Because Elections Are Determined by the Median Voter

The folks at the Census Bureau have just finished compiling the most recent data on income distribution. I’m betting that the following chart will get a lot of political play:


Yes, median real household income was lower last year than in 2000. And yes, it has in fact grown somewhat in the past few years.

My two cents: The real news (and the really sad news) is the rising poverty rate over the same period:


But I’m betting this will be ignored. With convention season underway, all data entering the public square will be viewed through the lens of electoral politics.

The talking points in the above chart are less crisp for either side (poverty today is higher than in 2000, but lower than in 2004). And more to the point: elections are determined by the median voter, not by the families on the brink of poverty.

There are pages of fascinating reading in the full data release, here.

Mark Thoma rounds up reactions from across the blogosphere, here.


Re: 44

"I’m amazed at how much people overestimate the influence of presidential politics on economic growth."

Many do this but I think the truth is in the middle. I have no doubt that Bush's tax cuts did not help matters. Part of letting the business cycle "work" is to prudently move out of the way; Bush did not do this. I also think part of this is related to 9/11; again, actionable intelligence before the fact was not acted on. And there was a downturn after this watershed event. In the end tho it matters not whether presidents have this pull as they all associate themselves with the economy. Bush sold his tax cuts by linking them to an economic upturn, for example. Think of Regan's "better off now" question. So you reap what you sow and voters have every right to judge the presidency as a whole based on economic trends. On the whole, Bush has failed and McCain looks like a lackey for stating the Bush years were good years. It's perception that makes reality here.



#19: Raman, to get to the median household income, you have to have to start with all the households and all the household incomes. If it worked the way you would appear to think it does, you could strip out a percentile group from the calculation without any effect.

If you think you're right, go ahead and try it.


@David (25): If you think Gooch was being serious, that single mothers marrying their child's fathers would end poverty, maybe you should go back and re-read it. The point he was making is that the way poverty is defined is ridiculous, to put it lightly.

I'm not sure about elections past, but it's sad if the median income is the one focused on during any election season. The economic goal of government shouldn't be to make everyone middle class *cough* SOCIALISM *cough*, it should be to assist those in true poverty only to the point where they can work to raise themselves to the middle class and beyond.



Marriage went rapidly out of favour in the last 5 years? Really?

It always surprises me when low-tax/small-gov/libertarian types are so gung ho for socially authoritarian measures, but state-enforced shotgun weddings are a new one...

(...and you say tax is coercion?)


@Chuck Simmins: "For the first seven years of Bill Clinton’s terms, it was 13.6%. Under George Bush, the average poverty rate for 2001-2007 is 12.4%."

That's a REALLY twisted way of looking at the statistics. Clinton inherited a very high poverty rate, it was reduced during his term; Bush inherited a lower one, and it increased. The average over that period is, in terms of assessing performance, almost irrelevant.


"I was actually more intrigued by the numbers from 1992/93 to 1999/2000. These were the Clinton years. Impressive. We need more of this, no matter who can bring it. "

Nobody brings it. It's a business cycle. I'm amazed at how much people overestimate the influence of presidential politics on economic growth.


I really hate it when people blame the President for the economic woes of a country. He does not have the power or influence to alter the economy. He can neither improve or damage it to any significant degree. The economic problems of the past 8 years are caused by the tech bust and the general debt/consumption pattern of the american public. If people would be wiser with their own money, it would have a more significant effect on the economy than a President.


Forget median income... look at purchasing power... which has plummeted.

Chuck Simmins

The average poverty rate for all Americans for the first seven years of the Reagan terms was 14.2%. For the first seven years of Bill Clinton’s terms, it was 13.6%. Under George Bush, the average poverty rate for 2001-2007 is 12.4%.

American women reached record wage parity with men in 2007. Women earned 77.8 cents for every dollar earned by men.

The average income for all blacks in 2007 was $46,631. This is the third highest average in history, exceeded only by the years 1999 and 2000.

The average income for Hispanics in 2007 was $50,828. While it is lower than last year’s all-time record, it is higher than all years preceding 2000.

The poverty rate for people over 65 was at a record low in 2006, 9.4%, and is at its second lowest for 2007, 9.7%.


Unless those Billionaires had a really really sucky year and ended up on the second chart.



Zero of the statements I made about medians were incorrect.

It turns out exactly as predicted:

Strip out the top %: median falls.

Strip out the bottom %: median rises.

Strip out equal top and bottom: median unchanged.

If you believe otherwise you are very wrong.



Kevins mistake, not Ted's. Oops, you're right I far overestimated Medicare and Medicade going by memory. I must have included social security.

However we both seem to agree on this point:

Less than 5% (about 3.5% by your data) of the federal budget goes to the wars in Iraq and Afganistan.

The projected $490B deficit spending this year equals about 18% of the projected $2800B spending.

With zero war spending we would not even cut deficit spending in half.

The war is neither the sole cause of our spending problem nor the largest part of it.


@6 and @7

Median - I don't think that word means what you think it means. It would be (near) impossible for the top wage earners to affect the median income. You may be thinking of mean, which isn't what this chart is showing.

Mark B.

@ #7 Dennis Garber

Actually a recent analysis of the increasing difference between the rich and poor suggests that this has been increasing since at least 1972. In fact the correlation between year and the Gini index was over .95. For the non-statisticians out there, thats a very very strong relationship (with a 1.00 being a perfect relationship). Over the same time period the GDP per capita has also increased with a similarly strong relationship.

What do these relationships all mean? I'm not sure. I'll let you guys debate that.


#9: jrnold: It doesn't work the way you think it does. As presented in Justin's top chart, the household median incomes are based on two things: aggregate household incomes and the number of households. Of these, the change in aggregate household incomes is the driving factor affecting changes in the median household income observation.

Two things contributed to forming the median income bucket shown in Justin's top chart. First, total aggregate income in the U.S. became compressed during the burst phase of the U.S. stock market bubble and coinciding recession from 2000 through 2003 (shown here, specifically the "All 100%" curve).

As you can see in this chart, the aggregate incomes of the Top 1% of household income earners was the most compressed (compared to 2000). That fact may be understood in that a lot of the people who had been high-fliers during the stock market bubble dropped out of the top ranks of income earners in the subsequent fallout.

The second thing you should see is that median household incomes recover as the household incomes of those in the Top 1% recovered, and we should note that many of the people in this Top 1% are not the same people as in the earlier years.

The reason you see the median incomes compress as they do is because the total aggregate income was compressed during this period. You just have to know how the numbers were determined to actually understand it.

#12: I don't disagree, Chris. Median incomes didn't dip by anywhere near as big a percentage of previous income level as did those for the Top 1%! Nor did they spring up as fast in the recovery phase. The incomes of those counted in the Top 1% are highly volatile.

#14: Close! As explained above, the key is total aggregate household income, which the Top 1% of household income earners disproportionately affect.



Justin - how can you say the following "elections are determined by the median voter, not by the families on the brink of poverty.". Median voter is different than median income earner, since income doesn't perfectly correlate with political affiliation. I don't think we will know who the true median voter is until after the election, but it could well be families on the brink of poverty.

Black Political Analysis

In a war year with a bad overall economy it doesn't surprise me in the least that poverty isn't mentioned on the campaign trail. Since the Great Depression, only when the middle-class is doing well can we turn our attention to poverty.


poverty line is a difficult assessment because it is the level of annual income below which a household is defined to be living in poverty. This is defined differently by different governments and institutions and, in spite of the great importance of its intent, is not in fact as meaningful as one might wish.

Because it is a specific threshold, say household earnings at 20% (don't know the actual number) of the average household earnings when the average wage increases this 20% increases and so too does the number of people below the poverty line, even though those people may not be any poorer than previuosly.

the disturbing thing here is it happened as the average wage was falling. So therefore it seems that the 20% has fallen yet there are more people below the line. This means that those in poverty are even poorer than before and there is more of them.


The chart above is based on the median income (half above, half below), not the average. If the top 1% had a really sucky year, it would only show in the chart if the cause was something that also impacted those in the $45,000-$55,000 income range.



You're flatly wrong. In FY06 Medicaid, Medicare, and SCHIP accounted for 19% of the federal budget. Defense 21%. Social security 21%. 9% interest on debt. 9% other safety net programs. 21% other random stuff.

About 1/6th of the defense portion went to support actions in Iraq and Afghanistan.