American income statistics aren't quite what people are saying they are

The latest numbers for incomes, health insurance and poverty are out in the US. This is one of those times when we've got to really understand how a statistic is calculated before we can work out what it's really telling us. And here what it is telling us is not what some people are saying it is. Take Catherine Rampell at the Washington Post:

One of the more striking parts of the U.S. Census Bureau’s 2016 income, poverty and health insurance report is its data on income stagnation. While incomes went up in 2016, the current median household nonetheless brings in about as much as its counterpart in 1999 did — that is, around $59,000 (in inflation-adjusted dollars).

If you look at men’s earnings alone, though, the trends are worse.

Today’s full-time, year-round male worker earns a median of $52,000. That’s roughly what his counterpart made in 1972.

This is equal to the statement, the same as the statement, that the lifestyle you can buy on the median male wage is the same now as it was in 1972. Which is a statement of such utter ridiculousness that merely restating it that way shows that it must indeed be wrong. Some of us were around and adult back then and it really wasn't that way.

One problem is that of how this income is being measured. Which is to leave out crucial parts of compensation - most notably the health insurance and pensions contributions that people gain from going to work. It's entirely true that health care has gone up in price over these decades but it's also true that it has got a great deal better. We do live longer lives, we live healthy lives for longer, now than then which is, at least we would assume it to be, an increase in lifestyle and therefore of any real measure of income.

But the biggie here is the measurement of inflation. These numbers are, of course, adjusted for such. But they are adjusted by a variant of the CPI. And we know, absolutely, that the CPI is wrong. By, depending upon who you talk to, anything between 0.5 and 2%. Always, but always, over-estimating inflation.

The biggest problem being, and we've not really got a solution to this either, practically or conceptually, the prices of new things. To get an inflation level we want an average inflation level. Sure, that health care rises in price but food much less so, we've even seen significant deflation there at times. Our method is to work out how much the average family spends upon food (say, 12% for American households today, around and about), how much for health care (umm, 20%?) and then multiply their respective inflation rates by that basket composition. If we consumed the same basket of goods over time then that would work just fine.

But of course we don't consume that same basket. The smartphone, with the iPhone, is a decade old now. When it first came out it wasn't in anyones' consumption basket. After a year or two it was in those of some people but not enough for us to say that it was a common enough purchase that we should be measuring that inflation - or crucially here, deflation - rate. By, say, 5 years in and the flood of Landfill Android a smartphone was an average purchase and so made it into the basket we use to calculate that average inflation rate.

But the iPhone started at $500 (??) and by the time a smartphone was in the inflation basket it was what, $50? And this is true of all new products. ABS started out only on the most expensive cars - is it even possible to buy one without it now? And at what point in the price drop for the technology did it enter the consumption basket we use to measure average inflation? The tumble drier, central heating, microwaves - all new technologies suffer from this problem and as above we've no real method of solving it.

But look what happens to real incomes if CPI is 2% overstated for 45 years - correcting for that means that what we're recording as the same has in fact more than doubled. Even just the 1% over 45 years gets us close to a doubling and that strikes us as much more representative, somewhere in between those two perhaps, of the development of the average lifestyle that can be afforded on the male American median wage over that time period.

It's entirely obvious that in any way that matters, looking at consumption possibilities, that US male wages have not stagnated for 45 years. Which just leaves us wondering why so many seem so gleeful in telling us that they have?