Imagine the following question on a PhD comprehensive exam:
"Using a macroeconomic model with monopolistically competitive firms, explain how an increase in the expected future price level will cause an increase in the current price level. Also explain whether there is an effect on real output.
Your answer must use words only, with no diagrams or equations. Be very precise about all the mechanisms that would be involved in this interdependent system of simultaneous causation. Your answer must assume no previous knowledge of economic theory or familiarity with economic concepts on the part of the reader. Try to make your answer as realistic as possible, using 10 real-world goods as examples. These should be goods that a homeowner with liquid domestic currency assets living in Sao Paulo Brazil in 1979 might want to buy in response to an increase in the expected future price level. Any transactions in your explanation must be shown to be consistent with double-entry bookkeeping. Please write clearly.
You have 2 hours to answer this question.
Your answer will be graded by a committee composed of: an accountant; a businessman; a Russian day-trader; an economics professor; and anyone else who happens to wander by. You must satisfy all these examiners in order to pass the exam, including follow-up questions in the oral.
Your answer, and the comments of your examiners, will be a permanent part of the public record."
I was foolish enough to try something like this (though I cheated a lot). (Sorry Winterspeak, but when I thought about what I was trying to do, it did seem rather funny, and I couldn't resist posting this, though the joke's really on me.)
Worthwhile Canadian Initiative 
simultaneous comments on the same question!
BTW, here’s something that may be interesting to you:
Steve Keen defines aggregate demand = GDP + change in debt
But he’ll define GDP as income without capital gains
Let me give you an example: suppose all production is apples. And there’s a new technology suddenly invented this year that doubles the number of apples each tree can produce per year. So each tree is now worth twice as much, in terms of apples.
So output of apples goes: 100, 100, 200, 200, etc. Assume the rate of interest is 10%, so each tree is worth 10 times the annual apple crop. In what ways would it be useful to say that income goes:
100, 100, 1,200, 200, etc.?
Nick:
”
I earn $100, and my house appreciates by $20 over the year. I decide to stay in my house, and spend the $100 on food.
”
How do you know that your house “appreciates by $20” ? You need to sell it before you know that. Until the transaction occurred, your house book value is exactly what you bought it at.
As another example of assuming that investment income is not income, consider a Wall Street trader. Apparently, he does not have any income, in the macro sense, from his activity. One wonders how the poor sod survives..
Somebody help me out. There’s a definition of “income” I’m trying to remember the name of. Hotelling? Change in present value of maximum sustainable consumption??? My memory has failed.
Sorry, Nick. I’m not following your example at all. That’s not a criticism. I just can’t understand the relationship between your verbal description and and your numbers. Can you restate?
vjk: you look at market prices. “Book value” is useless.
anon: suppose the economy consists of one tree, that produces 100 apples per year. The interest rate is 10%, so the tree is wroth 1,000 apples (it lives forever). Suppose in year 3 you come up with an invention that doubles the tree’s output. Annual GDP is 100, 100, 200, 200, etc. But in year 3 the tree is suddenly worth 2,000 apples, a capital gain of 1,000 apples. The invention is worth 1,000 apples, in present value terms. How should we define “income” in year 3? 200 apples? or 1,200 apples?
Google tells me that “Hicksian income” is maximum sustainable consumption. But some other economist refined Hicks definition later, and used it to analyse income taking natural resources, renewable and non-renewable, into account.
Nick:
Typically:
Sales
minus Cost of Good Sold = Gross revenue
minus sales/administrative expense/etc = Operating Income before depreciation
minus depreciation/depletion/amortization = Operating Income
plus non-operating income(e.g. Investment Income/Interest) = Pretax Income
minus taxes = Net Income
Is the “income” macro definition different from Net Income ?
If so, why ?
Year three income is 200 apples.
A separate issue might be what sort of apple flow the inventor was paid out of that prospective stream of 200 apples.
vjk
no problem as long as you don’t clutter investment income with capital gains
Nick,
Book value is the basis of GDP/income equivalence. You record the value of production at the point of sale or book value inventory value.
Nick:
”
you look at market prices.
”
Sorry, that’s not how accounting is done except some short-term/”ready to sell” investment bookkeeping.
Historical cost/book value is the current foundation.
So, if one is so adamant about “accounting identities”, it’s hard to see how one can ignore the rest of accepted practices. One can of course invent one’s own rules where some kinds of income are not income. There is nothing wrong with that. Perhaps, it makes more sense in the macro sense 😉
anon:
If I buy a share @10 today and sell it @11 tomorrow, are you saying that I do not have any income ?
Likewise, if I buy a house @100K today and sell it @110K in six months, would there be no income in any sense of the word ?
Nick, vjk,
I’m saying that capital gains don’t aggregate up to the income paid to the factors of production.
Then I’m saying its counterproductive to lump capital gains in with income at the micro level.
Here’s something interesting for you Nick (and you vjk in case you’re Canadian).
Drag out you’re last income tax return. Notice:
– Schedule 4 statement of investment income does NOT include capital gains
– Total Income at the beginning of the return includes NOT capital gains but TAXABLE CAPITAL GAINS – which as we know in Canada is 50 per cent of gross capital gains
The point is that you can’t even argue at the micro level on that basis that capital gains are part of income
The tax return treatment is simply a workaround to calculate taxes on capital gains according to some rule that can be integrated with income
So you have “taxable capital gains” under income which is an entirely different number than capital gains – ie. capital gains per se are not treated as income for income tax purposes in Canada AND I EXPECT THE SAME SORT OF TREATMENT HOLDS IN THE US
The micro treatment is only for tax purposes
And capital gains are excluded from macro income altogether
So from the government’s perspective, capital gains are not recorded as income, which means that can’t be a reference point for why they should be treated as income in microeconomics, let alone micro accounting.
So what is the argument as to why capital gains should be treated as income in microeconomics?
vjk: I know most accounting conventions use book value/historical cost. I think it’s wrong! But that’s a debate for another post.
anon: consider two cases:
1. The invention that doubled the trees’ output just flashed into someone’s head, and he told everyone for free.
2. Some inventor worked hard to figure out how to double the trees’ output, and was paid 1,000 apples (in apple tree stock) for his work.
In case 2, it would be counted as investment, and part of GDP. So GDP would be 1,200 in that year. But the only difference between the two cases is that one guy had a lot less leisure in the second case.
Maybe I should reconsider this.
As soon as you mention the word “invention”, it should trigger the thought of an investment good, as well as income earned in creating that investment.
If that’s the case, it wouldn’t be treated as a capital gain. It would be treated as an investment and part of GDP, as you say, with income earned.
In that case, if the investment is given away, that would become attributed income (I suppose) to the recipient of the investment. So it still wouldn’t be a capital gain.
So I should probably change my answer, based on the interpretation that the invention is an investment and the change in tree value is due to investment rather than a capital gain, and that income is earned from the production of the investment.
So income is 1200 rather than 200.
(hangs head in shame)
Have I contradicted myself in other ways?
Anon:
In the US, short-term capital are treated as ordinary wage/interest income (skipping a lot of details — I am not a tax expert by a long shot). Long-term capital gains are taxed at a lower rate, just as in Canada. The long/short division is quite arbitrary in the temporal sense, like legal “drinking” age.
The tax law, therefore, possesses its own quite convoluted “logic” and hardly can be used as an argument. I’d say taxation peculiarities are totally irrelevant for the notion of income.
Still, in my example with the trader, would the trader starve to death because he does not have any macro income ? Or should the notion of macro income be different from accounting income ?
Nick:
Re. accounting. You cannot take say some body parts you like (accounting identities) and chuck others (book value/historical cost) — you’ll get a corpse (or “creative” accounting).
anon: no shame in reconsidering one’s views.
“Have I contradicted myself in other ways?”
Probably. We all do. But not obviously (to me).
But why should it make a difference if someone worked (or was paid) to come up with the invention, or if the trees spontaneously doubled their fruit?
And is it different if the annual fruit output stays the same, but the trees double in price because the interest rate falls to 5%?
What are we really looking for, when we define income? It’s not obvious.
“Suppose I cave momentarily and allow you any definition of income you like, including any inconsistency of definition you choose between micro and macro.
Can you then summarize the point of economics you are making? And can you make that point independent of and beyond the accounting definition for income or anything else? That might be more productive.”
Damn.
Now that we are being reasonable I don’t have time right now. But I will come back!
Some quick points:
The whole issue under dispute here is whether when households save X, that there is X of investment in terms of “real” units of capital.
I.e. are asset bubbles possible, in which households grant each other excessive (nominal) incomes due to debt-financed spending?
Now, at some point, the real and financial sides have to reconcile.
But that could be a decade away.
In the meantime, when a household realizes a capital gain, that capital need not be capitalized in terms of NIPA GDP at all. It could just be a household receiving a gain because real estate prices went up, with no concomitant real investment recorded anywhere by NIPA.
Now when the household receives this income — and it is investment income, if they spend a portion of that income on goods, prices may go up somewhat, but not by the full amount, as you have downward sloping demand curves.
But to the degree that they go up, then real wages fall. You get an “asset based” economy in which households are relying on capital gains in order to fund consumption expenditures.
Now when you say “they have dissaved”, that doesn’t make sense. I understand where you are coming from — from the social planner’s point of view, a household sold an asset for more than what it “should” be worth and spent some of the proceeds on goods. The market price of capital went up more than the real interest rate and the quantity of capital. The household overconsumed.
But this is wrong. They spent only a portion of their income, not more than their income. They did not dissave.
To argue that they dissaved would be suffering from “real illusion”.
Households do not suffer from real illusion.
They understand full well that their own net worth can increase even though the aggregate capital stock remains fixed.
At the same time, the buyer of the asset — the one who overpaid — doesn’t think that he overpaid, because he thinks the house will appreciate even more.
So no one overpaid, according to their beliefs.
What happened was a shift, in which total output increased somewhat and real wages decreased somewhat, and both households together planted a time bomb for future demand and output.
When you want to model that in terms of representative households, you need to include a capital gain term in the euler equation.
The windfalls received from selling overpriced real estate are not “psychological”. If you want, you can take a bath in the 100 bills, to convince yourself that the money is real. Perhaps before the capital gain is realized, you can argue that it is psychological, but once someone writes the check, then it is real.
The endowments of those with capital go up and the endowments of those without capital decline somewhat, and the system equilibrates to a new set of prices in which national income is somewhat higher, the (real) investment share of income doesn’t need to change one way or another, but the resulting system is unstable when the bubble bursts.
But if you keep insisting that household net worth is some exact amount of assets capitalized during previous episodes of real investment, then you wont see any of this.
I will post a bit more later.
Nick,
“But why should it make a difference if someone worked (or was paid) to come up with the invention, or if the trees spontaneously doubled their fruit?”
Important question.
Basically, you want accounting to tell an accurate story about the numbers and hopefully the economics behind the numbers.
You have two different stories there.
One is an invention, where the best representation may be to include it within GDP, with corresponding income. In that case, the 1000 is an investment along with income (and GDP), but there is no capital gain.
The second is a spontaneous endowment of productivity and wealth. I’d say the best way to represent that is as a capital gain. In that case, the 1000 is a capital gain in the value of the asset AND the value of the equity of the owner of the existing investment. There is no new investment or income or increase in GDP.
At the end of the accounting period, you tally up the value of all the real assets in the economy, perhaps including unconsumed inventory if you like, and that will be the wealth or net worth of the economy. The number will be the same in both cases.
The number is the same, but two different stories – one of an invention that is recorded as newly produced investment and value, the other of a fortuitous endowment of productivity and value.
Two completely different (economic) stories, but with the same numercial value result – the story of how you got there is what you want the accounting to differentiate.
That differentiation of stories is what makes the accounting distinction useful. Somebody who didn’t know the story but saw the accounting could go back to you and say – “Nick , it looks like such and such happened here – is my interpretation correct?”
That differentiation and correspondence makes it easier to have a coherent discussion about the economics. Different example: just because one year’s GDP happens to be the same as the next doesn’t mean the story and the economics of how that happened are the same. Similarly in this case, a different story leads to the same outcome in terms of wealth or net worth, but via a different path, reflected in different accounting.
So that’s the answer to your important question.
And that’s the macro. I’m just saying that you have a choice about whether you want micro economics and micro accounting to correspond to macro, making it easier. If capital gain is distinct in the macro accounting story, why not mirror it in the micro story? It’s a choice, but the choice clarifies the story.
vjk,
“I’d say taxation peculiarities are totally irrelevant for the notion of income”
My only point was that anybody tempted to use the Canadian treatment of capital gains in their individual tax filings as an “obvious” rationale to justify a “common sense” approach of including capital gains in income really has no effective argument. I wasn’t making such an argument and I certainly wasn’t accusing Nick of making such an argument. But I would think many would be tempted by that example, even if not acknowledging it explicitly.
So my follow up question was – what is the argument?
Nick,
BTW, headlines show a $ 1.7 trillion decline in real estate value in the US this year.
That does not show up as a corresponding component of GDP, obviously. The GDP component will be positive, corresponding to new real estate produced and incomes flowing from it.
Do you really want to distribute that $ 1.7 billion as a subtraction of income in micro discussions? How does that help the discussion? Why not keep it as a loss in net worth. And while we’re at it, let’s note much of it is unrealized capital loss at the micro level, which also helps the discussion. Economic decisions flowing from unrealized losses won’t necessarily be the same as those flowing from realized losses. So it helps the discussion.
Nick,
Accounting tells a story about numbers, implicitly and/or explicitly.
You have a choice about what accounts to set up.
But you want the story to be as clear as possible.
That helps the economics discussion.
Including relating micro to macro.
BTW, you CAN’T avoid choosing your accounting, at least implicitly, when discussing economics.
So why not be explicitly careful about it?
RSJ,
“In the meantime, when a household realizes a capital gain, that capital need not be capitalized in terms of NIPA GDP at all.”
Exactly – which is why it makes sense to separate capital gains from the income generated by GDP and paid to the factors of production.
“Now when the household receives this income — and it is investment income”
No. See my first point. The household is not a factor of production in generating a capital gain.
Saving is defined as saving from income. It is a flow measure. Capital gains are not a flow measure in the context of income. They can be a flow measure in the context of the flow of funds – or flow of capital – but not the flow of income. Those who consume output equal in value to all of their income plus their capital gains dissave in the amount of their capital gains. They are consuming output beyond their income. Somebody else must save to balance the accounts at the end of the period.
I didn’t say that capital gains were only psychological; they obviously represent value, although not income. I said they provided psychological support for aggregate demand, not that they didn’t represent value.
Nick,
So my question again, based on our freedom to choose our accounting, and based on our discussion so far:
What is your argument then that micro accounting SHOULDN’T be consistent with macro NIPA accounting, in terms of separating capital gains from income?
anon: I agree with you that the micro and macro definitions of “income” should be consistent. It would make life very difficult if they were different. Stuff should add up.
So if we stick with GDP (actually, NNP, but let that pass) for macro income, we should do the same for micro income.
But is GDP obviously the best macro definition of “income”?
I would be tempted to say that income in year 3 is 1,200, regardless of whether someone worked to come up with the idea for doubling apple tree productivity. Are we measuring the results of the investment, or how much work we are doing to get those results? And should we capitalise the results of our investment today?
Actually, there was a real world example of this, just a few years back, in Canada. It takes a lot more work to get oil out of the ground with the oil sands in Alberta, compared to what it used to in the past.
Nick,
Doesn’t this all depend on whether or not you believe GDP should equal income at the macro level?
Isn’t that equation the starting point (for some) which in itself requires the exclusion of capital gains from income at the macro level?
Conversely, your starting point is that you have no problem with the non-equivalence of GDP and income at macro.
So your equation would be GDP + X = income at macro.
How do you define X in a holistic way?
Nick,
“should we capitalise the results of our investment today”
you have to be careful not to double count:
a) the present value of a future income stream as income today
b) the future value of that income stream as it enters future income
Nick,
I guess you’re just recommending that everything be marked to market across the board for purposes of accounting for income.
Would you like to see NIPA marked to market for income purposes?
If not, why not?
If so, do you “mark to market” GDP at the same time? In what sense do you interpret the marking of old assets as GDP output?
And if not, again, what’s X above?
anon: “you have to be careful not to double count:
a) the present value of a future income stream as income today
b) the future value of that income stream as it enters future income”
But, doesn’t GDP do this?
If we invest to make a tree that grows 10 apples per year forever, and the tree is worth 100 apples, GDP includes the investment of 100 apples, and also counts the 10 apples per year in all future years.
You’re right. That particular comment made little sense. It’s too early in the morning.
There is depreciation of course, which sort of takes back the original income associated with the production of the investment good.
anon: Or, maybe your comment made sense. But does GDP make sense?
Suggestion: define “income” as Friedman’s permanent income: income is the rate of interest times the expected present value of future consumption, discounted at that same rate of interest.
In that case, in my example where there’s an unexpected invention that doubles the productivity of the apple trees, “income” increases from 100 to 200.
And if the invention takes some years to “bear fruit”, income still rises in the year of the invention, but not as much, and slowly rises to 200 in the year it bears fruit.
But if the apple trees doubled in price simply because the rate of interest fell from 10% to 5%, income stays the same.
Not familiar with Friedman’s hypothesis.
But it sounds like its more consistent with taking the volatility of capital gains and losses out of the definition of income.
e.g. capital gain could be caused an interest rate decline
different from a productivity gain due to an invention that arguably could be capitalized as an investment, etc.
Defining income as permanent income has (I think) some nice properties:
1. It aggregates up.
2. If you consume more/less than your permanent income, your permanent income will be expected to fall/rise over time.
Nick,
I know you showed something above, but can you do a really simple example of a permanent income approach, making the calculation of permanent income clear, versus income defined otherwise, capital gain, etc?
and I forgot the third, nicest property:
3. A utility maximising person, who can freely borrow or lend at rate r, will choose consumption as a function of: permanent income; r; preferences, only. In other words, you get exactly the same mpc out of all sources of permanent income.
anon: assume your only asset is a tree that will produce 110 apples next year, then die. r=10% . Your permanent income is r.110/(1+r)=10 apples per year. Your tree is worth 100 apples this year, 110 next year, and 0 apples thereafter. If you sell your tree for 100 apples, lend the proceeds at 10% interest, and consume the interest, your consumption will stay constant forever.
Nick:
“But does GDP make sense?” — In my humble opinion, it does not. It’s a rather meaningless number whose information content is worse than, say, the level of retail sales.
If one were to follow the line of thinking that there is no investment income at the household level but just a swap of assets, then, inexorably, one would arrive to the conclusion that there is no income at all in any circumstances since any sale is nothing but a swap of assets too which makes a new house sale no different from an old house sale, or any I-pod sale no different from an old car sale.
Still, it’s curious that no one commented on how the wall street parasite can survive without having any macro income 😉
I think distinction between Operating Cash Flow and Free Cash Flow explained here would be of some help in thinking about this.
Roughly speaking, Operating Cash Flow corresponds to GDP.
And, Free Cash Flow equals Operating Cash Flow minus Capital Expenditure. Thus, Free Cash Flow circumvents double count. That’s why Free Cash Flow is used in evaluation of the corporate value.
When you apply this to the above apple example, it goes like this:
OCF of tree owner: 100, 100, 200, 200
OCF of inventor: 0, 0 ,1000, 0
OCF total: 100, 100, 1200, 200
FCF of tree owner: 100, 100, -800, 200
FCF of inventor: 0, 0 ,1000, 0
FCF total: 100, 100, 200, 200
You can see that FCF is better for aggregation use.
“You can see that FCF is better for aggregation use.”
Maybe I should have written, “You can see that FCF is better for value-aggregation use.”
As for aggregation of economic activities, OCF may be of better use. After all, an extra economic activity (transaction between tree owner and inventor) took place at t=3. I suppose that’s why GDP does not circumvent double count.
Adam
Re: whether NK models (or any mainstream model, really) allows for a divergence between the market price of capital and actual price of capital (as measured by the cumulative amount of investment), the answer is no.
If they did, they would not assume that the rate of interest that goes into the euler equation is the same rate of interest that clears the money markets.
What clears the money markets is the bond rate, but the actual rate that households optimize on is the total return = interest payments + capital gain.
Households do take capital gains into account — i.e. the price that they are willing to pay for a house is not just a function of the discounted stream of rental consumption, but is also a function of the expected appreciation of the asset (net of maintenance and taxes) over the expected hold time. In many cases, there is a legitimate reason to expect the asset to appreciate — e.g. you expect incomes in the neighborhood to go up.
To assume that the price of the house is just the rental stream is to impose a no-asset bubble restriction. I understand the desire of doing this, but its not what households optimize over.
Part of the problem here is that you are using infinitely lived agents and then imposing a transversality condition. I.e. you assume the hold time is infinite.
Again, I understand the logic, but that isn’t a reasonable condition that you can impose on households. It’s not just an issue of altruism, but imposing this condition rules out degrees of freedom that are economically important, just as requiring the risk free rate to be larger than the economic growth rate rules out economically important degrees of freedom for government spending.
In those cases where these degrees of freedom can dominate the trade cycle or make a material difference to the outcome of the model, then you can’t rule them out, the model needs to determine the constraints as arising out of voluntary exchange and normal utility maximization rather than imposing them a priori.
Wow, I’m behind here on the comments.
Anon/Nick,
I think you are missing the forest for the trees.
Income is the maximum that you can consume without decreasing your net-worth. An income event is defined as an event that triggers an increase in net-worth, and an expenditure is an event that triggers a decrease in net-worth.
Incomes – Expenses = change in net worth over the period.
An accounting system is just a consistent way of recognizing which events trigger a change in net-worth. The whole point is to make the balance sheet adjustment at the same time as you make the income statement adjustment, so that the change in balance sheets over a period of time is explained by the statement of incomes.
So Anon, you cannot say that households lost X Trillion in housing value from their balance sheets without saying that this was an expenditure. It was an expenditure under a mark-to-market accounting.
Under a different set of conventions, less would be deducted form balance sheets, and expenditures would be lower.
Under a cash-basis accounting system, a realized gain is income.
P.S.
This is not to minimize the validity of anon’s point that under more realistic accounting systems, households did not lose X Trillion of income because their balance sheet did not decrease by that much.
But that’s just an argument to use a better accounting system, not one system for the balance sheet and another for the income statement.
Nick/Anon,
Under any accounting system, income statements will not aggregate if the entities are allowed to transact with each other.
The assumption that for any 2 people, A, B, in your economy:
(1) Income{A} + Income{B} = Income{A,B}
(2) Expenditures{A} + Expenditures{B} = Expenditures{A,B}
Is equivalent to the assumption that none of A’s income is obtained from B’s expenditure and vice versa.
In general,
Income {A,B} < Income {A} + Income {B}
Because you need to subtract out transactions that do not cross the accounting boundary.
In other words, A and B do not transact with each other.
Only in that case can you say that (1) and (2) hold.
This has nothing to do with mark to market, or cash basis.
And it’s precisely this no-interaction assumption which gives you the wrong answers, or rather, which prevents the model from getting the right answers.
That’s simplifying too much.
“If we define “income” as “value added” (sales minus purchases from other firms) it aggregates up across firms”
Except that you cannot define income as “value added”, unless you assume that the firm has zero expenditures.
P.s.
And of course, if the firm has zero expenditures, then it is not transacting with anyone.
This only applies for autarky.