Tuesday, August 02, 2005

The Reverse Lake Wobegon Feature of the Distribution of Wealth

by Tom Bozzo

A CNN.com article on the latest savings rate data (via Economist's View) claims:
Even as a government report Tuesday showed the national savings rate at zero -- that's right nada -- the rise in the value of homes has given the average U.S. household a net worth of greater than $400,000, according to a separate report from the Federal Reserve.

The distribution of wealth has a long right tail, the very rich being, well, very very rich. So the average net worth — i.e., total net worth divided by number of households — in the 2001 Survey of Consumer Finances (*) was $395,500. But the median net worth — the amount such that 50% of households have greater net worth and 50% have less — was only $86,100. (Update: See here for the latest data from the 2004 Survey.) Indeed, in 2001 around 80% of households had below-average net worth.

The poorest 25% of households had an average net worth of zero and a median net worth totaling a whopping $1,100. This includes bank accounts and other financial assets, vehicles, houses and other real estate, and more.

Wealth isn't that democratized.

Additional Note: From 1992-2001, the mean and median diverged,with the median increasing by 40% ($24,800) and the mean increasing 72% ($165,000, both adjusted for inflation). The distribution data in the Federal Reserve Bulletin article allows some easy lies with statistics. The 12.5th percentile of net worth (1/8 of households are poorer) increased 83%, while the 95th percentile (19/20 of households are poorer) increased "only" 58%. But if you want to know who will be popping the Champagne, the former increase was $500, while the latter was $479,300.

Another important definitional limitation ofthe data is that the Survey of Consumer Finances wealth measurement excludes employer-sponsored defined benefit pensions. As a result, the shift from defined benefit to defined contribution retirement plans will make measured household wealth increase when there has actually been a subsitution of a measured asset type for an unmeasured type.


(*) This is the most recent data I could locate; the Fed web site notes that results from the 2004 update of the Survey are due early next year. I'm trying to figure out precisely what Federal Reserve release is referenced in the CNN.com article. (The summer Federal Reserve Bulletin is due soon, but not yet posted to the web.)
What is the historical trend for average and median net worth? I tried looking at the Federal Reserve and Census pdfs, but there are too many small numbers for a novice like me to figure out. Specifically, has the poorest 25% of households seen an increase in median net worth?

Hypothetically, if the average net worth matched the median net worth, would that satisfy that wealth is democratized? Or would we still have to take down those rich bastards?
Bryan, see the additional note in the expanded post.

Mere convergence of the mean and median isn't sufficient (they could converge at zero). While I do consider the rich to be undertaxed, the optimal tax would leave them rich.
If the mean and the median converge at zero, isn't wealth still democratized? There's always going to be a distribution of wealth, and regardless of where the number lies, the total amount of wealth will be shared in almost the ideal way possible???

I am naive on these economic matters, but I know from experience that Gaussian distributions show up in everything. There will always be poor and rich people, even if everyone received the same amount of money every year, due to the ways people invest and spend money.
BTW, I agree with you, in the current climate of fiscal insanity, the rich are undertaxed. My preferred solution would be to shrink the gov't (including the defense dept), but I don't see how we can continue to run up the deficit. If taxes need to be raised, then lets do so.
Usually, "democratization of wealth" is used to mean "everyone is (more-or-less) rich," which is also my intended meaning. A symmetric mean-zero bell-shaped distribution a la the Gaussian would imply a small proportion of rich in the right tail while half the population had negative net worth... not so great.

FWIW, there are nontrivial government expenditures that I'd reduce or eliminate were it up to me.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?