Silly political statistical graphics

The awful graph below showed up on facebook recently. It’s old (from 5/3/11), but it struck me immediately how silly it is, and I was curious what the numbers look like when plotted more transparently (i.e., honestly).

This is disingenuous (at best).

What’s the problem with the graph? Lots of things (not least of which is the fact that I’m spending a good chunk of time thinking and writing about a political graph I found on facebook). Among the problems: These presidents weren’t presiding for the same amount of time. The debt isn’t only (or even primarily) due to anything a president does (it’s been a while since high school civics, but I dimly recall that Congress has a role to play in budgetary matters). Oh, and the increase expressed as a percentage of the previous debt is not the only relevant way to look at debt numbers. Given that the debt has been increasing for so long, it makes sense to look at it in terms of actual (billions of) dollars and as a percentage of GDP, for example.

So, I went to whitehouse.gov and got the excel file “Table 7.1—Federal Debt at the End of Year: 1940–2016” to see if I could unpack these numbers a bit. First, I tried to roughly recreate the above graph. Here are the percentage increases relative to the debt in the previous year from 1982 to 2010, with red indicating Republican administrations and blue indicating Democratic administrations:

Percentage debt increase from previous year

I went year by year and calculated the raw increase (in billions of dollars), divided that by the value of the previous year, and multiplied by 100 to get a percentage. If you squint-blur your eyes, this kind of looks like the facebook graph. The debt increased by a lot under Reagan, less so under Bush The Elder, still less so under Clinton, more so under Bush The Younger, and at Reagan-esque levels under Obama. Wait. What? Debt increased less under Bush The Younger than it did under Bush The Elder? And it has increased more under Obama than under Bush The Younger? What’s going on?

If you sum up the percentage changes over each administration’s time in office, you get numbers that look a bit more like those in the first graph. Reagan: 103%, Bush The Elder: 46%, Clinton: 35%, Bush The Younger: 60%, and Obama: 33%. Close, but still not the same. I don’t know what numbers they used in the first graph, but they don’t agree with numbers available on the White House’s website.

In any case, percentage of previous debt doesn’t tell anything like the whole story. The actual dollar amount of debt, and the increase from year to year in dollars, seems to me to be pretty important, too. Here’s a graph showing the increase from previous year (over the same time frame), in billions of dollars:

Billions are big.

Hey, now. This one makes Bush The Younger (and Congress during his two terms) look incredibly irresponsible, and it makes Obama (and Congress during his time in office) look even worse. Now, these numbers aren’t adjusted for inflation or, probably more importantly, for the size of the economy as a whole. Here’s a graph showing the overall debt as a percentage of GDP:

Debt as % of GDP

So, debt only decreased as a percentage of GDP under Clinton, increasing under everyone else in this time frame. Here’s the year-to-year increase in % of GDP:

Last graph, I promise.

Once again, Bush The Younger and Obama (and Congress blah blah etc…) come off looking pretty bad, Obama (and cohort) even worse than BTY (and cohort), despite what I hope were not the best efforts of the Office of the Democratic Leader (see the credit at the bottom left of the silly first graph).

The overall point here is the dog-bites-man message that it’s easy to deceive with numbers and fancy graphs. It’s not surprising that politically motivated people would produce politically slanted graphs like this. It is kind of annoying that they do, though. And it’s distressing that this kind of thing gets praised as informative and important, either because people can’t or won’t think hard about what numbers are being shown. Here, for example, are some of the comments that accompanied the graph on facebook (anonymized, of course):

“Thanks for this data which belies the Republican debt narrative. Facts do matter!”

“Thanks. If only voters more often read and researched and thought for themselves….”

“This is wonderful, thanks so much, [name]! I’m looking forward to the Republicans and Foxers trying to spin this…. LOL… In any case, thanks!”

“You’re welcome. My grandmother used to say, “Figures don’t lie, but liars do figure~!” I trust the chart prepared by the US Dept. of the Treasury.”

First, the graph wasn’t prepared by the US Treasury. Second, 10 minutes of research ‘for myself’ shows that the graph is exceptionally dishonest. Third, facts do matter, but the original graph is short on facts.

Fourth, and just to be clear, my goal isn’t to “spin this.” I’m not a Republican (or a Democrat), and I’m no fan of any of the presidents pictured above (nor of their Congressional cohorts). My goal here is to show that even a little bit of careful thinking (and a couple quick google searches) can do wonders for finding out what’s going on in a set of numbers.

UPDATE: I forgot to mention that the original graph is aesthetically awful, too. It’s ugly and dishonest. Didn’t want to let a gratuitous dig fall by the wayside.

UPDATE 2: A friend (on facebook!) pointed out that GDP is a moving target, which was something I meant to mention in the first place. This is important because, e.g., there was a nice big economic boom during Clinton’s administration, and a nice big recession during the end of BTY’s and much (all?) of Obama’s so far.

The same friend pointed out that it would interesting to plot with color encoding congressional control, so here’s a figure of the billions-of-dollars increase with that color scheme. I got the relevant information here. The two houses were controlled by the same party over much of this time period, but when they weren’t, the outer ring indicates senate and the inner the house.

Billions of increase, coded for congressional control.
This entry was posted in mildly informative filler, statistical graphics. Bookmark the permalink.

4 Responses to Silly political statistical graphics

  1. It’s not just Obama’s bust and Clinton’s boom, your second-to-last graph – the one misleadingly labeled “last graph, I promise” – tracks recessions and their severity pretty well (possible exception: BTE’s recession was not as severe as the 1980-82 recession and yet leads to a higher %GDP increase). So, whether and how deep a recession there was is clearly a confounding factor.

    Are you aware of Hauser’s Law? It’s the idea that govt. spending as a percentage of GDP has held more or less constant at 18% in the US since the end of WWII. The Obama Administration marks the first long-term projected deviation in the direction of – you guessed it – more government ownership of the economy.

  2. Damien RS says:

    Bush’s debt went up because he cut taxes on the rich.

    Obama’s debt went up because there’s a massive recession and massive unemployment and thus tax revenues are down because people aren’t working, while entitlements like unemployment insurance and food stamps and Medicaid go up because people aren’t working. Note that happens automatically, without Congressional action, though they can and did expand it with unemployment benefit extensions, stimulus, and tax cuts. It’s not a bad thing, unless you have a fetish for balancing the budget on an annual basis, unlike households or businesses. It’s also not “government ownership of the economy”.

    An interesting to see might be the amount of inflation adjusted debt incurred along with the interest rate at the time of borrowing. Or relatedly, estimated lifetime cost of the debt.

  3. Bush’s debt went up because he cut taxes on the rich.

    The recession-tracking graph suggests it has more to do with falling revenue as a result of recession. There was a recession in the US in 2002-2003; this coincides with debt/gdp ratio increases during the Bush Jr. years. The same is true for Reagan 1980-82, Reagan 1987, Bush Sr. 1990-91, Bush Jr. 2008, Obama every year in office. (The one year on that graph that doesn’t seem to track the pattern is 1986.) It’s fine to say that tax cuts exacerbated the problem by cutting revenue at a time when it was falling anyway, but the overall trend suggests that tax rates have less to do with it than the general direction the economy is heading. In any case, it is certainly not true that had the Bush tax cuts not gone into effect, the deb/gdp ratio would not have grown by comparable amounts. The primary cause was the recession.

  4. Pingback: Does calculus matter? » Source-Filter

Comments are closed.