Tuesday, October 11, 2011

Income Inequality in the U.S. and Connecticut: A Statistical Thought Experiment

By Stephan Adair, Professor, Central Connecticut State University

Our current economic system works for some, but not so well for others. To hold it accountable, we need some measures to assess how well it meets the needs and interests of both the haves and the havenots. One step toward that goal is to take a careful look at the distribution of household income.

If the rich always got richer and the poor always got poorer, it would be inconceivable that a society could ever have a sizeable fraction of its population in a middle income category. In the “Golden Age” that followed the WWII in the US, the size of the middle-income group expanded, the poverty rate was reduced, and while the rich got richer, they did so at a slower rate than the gains for the rest of the population.
[i] Since the mid-1970s, the fortunes have reversed with all the gains going to upper earners. The chart illustrates the percentage of total taxpayer income earned by the top ten percent and the top one percent over much of the twentieth century.[ii]



 

 In 1970, the top 10 percent of earners captured a third of the total income (i.e. 32.6); by 2007, this was nearly half (49.7 percent). Most of that increase went to the richest one percent whose portion of the income pie rose from 9.0 percent in 1970 to 23.5 percent in 2007.

The Gini coefficient, named after an Italian sociologist Corrado Gini, is the most commonly used measure of variation or dispersion for understanding how unequally income is distributed. The Gini coefficient ranges from 0, a condition of perfect equality where all members of a population have identical amounts, to 1, where one member receives it all. Nations in the European Union have Gini coefficients for household income between .25 and .40.
[iii] The recently released 2010 US Census data on household income reports a Gini coefficient of .469 for the US.

Between 1970 and 2010, every state in the U.S. experienced an increase in inequality, but non greater than Connecticut, which went from the 36th most unequal state to the 2nd most unequal .
[iv]

The colored graph presents a statistical thought experiment for Connecticut to illustrate the consequences of this increase. The same “experiment” could be conducted for any state, but Connecticut provides the sharpest contrast. The black bars in the graph represent the distribution of household income based on the 2010 Census data for Connecticut’s 1.36 million households.

The red bars (Scenario A) represent a hypothetical redistribution of income keeping the overall average ($92,990) the same, but changing the Gini coefficient from.486 to what it had been in 1970 (.337).
[v] This change reduces the percentage of households making below $30,000 and increases the number around the average. Scenario A shows fairly significant increases in the percentage of households making between $100,000 to $150,000, and even a small increase in the number of households making over $200,000.

Scenario A depicts a Connecticut in which the overall size of the income pool is the same, while hundreds of thousands of people experience significant upward mobility. This upward mobility is “achieved” by lowering the average value of those making over $200,000 from $387,650 to $235,000. It is not mathematically possible to keep the average household income the same and reduce the Gini to .337 without lowering this value. Scenario A illustrates a zero-sum game in which a decline in the incomes of the richest 8 percent “pay” for upward mobility for others.

Scenario B (the blue bars) maintains the Gini coefficient of 2010, but imagines a 10 percent increase in income levels by raising the household mean income to just over $102,290. Given the current distribution nearly half of the new income went to the top 10 percent, such the average income of households making over $200,000 went from $387,650 to $440,400.

Scenario B yields small increases in the number of households in each category above $45,000 and some small decreases in the lower income categories. There are, however, significantly greater reductions in the low income categories in Scenario A than in B, and greater increases in most of the upper income categories. .

Well over ninety percent of households in Connecticut would be more likely to experience an improved economic condition by returning to the rates of inequality in 1970 with no economic growth than they would with a 10 percent overall increase in the income pool with no change in the degree of inequality.

Connecticut today, where half of all income goes to 10 percent of the population, it is difficult to conceive how economic growth could address the poverty and lack of opportunity that plague thousands. By contrast, in a society where tens of thousands of households are lifted from poverty into middle income groups (Scenario A), it is easy to imagine many small businesses finding new entrepreneurial opportunities and new markets. A Gini coefficient of .337 is still far from a socialist vision of a just society. Rather, it is merely a glimpse of our recent past when middle income groups were expanding, poverty was declining, and the top federal marginal tax rate was 70 percent.
[i] See Mishel, L. Bernstein, J. and Allegretto, S. 2007. The State of Working America 2006-2007. Cornell University Press: Ithaca, NY. [ii] The chart was created based on data collected by Picketty, T. and Saez, E, 2003, “Income Inequality in the United States, 1913-1998.” Quaterly Journal of Economics 118(1):1-39. The data have been updated to 2007 and are available at http://elsa.berkeley.edu/~saez/TabFig2007.xls. Picketty and Saez collect their data through the use of Internal Revenue Service data. In recent years, they have provided some of the most reliable and most cited data on income distributions. [iii] United Nations Human Development Report, 2009. New York: Oxford University Press. The data on the Gini Index are also available at http://hdrstats.undp.org/en/indicators/161.html (retrieved on October 20, 2010).[iv] The US Census calculated Gini coefficients for each state every 10 years for 1969, 1979, 1989, and 1999. Since 2006, the American Community Survey available at http://www.census.gov/acs/www/ offers Gini coefficients for each state every year. Galbraith, J.K and Hale, T. developed a method from data on wages and employment collected by the Bureau of Economic Analysis in the US Department of Commerce to estimate Gini coefficients for the fifty states. For more information on their method, see their 2006 paper, “State Income Inequality and Presidential Election Turnout and Outcomes,” University of Texas Inequality Project (UTIP) Working Paper 33, available at http://utip.gov.utexas.edu/papers/utip_33.pdf. The graphs presented here are based on combining the data from these three sources. No data for 2005 are available. [v] To create this statistical thought experiment, a dynamic Excel spreadsheet was created so that as people are “moved” from one income category to another, the mean income and Gini coefficient are immediately calculated. Smaller and smaller iterative “adjustments” were made until the predetermined mean and Gini values were reached. An electronic copy of this dynamic spreadsheet is available on request from the author at adairs@ccsu.edu

4 comments:

  1. This article was so interesting to read because I have never heard of poverty being measured by the Gini coefficient. According to this number, the U.S. is at .469, meaning we are halfway between total equality and total inequality. This is surprising because I thought our country’s wealth and income was much more concentrated than this number suggests (more inequality present). However, I still don’t think that this inequality is a good thing; something still needs to be done to fix it and distribute the wealth and income more evenly. The interesting question would be, how does the U.S. Gini coefficient compare to other countries we closely associate ourselves with?

    ReplyDelete
  2. This article is extremely interesting and informative. Not only does Adair explain in great detail the Gini affect, but he does so by showing graphs to make the point he is proving that much more drastic. He shows how middle income America these days, especially in Connecticut, is declining significantly as the rich are getting richer and the poor and getting poorer. This leaves no room all for the existence of middle working class Americans. Adair explains that since the 1970s our country has gone downhill economically when the middle class grew smaller and smaller. In post WWII, though, the economy’s rich got richer, but the middle class also prospered and more jobs and opportunities were available to those who were not as well off. Clearly, our economy has shifted over the past 40 years to favor the rich and until we can get to an even gap, our economy may not ever be as successful as it was 60 plus years ago. This article was very interesting and the most informative one I have read so far on this website. Well done!
    Pat Jenkins Intro to Sociology

    ReplyDelete
  3. I truly enjoyed reading Professor Stephan Adair's article about income inequality. His visual aids helped me understand the significant increase of inequality in the United States, specifically in Connecticut (in the second and third graphs). I think his use of the Gini coefficient is a great way to explain percentages of income distribution, and it furthered my understanding of how a majority of the states today have about half of the income received by the top ten percent. Professor Adair's article demonstrates how the United States cannot afford for this increase in inequality to continue, and we have to find a way for the distribution to even out. If states like Connecticut continue this horrendous growth, the lack of opportunity for thousands of people will also increase, and the future for many will be in jeopardy.

    ReplyDelete
  4. I don't think that the consequences of great inequality can ever be understated, so it is critical to understand the situation we are in. Professor Adair's article helps to do this, along with the Gini coefficient and the supplementary graphs. The first graph, in particular, caught my attention. The notable thing here is the years in which inequality undergoes a large decrease. The first, and only, major drop in inequality happened following a decade of extreme inequality (the 1930's). Not coincidentally, I feel, this was the time now known as the Great Depression. Only after a decade of extreme poverty and suffering did the country recover. Our economy is now as unequal as it was before the Great Depression. and the latest recession was proof that our economy is unstable. If history is any indicator, we will likely experience another depression, far worse than the last recession, before inequality once again decreases and the nation gets back on its feet.

    ReplyDelete