Income vs. consumption inequality

by on January 25, 2007 at 5:33 am in Economics | Permalink

Some time ago, John Quiggin became apoplectic at libertarians citing TV and Playstation purchases as evidence against American poverty being a serious problem; Henry Farrell chimed in too.  John asserts that consumption data "tell[s] us precisely nothing about what’s happening to inequality."

I cannot agree with this claim:

1. Consumption is robust for many categories, not just fancy TVs.

2. The data indicates that the people buying the stuff are not miserable, or at least not miserable for economic reasons.  There are plenty of historical episdoes where consumption does fall, and we know that is not a pretty state of affairs.

3. The demand for flat TVs and the like is not just a relative price effect, it is also a wealth effect.

4. If robust demand for fancy TVs and PlayStations is not convincing, what kind of consumption data would be?  Let’s say there was a robust demand, among the middle classes, for medium-size yachts.  Rembrandt etchings?  Wouldn’t that show something?  It can be argued that "TVs are not enough," but that is not reason to reject consumption data out of hand.  It is a reason to look at more categories of consumption.

Consumption studies do have the following defects:

1. They sample smaller numbers of people than do good income studies, and they cannot pin down the consumption patterns of definite percentiles very easily.

2. Money spent is not always money well spent.

3. The data series do not go far back in time and there may be problems of consistency over time.

4. People may be borrowing and accumulating large debts.  Note that in this case, however, the comeuppance, however bad it may be, has yet to come.  It could instead be argued that "inequality will (someday, when the debts come due) be a serious problem."

Mark Thoma surveys some interesting pieces.  Here is a very detailed study of the topic.  Here are many excellent slides on the topic.

Consumption data, even if sometimes misused by zealous libertarians, are not a means of dismissing the poverty problem, but they do put that problem in another light. 

First, they show that income and wealth data overstate poverty and inequality problems.  Second, a focus on income data leads one to conclude that the elderly require most of the assistance.  A focus on consumption data lead one to conclude that helping parents with children is, in many cases, more important.  That sounds right to me.

1 tom s. January 25, 2007 at 7:41 am

For some comments on consumption vs. income measurement in a Canadian context, see a posting at “Relentlessly Progressive Economics“:

[Neil] Reynolds [of Report on Business]seems unaware of the fact that relentless pressure from the right led the federal and provincial governments to develop a consumption based measure of poverty, the so-called Market Basket Measure. This was intended to calculate what income would be needed to purchase the basics of life for representative families in different communities, apparently in the hope of lowering poverty rates by statistical fiat.

While the MBM poverty rates generated by this exercise differed from the after tax LICO rates in terms of detail, they were not really all that different in terms of the overall dollar amounts which define poverty thresholds. That’s probably why we have barely heard of the MBM since.

In summary – shifting the focus from income to consumption does not fundamentally throw into question the extent of poverty in Canada.

2 John Quiggin January 25, 2007 at 10:38 am

I really think you’re missing my point here. Of course, income and price effects are relevant in general, as is the fact that consumption has grown more than income (that is, saving has fallen), and I link to an extensive discussion of these issues in the post you link. But picking items where price effects are clearly dominant, like flat screen TVs, and observing growth in consumption of those items, just obscures the issue.

Also, I don’t think saying something annoys me amounts to getting apoplectic.

3 John Quiggin January 25, 2007 at 1:11 pm

Umm, maybe you should read the quote in context before claiming that I’ve backed off from it. My statement in the comment that “But picking items where price effects are clearly dominant, like flat screen TVs, and observing growth in consumption of those items, just obscures the issue” restates my observation in the post, that looking at consumption changes driven by price effects tells us nothing about inequality.

Honestly, there’s enough substantive points to debate that we shouldn’t be playing this kind of game. I’d be much more interested in a discussion of my actual views, put forward in this post which I’ve already linked to several times.

4 nelziq January 25, 2007 at 2:31 pm

Consumption or even income data can’t take into account inequality without considering wealth and hours worked. Is a household with both adults working two jobs making 100k a year the same as a household where one adult works one job and makes 100k or one where single adult makes 100k a year from a trust fund? Can we consider them to be in an equal economic position?

5 Michael Sullivan January 25, 2007 at 2:57 pm

John asserts that consumption data “tell[s] us precisely nothing about what’s happening to inequality.”

It’s probably already clear from John’s comments, but I’ll be more blunt than he can afford to be: This ripping of context is entirely unfair to JQ, and looks like a pure straw man attack, although I’m sure that was not your intention.

He did not assert that about consumption data in general, but about a particular point of consumption data which happens to be an outlier (playstations and flatscreen TVs). Your points do not speak to that premise at all, only to the one you falsely infer here, and even suggest that you would largely agree with his actual point in that line, though I’m certain you’d have other large differences in philosophy once you start engaging his actual arguments.


6 aaron January 25, 2007 at 5:16 pm

From BLS data:
Annual Change 2000-2004, Expenditures by EBIT quintile.
$17,940 $26,550 $34,716 $46,794 $75,102
5.3% -0.2% 2.7% 4.1% 2.7%
0.9% 2.4% 3.4% 3.5% 2.7%
-3.0% -1.5% -1.8% 0.1% 3.2%
-3.5% 2.5% 2.1% 1.0% 2.4%

Expenditures as % of EBIT (2000-2004).

233.5% 139.2% 105.5% 87.8% 68.2%
237.6% 130.4% 100.3% 85.6% 66.1%
229.2% 128.2% 99.7% 85.2% 65.3%
225.5% 124.4% 96.5% 82.6% 64.3%
194.6% 113.7% 88.9% 78.3% 63.3%

7 aaron January 26, 2007 at 8:37 am

2000-2004 income growth by quintile (lowest to highest):


The lowest group is falling behind, looking at just 2000-2004. But 2005 and 2006 probably saw faster growth at the low end (2005 and 06 had much higher job growth than previous years).

The middle seems to be catching up to the rich.

8 Ian Gregg February 12, 2007 at 11:32 am

This a prime example of too much of a good thing. It’s a part of an economic problem which is scare resources and unlimited wants. It’s always going to be the same story and peoples spending habits will never change. Plus income should affect your consumption but thats not always true.

9 不動産投資 July 15, 2008 at 2:53 am

資金を増やそうとするのに不動産投資をするのが手っ取り早い。日本で不動産で東京 賃貸をさがすのはきわめて難しくシステム開発は日本の会社が良い。

10 battery May 11, 2009 at 12:29 pm
11 r4 ds carta February 10, 2010 at 11:55 pm

Inequality in terms of consumption utility (rather than consumption dollars) is less still. Seriously, how much more utility is provided by driving a Ferrari instead of a Toyota? Survey after survey show that the rich as only marginally happier than everyone else.

Comments on this entry are closed.

Previous post:

Next post: