There's an idea floating around out there that I fear may be influential. And that idea is horribly wrong. Which makes it dangerous. And I want to try to kill it. But macro is hard. And it's not easy to explain clearly and simply.

I can only try. And I can only hope that others who are more influential than me, or can explain things better than me, will do the same.

"Sure, there's a risk that inflation is falling below target, and a risk that recovery will be delayed, and it would be nice if the Bank of Canada (or Sweden or wherever) could loosen monetary policy to prevent this happening. But monetary policy works by lowering interest rates and encouraging people to borrow more and spend more. And that creates a problem for financial stability, because some people are already borrowing too much. So there's a trade-off between monetary stimulus and financial stability, and monetary policy needs to take both objectives into account."

I made up that quote. But I don't think I made up the influential idea it expresses. And it's horribly wrong. It's almost the reverse of the truth.

One way to attack that idea would be to say that reducing unemployment is a much more important goal than financial stability. That might well be true, but that's not my argument here.

A second way to attack that idea would be to say that there are other and much better ways to address the concerns of financial stability and prevent some people borrowing too much. That too might well be true, but that's not my argument here.

Instead I'm going to argue that the idea is fundamentally flawed.

It rests on a fallacy of composition.

It confuses a consequence with a cause.

It misunderstands the relation between monetary policy and interest rates.

It's just horribly wrong.

1. It rests on a fallacy of composition.

Suppose the Bank of Canada wanted me, Nick Rowe, to spend more, to do my bit to promote recovery and prevent inflation falling below target. So it offers a special rate of interest just for me. Or gets the Bank of Montreal to offer a special rate of interest just for me. If it lowered the Nick Rowe rate of interest I would save less if I were a saver, and borrow more if I were a borrower. Either way I would spend more. (Unless maybe if I were a lender and the income effect of the cut in interest rates making me poorer were bigger than the substitution effect.) The magnitude of the effect on my spending would depend on my interest-elasticity of demand for consumption and investment.

It is awfully tempting, but horribly wrong, to think that if we multiply that individual experiment by 35 million (assuming I'm the average Canadian) we get the macroeconomic effect of a cut in interest rates.

It's horribly wrong because, just for starters, it ignores the Old Keynesian multiplier. If I spend more that means I am buying more from other people, so their incomes will rise, and they will spend more too, which means still other people's incomes will rise…and so on. The macroeconomic magnitude will depend not just on the interest-elasticity but on the marginal propensity to spend (equals marginal propensity to consume plus marginal propensity to invest). And, depending on how long people expect the increased spending and income to last, there is nothing to prevent that marginal propensity to spend being greater than one, which would mean an infinitely big multiplier (until the Bank of Canada sees it needs to reverse course to keep it finite).

It's even more horribly wrong when we think about another accounting identity that holds in aggregate: for every $1 borrowed there's $1 lent. To keep it very simple, imagine an economy where everyone is identical. If I wanted to borrow then so would everyone else want to borrow, which means nobody would want to lend to me, so I wouldn't be able to borrow from anyone else. And to keep it even simpler, imagine that people pay for everything with central bank cash, and that the velocity of circulation is almost infinite and so the amount of cash in circulation is vanishingly small. (Yep, like Woodford's New Keynesian model). If the central bank lowers the rate of interest, people borrow a tiny amount of extra cash from the central bank (they all want to borrow from each other but nobody wants to lend), and spend that cash, and their incomes increase as others spend, which increases their spending even more, and incomes and spending keep on rising until it reaches a level where nobody wants to borrow from other people or from the central bank (or the central bank decides it has increased enough and raises interest rates again).

This is a world in which a cut in interest rates makes people want to borrow, and monetary policy works by making people want to borrow, but there is never any actual borrowing (except for a tiny amount of borrowing cash from the central bank).

Monetary policy does not work by increasing actual borrowing. That is not the causal channel of the monetary policy transmission mechanism. Monetary policy works by increasing spending, not borrowing. And one person's spending is another person's income, so people in aggregate do not need to borrow more in order to spend more. Their increased spending finances itself.

Yep. Macro is hard. You can't just sit back and think "how would I react if my rate of interest fell?". You have to think about how my reactions would affect others, and how their reactions to my reactions would affect me, and so on.

2. It confuses a consequence for a cause.

"Dangerous roads cause drivers to slow down. If drivers slowed down there would be fewer accidents. Therefore, if we made roads more dangerous, there would be fewer accidents."

That argument is clearly invalid. The premises do not entail the conclusion. (OK, the conclusion might conceivably still be true, if safe boring roads and auto trannies lead to distracted drivers texting…). What causes accidents is roads that seem safe, so drivers speed up, but that suddenly and unexpectedly become dangerous.

People are not all identical. At any given rate of interest, some will want to borrow and others will want to lend. So some people will actually borrow from other people. Maybe via financial intermediaries. And some of those financial intermediaries issue liabilities that are used as media of exchange and so are called "banks".

And sometimes some people will borrow too much. Which means, of course, that sometimes some other people will lend them too much. Accidents happen, even on safe roads, because some people drive too fast or aren't paying enough attention. But the biggest accidents happen when an apparently safe road suddenly and unexpectedly becomes unsafe. Because the drivers can't slow down quickly enough. Or even if one driver can slow down quickly enough, another driver who can't simply ploughs into the rear of the slowing car.

OK, that's just an argument by analogy. But I think you can see the link. If central banks keep nominal income growing smoothly, most people will adjust their borrowing and lending (speed) to the prevailing conditions. And some won't, of course. But if the central bank lets nominal income fall, without giving people lots of advance warning, that means that even some otherwise safe borrowers and lenders suddenly become risky, and they crash too. And the best palliative is to get nominal income back onto as close to its previous path as possible as soon as possible. Even if that does cause drivers to speed up, now that the roads are safe again.

3. It misunderstands the relation between monetary policy and interest rates.

As Scott Sumner echoes Milton Friedman: low interest rates does not mean loose monetary policy; low interest rates are a consequence of past and expected future tight monetary policy. That's true both for nominal and real interest rates. If tight monetary policy means actual and expected inflation is low, then nominal interest rates would be low for any given equilibrium real interest rate. And if tight monetary policy means that current and expected future demand and real income are low, then people will want to save rather than counsume, and firms won't want to invest, and so the real interest rate that would equilibrate desired saving and investment will be low too.

So if you want higher nominal and real interest rates, and you want them higher not just today but for the longer term future as part of a sustainable equilibrium, then you need to loosen monetary policy. If inflation and real growth, and hence nominal income growth, are in danger of falling, then you need more monetary stimulus to raise interest rates in a sustainable way.

Microeconomists are fond of saying that the best cure for high prices is high prices. Macroeconomists should be fond of saying that the best cure for low interest rates is low interest rates.

4. It's just horribly wrong.

The idea that more monetary stimulus might be desirable but would unfortunately reduce financial stability, so we shouldn't do it, is a bad idea. It's not enough to say "yes but..", and talk about the costs of unemployment and about other ways of promoting financial stability. There's no "yes" about it. Our answer should be: "No. That idea is horribly wrong".

[I had been thinking of writing something along these lines for some months, ever since I reviewed a draft paper on the idea. Then reading Simon Wren-Lewis' good but depressing post, about Lars Svensson's resignation from the Riksbank, (plus the fact I've now got my grades in), pushed me to write it now. While I generally agree with both Simon and Lars on this issue, I fear that both concede just a little too much to the other side.]