead of birth rate.)  If 2009 had the same birth rate as 2005, there would be an additional 2.25 million babies born  153,000 births per year. Over the course of the past several years, it is likely that there are 5-8 million babies not born as a result of this prolonged recession.

Marriage is, oddly, not correlated with unemployment increases by state.  Is it because marriages are income risk-reducing instead of income risk-increasing events like divorce and having a child?  Is the average cost of a wedding going down in this recession?  Segmentation in who is getting married, particularly by education (Sober Look argued that here)?

In some ways this linkage makes me a bit more worried about the “recovery winter” argument happening.  I read household formation as primarily a young person issue.  Households are doubling up at high numbers, and there are record numbers of adult children living at home.  Census blog: “All in all, 61.7 million adults, or 27.7 percent, were doubled-up in 2007, rising to 69.2 million, or 30.0 percent, in 2011.  Young adults were especially hard-hit…That left 14.2 percent of young adults living in their parents’ households in March 2011, up more than two percentage points over the period.”

My previous worry was that young people will be last in line to get jobs as job growth happens, keeping a lot of potential household formers on the sidelines, thus keeping the recovery in check.  I’m not sure if they are actually “last in line”, but if so there’s a problem there.  Now I’m a bit more worried that household formation dynamics are being equally driven not by something that requires a bit of extra money – like being able to pay rent – but a lot more extra money.  Having a child is expensive, particularly since it requires adding cost household labor hours – if household formation recovery needs to expand the birth rate to truly engage a multipler-accelerator “winter recovery”, that could be quite problematic.

(On the other hand, I can start prepping my #slatepitch: “How the Upcoming Divorce Wave Will Save the Economy.”  Maybe Newt Gingrich can tell us all something about how to have a strong economy and reduce unemployment after all….)

UPDATE:  The estimated number of babies not born was off in calculation; it has been fixed and changed in the post above with previous estimated struck-out in text.

This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.

18 Responses to Household Formation: Divorces, Births Correlated with Unemployment Across States

  1. Deschain says:

    Question: if the population growth rate slows, isn’t a lower rate of economic growth also ok? IE, GDP/capita can keep increasing at the same rate.

  2. Pingback: Secondary Sources: Dovish Bernanke, Fed Market Movers, Household Formation - Real Time Economics - WSJ

  3. Pete Murphy says:

    Economists are wrong. A wave of household formations will not come to the rescue of the housing industry. The fact is that, as populations grow more dense, beyond some critical level, per capita consumption of nearly everything, except food and clothing, begins to decline. Nowhere is that decline greater than in the housing sector. As an example, consider Japan, a wealthy nation but one that is ten times as densely populated as the U.S. The average Japanese citizen lives in a dwelling less than one third the size of the average American’s, not because the Japanese like living in cramped quarters but because there is no room for anything larger. Now, imagine the impact on the per capita consumption of everything that goes into building, furnishing and maintaining homes.

    The real problem here is that, since per capita consumption and per capita employment are inextricably linked, worsening unemployment and poverty are the inescapable result as long as the population continues to grow.

    Our unemployment problem is driven by our ever-worsening overpopulation and overcrowding, the same force that is steadily eroding the per capita consumption of housing and nearly every other commodity. So, no, it’s not possible that unemployment will give rise to a new wave of household formation.

    Pete Murphy
    Author, “Five Short Blasts”

  4. Will Smit says:

    Where are you getting a change in births of 2.25 million below annual trend from? The graph on the first page of makes the delta look much tinier than your number – a drop of 2.25 million would suggest that less than half as many babies were born in 2009 as 2005, which none of the statistics I can find confirm. I wonder if you are not mixing percentage changes to the birth/fertility rate with absolute changes to the rate.

    • Mike says:

      Will Smit,

      Thanks for catching an unfortunate error – I had messed up the calculation when I estimated estimated the impact of the drop in the birth rate. Fixed now, apologize for the confusion.

  5. Barry Branton says:

    For the first graph it would be nice if you told us what was the level of significance and the r-squared value. The r-square looks very small but I would like to know what is the level of significance for the independent variable on the x-axis

  6. Steve Sailer says:

    The Hispanic birth rate has collapsed since the end of the Housing Bubble.

    • yorksranter says:

      have you run the numbers for tiresomely unproductive fascist trolls, by the way? what is Mr Sailer’s fucking job, may I ask, having seen your bullshit all these years?

  7. Pingback: Daily Digest for January 30: Married to the Recession » New Deal 2.0

  8. lee wolf says:

    Lee Wolf asks

    Please comment on total household formations vs. family formations during the same time period.

  9. john putzivokium says:

    How can you arrive at any conclusions using Japan relative to the US in overcrowding and density of population? The area of the US in square miles is 3.79 MILLION with a population of 312 MILLION, a ratio of a about 1.2 people per square mile, vs. Japan with an area of 142 THOUSAND square miles, 2,667 times SMALLER than the US, with a population of 127.5 million, or approximately 90 people per square mile, with a population almost half the size of the population of the US? The Japanese have no where to go or expand but on top of themselves. Sounds like fuzzy math to me was utilized in arriving at your final conclusion. Don’t bet against Americans or the US, you’ll lose!

  10. john putzivokium says:

    Sorry, it’s approximately 900 people per square mile, not 90, even I can make a mistake.

  11. Alex says:

    interesting. i bet domestic abuse is also going to rise as a result

  12. Pingback: Daily Digest for February 1: Leaving a Deutsche Mark » New Deal 2.0

  13. Freddie says:

    Very disappointing to see a guy with such great quantitative chops accepting the (totally useless) divorce rate as indicative of actual failing marriages.

  14. Pingback: The Coming Boom in Divorces? « Herston on Tennessee Family Law

  15. Pingback: Divorcing our way to prosperity « Family Inequality

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s