Trouble with Charts: Debt/Income and Retail Sales

The following two charts are hugely appealing to my inner contrarian. One “shows” that the U.S. is less reliant on retail sales than commonly supposed; the other “shows” that debt loads of U.S. consumers by income decile isn’t altogether troubling. I like both arguments and find them intriguing, but the graphs don’t do it for me.

Here is the first one, from Bloomberg. The trouble with it, as must be obvious, is the y-axis. Seen properly, retail sales as a percentage of GDP has essentially flat-lined over the period, moving up or down around one percent. Granted, one percent either way of a $13-trillion economy is a sizable number, but the axis choice makes the variation seem much larger than it is.


And here is the second, from Credit Suisse. On this one I have a couple of issues. First, there are overlapping deciles/quintiles, which is an odd decision. Second, it’s not clear whether these are income and household quintiles. In a subsequent discussion a CS analyst said that they are are household, but the graph makes it tough to tell and generally creates a confused story – even if an intriguing one.