The numbers don't lie. Except when they do.

On The Rachel Maddow Show today, they did a cute little segment where they played a fake game show called Pin the Debt on the Donkey whose purpose was to point out that Republican presidents in recent history have added much more to the national debt than Democratic presidents. To that end, they showed this graph.



I don't want to argue about this from a political standpoint, or imply that Rachel Maddow's staff got the numbers wrong* or anything like that. My issue with this graph is that the way the graph is designed deceives you into thinking Reagan spent way way more than anyone else. It's the kind of simple misunderstanding of math that Good Math, Bad Math is all about. First of all, some of these presidents were in office for 4 years, and others for 8. They put this in the "fine print" so to speak, but this is a graph! The point is to show information visually. They could have used the average increase per year, which is also readily available on PresidentialDebt.org. That graph would look like this:



(Sorry it doesn't look as fancy as Rachel's. Upper Harmonics' graphics department consists of, well, me and OpenOffice.org.) In this graph, things look a little more balanced, and you could argue that if Carter and Bush Sr. had gotten a second term, they would have increased the debt by about as much as they did during their first term, so it's a little more fair.

I don't think this is the biggest issue though. The real problem with this graph is that the numbers are percentages, not absolute numbers. This would be of no consequence if they were percentages of the same thing. But each number is a percentage of the national debt at the start of that president's term. So if you happened to be president after someone who (as Rachel points out) nearly tripled the national debt, you look much better in comparison. Not just because you're being compared to someone who spent a lot of money, but because your spending is being reported as a percentage of a higher number. If we look at each president's debt increase in actual dollars (actually, trillions of dollars), we get this:



Wow! You mean the first President Bush and President Clinton actually increased the national debt by the same amount? Yup. And both of them increased the debt by just a little less than Reagan did? Yes. But that's not really fair, you say. Clinton had two terms, and Bush only had one. Okay, let's look at each term, rather than each president, shall we?



This still illustrates the general point Rachel Maddow was trying to make: The worst four-year presidential terms of national debt increase in recent history were mostly during Republican administrations, and the best were mostly during Democratic ones. But I think the graph they chose to create didn't show the whole story as clearly as it could have.

You don't have to be an economics or math major to understand this. This is very simple math. A given amount of money may be a small percentage of one value, and an enormous percentage of another value. Comparing percentages the way they did on the show distorts the facts.

To be clear, I don't think Rachel Maddow or her staff were intentionally deceiving anyone or trying to distort the truth -- it sounds cliche but I'm really not trying to make this into a politcal thing. I suspect they just took the five easiest-to-spot numbers off of PresidentialDebt.org and quickly threw them onto a graph. So some of the blame should lie with that site, which made those numbers a bit more prominent than they should have been.

* Although, they rounded Bush Sr.'s number from 55.6% (direct from PresidentialDebt.org) to 55%. We could argue about the .5 rule all day, but I think we can all agree that 55.6 rounds to 56, right? Also, during the segment, she says George W. Bush grew the national debt by $4.9 trillion. The number is actually more like $5.04 trillion, which I assume is due to another rounding error. Either that or they got that number from another source that uses the actual inauguration days, instead of the end of the year, as the delineation between presidencies.