Bad banks, and the effectiveness of fiscal and monetary policies

I'm going to take another crack at this topic. Do bad banks (and a bad financial system) reduce the effectiveness of fiscal and monetary policies in shifting the Aggregate Demand curve to the right? The answer matters, because if they do reduce the effectiveness of fiscal and monetary policy (a lot), then we need to fix the banks and financial system first. If not, then we shouldn't wait. Plus, using fiscal and monetary policy to shift the AD curve to the right might be one of the most effective ways to help fix banks and the financial system. The risk of deflation and depression is presumably one of the most important things making loans riskier.

In my previous post, I was looking at the broader question of whether bad banks shift the AD curve, AS curve, or both. Just to be clear (because I am really not sure whether Brad DeLong thought I might be saying something different), I believe that the current crisis was caused by the AD shifting left, not AS shifting left, (and that bad banks and bad financial system are a large part of the reason why AD shifted left). The empirical evidence is clear: we have decreasing not increasing inflation; and lots of stories about firms unable to sell goods, and no stories about anybody being unable to buy goods. It is possible that bad banks caused the AS curve to shift left as well, though by a lesser amount than AD shifted left. Output is currently demand-constrained, not supply-constrained.

Let's take a really extreme and therefore simple case. Suppose that banks, and the financial system generally, go into a total freeze. No new loans.

OK, I need to make an exception to my assumption of "no new loans", otherwise it is hard to even talk about fiscal and monetary policy. The government and central bank can still borrow and lend. I mean no new private loans.

Obviously, this is not good news. AD (and AS) would shift left, and a lot of bad microeconomic things would happen too. We know that. But starting from that new and very nasty equilibrium, would it prevent fiscal and monetary policy from shifting the AD curve to the right?

In other words, we are not talking about the effect of bad banks on AD; we are talking about the effect of bad banks on the effect of fiscal and monetary policy on AD. It's too easy to get those two things muddled: confusing a partial derivative with a cross-partial derivative.

In the new nasty equilibrium, households who want to spend more than their current income can only do so by reducing their holdings of cash and government bonds. Those who do not hold cash or government bonds will be constrained to spend no more than their current disposable income. Firms likewise will be constrained to invest only their retained profits, unless they (implausibly) can reduce their stocks of cash and government bonds. Lots more households and firms will be borrowing constrained.

Those borrowing constraints may be bad news for AD, but that's not the question. Can fiscal and monetary policy still help make things less bad? That's the question.

The borrowing constraints are actually very good news for the effectiveness of one type of fiscal policy -- tax cuts. Ricardian Equivalence is false for borrowing-constrained households and firms. They are spending all their disposable income, so a small cut in taxes (one small enough so the constraint still binds) will be spent 100%.

Households and firms who are not borrowing constrained, either because they hold stocks of cash and government bonds, or because they don't want to spend more than their disposable income anyway, will presumably respond to a tax cut in the same way as they would if private credit markets were still open. So the net effect of frozen private credit markets would be to increase the effectiveness of tax cuts as a fiscal policy.

What about the effectiveness of (bond-financed) increases in government expenditure? Borrowing constrained households and firms have a marginal propensity to spend of one. With private credit frozen, a higher percentage of households and firms will be borrowing constrained. If the marginal propensity to spend of unconstrained households and firms is less than one, that increases the fiscal policy multiplier.

Now it's possible that the marginal propensity to spend of an unconstrained household or firm is greater than one. An unemployed worker who gets a job borrows to buy a new car; a firm with increased sales borrows to finance new investment. In ECON1000 we normally rule out marginal propensities to spend of greater than one, and an infinite multiplier, because it makes the Keynesian Cross equilibrium unstable. Successive rounds of induced spending get larger and larger. "Who has ever seen an egg standing on end?". But if interest rates increase when output increases (as they normally do, and do in the normal ISLM), a marginal propensity to spend greater than one, and an upward-sloping IS curve, does not give an unstable equilibrium (provided LM is steeper than IS, and interest rates adjust to satisfy LM before output adjusts to satisfy IS).

We can't rule out the possibility that bad banks would reduce the marginal propensity to spend, and so reduce the fiscal multiplier. But it could only happen in a world where typical households and firms had marginal propensities to spend greater than one, which would be a world where the fiscal policy multiplier would be very, very large anyway, if central banks kept the interest rate fixed. So I'm not going to worry much about that possibility.

What about monetary policy? Would a freeze on private lending reduce the effectiveness of monetary policy?

A freeze on private lending creates one immediate problem for monetary policy. The central bank can expand the quantity of currency in circulation, but demand deposits are loans from households and firms to banks. If private loans are frozen, demand deposits cannot expand by definition, so a major portion of the money supply cannot expand.

That was a problem in the 1930s, but is not a problem now. Putting money in a chequing account is not a purely private loan to a private bank if the government and central bank insure demand deposits and act as lender of last resort to prevent bank runs.

So demand deposits can expand if the central bank buys government bonds from households and firms with currency, and they in turn deposit that currency in a bank, covered by government and central bank guarantees.

But what if the central bank instead buys government bonds from banks? Normally, those banks would increase lending to households and firms, and increase demand deposits by doing so. That can't happen in this case, because we have assumed no new private lending. The banks would have to just sit on the extra cash.

Unless I am muddled on this (always possible): a freeze on new private lending totally blocks the normal channel through which monetary policy operates (via increasing bank reserves and loans); but central banks can still use any other way to get new cash directly into the hands of households and firms.

Suppose the central bank buys a government bond from a household, for $100 cash. The household deposits the cash in a bank, so demand deposits increase by $100. Then it stops. The extra $100 in the money supply can be spent, and pass from hand to hand increasing AD, but the money supply cannot expand by any more than that original $100 cash injection, because bank lending is frozen, by assumption.

So an assumed total freeze on new private lending would reduce the effectiveness of monetary policy. The central bank could only expand the money supply by injecting new cash directly into public hands, not via the banking system. And it could only expand the money supply 1:1 with new cash injections.

Comments

[TK]

>But what if the central bank instead buys government bonds from banks?...The banks would have to just sit on the extra cash.

Which is exactly what is happening - 724.6 billion excess reserves from 1.8 billion a year ago.
http://www.federalreserve.gov/releases/h3/current/h3.htm

So how does the central bank buy bonds from households?

[Nick Rowe]

TK: (TKR, by any chance??) That is exactly what I was thinking when I wrote that bit. And that was the question I was asking myself. But then my brain got tired.

But "household" does not literally mean household. It just means anything other than a bank (a financial institution whose liabilities are media of exchange, like demand deposits). So "household" could mean "mutual fund", or "pension plan", for example.

[Adam P]

But Nick, at least in the UK, the BoE does buy gilts from hedge funds, real money and other institutional investors. They're not just buying them from banks.

[Nick Rowe]

Adam: that makes sense to me, but I did not know if it were true. Did "QE" make a difference in who sold the Bank of England the gilts?