Question about economics – how to improve GDP as an economic indicator

Great Debates
1

I tried to find the answer to this on the internet, but I couldn’t find much.

As BrainGlutton mentioned in this thread, the GPI (genuine progress indicator) as a more progressive alternative to the GNP (gross national product) for measuring the well-being of people in a country. That got me to thinking about what the single best indicator of a country’s economy and standard of living would be. There are various problems with GDP, arising mostly from externalities, or things that are outside the monetary system, which various people have pointed out over the years.

The problem with the alternative systems like the GPI or PPP (purchasing power parity), though, is that they make assumptions about what people value. Since GDP is the most comprehensive indication of (monetary) wealth, the only assumption that it makes is that money is an indication of value. This assumption isn’t perfect, but most externalities eventually get incorporated into the monetary system, raising the GDP as the society becomes more progressive and living standards start to genuinely rise.

With the caveat that no one indicator can ever tell everything about a country’s economy, then, GDP seems logical as a first place to look. In order to process that information, though, you would have to divide that by the country’s population. There are two ways to do that – if you are concerned about what the dollars (or whatever the unit of currency is) are doing then you would use the mean (average), or if you are concerned about how their owners are doing, then you use the median. In social science, you’re studying people, so the median is usually the best measure. (I was going to think of a counterexample of something better analyzed using the statistical mean, but I actually couldn’t come up with any, outside of abstract physics problems. Football coaches might care about the average number of yards per play, except for turnovers on downs. Maybe someone can help me out here.)

Despite this, GDP, when it is divided by population, is just about always done using the mean.  See the [the CIA World Factbook](http://www.cia.gov/cia/publications/factbook/rankorder/2004rank.html), for instance.  (Per capita is presumably the same thing as mean).  One reason for this fact is probably that the United States has a large disparity in income which doesn’t come out in the mean GDP figure. Measured using the median, the US’s GDP would almost certainly be far from the place it is listed at here, second only to Luxembourg’s.

Looking on the web, I tried to find more progressive economists advocating for this approach, without success. Some people have looked at median income, but that doesn’t count other non-income sources of wealth that are counted in GDP (like investments). Also, it doesn’t count people who make no income, which makes it pretty useless as indicator of a country as a whole. I’m assuming that this median figure is possible to calculate, since every dollar is ultimately traceable to a person, even if maybe the methods they presently use to calculate GDP don’t collect the right data. Maybe someone who knows more about economics than I do can correct me if I’m wrong.

I tend to abide by the philosophy that it is best to change things the least amount possible, which is why I’d propose a less radical solution than the GPI, or other similar systems, to cope with problems with GDP. Disparity in wealth is one of the biggest externalities usually cited, and this system would help to counteract that. This problem is also often symptomatic of other such problems, so measuring it would indirectly measure those as well. So does anyone know why this isn’t more widely accepted, or does anyone know of anyone using this system at all?

If anything here is unclear, please ask me.

2

Well, for one thing, I think the implicit assumption in GDP that individual wealth and progress scale linearly is grossly mistaken. Under GDP, if you have 1 person with $90,000 and 1 person with $10,000, it’s exactly the same GDP as 2 people with $50,000. I would like to see an indicator that sums up say the square root or a log of each persons individual incomes and see how that compares. Problem with that is you need a lot more data to compute it. GDP might be so overwhelmingly popular because it’s one of the few data points that we can compute about a country reliably.

3

Oh, btw: I’m reminded about an article by Ross Gittens, an Australian economist who pointed out who non-sensical it is for household debt figures to be published in terms of per capita income (ie: The average american now has a debt of 110% of his per-capita income or whatever). Debt is a measure of wealth, income is a measure of change of wealth. It’s about as sensible as saying the distance from LA to SF at 150% of my average speed.

However, the reason why it was so ubiquitious was because income figures are vastly easier to get a hold of than wealth figures so lazy economists would just use the first. After looking at debt from a wealth perspective, the problem is actually much milder than it appears.

4

I think in your first post you’re basically going at the same idea I am, that it’s the owners of the money that’s more important than the money itself. I’ve read that the distribution of wealth tends to be logarithmic, so in theory taking the log of each person’s wealth, averaging the result, and then bringing the base of the log back up to that power should give a similar result as using the median. In fact, that method might even be better, because it factors in every single person’s wealth, rather than just the one person who is exactly in the middle.

You’re also probably right that computing people’s actual wealth, rather than movement thereof, would be very tricky. You’d have to get people to look behind the couch to find everything they own, then go on Ebay to find out what its value is.

FWIW, I thought of an example where you would want to use an average – when computing lottery or gambling winnings. In that case, you’d be concerned about how much money you are making, rather than how many times you go without winning anything. With economics though, you are concerned about how many people don’t have any money at all, regardless of who has lots of it.

I was thinking about how you would compute this number, and there would be a problem figuring out how to account for government spending, since it isn’t tied to any one person. I would imagine theoretically you would want to figure out how much say a person had in government decisions in order to calculate their gains or losses from government sales and purchases. Practically, the best thing would be to assume that everybody has an equal say, and the worst possible result would be getting the same number as the GDP calculated per capita (which already assumes that everyone has the same wealth). If someone wanted to get a more accurate number, there are probably ways of calculating people’s sway over government in a reasonably objective fashion. For instance, I’ve read studies that federal disaster money is more likely to be dispensed close to election years, and quantifying those kinds of increased likelihoods would provide a reasonable estimate of how much being in a swing state or a key congressional district, or being wealthy or owning parts of influential corporations, would matter. It would get pretty complicated, which maybe could be part of the reason this measure isn’t used – but complexity hasn’t really scared off economists in the past, and those measures still wouldn’t be neccesary.

5

The reason why complexity is bad is that complexity breeds subjectivity which allows people to insert their own agenda into the figures. Why not provide raw income and wealth data free to everyone so people can come up with their own indicators?