Home > Economics - General > Income and Consumption Inequality (crossposted at CT)

Income and Consumption Inequality (crossposted at CT)

December 9th, 2005

I’ve spent a lot of time trying to work out what’s been going on with income and consumption inequality in the United States. Partly that’s because the subject is of interest in itself and partly because social and economic developments in Australia often (not always) follow the lead of the US.

However, there seem to be lots of contradictions in the data, and between data and popular perceptions, for example regarding social mobility and consumption inequality. I’ve finally managed to sort out what seems (to me, at any rate) to be a coherent account of what’s going on. A list of the main points follows, with supporting links, some of which may require registration/subscription. I’ve tried to indicate which bits of the story reflect my judgements, and which are drawn from the literature.

Comments and criticism on this are most welcome.

Here’s the outline

  1. Measures of inequality, mobility and so on reflect a combination of transitory variations and more stable effects arising from social class, occupation, education and so on.
  2. Although the US displayed exceptionally high social mobility in the 19th century, social mobility declined from the beginning of the twentieth century the US and is now comparable to that in European countries.
  3. Because of greater income inequality in the US, mobility within the income distribution is lower in the US than elsewhere.
  4. Wage inequality in the US has grown greatly since 1970. Income inequality has also grown, but not as much since low-wage households have increased hours worked.
  5. (Annual) Consumption inequality has not changed much since 1970. In my judgement, this reflects increased use of credit markets to smooth out short term fluctuations in income, which offsets increased long-run inequality
  6. In my judgement, increased reliance on credit has been the main cause of the dramatic increase in bankruptcy, particularly evident since 1990.
  7. Bankruptcy laws act as a kind of income insurance, and generous (to debtors) bankruptcy laws are a substitute for redistributive taxation.

Obviously, More Research Is Needed™

Categories: Economics - General Tags:
  1. Terje Petersen
    December 9th, 2005 at 19:16 | #1

    QUOTE: A list of the main points follows

    RESPONSE: Where?

  2. jquiggin
    December 9th, 2005 at 19:19 | #2

    It’s points 1-7, hopefully visible now

  3. SJ
    December 9th, 2005 at 19:31 | #3

    Because of greater income inequality in the US, mobility within the income distribution is lower in the US than elsewhere.

    This is a non sequitur. There may be both greater income inequality, and lower mobility within the income distribution, but you’re assuming a causal relationship that you haven’t demonstrated. Not saying that it’s you’re wrong, but the causality is a separate point that must be proved.

  4. Terje Petersen
    December 9th, 2005 at 19:36 | #4

    Yes thats better.

    1. Surely age must be significant also. One would expect most 20 year olds to be less wealthy than most 50 year olds.

    2. Can we blame the rise of the welfare state and progressive income taxes? These work to ensure that the social climbers are cut down to size.

    3. You assert that inequality leads to less mobility. I would imagine that a perfectly equal society would have no mobility. So how does your logic flow.

    4. Interesting that year 1970s. So many things came unstuck at around that time.

    5. I have trouble seeing how credit can help much in this regard. Ultimately the cost of interest eats into a persons ability to consume.

    6. Dido.

    7. Bankruptcy reached a peak towards the end of the 1990s which was consistent with what the supply-siders in the USA were saying about the deflation monster (brought on by the strong dollar policy). Remember gold was in the doldrums and in 1997 the Asian nations that were linked to the US dollar (and then Argentina) could not bear the deflationary pressure any longer.

  5. fatfingers
    December 9th, 2005 at 19:40 | #5

    QUOTE: 6. Dido.

    RESPONSE: A vaguely sexy British artist, one of the few capable of breaking through to a global audience in the new millenium? Or ditto? ;-)

  6. Terje Petersen
    December 9th, 2005 at 19:48 | #6

    ditto.

  7. conrad
    December 9th, 2005 at 20:13 | #7

    In terms of point 2) The US gets many more poor immigrants than Europe, and has for large amounts of time (including many that don’t have a great command of the native language). What proportion of the income inequality is due to this ? There is presumably a direct effect, in that poor people with poor education are not exactly going to be on top of the heap and an indirect effect, in that if you accept large amounts of poor immigrants, it is going to be difficult to convince the population to pay for European-style benefits that might help social mobility.

    If the US and Europe (which countries ?) have similar levels of social mobility, then the US might be doing a better job, but the conditions under which it is trying to do them are more difficult, hence wiping out a direct effect that might potentially be otherwise observable.

  8. December 9th, 2005 at 20:37 | #8

    Terje commented:

    3. You assert that inequality leads to less mobility. I would imagine that a perfectly equal society would have no mobility. So how does your logic flow.

    Consider a society of 2 people. One is rich, the other is poor. In a perfectly unequal society neither would be able to change their wealth, so the rich person stays rich and the poor person stays poor. There is no social mobility.

    Consider another society of 2 people, where both have the same level of wealth. Again, there is no social mobility.

    Consider another society of 2 people. One is rich, the other is poor again. In a society in which mobility is available, and certain reallocative policies are in place. Where the policies are *perfect* we should see both people’s wealth levels approach each other until they are equal, and then no change. So the level of social mobility starts at a high level, and then tapers off as they approach equality.

    Now, if the policies are not perfect, then you might expect some cycles to occur: the reallocation could overshoot, taking it longer for both parties to achieve wealth parity. If the policies are completely messed up, you might expect the society to become more and more unequal as the reallocation mechanism oscillates between the two parties, giving more and more each time period and making the whole situation worse (from an equality standpoint).

    Anyway, to make a long story short: I would think it would be possible to observe less mobility in a society, but diagnosing whether this is due to a society becoming more equal, or one that is remaining inherently unequal is a far more difficult task.

  9. Terje Petersen
    December 9th, 2005 at 21:06 | #9

    Conrad raises a good point. Although Europe clearly has more language issues to contend with. I wonder how more of Eastern Europe joining the EU might change the dynamics.

  10. December 9th, 2005 at 21:19 | #10

    conrad commented:

    The US gets many more poor immigrants than Europe, and has for large amounts of time (including many that don’t have a great command of the native language). What proportion of the income inequality is due to this ?

    On the whole, this would be enormously (if not impossible) to factor into social mobility statistics and policies, but you might get a good idea by looking at those states which have a low immigrant intake.

  11. Terje Petersen
    December 9th, 2005 at 21:20 | #11

    QUOTE: In a society in which mobility is available, and certain reallocative policies are in place.

    RESPONSE: There would appear to be an assumption creeping in that reallocative policies are a necessary pre-condition for mobility.

    JQ stated earlier that in the USA mobility in the 20th century was less than in the 19th century. However reallocative policies were far more prelevant in the 20th century.

  12. December 9th, 2005 at 21:23 | #12

    I apologies for that. I in no way meant to imply that reallocative policies are a condition for mobility, but rather that some level(s) of government were implementing reallocative policies.

  13. Ian Gould
    December 9th, 2005 at 22:45 | #13

    “JQ stated earlier that in the USA mobility in the 20th century was less than in the 19th century. However reallocative policies were far more prelevant in the 20th century.”

    Ever hear of the Homesteading Act of 1862?

    The forcible dispossession of hundreds of thousands of Native Americans from their land and the distribution of that land gratis to whites seems like a rather large act of redistribution to me.

  14. conrad
    December 10th, 2005 at 06:36 | #14

    I think it would be extremely hard to look at social mobility of immigrants in some countries that basically deliberately try not to find the statstics (like France). Alternatively, I’m not sure why it would be hard in others. It should be easy to visualize as well, versus end up with one number (like “more or less social mobility”).

    You could just construct a distribution based on data from the various censuses that countries have done over time (which often go way back), and overlay the distributions at different time points. That could be done overall or for different relatively specific groups that you might be interested in (say, like “Mexican immigrants at census points A to B to C, and their children at points B to C”).

    The distirbution would be handy because it seems easy to imagine that some countries do extremely poorly at the extreme tail of the distribution but not so badly over the entire distribution (blacks in Australia would be a good example that might cause this) and other countries might do more poorly overall, but better on the tail. THe effect of both of these things might be interesting in their own right. Its not my area, but I’m sure this must have been done already, incidentally.

  15. Ernestine Gross
    December 10th, 2005 at 12:53 | #15

    JQ, you invited comments. Given the length of the following, you may wish to retract your invitation.

    1. Measures of inequality, mobility and so on reflect a combination of transitory variations and more stable effects arising from social class, occupation, education and so on.

    2. Although the US displayed exceptionally high social mobility in the 19th century, social mobility declined from the beginning of the twentieth century the US and is now comparable to that in European countries.

    Comment: The article by Joseph P. Ferrie (2005), which you referenced, contains data that straddles the US civil war (1860-65). This was a period of great social upheaval. By social upheaval I mean a break-down and rebirth of social structures (education, class, wealth, culture) and geographic relocation, and political changes. I don’t have access to the full article. Hence I don’t know whether Ferrie has treated the civil war as a possible cause of ‘social mobility’.
    The empirical literature on multinational firms (and possibly many other bodies of literature about which I don’t know) contains studies that estimate the decay rate of perceptions (stereotypes) about countries. The last time I looked at this literature was in the mid-1980s. At that time the overall conclusion was that the decay rate is very slow. In terms of this literature, it would not be surprising that recorded empirical episodes in the history of a country, such as ‘high social mobility’, linger on as unchecked beliefs for many decades.

    3. Because of greater income inequality in the US, mobility within the income distribution is lower in the US than elsewhere.

    Comment: You provide a reference to the book by Robert Goodin, et al (1999), The Real Worlds of Welfare Capitalism. I have only read the synopsis. Without knowing the conceptual framework and the empirical measurements, it is not possible for me to say very much. However, there is one item in the synopsis, which does not make sense to me. The authors categorise the German system as corporatist. Which time period do these authors consider? Nazi Germany had corporatist features (big corporations cooperated with the Nazis’ in several ways). I would not call the Federal Republic ‘corporatist’. I would say, starting with Erhard (“welfare for all�) in the 1950s, the post WWII Federal Republic has gone through versions of a ‘social democracy’. During Kohl’s (CDU) long chancellorship, the 1950s notion of social market economy (the term capitalism is not popularly used in Germany; capitalism is generally thought of as something different from a market economy) was expanded to ‘social and ecological market economy’. That is, explicit policies on environmental sustainability were formulated as a consequence of the pressure exerted by the Green party (as well as the environmental degradations – such as the river Rhine – that could hardly be ignored by anybody). Unless I am missing something really big, the foundation of the Federal Republic remains unchanged. For example, the current Chancellor, Dr Angela Merkel (CDU), intends to increase the top marginal income tax rate as one of several measures to reduce the government debt which accumulated since the unification of Germany in 1990. Her coalition partner, the SPD, is most unlikely to object to this measure.

    As to the causality aspect of your statement, I would suggest that without further detailed study, the argument would hold only at or close to the ‘survival constraint’ (roughly speaking, working the maximum number of hours that are physically sustainable and having exhausted borrowing possibilities). Comparisons of income inequality measures (at least those I know), do not contain sufficient information on the survival constraint.

    4. Wage inequality in the US has grown greatly since 1970. Income inequality has also grown, but not as much since low-wage households have increased hours worked.

    Comment: Your refer to a working paper by Krueger and Perri (2003). I suppose your argument is that, given the data in this paper, the income inequality would be greater if people in ‘low-wage’ households would not have increased the number of hours they work.

    The working paper by Krueger and Perri (2003) contains statistical details which raises the following questions in my mind:
    a) How sustainable is the apparent process of people in the bottom of the income distribution maintaining their consumption patterns by means of working longer hours? There is an absolute limit of 24 hours per day and, without requiring exact medical data, I can safely say that 24 hours per day is not sustainable (sleep, eating, washing, shopping). Without further data, the only general statement I can make is that the process of working longer hours as a means of maintaining expenditure is unsustainable.
    b) How useful are macro-economic models of ‘the economy’, characterized by a few aggregates on ‘employment’, ‘unemployment’, ‘GDP’ and one price ratio, interest rates on home loans? These numbers can behave very ‘nice’ without giving any indication of an imminent and sharp decline in say consumer spending
    c) Why do people work longer hours to maintain their consumption? Is it because they are close to a survival constraint (including borrowing constraint) and hence have no choice, or is it because they want to keep up with the Jones or is it a struggle against all odds to keep away from the survival constraint.
    Other: The results in this study are quite consistent with the data provided by the CIA on the USA economy (12% under the poverty line; all productivity gains during the recent past have gone to the top end of the income distribution).

    5. (Annual) Consumption inequality has not changed much since 1970. In my judgement, this reflects increased use of credit markets to smooth out short term fluctuations in income, which offsets increased long-run inequality .
    Comment: Without further data, I would be sceptical about the above judgement. The above judgement seems to rest on the permanent income hypothesis. Looking at it from a general equilibrium model perspective and taking item 4 above into account (real resource constraint), I can’t see how an ‘increased long-run inequality’ is ‘offset’. But I can see that an increase in consumption inequality over time is likely to occur. (I assume here that ‘consumption inequality’ is some measure of disposable income spent on consumption.)

    6. In my judgement, increased reliance on credit has been the main cause of the dramatic increase in bankruptcy, particularly evident since 1990.

    Comment: There is one thing for sure, bankruptcy presupposes debt.

    7. Bankruptcy laws act as a kind of income insurance, and generous (to debtors) bankruptcy laws are a substitute for redistributive taxation.

    Comment: I wouldn’t call bankruptcy laws a kind of income insurance. However, I would agree with the proposition that bankruptcy laws are an effective means of wealth redistribution. Bankruptcy laws differ between countries. Setting these differences in bankruptcy laws aside, I would concur that redistributive taxation (not only income but also wealth taxes) can result in a similar wealth redistribution as that which can be achieved by bankruptcy laws. I would be very interested in the parameter values involved because it seems to me the wealth redistribution achieved via taxation would have to be ‘huge’ to make up for the wealth redistribution achieved via bankruptcy laws (particularly in the corporate sector). But this is not all. While I am quite happy to treat various types of labour services as ‘commodities’ in very abstract theoretical models of economies that aim to check out whether philosophical ideas make any sense at all and, if so, how robust the results are to changes in the institutional environment, I would hope that nobody in the applied areas of economics would treat humans like commodities such as pebbles, used in the production of concreted surfaces of swimming pools or gold, used in the production of necklaces. I would imagine that the stress caused by living ‘on the edge’ financially is very great, although I don’t know whether people can become so brutally objective that they simply ‘play the game’ in the knowledge that they can’t influence the making of the rules of the game. Assuming this were the case, I wonder what the work ethics would be like. I find it difficult to believe that people could work up an interest in acquiring skills – however humble – and to be proud in what they are doing – however humble – if they know they are treated like pebbles or gold bars – a ‘commodity’. To the best of my knowledge, gold doesn’t care whether it contributes something to the production of a necklace or a wedding ring or just sits above or below ground. I am not convinced people don’t care whether or not they are productive. I am very confident in saying that at least some gold smiths take pride in their work. I am afraid, one can’t take these important differences between various types of ‘commodities’ into account in very abstract theoretical models – one kind of relies on humans not being silly (ignore all other information about humans, about societies, about politics, about medicine, etc, etc) when interpreting these models.

  16. Fred Argy
    December 10th, 2005 at 13:15 | #16

    John, I think it is important to distinguish between iner-generational and intra-generational social mobility and between the continental Europeans and the Scandinavians. There is very little literature on intra-generational mobility (apart from Goodin et al and some recent work by Andrew Leigh on Australia, US and Germany and HILDA on Australia) but most of it is fairly short range. On the other hand there is much good literature on levels and long range trends in inter-generational mobility and on this criterion US mobility is definitely lower than in Scandinavia and the gap seems to be widening, but the US is comparable to the continental Europeans. As for Australia, Andrew Leigh has some tentative estimates if you are interested. There are also differences between income mobility, occupational mobility and educational mobility which I have not fully sorted out.

  17. December 10th, 2005 at 20:11 | #17

    The “their land” idea hurt the Indians whether they were being deprived of land or allocated it according to western concepts. The American (not the UK) allotment movement aimed at solving their problems by giving them land. Unfortunately it backfired because without a transitional tutelage period during which they could have westernised (also culturally damaging, of course) they ended up losing much of their land since they couldn’t adapt to western ways of using real estate.

    Incidentally, this is what is wrong with much of the thinking behind aboriginal land rights, including a misunderstanding of what “terra nullius” was supposed to mean (I’m not referring to how Europeans used it, i.e. as suggesting that the land was up for grabs). But that’s another long story…

  18. December 11th, 2005 at 20:29 | #18

    As Fred mentioned, I’ve been working on 2 & 3 for Australia, and am happy to share rough drafts of papers with you if you’d like. Looking at social mobility or income mobility, the US looks a bit less mobile than Oz.

  19. December 11th, 2005 at 21:22 | #19

    Because of greater income inequality in the US, mobility within the income distribution is lower in the US than elsewhere.

    The work I did a few years ago with colleagues on short/medium-term mobility among families with children didn’t show a huge difference between the US and other countries – though where there were differences the US was indeed less mobile (www.unicef-icdc.org/publications/pdf/iwp78.pdf).

    Irrespective of the actual patterns, I’m particularly interested in the ‘because’ in the above quote. If there is more inequality, then a greater change in income is required to change one’s position on the income ladder – so this is a plausible hypothesis (see footnote 11 in our paper).
    However, one could also think of income generating processes where greater inequality implies an equally greater income change – and hence no relationship between inequality and mobility. For example one might think of economic institutions creating a structure of incomes, with individuals competing for places in this structure. In this case, swapping places with the person above you would lead to a greater income change in the country with the greater inequality.

    Has anyone been able to provide evidence for the ‘because’?

  20. Ernestine Gross
    December 12th, 2005 at 00:53 | #20

    Bruce Bradbury: Thank you for making available your paper.

  21. jquiggin
    December 12th, 2005 at 06:50 | #21

    Thanks to everyone who’s commented, and made work available. Andrew, if you could send me your draft papers that would be great, and I’ll try to reciprocate with comments.

  22. derrida derider
    December 12th, 2005 at 13:11 | #22

    I haven’t had time to read or comment on this yet, but commenters oughtta know that when John says “… there seem to be lots of contradictions in the data …” he’s talking about the US, but this problem is worse in Australia – so much so that IMHO anyone who tries to draw stories about poverty, etc from the ABS household income surveys does so at their peril. OTOH, the gap between public perceptions and what the science (such as it is) shows on these issues is probably not quite as massive in Australia as it is in the US.

  23. Graeme Bird says:
    January 29th, 2006 at 03:18 | #23

    I’d refer you to a book by the American Historian (not an economist) David Hackett Fischer. The book is “The Great Wave (Price Revolutions And the Rythyms Of History).

    It seems that the period between World War II and the 1970′s is a bit of an anomaly. In that its about the only time when income and wealth inequality has lessened concurrent with high monetary growth.

    If we want greater equality we have to get used to “Growth Deflation”.

    Its pretty easy to see how the currency being debauched the whole time allows the rich to get richer without a great deal of effort. For example if you started out with a net wealth $20 million in 1970 you would have had to have been pretty lazy or powerfully stupid not to have multiplied it many times over by now.

    This would not be the case under very hard money. What I would suggest is neutral monetary policy is NOT zero inflation. I would say it is where nominal GDP is rock solid. And prices fall all the time to reflect increases in real GDP growth. If this was maintained for some decades we can expect inequality to lessen. Supposing we have a government sector that is both minimal and is kind to people who are struggling.

    A smaller government has more chance to be kind to poorer workers since it has a greater ability to raise its funds primarily from people who hold greater wealth.

Comments are closed.