There’s been a running discussion of this on various blogs (sorry if I missed linking some!), inflated simultaneously by Krugman and by magisterial and mysterious commenter JKH’s “paradigm riff,” here.
That discussion has brought me to the following conclusions.
Assuming you have a coherent and accurately representative System of National Accounts*:
• Accounting, and accounting identities, do (or should) impose a constraint on our economic reasoning and predictions.
• If some piece of economic reasoning predicts something that simply can’t happen according to the accounting (things don’t add up, balance), that reasoning/prediction is wrong.
• Accounting can’t tell us whether a piece of economic reasoning is right. It can only tell us if it’s wrong.
• Accounting won’t necessarily tell us that a piece of economic reasoning is wrong. There are plenty of economic ideas out there — behavioral notions about how people (will) respond to incentives and constraints – that conform to accounting identities and balances, but are nevertheless wrong.
• Accounting tells us exactly nothing about how people will behave, nor can it cause or constrain that behavior. It can only tell us that that it’s logically impossible for them (all) to behave in a given way.
Takeaway: Conformance to the rules and balances of accounting is a necessary but not sufficient condition for economic reasoning and predictions to be correct.
Or to put it another way: Accounting is a constraint on economics, not economies.
Simple example: if somebody suggests that all countries should/can get out of depression by increasing their net exports, it’s false/bad reasoning. Because global imports equals global exports; the books can’t add up that way.
Or suppose someone says:
1. We should reduce government debt.
2. There’s not much we can do about net imports, the trade imbalance. (Exports are determined largely by international demand, and we don’t want to use trade policy to deny our people the benefits of cheap imported goods.)
3. People should save more.
This is impossible, by accounting identity. The only way to increase private savings (the stock of net financial assets) without changing imports is to increase exports or run government deficits.
People, institutions, and policy makers could certainly try to achieve these mutually incompatible goals. They could even believe that it’s possible to achieve them all. But the arithmetic of stock-and-flow accounting tells us that they will fail — and that if they believe they will succeed, they’re wrong.
* Even though I have real qualms about the conceptual structure of the current system — I find it much easier to do good thinking using Wynne Godley’s modification of that system — the current system is coherent and accurately/usefully representative. It’s coherent in that all the stocks and flows balance out, and representative in that it covers most of the important stocks and flows. No system could be perfectly representative, of course; the map is not the territory. In both systems there’s a great deal that’s not considered — nonremunerated work, for instance. But that doesn’t discredit, is peripheral to, the logical thrust of this post.
Cross-posted at Asymptosis.