EconTech

He is kind of late to the ball game on this particular article, but I definitely want to post his thoughts (see below). When Andrew is confused, I usually consider myself lost. This is not necessarily an exception, but I think I see a way out. While there are definitely measurement issues, if, after a discussion of utility of money and the lifetime consumption smoothing hypothesis, all I got was a note about the difficulty of measuring income and consumption (misreported sales data, non-reporting of non-taxable income, etc.), I MIGHT give you a B- (at least insofar as this sounds like a stats class targeted at social scientists). The issue is much deeper than that, and in my mind, it is just as important for the statistician to step away from the data and figure out what and why they are measuring something as to be aware of the problems in measuring it.

Also, as I think I’ve said before, I do not find it so obvious that we want the data adjusted for government transfers and such. Sure, I want that data set analyzed as well, but the idea of welfare is linked to more than just a snapshot of what one was able to consume in the past year. I think Andrew realizes this when he chooses (a) over (b), but I thought I’d make it more explicit. Read the linked posts above for more.

You are what you spend

I just ran into this article by W. Michael Cox and Richard Alm on the comparison of incomes and spending of rich and poor:

. . .

This would be a good example for an intro statistics class when the topic of measurement comes up. The challenge for a stat class is to focus on measurement issues–how to design a survey to estimate people’s income, assets, and spending patterns, or how to design an experiment or observational study to estimate the effects of changes in income on spending.

From the economics perspective, the example confuses me–on one hand, it makes sense to use consumption, not income, as a measure of well-being. On the other hand, if I were given the choice between two options:

(a) Earning $100,000 next year and spending $50,000, or
(b) Earning $40,000 next year and spending $60,000,

I’d prefer option (a). So I don’t really know how to think about this. This sort of thing always confuses me in discussions of the utility of money (which I teach in my decision analysis class): it’s good to have more money, but, usually, it’s not money that brings joy, it’s the things that money buys that do it.

In the example above, it would certainly make sense to adjust income for taxes and transfer payments and probably for household size (even if not by simply dividing by the number of people). It’s harder for me to think how whether to adjust for savings or for non-cash benefits such as health-insurance.