Marriage In The United States Of America

Last month I made a claim that one of the reasons the GINI coefficient, a measure of the disparity in income, is not telling the whole picture in America is that it doesn’t reflect the true concept of households.  I made the case that:

For example, you could take 4 people with incomes described as:

  1. $24,000
  2. $30,000
  3. $50,000
  4. $75,000

The Gini coefficient for the above data is .24162

Now, marry two of those wage earners:

  1. $24,000
  2. $50,000
  3. $105,000

The Gini coefficient for THAT data is .301676.  Without ANY income changing at all, the Gini increases by 25%.  In other words, the same number of people are working the same number of jobs and earning the same number of dollars.  The only difference is the method by which they calculate the Gini.

Nickgb over at Poison Your Mind called shenanigans.

You are taking a nation-wide economic statistic, applying it to a population of four, then three, and drawing conclusions as to its usefulness?

Certainly my math was simple.  It consisted of a population of 4 individuals.  Clearly this was just a demonstration of what could or might occur in a larger group.

However, since then, we have learned that a record number of Americans are unmarried:

Barely half of all adults in the United States—a record low—are currently married, and the median age at first marriage has never been higher for brides (26.5 years) and grooms (28.7), according to a new Pew Research Center analysis of U.S. Census data.

In 1960, 72% of all adults ages 18 and older were married; today just 51% are. If current trends continue, the share of adults who are currently married will drop to below half within a few years. Other adult living arrangements—including cohabitation, single-person households and single parenthood—have all grown more prevalent in recent decades.

Certainly there is nothing wrong with people deciding that they would rather enjoy life’s treasures as a single person rather than a married person, but it is also true that when measuring household income using the GINI coefficient, a drop in marriage rate of 33% will impact the results.


3 responses to “Marriage In The United States Of America

  1. I assure you if something like this made the US GINI coefficient way off base compared to other states, statisticians would have caught it and published the data. If there were such a glaring error in a statistic used by the CIA and almost everyone, it’s not something that the world’s best economists and analysts would miss. There are problems and the way the index is determined can be manipulated, but its a well scrutinized measure.

    So what it would take to convince me there may be problems is not speculation or thought experiments that “maybe this will effect it,” but some analysis that demonstrates there is some problem, especially in making comparisons to other OECD states. I do think that overall wealth differences and other social differences makes comparing first and third world GINI scores of limited value. I haven’t seen any analysis which suggests comparisons to OECD countries are off.

    • I assure you if something like this made the US GINI coefficient way off base compared to other states, statisticians would have caught it and published the data.

      The reason that statisticians don’t point out anything wrong with the GINI is because the folks who publish the GINI are using it correctly. The GINI is nothing more that a statistical tool that measures the dispersion of a thing. Dandelion seeds from the flower in inches. Touchdowns per game. And income.

      The calculation is accurate.

      The applicability is the inaccurate part.

      We see the same phenomena when things like infant mortality and life expectancy are used as comparison as well. For example, consider life expectancy. In the United States, people of foreign decent trend to the same life expectancy as those native born foreign citizens. Specifically, Japanese-Americans have the same life expectancy of native Japanese. Same for Swedish-Americans and Mexican-Americans. However, because America has such a mix of cultures, the comparison to more homogenous society paints us in a bad light.

      Further, life expectancy stats, commonly used or published by CIA, don’t go into cause of death. No one considers that American’s suffer much higher death rates due to sports, car accidents and gang land violence than do the Nordic nations. However, when such variables are isolated, America ranks #1 in the world for life expectancy.

      The same can be said of infant mortality. The United States has significantly higher rates of IVF pregnancies, multiple embryos, single mothers, at risk mothers, women with high blood pressure, incarcerated fathers and low birth weight babies. Again, when those factors are considered, the United States scores much MUCH higher.

      No one cites that data.

      I have no faith whatsoever that today’s media looks into the statistics further than asking if a high GINI as it relates to income distribution is good or bad, and then simply prints the list.

      I haven’t seen any analysis which suggests comparisons to OECD countries are off.

      I’ll keep digging.

  2. I can make anybody number one in life expectancy if I’m allowed to remove some variables. That’s why it’s illegitimate to say “if you take X, Y, and Z out we’re number one.” Essentially if you have to remove those factors it’s a sign that income inequality in the US is creating a lot of death and suffering for those on the lower rungs — a stinging indictment of US policies and culture. It proves that the US has a problem with income inequality and lack of class mobility (something other studies show we’re very poor at, especially the bottom 20%).

    Again, if GINI had real flaws that totally altered the data of comparing the US vs. the OECD, there would be evidence out there — people would be proving it and showing it. I also note that pre-tax and transfers the US is pretty good .46, more equal than Germany’s .51. Post tax and transfer Germany is at .30 while the US is at .38. Other OECD states are similar to the US pre-tax and transfer, but post tax and transfer range from .23 to just over .30, at least western states (we’re more like East European states). That’s strong evidence about the fact that our tax and transfers are much less progressive than those in other states. Nothing you bring up would impact that — we’re looking at differences pre- and post- tax and transfer.

Leave a Reply