Wednesday, December 18, 2013

The underconsumption theory and unemployment

I strongly recommend the comment thread in this interfluidity post where Mark Sadowski does a good job clarifying the difference between wealth and income, and what the data says in terms of distribution and consumption in the context of an underconsumption theory of unemployment.
In underconsumption theory recessions and stagnation arise due to inadequate consumer demand relative to the production of new goods and services. Consequently it is precisely the income that is derived from the production of new goods and services that is connected to the matter of whether underconsumption is actually taking place. The income derived from the buying and selling of assets is completely peripheral given the very nature of underconsumption theory.
“About your IRS example, if you do not want to pay taxes all at once as a result of selling the asset outright, you can borrow to pay the lump sum tax payment and spread the payments out over time, just as you would have a time series of tax obligations by renting out the property.”
I never mentioned “taxes”. The only reason why I brought up “IRS income” is because that is the only context where capital gains are even referred to as “income”. Capital gains only relate to the buying and selling of assets and hence are not income in the conventional sense of that which is derived from the production of new goods and services. My point in making the comparison was that referring to capital gains as income will produce a measure of savings that is wildly inflated compared to the more normal measure of savings out of the income derived from the production of new goods and services, which is of course the measure of savings relevant in this context.
“Moreover, I hope that this can be internalized as a truism: no one receives NIPA income.”
Frankly, this is bordering on the ridiculous. Nearly everyone receives NIPA income, and given that the income derived from the production of new goods and services is the measure of income that is relevant to underconsumption theory, that is the measure of income that I think we obviously should try and stay focused on.
I think SRW's interpretation of inequality and growth has certain causal chains backward, but propensity to consume is an important factor to consider when thinking about what impact distribution may have on demand. If, when looking at income and not capital gains (as Mark convincingly argues), high income earners spend as much as lower income earners, then that's good empirical data to have. That said, I also think there is a budget constraint/wealth effect, but that may paradoxically have more of an impact on low wealth households. In this sense, what drives demand is not inequality per se, but the sensitivity of low wealth households to spend out of income as a function of their wealth -- including capital gains.

I'll need to read the original paper to see if this effect is there, as the comments are focused on the upper end where there turns out to be little to see.


Blogger JKH said...

Tour de force by Sadowski in those comments.

Such as:

"Moreover this is in keeping with the fact that capital gains “income” is purely a tax construct arising from an asset swap. Fundamentally, the capital gains tax is a wealth tax (not that there’s necessarily anything wrong with that), not an income tax, and this is precisely where the wildly exagerated estimates of the savings rates of those with wealth/high income are coming from."

Good because it gets the economics right in the context of proper macro accounting.

If we get get a million similar insights on that sort of relationship,, ecnomics might make some progress.

Also, this:

"You’re getting fixated on semantics. In order for the underconsumption thesis to have merit it must be shown that income derived from new production is not being spent. It does absolutely no good to focus on the fact that an antique painting somewhere may or may not have appreciated in value."

And this:

"Capital gains is a wealth tax that has caused confusion as to the difference between income and wealth."

We need more guys like this - with bullshit cutters - prowling the blogosphere, and snipping out the nonsense of discussions that are incompatible with basic accounting. Given that about 90 per cent of the discussion there focused on differences of "opinion" on this sort of thing, it shows how important it is.

3:58 AM  
Blogger Nick Edmonds said...

I missed the original exchange there and it's going to take me a bit of time to plough through it. However, it does appear that this is confusing an accounting issue with a behavioural relationship.

It is true that what we are ultimately concerned with is the ratio of consumer spending to GDP. However, when we want to look at what determines consumer spending, we don't look at GDP, we look at the resources of the household sector, which is not at all the same thing. As an obvious point, we adjust for taxes and benefit payments, even though these do not constitute GDP. In contemplating the savings ratio, we don't ignore the impact of tax cuts on the basis that they don't reflect a change in actual production (unless we're hardcore Ricardian Equivalence advocates). Likewise, if we ignore the potential impact of capital gains, on the same basis, we risk missing a trick.

This is particularly so, if we're looking at heterogeneous agents, which I believe was the whole context of the discussion. My capital gain might be nobody else's loss but it still changes the balance of spending power. If I'm trying to identify the impact of changes in distribution of resources, that matters.

But I'm going to have to go and read this exchange, I guess. I'm pretty interested in this issue.

7:25 AM  
Blogger JKH said...

Absolutely look at GDP plus capital gains of all sorts as the basis for a decent behavioral function. Just be sure to set up the behavioral function with consistent accounting time periodicity - so that accumulating non-GDP wealth has an effect on future period aggregate demand and GDP.

8:55 AM  
Blogger Nick Edmonds said...

The distinction between time periods is an important one, I think. My capital gains are real for me - they allow my spending to exceed my income over my lifetime, even though when we aggregate individuals within a period income and expenditure must be equal.

9:11 AM  
Blogger winterspeak said...

JKH: Agreed, Mark did a fantastic job. He also managed himself very well in the exchange and was consistently clear and non-personal.

I learned a lot, which is one of the highest compliments one can get, I think.

9:34 AM  
Blogger JKH said...


Godley and Lavoie do a superb job of integrating/ interfacing GDP accounting with capital gains accounting, developing a really solid behavioral modeling approach. But I think you are quite aware of that with your connection to Godley.


Right. And I didn't mean to infer anything negative about other comments there. I only looked at MS's comments, and he just keeps making points that need to be made.

9:59 AM  
Blogger Ramanan said...

In general I think - there is a denial that people with low incomes have a higher propensity to consume and rich have a low propensity.

Mark Sadowski says xyz is peripheral to the discussion but according to his own source, it matters.

This is because consumption is a function of both income and wealth and if you see it as a moving picture, people who accumulate a lot of wealth in some periods such as periods of rising stock markets and will consume more around those periods. Observations will compare it to income and report a higher propensity to consume out of income for them but from a consumption function (which depends on both income and wealth) viewpoint, the propensity has not changed and high consumption/income in one period is just because of the accumulation of wealth over several previous periods .

So people may have high capital gains etc and all this matters.

The arguments around these are clouded by the incorrect notion that capital gains should be counted in income.

I guess people who want to defend the Kaleckian notion of propensity to consume for rich vs poor seem to struggle in that thread but Sadowski is erroneous.

10:41 PM  
Blogger Ramanan said...

To illustrate this assume that Mr M's propensity to consume out of income is 0.6 and out of wealth is 0.1.

Let's say Mr M makes $1mn per year in income and his wealth at the end of 2012 was $5mn. In 2013, there is a massive rise in wealth because of gains in the stock market of about 20% ie $1mn.

In 2014 Mr M's consumption according to the consumption function will be

0.6 times $1mn plus 0.1 times $6m which is $1.2mn and this would be higher than his income of $1m.

So Mr. M's consumption/income is 1.2 - higher than poor people whose will be close to 1 but nonetheless this preserves the notion that rich's people's propensity to consume out of income is less than poor.

10:51 PM  
Blogger Ramanan said...

Okay the above calculation has some corrections due to things in 2013 itself, but I guess my point is still valid.

11:00 PM  
Blogger Ramanan said...

Also, I think the data says that poor people's propensity to consume is higher than the rich and shouldn't be surprising since the live in subsistence.

"It’s true that at any given point in time the rich have much higher savings rates than the poor."

h/t Steve Roth's post at Asymptosis.

but in the next sentence Krugman says it is a statistical illusion!!

I'll go further and say that economic studies obfuscate this basic fact.

Man never trust economists!

12:05 AM  
Blogger Ramanan said...

This comment has been removed by the author.

12:46 AM  
Blogger Ramanan said...

In summary, my comments was to suggest that Mark Sadowski's comments are quite wrong. The fact that those arguing against him seem to struggle do not make Sadowski's analysis correct.

His analysis is normative disguised as positive analysis.

The fact that in some periods, rich people's consumption is high doesn't mean that it will be high everytime, in fact it is lower in the sense meant by underconsumptionists.

This debate has a policy background. Those who argue that distribution of income matters argue that saying if there is better distribution of income, it is good for the world economy as a whole because the poor will consume more and via the multiplier effect will lead to higher output. Sadowski's comments seem to constantly deny this basic notion.

1:05 AM  
Blogger Greg said...

Sadowskis arguments feel wrong, but they sound right. He's a smooth talker and has lots of charts. Ram is right though no one seems to know exactly how to puncture them yet.

One of the things that has occurred to me in much of that discussion is that the underconsumption theory, by allowing itself to be reworded as 'The rich aren't consuming enough" has made itself easily assailable. Look around us, does anyone really think the rich aren't consuming enough? They are consuming all that they want and in most instances even more than they need.

Additionally they are doing most of this consumption without taking on debt. Its the rest of us that have had to use credit cards/home equity loans to consume at the level which the 2000s became used to. When the rest maxed out we saw credit markets slow and near ZIRP become standard fare.

We don't need the rich to consume more. I think we need them to pay more to their employees, stop the incessant whining about their own taxes while pushing for more consumption taxes and stop simply trying to extract "rents" from all our income streams but actually invest in new production.

8:29 AM  
Blogger Nick Edmonds said...

I think Sadowski actually makes some good points and also highlights the issue that the question of whether the rich save more is not equivalent to the question of whether a more equal distribution of income will reduce the aggregate savings ratio.

However, there do seem to me to be points of confusion there, such as equating the consolidated capital gains of the national economy with gains in the hands of households.

8:49 AM  
Blogger Ramanan said...


Right. Good point about rewording.

Also I think his point about convexity/concavity is another deceit.

In fact after reading the comments the Robinson quote "The purpose of studying economics is not to acquire a set of ready-made answers to economic questions, but to learn how to avoid being deceived by economists." became more firm!

9:28 AM  
Blogger Ramanan said...

Blogged on this here:

4:01 PM  
Blogger Ramanan said...

Nick also wrote on this:

8:55 AM  
Anonymous Anonymous said...

I've copied this comment that I put on Nick's post in case any of you can put me straight on it:

I think Mark A Sadowski had a point that income and saving need to be precisely defined and I guess I was being exasperatingly woolly about that but I also think he was failing to see a real phenomenon that is occurring.

I think we really need to get a good definition of the underconsumption hypothesis that encompasses what we mean.

To me the underconsumption hypothesis means that that there is less production and consumption than there would be if global wealth were equally divided up amongst all seven billion people on earth. The crazy thing is that Mark A Sadowski actually said that he agreed that that would cause more consumption at least in the short run. Perhaps the real argument is simply about what is meant by the term, "underconsumption hypothesis".

To be honest, I'd be delighted with everyone agreeing that some "classic underconsumption hypothesis" (that I'm unfamiliar with) is bogus whilst the "newfangled underconsumption hypothesis" (as defined above) is a real phenomenon holding us back.

IMO Mark A Sadowski was doing more than policing sloppy definitions. My impression was that he was challenging the whole idea that inequality is causing a less productive economy. He was basically trying to make the case that inequality is only a problem to the extent that it is "unfair" rather than it being something that will decide whether the economy evolves over coming decades to provide for future advancement and food, water and energy security etc. To me that is the crucial issue.

2:16 PM  
Blogger winterspeak said...

Ramanan: Nick's post on this was a good one.

Nevertheless, I think the under consumption theory is poorly supported and makes a weak case for redistribution. There are better. There are also better policy options for increasing AD in general, and tilting the scales in favor of low income households. To the degree that the under consumption discussion takes the focus away from these, it may be doing more harm than good on net.

11:09 AM  
Blogger Ramanan said...


Let me caricature your argument: Income does not include capital gains in national accounts and hence underconsumption doesn't hold.

As admitted, that is a caricature, but then what is the argument?

Of course fiscal expansion can expand demand and output but it doesn't mean that it is the only way to do it. For example, workers if paid higher may feel they are contributing more and be more productive and this improves the supply side.

I am not sure how you hold the belief that capitalist economies are demand constrained on one hand but then claim that the distribution of income is sort of not relevant to aggregate demand. Your two views are inconsistent.

I agree with you that the discussion may somewhat shift the focus from fiscal expansion but it's not that it is not important.

2:46 PM  
Blogger Ramanan said...

And these things have important implications when one considers complications due to open economies.

For you, I assume balance of payments doesn't matter, so it is as simple as fiscal expansion.

But as long as the external sector matters (which it really matters!), low wages is a race to the bottom because nations feel pressured to improve their external sector and may want to keep wages low and this is deflationary for the world as a whole.

As Robert Blecker says:

A situation in which competitive wage cuts (or ‘wage restraints’) are pursued in all countries will potentially harm the interests of workers everywhere: real wages will be sacrificed, as long as mark-ups are flexible; but employment will not increase, as long as the competitive gains cancel each other out. In this case, the regressive effect of multilateral wage cuts on income distribution could well lead to a world-wide depression of demand and employment. On the one hand, if workers in all countries increase their money wages, and if the international competitive effects roughly cancel out, then the world economy as a whole can potentially enjoy wage-led growth – provided that firms still feel sufficient competitive pressures to compel them to cut their mark-ups in response to the wage increases

3:07 PM  
Blogger winterspeak said...


First, I don't have any strong beliefs about capitalist economies being demand constrained.

Second, I think aggregate demand in the context of income distribution is an unhelpful way to think about both aggregate demand and income distribution.

Third, I think balance of payments matter very much, and in some ways more than fiscal expansion.

5:18 PM  
Anonymous Anonymous said...

Winterspeak, I've also come around to realizing that it could be just a distracting straw man argument when the "underconsumption hypothesis" is framed to mean that recessions result from underconsumption by the rich.
Its dawning on me that the way to frame the issue might be in terms of real GDP being below what it would be if no one had any financial constraints at all. I think Mark A Sadowski sort of eluded to that when he brought in the whole NAIRU constraint subject. BUT as I think he conceeded, NAIRU is totally dependent on distribution. If all of the money is with a handful of oligarchs who only have need for staff amounting to 1% of the population but have very select needs (eg multilingual expert butlers, clockwork watch craftsmen etc) then the NAIRU could be >90% and GDP could be <10% of potential even with unemployment being less than the NAIRU.
Perhaps the argument against inequality could even be distilled down to 'The NAIRU is higher when wealth inequality is greater' .
There is also the political economy dimension since if a few people have a mountain of government bonds, then they will fight for ensuring that all policy has inflation busting as the top priority

9:07 AM  
Anonymous Anonymous said...

It is also worth pointing out that Steve R Waldman never actually mentioned under-consumption once in his post. Mark A Sadowski however said that Steve's overall thesis about inequality limiting growth was rubbish because the under-consumption hypothesis was disproved by the facts.
I think the case needs to be made that the phenomenon of inequality limiting growth is not contingent on the Mark A Sadowski style straw man under-consumption hypothesis.

9:36 AM  
Blogger Ramanan said...


Private consumption is one component of demand and if low income earners have higher propensity to consume, it matters to aggregate demand.

So I don't know what you mean by "unhelpful".

11:02 AM  
Blogger winterspeak said...

NAIRU is dependent on a great deal more than distribution. But if all you have is a hammer...

1:59 PM  
Anonymous Anonymous said...

Winterspeak, you are right that I think distribution is a lot more important than most people think it is. BUT I'm interested and willing to learn, So please give a snippet as to what you think is more/most important for determining NAIRU.

By the way, I think it is bogus to say that high NAIRU is about lack of education. To me the most stunning thing is that after the soviet union collapsed, the former soviet countries had one of the most highly educated populations there has ever been. Still didn't stop mass under employment and unemployment.

2:15 PM  
Blogger winterspeak said...

I agree, I don't think high NAIRU is about lack of education.

I think that NAIRU is fine as a way to build models built to have a hard and fast tradeoff between inflation and unemployment. But that begs the question.

I think you can have high inflation and high or low unemployment, just as I think you can have low inflation and high or low unemployment.

Both inflation and unemployment are aggregates made up of lots of different factors, no reason everything has to pull in the same direction.

Two additional thoughts you may find interesting. First, if AD is too low so redistributing away from high earners is the cure, if AD is too high should we take from the poor and give to the rich to fix that? Second, if the rich aren't consuming, doesn't that leave more goods and services available for the poor? If the rich start buying, don't we just get gentrification?

Not advocating any of the above. Just some ideas to consider.

Merry Christmas!

10:47 PM  
Anonymous Anonymous said...

Winterspeak, Merry Christmas to you too!

I completely agree that it is very dodgy to see inflation as something that could lower unemployment. Zimbabwe shows you can have >1000% inflation and 80% unemployment.
I also think the NAIRU needs to be seen as something that needs to be kept down to zero. Basically having a NAIRU greater than zero to me is symptomatic of an administrative failure. Inequality is part of that IMO.
I'm not sure that I can picture the scenario you give of AD being too high in a normal peacetime economy. I suppose Norway makes an effort to continuously run budget surpluses so as to mop up oil money. To be honest though, I really don't see why they want the state to be buying foreign stocks with the surplus rather than leaving the population to deal with the excess oil money. It seems a bit like paternalism to me but it is their country and up to them of course.
The other example I can think of with AD demand being too high was in the UK in WWII and for the rest of the 1940s. BUT that was only because of the disruption due to the War. The way that was dealt with was by rationing. In such circumstances I think rationing is the only option. The choice is basically between rationing and total collapse perhaps with hyperinflation. In India in WWII the authorities (British) failed to introduce rationing in similar circumstances and the result was price spikes, food hoarding and the Bengal famine of 1943 with several million dead.
Basically as I see it AD being too high is only a problem if their is an insolvable pinch point in the supply of essentials and in such circumstances it is an economic breakdown and rationing is the only viable option.

12:04 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home