Macroeconomics and the Celestial Emporium of Benevolent Knowledge

This is what we teach, right at the very beginning of macro:

1. Macroeconomics is about GDP.

2. On the expenditure side, GDP is divided into Y=C+I+G+NX

3. On the income side, GDP is divided into Y=C+S+T

Now read what Borges said on a closely related subject.

We don't have to think of macroeconomics that way.

1. I can imagine an economy in which GDP is a technological constant, but in which monetary shocks cause fluctuations in the amount of market activity and human welfare. Sure, my imaginary economy is "unrealistic", in that it leaves out a lot of what happens in the real world. GDP does fluctuate. But an exclusive focus on fluctuations in GDP is also "unrealistic", in that it leaves out a lot of what happens in the real world trade cycle that matters to people and yet is not related to the production of newly-produced final goods and services (GDP). It is not obviously wrong to define (short run) macroeconomics as the study of fluctuations in all market activity, for new and old goods, for final and intermediate goods, for produced goods and goods that are not produced goods at all. It is not obviously wrong to distinguish between production for the market and production for own use, and to say these behave differently over the trade cycle. It is not obviously wrong to distinguish between markets with monetary exchange and barter exchange, and to say these behave differently over the trade cycle.

But when we start out with GDP, and ask why GDP fluctuates in the short run, and grows in the long run, we have already stacked the deck against any different vision of what macroeconomics is all about. The question presupposes a particular type of answer. A different answer, which might be a much better answer to a slightly different question, won't look as good as it should. Getting the best answer requires asking the best question.

If we think of macroeconomics as being about GDP, we may miss seeing the effect of monetary disequilibrium on all monetary exchange.

2 and 3. A feminist might want to divide expenditure into: expenditure by men; expenditure by women. And divide income into: men's income; women's income. C+S+T=Y=C+I+G+NX just doesn't cut it for her.

If I am interested in the burden of the debt on future generations, I will not want to aggregate all generations' incomes and expenditures during a particular year. Instead I will want to aggregate all years to look at the lifetime incomes and lifetime expenditures of a particular cohort. If I aggregate all generations for a particular year, which is what GDP does, I will totally miss seeing the burden of the debt on future generations who have a lower present value of lifetime consumption. [Preceeding sentences edited for clarity]. An awful lot of macroeconomists have missed seeing something rather obvious simply because it couldn't be described within the categories of their Celestial Emporium of Benevolent Knowledge. Some even described it as "especially nonsensical" because it didn't exist in their Celestial Emporium of Benevolent Knowledge.

If I have a monetary disequilibrium theory of the trade cycle, I don't care about the distinction between new and old goods, and between final and intermediate goods, and between consumption and investment, and between consumption and saving. They just don't cut it for me. I am interested in the distinction between spending money and not spending money; and between selling goods for money and not selling goods for money.

Why should I try to write my theory using the categories of the Celestial Emporium of Benevolent Knowledge written by long dead National Income Accountants for some totally different purpose? I don't have to share their order of things. I can see the world their way if I want, because I've been taught to see it that way, and I've taught thousands of students to see it that way too. But I don't have to see it that way.

Though it is rather hard to escape, when you are so used to it.

Comments

[Joe]

It seems to me that it would be better to say that macroeconomics is about the supply and demand for money.

[Nathan Tankus]

Actually, most feminist economists I'm aware of spend more time discussing "the care economy" and unpaid household (and social) work. here's some good statistics on that.

http://www.bls.gov/opub/mlr/2009/07/art3exc.htm

[Nick Rowe]

Joe: that would be tempting, but perhaps a little too narrow! We ought at least add that it is about all the consequences of changes in the supply and demand for money.

Nathan: What you say sounds right. I made up that bit about Y=M+F, but it is not totally unrelated to what you are describing. They don't like GDP, and Y=C+I+G+NX, because neither capture the aggregates and distinctions they are interested in.

[Greg Ransom]

This is excellent.

But note well, Hayek also has a monetary disequilibrium theory of the trade cycle -- a heterogenous production goods and out-of-kilter relatice price theory -- and he is concerned with much more than you are, because many of the large and systematic causal pathways effected by monetary disequilibrium involve much more than is imagined or conceptually allowed in the stark distinction between "spending money and not spending money; and between selling goods for money and not selling goods for money."

Nick writes,

"If I have a monetary disequilibrium theory of the trade cycle, I don't care about the distinction between new and old goods, and between final and intermediate goods, and between consumption and investment, and between consumption and saving .. I am interested in the distinction between spending money and not spending money; and between selling goods for money and not selling goods for money."

[Bill Woolsey]

Employment, production, and consumption are really important.

Perhaps they are less important now than in the past. Perhaps one day they will be even less important.

If you imagine a world where no one (or hardly anyone) needs to work, then it would be different. Current consumption is trivial as the near immortals direct their robot factories to maintain centuries of future consumption.

In the real world, it seems to me that you are limited to the housing market.

[Ken]

Great stuff! Two comments:

1. Did you accidentally type "If I aggregate across generations, rather than across years" rather than the other way around? It seems backwards...

2. I'd also like models that make a distinction between debt "good debt" and "bad debt" rather than just aggregating all debt and assuming that it is repaid.

[Nick Rowe]

Greg: Thanks! Yep, the distinction between spending money vs not spending money is a basic starting point. That's what gives you real consequences of monetary disequilibrium. Where you go from there, for a finer-detailed understanding of those consequences, will require other distinctions. I read Hayek as saying the simple C vs I distinction doesn't really cut it. There's more of a continuum of goods of various orders, or lengths of production processes, from immediate C at one end to the longest/highest order I at the other end.

Bill: if you believe that employment, production, and consumption are really important, you should not be looking at GDP. Because GDP ignores lots of employment, production and consumption. (Home production). What happens over the trade cycle is that the composition of employment, production, and consumption changes. Employment, production, and consumption that is bought and sold for money falls, and E,P,C that is not bought and sold for money rises. And we are worse off because trade makes us better off, and monetary trade is easier than barter trade.

To the extent that GDP gets it right, it gets it right by sheer accident, because it's too hard to measure a lot of non-monetary employment, production, and consumption.

[Steve Roth]

As you can imagine, I love this post. The NIPA model, along with the textbooks' "these are accounting identities; they're just true!", is at least a potential conceptual trap. And the conceptual magnetism is strong. I think some of my difficulty may arise from trying to use the conceptual constructs of the NIPA model to think in terms that are not (well) supported by that model. Incorporating financial assets, capital gains, and such.

Presumably all transactions for real goods/services/assets result in a surplus from trade -- otherwise they wouldn't have happened. For intermediate goods this is accounted for by the value added approach for newly produced goods. But for antiques? Houses?

I don't *think* that excluding non-remunerated work invalidates the model itself, or that including it would alter the bookkeeping setup. (?) It would certainly alter the numbers in important ways.

http://www.asymptosis.com/home-work-25-of-gdp.html

I do wonder, though, to what extent the NIPA conceptual model is constructed based on what it's feasible to measure (or what was feasible in the '30s).

[Steve Roth]

Oh and thanks for the Borges ref. Perfect. What an Aha! moment when I discovered him in college...

[Frances Woolley]

Nathan, Nick - on feminist economics.

You're both right. Although much feminist economic research, e.g. Nancy Folbre's work, looks at unpaid work/care etc, the type of exercise that Nick describes is used in, for example, the United Nations Development Program's Gender Empowerment Measure and Gender Development Index. One element of that is based on women's incomes relative to men's incomes (and I mean income in the sense of share of GNP, not income in the sense of hourly wages). This was developed by/in consultation with feminist economists.

[Nick Rowe]

Ken: Thanks!

1. I kept wondering myself whether I was saying what I wanted to say correctly. Since it looked wrong to you too, I have totally re-written that whole bit. Hope it's clearer now.

2. And debt, good or bad, doesn't appear at all in GDP.

Steve: Thanks!

"conceptual trap" "conceptual magnetism" Yep.

"Presumably all transactions for real goods/services/assets result in a surplus from trade -- otherwise they wouldn't have happened. For intermediate goods this is accounted for by the value added approach for newly produced goods. But for antiques? Houses?"

Yep, there are presumably gains from trade, otherwise those trades wouldn't happen (sure there's also externalities etc.). But what GDP measures is the total revenue from some trades, not the gains (consumer+producer surplus) to all trade.

[Nick Rowe]

(The "value added" by a firm is not the gains from trade; it's the total revenue of that firm minus the total revenue it pays to other *firms*. It measures rectangles, not triangles.

[Scott Sumner]

Great post. I've always thought that if we are worried about employment, then we ought to talk about employment, not GDP. If I'm not mistaken, a significant chunk of US GDP is the imputed rent on owner-occupied housing. Suppose this falls because more houses are empty, people are doubling up. Is there any change in employment?

On the other hand, in my example consumption does fall. But if we are worried about consumption, why not measure consumption?

On the other, other hand, I wouldn't really want to argue against someone who claimed that RGDP is a reasonable proxy for the problems that we instinctively believe are associated with business cycle instability.

[Lord]

Nonsensical for another reason. The story of debt is the story of war. What makes war good or bad? Who is better off because of it? Is winning success or fighting failure? What value do we put on the outcome? Is there any benefit or only costs? Is it avoidable or inevitable? Is there a choice of costs now or later or are costs now just expenses later? What is the utility of it to whom?

[Nick Rowe]

Scott: Even employment is dodgy, because there's employment and then there's selling your labour for money. And then there's selling your labour for money in its best use. (Imagine if the unemployed got employed walking around wearing billboards for $1 per day to try to sell stuff for firms in a recession).

Yep. RGDP is going to be a good empirical proxy. But we don't build our theories based on proxies. We build our theories first, and then look around for reasonable proxies afterwards.

[Dan Hirschman]

As a historian of national income statistics, and Borges devotee, I find this post absolutely delightful! Interestingly, many of the problems with GDP (and its predecessor statistics) were known in the 1920s and 1930s, and debated by the statisticians who designed them. The person most associated with national accounting in the US, Simon Kuznets, actually hated the idea of putting GDP into an accounting framework at all.* He argued that it would cover up much of the arbitrary nature of the calculations by making it seem as if all the accounts balanced. Rather, he argued that the subtotals should be the emphasis of the work, and investigators would need to judge for themselves when it made sense to include particular calculations or estimates (for owner-occupied housing, unpaid labor, or even the government, which was hotly debated at the time as it's not at all clear that all government output should be valued as final consumption, but rather much of it looks like intermediate production).

All this to say, I think that many of the most influential figures in the early formalization of national income statistics would agree with much of this critique. Early macroeconomists and development economists began to rely on GDP as a sufficient statistic for the economy, but they did so against the express advice of many of the statisticians feeding them numbers.

* See especially Kuznets' debates with his successors in the Commerce Department ( http://www.jstor.org/pss/1926746 ) and their responses.

[Steve Roth]

Nick: "GDP measures is the total revenue from some trades, not the gains (consumer+producer surplus) to all trade."

Right. Of course. That surplus is perhaps what's most interesting for macro, no? Or one key thing in any case.

Semi-aside: You might be interested in the comment thread starting here, mainly JKH:

http://www.asymptosis.com/an-mmt-thought-experiment-the-arithmetic-and-political-mechanics-of-net-financial-assets.html#comment-3450

Dan Hirschman:

Excellent. Thanks for that thinking and the Kuznets link.

[Nick Rowe]

Lord: but in this case the good/bad debt distinction is about the probability of whether the debt will be repaid, I think.

Dan: neat! I didn't know much about the origins of GDP. But from what you are saying, those "long dead National Income Accountants" would basically be on my side here!

[Greg Ransom]

For Hayek the basic starting point is the expansion and contraction of finance, money, credit, leverage, e.g. the way asset values can become highly liquid, can expand quickly in value, can be used as reserves for further credit expansion and in leveraged investment, or the way the demand for reserves can expand or contract or how the degree of leverage in transaction can expand or contract. (And note well none of this is premised on the existence of central banking, which simply adds an addition causal factor to an existing causal degree of freedom or "loose joint" in the system). The banking system is able to create money and increase the size of finance and change the cost of finance across time -- and sometimes it cannot help but do the opposite. The basic "spending money vs not spending money" most significantly implicates the dimension of the time structure of the economy -- the whole damn economy is wrenched -- imagine a net that is stretched one way, then wrenched suddenly and violently in a different direction with such force that it is ripped apart in different sections. (The net is all price relations related by the unavoidable choice of more time/superior output or less time/inferior output across time, i.e. the rate of "profit" or "interest" depending on which generation of economists is doing barter economy economics). When the net is wrenches, many production imputs become either non-economic goods or radically less valuable goods, and some never regain an "economic good" status.

Nick writes,

"Greg: Thanks! Yep, the distinction between spending money vs not spending money is a basic starting point. That's what gives you real consequences of monetary disequilibrium. Where you go from there, for a finer-detailed understanding of those consequences, will require other distinctions. I read Hayek as saying the simple C vs I distinction doesn't really cut it. There's more of a continuum of goods of various orders, or lengths of production processes, from immediate C at one end to the longest/highest order I at the other end."

[JP Koning]

I am grokking your line of thought, Nick. If you get the chance click through my name, curious what you think.

A few questions.

In your previous post, you advocated adopting the most "useful" way of dividing up the world into categories. But what does that mean? Useful in terms of teaching students, reaching the layman, articulating theory amongst other economists, calculating statistics, conducting policy?

Are you saying that economists should have multiple "categorical universes" in their head? Or is their "one ring to rule them all" way of dividing up the universe that you are trying to tease out?

One useful reason for having one standardized way of splitting the world into categories is it makes conversation easier. Having multiple ways adds subtlety but confusion.

I am learning that the core of economics (or at least near to it) is about words. We create meanings for things in the real world using some of our existing words, and this creates categories, but since other words can also be used to create similiar-though-not-overlapping categories, we have to choose between categories. This process comes prior to mathematical econ, since the math needs categories like C and I before mathematical phrases can be put together.

[Lord]

Have always been fond of Napoleanic gilt perpetuities myself. No pretense of repayment. Growth has always diminished it for winners and default for losers. Repayment makes for good counter cyclical policy though.

[Jim]

The problem with GDP as defined is that it is totally wrong. It is a hapless blend of spending and investment. It is a total denial of basic accounting and finance, and therefore the concepts of ROI, investment and debt.

No wonder macroeconomics is not really a science. most macroeconomic models do not even take debt into account as a potentially devastating shock to the system. The mistake begins with GDP measurement.

It is time to scrap macroeconomics in favor of network theory and good old accounting measurements of the economy, including government expenditures. Then maybe, just maybe, it would become clear that we can not redistribute, print and borrow our way to prosperity.

[Brent]

The System of National Accounts inclues a lot more than GDP. For example, the recent versions of SNA (1993 and 2008) include barter transactions, transactions in used assets, financial transactions, and balance sheets. Non-market home production isn't included, though it could be added as a satellite account. Similarly, the SNA's standard sectors don't include generations, but the framework could be adapted to develop a generational accounts as a satellite account. The SNA is really a pretty general framework that isn't given justice by the typical textbook.

[Anders]

Interesting post. Can anyone explain why it is that we like the C+G+I+NX = C+S+T formulation?

It seems to me that C+G+I+X = C+S+T+M would be neater, ie "GDP plus imports", as it would group all injections and leakages together.

I haven't managed to get an economist to explain why we don't focus on this instead of GDP.

[Jacques René Giguère]

Anders: in French ( from France, not Québec) books, G+I+X=S+T+M is still known as "équation des grands équilibres) (great equilibriums equation".
It was standard formulation when we still teached injections-withdrawals models in the '70's, or what Nick calls the barbarians times ;-)

[pe]

Is it right to say that GDP is a Keynesian construction as the Austrian economists claim ?

Mark Skousen's book "Structure of Production"
"GDP is a Keynesian concept that measures only the output of final goods and services and excludes intermediate production. Second, government spending is included in GDP data, an autonomous addition to national output.

Both peculiarities of GDP have led to much mischief. In the first case, by focusing solely on final output, many economists and commentators in the media have concluded that consumer spending is more important than capital investment in an economy, based on the fact that consumption expenditures usually represent about two-thirds of GDP. In the second case, including government spending in GDP has led many pundits to believe that an increase in that spending—even if accomplished through deficit spending—will automatically increase economic growth

Most students of economics are unaware of the fact that GDP was created by Simon Kuznets during World War II to quantify final aggregate demand according to the new economics of Keynes."

[Nick Rowe]

JP: It was reading your ("this is not a blog") blog that made me think I should write this post, and collect my thoughts together.

I'm not sure on the answers to all your questions. Even if we all agreed on what the true theories were, we might want one theory for one purpose and a different theory for another purpose. And each theory would have its own most useful set of categories.

Yep, it might make conversation harder. But maybe some conversations would be impossible if we had to stick to one set of categories that didn't fit the topic at hand.

Brent: neat. I didn't know that.

Anders: "It seems to me that C+G+I+X = C+S+T+M would be neater, ie "GDP plus imports" "

Well, it's bigger than "goods produced in Canada". There might be some theory for which that would be a useful aggregate, but I can't think of one off the top of my head.

Jacques: Yep. G+I+X=S+T+M is what I learned at skool, and as an undergrad in the UK in the early 70's. But that's a little different from what Anders has, because he has added C to both sides.

pe: Good question. I have heard that too, but don't know if it's really true. Maybe they were both coming into existence at the same time.

[Jacques René Giguère]

Nick: the "grands équiibres" concept was used in ( continental only?) Europe to link gonernment first-order policy with the balance of payments , known as "the dollar shortage problem", the perpetual fixation of European economies about foreign exchange.
1947 vintage British joke: a woman gives birth to twins. The nurse ask her which one she wants to keep. "Why do I have to choose? cries the poor woman. "One for you, one for export."

Adding C on both sides was irrelevant as consumption was constrained.

[Jeremy Fox]

I have the vague sense that there are some interesting analogies and contrasts here between accounting identities in macroeconomics and in ecology and evolution.

I'm thinking in particular of the Price equation in evolution (http://en.wikipedia.org/wiki/Price_equation), which is an accounting identity that carves up total evolutionary change over a generation into components attributable to different classes of evolutionary process. It's widely seen as a very useful conceptual tool, but seen by a minority as a trivial tautology (although both those groups are actually outnumbered by people who don't know what the Price equation is or don't understand it!) The Price equation also has a couple of alternative forms which carve up total evolutionary change in slightly different ways, and there is some ongoing debate about whether one form is always or sometimes preferable to the other. All of this seems broadly analogous to the situation in macroeconomics. On the other hand, nobody would ever try to justify some substantive claim about the kinds of evolutionary dynamics that are possible just based on the fact that the Price equation must be true. But I have the impression that certain macroeconomists sometimes try to do just that.

In ecology, I suppose there might be an analogy to the fact that population growth rate always must equal births - deaths + immigration - emigration. So that, for instance, if a population is currently in equilibrium, births + immigration necessarily must be balanced by deaths + emigration. And if, say, births jump, one or more of the other three quantities must change if the population is going to remain in equilibrium at its current size. But this seems rather less analogous, in that there's never any conceptually interesting debate to be had about alternative partitionings of total population growth. Not that alternative partitionings aren't possible--one could, for instance, subdivide each of the four components of population growth in all kinds of ways. But the appropriate choice of partitioning is generally quite obvious (e.g., if you're interested in deaths imposed by predators, you're going to want to track that separately from other sources of death). And again, no ecologist would ever argue for some substantive claim about population growth just because population growth necessarily equals B - D + I - E.

These are quite ill-formed thoughts, and I'm sure they're pretty opaque to the 99.9% of readers here who aren't ecologists or evolutionary biologists. Sorry about that. Just wanted to get the thoughts down, in the hopes of turning them into something better-developed and more broadly-interesting in future...

[Nick Rowe]

Jeremy: I saw you briefly discuss the Price Equation on your blog, and saw you say it's an identity, but a useful one. And I thought of the analogy to MV=PT and Y=C+I+G+NX. I even had a brief search on your blog to find out more about it, but it was over my head.

From a quick skim of that Wikipedia, I would say that the price Equation looks like a lot more sophisticated identity than those two economics ones. (Probably more useful too).

I was vaguely wondering how someone like you would generate a theory of the equilibrium number of animals on earth if you started out from the Borges taxonomy! You would presumably be much better off tearing it up and starting afresh with some sort of biomass predator/prey model. Or you might change the question completely and ask what determines the equilibrium mass of all the animals on earth. I was going to put my musings on this in the post, then decided against.

[Jeremy Fox]

Thanks Nick. So perhaps the notion that population growth rate = births - deaths + immigration - emigration is slightly closer to the sort of accounting identity that economists think about? Especially since immigrants arriving at one location necessarily must've been emigrants from some other location. So that, at a global level (i.e. summing across all populations), immigration and emigration must cancel out, leaving only births - deaths. There are obvious (but perhaps rather trivial?) analogies here to economics--every person's debt is someone else's credit, every purchase is also a sale, etc. And some ecologists sometimes trip over the differences between "micro" and "macro" in ways that seem at least loosely analogous to the way economists sometimes do. I even did a post on the "fallacy of composition" in ecology. It's aimed at ecologists, but I think an economist would get the gist of it:

http://oikosjournal.wordpress.com/2011/11/22/species-pools-and-the-fallacy-of-composition/

[JP Koning]

"It was reading your ("this is not a blog") blog that made me think I should write this post, and collect my thoughts together."

Aha, so that's why this post hit a chord with me. I don't disagree with any of your points and don't have much to add, these issues are complex. Will definitely be tuning into future posts on this issue.