And Scott Winship responded with a thoughtful post addressing my argument. I want to finish my argument and then respond to Winship.
There are two things I want to mention before I reply to Winship’s response.
IV: The Median Person Might Not Be Middle Class
This is, of course, up for debate, though talking about class online often becomes very heated so quickly it isn’t very productive. This is a my fault issue – it’s something I assume without any evidence; I was talking about the “middle class” and using median data, and this often isn’t useful. Assuming that the “middle-class” falls where the median person is assumes a rather flat income distribution, and our income distribution is tilted towards very high inequality. (Though average household income also peaks in 1999.) Though then that requires a theory of post-industrial middle-class-ness, which you aren’t getting in this blog post. This was the error Thomas Frank pointed out in Bartels’ critique of What’s the Matter with Kansas?, and it’s always something I keep an eye out for.
From the data of male wage growth by percentiles provided by Winship, every percentile above the median had wages increase over this time period, so if you tend to think of what we culturally think of the “middle-class” as hovering around the 70th percentile of the distribution, then they are definitely better off. Do check out the chart: if you are in the business of thinking our inequality and current financialized capitalism aren’t a big deal as long as they benefit the least-advantaged in our society, every percentile below the median has lost wages over this time period. This is open to a critique of how to deflate these numbers, which we’ll get to in a minute.
V: Two-Income Trap
Reihan and Waldman both brought it up, and I’ve blogged about it before: The Two-Income Trap. Let’s grab a chart from Elizabeth Warren’s Two-Income Trap:
Though households make more money than they did in the 1970s, mostly through putting a second worker in the workforce, they have less discretionary income than they once did. And they are also maxed out on available labor, leaving them extra vunerable to any additional shocks. That chart is from this blog post of mine, where we try to get a quant point-of-view on what a credit default swap on the middle class would look like (one of my favorite entries).
I suppose what you make of that chart depends on your priors. If you are comfortable turning the rational-expectations style logic all the way up to 11, you could say that this is a sign that households feel less risky about their futures since they are choosing to both take on leverage and commit themselves to fixed and difficult-to-adjust expenses. What jumps out at me, as a finance geek, is that if this was a business I was reading the statements and sheets for, and I saw an operating costs structure that was trending this way, I’d have to think that something was going wrong. I blog about some rough sketches of ideas on how inequality and policy factors into this here.
Again I like to thank Winship for taking the time to respond. A few quick points.
Rates and Levels Taking the rate of borrowing is what is natural to the argument here – we want to see reactions in borrower to the increase in wages going to zero. “If you can find a relationship there, you are more creative than I am” – maybe it is just me, but it is obviously clear to me that the first derivative of total credit borrowing skyrockets in 2000. The relationship is there in comparing both rates.
And I take the argument that credit among consumers started to balloon in 1999 and 2000 for granted. There’s a lot of ways to see that, take for instance this chart on income over debt that is at Ezra Klein’s blog here.
We don’t know why that happened. Maybe it is the result of China’s One-Child Policy and a global savings glut. Things in this part of the world are complicated; what I wanted to make sure was that we had wage stagnation during the 2000s, with a peak at 1999-2000, because that gives us the potential for a demand movement in addition to the supply ones.
Housing I think the housing bubble has left a fragile middle-class in a terrible place, something that isn’t just rounding errors. Some numbers:
More than two million families have lost their homes to foreclosure since this crisis began in late 2006…Another eight to 13 million additional homes are projected to go into foreclosure in the next five years. Nearly 10.7 million homeowners are underwater in their mortgages…Roughly half those borrowers owe more than 120 percent of their home’s value. American homeowners lost an astonishing $3.6 trillion in equity in 2008, and are projected to have lost another $500 billion in 2009.
And as Kai Wright has reported, this has devastated the precarious black middle class.
Effects of Increased Work Among Women This is a good point. I will do a literature review before I write about this again; I remember the Acemoglu paper Women, War, and Wages that started convincing me that women’s labor were more of a substitute to men’s labor during the midcentury than I would have normally thought; no doubt lots of follow-up work has been done on that (getting the current citation list for that paper from google scholar makes me love the internet).
Deflator A lot of Winship’s argument is whether or not to use the CPI deflator versus the PCE deflator. I prefer the CPI definition, and I also think that the arguments about how consumers should feel richer based on the fact that they have no additional wages but the quality of goods has increased loses the entire “option value” of money. If only for risk mitigation purposes, I suspect this option value has been increasing in the past few years. But this is an argument that requires more thought, which is something I’ll be working on in the near future.