The grain of truth in the Junker Fallacy

I saw an example of the Junker Fallacy recently (it doesn't matter where). You have probably seen it too somewhere.

"Investment is low because people (or firms) spend all their savings on speculation rather than investment -- bidding up the price of houses, shares, land, or in corporate takeovers; there is no money left over for real investment that could increase future productivity".

It is indeed a fallacy, in that the argument makes no logical sense. It's an example of a fallacy of composition: what is true for each individual is not true for the whole. But there can be a grain of truth in it, in some circumstances, if it's re-stated correctly.

Start with some standard (to economists) definitions. "Income" means income from the production of new goods and services (it excludes capital gains for example). "Consumption" means spending on newly-produced consumer goods and services. "Investment" means spending on newly-produced capital goods and services (think buying a new machine, not buying stocks and bonds). "Savings" means that part of income that doesn't go either to taxes or to consumption (yep, it's a residual, and can mean almost anything you do with your income, except taxes or consumption). These are all flows, per year.

Yes, those definitions are very different from some ordinary meanings. But it's not just economists being prissy, ornery, or "special". They help us keep our heads straight in thinking about things like the Junker Fallacy. (But they also maybe stop us seeing the grain of truth in the Junker Fallacy).

Assume a closed economy. Ignore government spending and taxes (or else include government savings and investment spending along with private).

National Income Accounting identities divide all production and expenditure into consumption and investment goods, and divide income into consumption and saving, so we get: C+I=Y=C+S, and rearranging gives us S=I.

That accounting identity, S=I, savings equals investment, means that there is something terribly wrong with the Junker Fallacy, when seen through the semantic lens of economists' definitions. How is it even logically possible for part of savings to be diverted away from investment, whether into land, stocks, or whatever, when savings and investment must be the same? The Junker Fallacy must indeed be a fallacy. It's logical nonsense!

The Junker Fallacy seems to make sense at the individual level. I can spend my savings on newly-produced real investment goods (like an education or a newly-built house), or lend it to a firm to spend it on newly-produced real investment goods (like a machine), or I can divert my savings away from investment to buy land, old houses, existing shares, existing commercial bonds, or whatever. But if I buy land, someone else must have sold me the land. We can't all spend our savings on buying more land. In total, we can't spend any of our savings on buying more land. They aren't producing it any more (OK, except maybe in Holland etc., where producing new land really is real investment). And it's the same with all the old houses, old shares, old bonds, etc.

At root, the Junker Fallacy is a fallacy of composition. What's both possible and true for the individual (buying old stuff diverts my savings away from real investment) isn't necessarily even possible, let alone true, in aggregate.

The reasoning behind the Junker Fallacy is logically invalid.

But that invalid reasoning doesn't necessarily make the conclusion false. And accounting identities like S=I can't tell us how the world works. Let me restate the Junker Fallacy so that it might be true.

Suppose people have a target level of desired wealth W*. If their actual wealth is less than W* they save part of their income to increase their wealth; if it's less than W* they dissave. W* is defined in real (inflation-adjusted) terms, deflated by the CPI. (People aim for a target level of wealth that can provide sufficient consumption for their retirement or temporary loss of income, for example).

Wealth consists of capital plus land. Suppose we are initially at the target level of wealth, so W=W*, and savings is zero. We are in equilibrium. Now suppose people decide they want to hold a higher proportion of their wealth in land, relative to capital. Each individual tries to buy more land. They fail, of course, because they can't all buy more of a fixed stock of land. But in trying to buy more land, they bid up the price of land.

A rise in the (real) price of land (relative to the CPI) means there is less room in target wealth for capital. In the short run the price of capital falls (as people try to sell capital to buy land). In the long run the real stock of capital falls, as existing machines wear out, and are not replaced, since the price of machinery is below replacement cost.

A desire by individuals to hold a greater proportion of their stock of wealth in the form of land has displaced real investment in capital.

Economists' definitions of income and savings can sometimes be misleading. The rise in (real) land prices does not constitute income, since it does not come from production of new goods; it's just a capital gain. So savings go negative as production is diverted away from producing capital goods into producing consumption goods instead. But from the perspective of an individual, savings does not go negative, since the capital gain on land counts as income, and indeed their real wealth stays the same as the price of land rises, so they can't be consuming their wealth.

It all depends on the consumption (or savings) function. If people have some sort of target level of wealth, there is a grain of truth in the Junker Fallacy. If instead they have a target ratio of present to future consumption -- an Euler equation that determines savings by comparing the marginal utility of present and future consumption, over a long horizon, the Junker Fallacy is indeed a fallacy (because if you reduce the capital stock future production and hence consumption will eventually fall).

[Junkers were Prussian landowners. Austrian economist Fritz Machlup discussed what he saw as a fallacy by those who argued that the Junkers' "investment" in land is what caused Prussia to have low investment in capital. H/T Bryan Caplan and Tyler Cowen.]

Addendum: It works for land. It might partially work for houses, which are both land (the lot) plus capital (the bricks). I don't think it works for shares, since who cares whether they own new or old shares? I will maybe explore just how far I can push this argument in a second post, when I come back from a few days at the lake.

Comments

[Luis Enrique]

this is tantalizing, because it comes close to a half-formed idea I've had about the consequences of fixed-supply asset price inflation, but I'm still not sure it works. How reliant are you on this mysterious preference toward land? Why would this sudden shift in preferences regarding the composition of wealth occur, unless something underlying was happening to change expected returns from land versus capital? If the prices of produced output aren't affected by this preference shift, and neither is total expenditure on produced goods - nor production technologies - then the real returns to owning capital aren't going to move, so why would the price of capital fall? And what's going on with the returns to holding land? Unless this shift in preference for land over capital takes the form of suddenly requiring higher returns from capital and tolerating lower returns from land ... but if you're going to suppose a sudden increase in the required returns from capital, then don't you trivially get a reduction in the quantity of installed capital?

[Don the libertarian Democrat]

"A desire by individuals to hold a greater proportion of their stock of wealth in the form of land has displaced real investment in capital."

There seem to be a lot of such conflicts or divergences between Individuals and the Greater Society. One I've come up against concerns CDSs. From the point of view of a lender on a mortgage, buying a CDS can make sense as a form of Insurance. It lessens the risk that the lender will lose a large percentage of his loan in the case of default.

On the other hand, if every lender buys one, and there are a large number of defaults, or even possible defaults, the system can explode, since the insurers can't pay everybody. Consequently, I've had people tell me that we don't want individual lenders decreasing their risk, since that entails the greater society taking on more risk. We want individual lenders to accept a lot of risk. Of course, that could drive interest rates higher and cause a lot less lending, or increased risk for the borrower, but that's still better than increasing risk for the general economy.

I think that's one reason that Gorton wants the Govt to basically guarantee everything. We don't want individual lenders increasing their risk.

In your case, I suppose we should ask if we want the govt to discourage the investment in land, and encourage real investment in capital.

[Min]

I have a different take on this, at least, if we do not restrict speculation to the examples you give, and if we interpret the "fallacy" in the vernacular, rather than in economic jargon. (Something is not a fallacy if it crucially depends upon interpretation.)

"Investment is low because people (or firms) spend all their savings on speculation rather than investment."

Of course, "all" is an overstatement, but it is normally understood as such. In the vernacular, "savings" is a very broad term. Basically it means money that you put away for later use. It does not have to be money, it could be something like a comic book collection that could be sold (or borrowed against) if need be. A lot of people consider the equity in their house as a form of savings. The term can also include potential savings, in this case, money that they would ordinarily save if they did not speculate with it.

What is speculation? It is risking money in hopes of getting more money back. In a word, it is gambling.

Suppose that everybody took all their extra cash, emptied their bank accounts, borrowed against their comic books and houses, and sat down to perpetual games of Fan Tan. There would be no investment, as all the money would be in The Game. Isn't this the prototypical idea behind the argument? Instead of expending time, energy, and financial resources in productive activities, people are making bets in the attempt to get rich quick.

Now, in real life, as you say, if you buy a house as a speculation, somebody else gets the money, and they do not necessarily speculate with it. But in the prototype of The Game, they do. Speculation is money chasing money.

Now where do we put speculation in terms of National Income Accounting? I suppose that it falls under consumption. The service being consumed is The Game, I guess. (BTW, what does putting money under the mattress count as?)

Certainly this interpretation does not seem fallacious. There indeed have been places and times when far too many human and material resources have been devoted to risky, unproductive behavior. :)

[Bill Woolsey]

I don't agree with your "correction" with present and future consumption. What is it each individual intends to do with this wealth they are "targetting?" As they earn capital gains on land or stock or whatever, they expect to be able to consume in the future out of that wealth, and they consume more now, save less, and this results in less capital accumulation.

Your Euler equation argument amounts to assuming that everyone knows that reduced capital accumulation will reduce their future income and so they will to save even more to maintain future consumption.

In other words, you are assuming omnisicent representative agents.

Exactly how different rates of investment will impact the growth path of real income in 20 years is not obvious except to someone working a simple model.

I am not sure about "junkers," and consider higher stock prices to be ambiguous (as you suggest, selling new stock can fund investment) but investment in gold bars and diamonds as opposed to capital accumulation is one of the problems with threats of expropriation.

Also, think about the pigou effect, and how that works when it isn't base money being hoarded but rather diamonds or gold, when there is no gold standard. It also applies to negative nominal interest rates. (What, negative nominal interest rates... I'll just buy gold... well, how does that impact the economy?)

[Nick Rowe]

Luis Enrique: I haven't quite gotten my head clear on it yet (my head's split between university admin and cottage), but I think elasticity of supply is a crucial ingredient. If land could be newly-produced, with an elastic supply schedule, then land and machinery would just be different forms of produced inputs, i.e. capital. And a switch in preferences away from owning machines towards owning land would change the composition of capital, but not the total amount held (for a given desired stock of wealth W*).

The shift in preferences was a deus ex machina, just so I could illustrate the argument most simply. Perhaps something else could substitute for the shift in preferences. Optimism (rational or not) about future land prices, a bubble, maybe? Dunno.

"If the prices of produced output aren't affected by this preference shift, and neither is total expenditure on produced goods - nor production technologies - then the real returns to owning capital aren't going to move, so why would the price of capital fall?"

NO! That's the conclusion you get out of the simple one-good Neoclassical growth model, where the capital and consumption good are identical (or have infinite cross-price elasticity of supply). This was one of the mistakes (from the US side in this case) underlying the old Cambridge(UK) vs Cambridge(US) capital controversy. (Time) preferences matter. A change in the rate of time preference for example will change the real (relative) price of capital goods in terms of consumption goods. In the very short run, when the stock of capital goods is fixed, the real price of capital goods (and the natural rate of interest) is a purely subjective phenomenon (the Austrians had this figured long before). In terms of the Irving Fisher diagram, the intertemporal PPF is not a straight line. So technology does not determine the real price of capital. Preferences (demand) matters too.

"... but if you're going to suppose a sudden increase in the required returns from capital, then don't you trivially get a reduction in the quantity of installed capital?"

I think I'm (implicitly) imposing an increase in the required *relative* return to capital (relative to land). Since the desired stock of wealth doesn't change. Not sure. My head can't handle it right now.

Don: "In your case, I suppose we should ask if we want the govt to discourage the investment in land, and encourage real investment in capital."

You can ask that question, but the answer is not obviously "yes". There is an optimal quantity of investment, and it's not 100% of GDP. You can build models in which there's too much capital and investment. ("Dynamically inefficient" equilibria in overlapping generations models for example, in which everyone can be made better off by the introduction of land, or even a useless asset, which serves as a stable Ponzi scheme because the rate of interest is less than the growth rate of GDP. This was the original theoretical foundation for US Social Security(?) We call it "The Canada Pension Plan" here.)

Min: The Game. There are winners and losers, but in aggregate nobody wins or loses anything in games of chance (except for lost labour time etc.). There is no aggregate wealth in The Game. So I think that "everybody invests in gambling instead of capital" is an example of the Junker Fallacy that does not have a grain of truth. (Though it does seem to be a waste of willingness to undertake risk, which is a different, but interesting, story). The Game does not appear anywhere in National Income Accounting (except the croupiers' salaries and the bricks and mortar of the casino).

Bill: I hope I'm not violating the assumption that agents play Nash (which I think is what you're getting at). They only care about their own future income. But if each agent owns the same amount of land (even if the real price has increased) and less capital, their income from land rents stays the same, but their income from capital rents falls, so their future investment income falls. If they have an infinite horizon Euler equation approach to consumption/savings, they care about their future investment income, and don't care about their current Wealth. If they have a short horizon "build up a big enough stock of wealth so it will see me through my next retirement years when I spend it down" approach, they do care about W, and care little about future investment income.

There used to be a big literature on "money and growth", in the 1960's and 70's. Did higher inflation, and lower real M/P, cause higher investment? That whole literature got wiped out by Sidrauski, who introduced an Euler equation (or similar) infinite horizon approach to consumption/savings, and got superneutrality. But OLG models tell a different story. I think that people hording chests of gold coins for their retirement pension *did* mean a lower capital stock. Back to "Banks, Money, and Debt", only with a target level of wealth, plus a desire for liquidity. Banks, when they do their job right, make capital as good as money. So we hold more capital. That's the history of financial civilisation.

Sorry. I'm rambling.

[Too Much Fed]

"It's an example of a fallacy of composition: what is true for each individual is not true for the whole."

How about what is true for the many may not be true for the very few?

[Too Much Fed]

"Investment is low because people (or firms) spend all their savings on speculation rather than investment -- bidding up the price of houses, shares, land, or in corporate takeovers; there is no money left over for real investment that could increase future productivity".

How about investment by the few is low to increase supply (degree matters) while investment by the few is high to reduce labor/wage income?

[Too Much Fed]

I don't put much faith in National Income Accounting because of excluding capital gains and excluding used goods (and probably other exceptions).

By National Income Accounting, shouldn't the Citigroup energy trader, who's whining about his 100 million dollars, be in debt to everyone else who makes stuff? INSTEAD and out in the real world, this person has excessive wage income and probably is trying to get everyone else into debt (demoninated in currency and probably bank loans) to him.

[reason]

Your on to something here - very good post. If we add in the wealth effect on consumption and financing (total debt) perhaps we can start to explain reality.

[Too Much Fed]

"That accounting identity, S=I, savings equals investment, ..."

How about trying to break down savings and investment into three categories; the rich, the gov't, and the lower and middle class?

[Too Much Fed]

That accounting identity, S=I, savings equals investment, ..."

How about trying to break down savings and investment into three categories; the rich, the gov't, and the lower and middle class?

Sorry if this is a repeat.

[Too Much Fed]

For Nick on debt and inequality:

http://rortybomb.wordpress.com/2009/08/04/debt-and-inequality/

Here is one quote: "Now remember our original assumptions: Income inequality is large. Consumption inequality is smaller. So savings inequality must be huge."

And,

http://baselinescenario.com/2009/08/18/united-states-inequality-in-the-recovey-period/

"The report hammers home what you might already suspect: The consumer debt problem in the economy really is a debt problem for the middle class. The need to work off a chunk of that debt will sap middle-class families’ spending power for perhaps years to come.

By contrast, the upper 10% of income earners face a much smaller debt burden relative to income and net worth. Those people should have ample spending power to help fuel an economic recovery."

My comment: The rich have ample spending power but do NOT need to spend. They are excess savers.

[Luis Enrique]

Nick,

I think I can understand the role of time preferences - after all, buying capital can be thought of as buying future consumption, so if your preferences for current versus future consumption change, then it's natural to think the price of capital versus (current) consumption changes .... I'm surprised that has anything to do with Cambridge (UK), it makes sense from within my narrow and (poorly grasped) picture of a basic (Ramsey model) neoclassical world.

"In the very short run, when the stock of capital goods is fixed, the real price of capital goods (and the natural rate of interest) is a purely subjective phenomenon (the Austrians had this figured long before). In terms of the Irving Fisher diagram, the intertemporal PPF is not a straight line. So technology does not determine the real price of capital. Preferences (demand) matters too."

I'm not sure I really understand this ... any time I ask my professors about capital controversies, they advise me to invest my time elsewhere .... can you recommend some easy, clear expositions of what you are writing about here?

[Patrick]

"By National Income Accounting, shouldn't the Citigroup energy trader, who's whining about his 100 million dollars, be in debt to everyone else who makes stuff"

National income accounting isn't applicable to individuals.

[Too Much Fed]

"National income accounting isn't applicable to individuals."

If true, I would call that another flaw for national income accounting. Is that an "almost perfect" economic model to ignore wealth/income inequality?

[Nick Rowe]

reason: thanks! If people save towards a target level of wealth, then there must already be a wealth effect on consumption: if W drops to less than W*, then cut consumption and increase savings, to restore wealth towards W*.

Too much Fed @2.44: If the inequality is a result of fluctuations in transitory income, that would certainly explain differences in saving/dissaving, and debt. But the inequality we normally care about, for those who want more govt. redistribution, etc., is more inequality of permanent income.

Luis enrique:
There ought be be a good simple book on capital theory, but I

Sorry Tstorm at cottage.

[Too Much Fed]

Nick, can you give me your definitions of transitory and permanent income, including but not limited to whether it means wage income?

[Too Much Fed]

I believe that I am referring to permanent wage income, and I think the 2 articles are referring to permanent wage income.

[Nick Rowe]

There was a magnificent thunderstorm here at the Lake. Had to shut down the computer, for fear lightning would damage it.

To continue, in answer to Luis Enrique:
I don't know of any good simple books on the Cambridge capital controversy. Your professors were right in advising you against wasting time on it. Too much "he said, she said". But reading something on capital theory could be worthwhile. I can't remember any good book to recommend. Michael Bliss (?) Capital Theory(?)?

The best simple model of capital in the Irving Fisher diagram.

Start with a standard graph of a curved PPF with apples on one axis and bananas on the other. Add an indifference curve, tangent to the PPF. Draw in a budget line tangent to both. The point where all 3 kiss is the equilibrium. The slope of the budget line is the relative price of the 2 goods.

Now, just re-label the axes. Instead of apples and bananas, it's consumption of apples today and consumption of apples in the future. By reducing consumption today, and investing, we can produce and consume more apples in the future, moving along the PPF. The slope of the budget line is 1+real rate of interest. The diagram shows that the rate of interest, and investment, are co-determined by technology (PPF) and preferences (I curve).

The only thing wrong with the diagram is that it only has 2 dimensions, so only can handle 2 years.

Take two extreme models of capital:

1. The "Schmoo" model. Schmoo is a plant that grows at rate r. K is the stock of Schmoo. Annual output is rK. You can either eat it, or let it continue to grow. rK=C+I I=deltaK. The rate of interest is determine by technology (biology). Straight-line PPF. One good model.

2. At the other limiting extreme, the transformation of consumption into capital or vice versa becomes impossible. Capital is like land. You have a fixed stock of land, and it grows apples. The PPF is like an upside down L. Preferences determine the rate of interest.

The real world is somewhere between those 2 extremes. In the long run it's more like 1, and in the short run more like 2.

Too much Fed:
Loosely speaking, your "permanent income" is your average expected level of income over your lifetime. "Transitory income" in any year (which can be positive or negative) is the difference between your actual income in any year and your permanent income. There's no distinction between wage and other sources of income.

Take my model of red and green apples growers. Reds and greens have the same income on average, but in even years reds are temporarily rich and greens temporarily poor, and in odd years vice versa. They borrow to offset negative transitory income, and lend to offset positive transitory income.
That's rational, so you can smooth out your consumption over time.

[reason]

Nick,
yes your point is correct, but surely you need a "permanent wealth" concept as well as a "permanent income" concept, since asset prices are volatile.

[Too Much Fed]

More on permanent wage income, saving/dissaving, and debt denominated in currency later.

[h.e.]

Luis,

Cohen and Hardcourt wrote a straightforward introduction to the Cambridge Controversies a few years back, which is available here:

http://econ.yorku.ca/~avicohen/Linked_Documents/JEP_Cohen_Harcourt.pdf

For a related paper on the difficulties of aggregation, which is related to the conceptual difficulties of "capital" as an analytic construct, see Felipe and Fisher:

http://riscd2.eco.ub.es/~josepgon/documents/Felipe_Fisher.pdf

[Nick Rowe]

reason: interpreted broadly, so that wealth includes human capital, wealth and permanent income are very closely related. If r is the rate of interest, W wealth, and Yperm permanent income, then (for an infinitely-lived person): Yperm=rW (and you just need to include a life-annuity factor for finite-lived agents).

But this misses your point about transitory fluctuations in wealth. I need to think about it some more.

[Too Much Fed]

Nick, you need to break your "apple income" down into at least corporate income and wage income.

What does the model look like if productivity and/or cheap labor shift the supply curve for apples to the right, the fed adds bank reserves to shift the supply curve of loanable funds so that interest rates come down, the workers have a price inflationary attitude of 2% per year for apples and overestimate their permanent wage income (expectations are too high) so that they borrow using bank loans denominated in currency (NOT APPLES) based on that overly optimistic permanent wage income leading to the demand curve for apples being shifted to the right?

In the end, does the "apple corporation" get operating margin growth, quantity growth, and price inflation?

Do the workers get negative real wage growth and more bank loans denominated in currency, and the "apple corporation" gets positive real corporate income and possibly excess savings?

Does that scenario lead to more wealth/income inequality?

[Luis Enrique]

thanks Nick, h.e/

[Too Much Fed]

About the savings rate and wealth/income inequality, see this link:

http://www.nakedcapitalism.com/2009/08/guest-post-the-savings-rate-has-recoveredif-you-ignore-the-bottom-99.html

"Fortunately, there IS some pretty good data on income stratification in the United States, and a few assumptions can help shed some light. Economists Thomas Piketty and Emmanuel Saez have made careers of studying US income inequality using IRS data, which goes back to 1913. The most recent data available (for 2007) showed that the top 14,988 households (0.01% of the population) received 6.04% of income, the highest figure for any year since the data became available. The top 1% of households received 23.5% of income (the second highest on record, after 1928), while the top 10% received 49.7% of income (the highest on record)."

And, "If we expand our survey to the top 1% of all households, we find an average income of $1.36 million for 2007. These folks had an average federal tax burden of just under 33%, so their after tax income averaged $916 thousand. If you assume this group had a savings rate of 33%, you get total savings of $452 billion (remember, $171.5 bn of this comes from the top 0.01%, we’re assuming a savings rate of around 25% of after tax income for the “poorer” 99% of the top 1%) This is more than 100% of the personal savings of the entire population, according to the BEA data. It implies that 99% of the US population still has, on average, a negative savings rate of around 1.3%. If you subtract the next nine percent, which likely still has a positive savings rate, the data for the bottom 90% becomes even more depressing, implying a negative savings rate of close to 5%."

[Too Much Fed]

International trade, offshoring, and US wages

http://www.voxeu.org/index.php?q=node/3920

Begins with:

"This column revisits the heated debate over international trade, offshoring, and US wages using new data. It says that increased international exchange with low-income countries has depressed US wages. That effect only arose during the 1990s, suggesting a different conclusion about trade, offshoring, and income inequality than the previous round of debate."

Hello nafta and china into the wto?

[Too Much Fed]

More on cheap labor:

http://www.voxeu.org/index.php?q=node/261

"But along has come China, which is far more labour-abundant now than the NIEs were then. A simple indicator is relative wage rates: in 1990, according to the US Bureau of Labor Statistics, the original four Asian NIEs had hourly compensation costs that were 25% of the US level. Now the BLS estimates that China’s labour costs are only 3% of US levels."