To borrow a disparaging phrase from Paul Krugman, the concept of ‘permanence’ in economics seems “to have a special ability to create intellectual confusion, even in smart people.”
Personally, I don’t really like Krugman’s “smart” category. Often, innocent questions cause economists the greatest problems (i.e. “What exactly is ‘money’?”).
But the concept of permanence appears to be tripping up extremely astute and thoughtful economists. We should know better: nothing is permanent.
The idea rose to prominence in economics with Milton Friedman’s permanent income hypothesis, which focuses on lifetime income when considering the effects on saving and consumption of changes in income. The concept has usefulness when thinking about taxation – although it’s empirical track record is debated. It is reasonable to expect a tax cut to have more impact on spending, if it is enduring, rather than one-off.
In most conventional macroeconomic models, “permanent” means something. Most models try to identify the effects of policy-changes on forward-looking economic agents – because investment and consumption decisions are made in the context of the present and the future. Often, macro models of consumption behaviour simplify by considering behaviour of individuals over two periods or over infinite periods. The mathematics is tractable, and there can be useful insights.
“Permanence” in these models is relevant because consumers and firms are typically modelled to behave very differently if something affects them in the first period only, or if its effects are enduring – which is modelled by “permanent” changes, to incomes, inflation, etc.
In this context, either when thinking about the effects of changes in taxation, or in simplified models of decision-making over time, “permanence” means something and it’s useful.
Where Krugman’s ‘smart’ people seem to go slightly crazy is when applying this to empirical reality, and helicopter drops in particular (for clarity, read Simon Wren-Lewis).
It was also Milton Friedman who first used the metaphor of a “helicopter” dropping dollars from the sky, in his abstract paper on the optimal quantity of money. Friedman’s model in that paper is extremely simple – it is designed to produced a relatively mechanical link between money and inflation. So money demand is fixed by a rule, there is only one financial asset, and the only money that exists is that which is dropped out of helicopters.
Now, Friedman’s paper on the Optimum Quantity of Money has nothing to do with what to do at zero interest rates or in a liquidity trap. It’s a model for thinking about the opportunity costs of holding money (and a simplistic one at that).
When Friedman actually discusses how monetary policy can work in a liquidity trap he makes no mention of “permanence” – because it is not relevant. In his 1968 American Economic Association Presidential address, Friedman discusses what to do when conventional monetary policy no longer works. He discusses QE, saying swapping one asset for another will have no affect – it leaves wealth unchanged. He discusses secular stagnation – that investment spending may no longer respond to lower rates; and concludes that in these conditions, conventional monetary policy is broken. In fact, Friedman provides a lucid and perceptive summary of everything that is supposedly novel about the economic environment of the last ten years.
He then proceeds to describe how it is recognised (in 1968!) that monetary policy still has potency. He doesn’t actually use the word ‘helicopter’, but he aptly describes the necessary policy: transfer payments financed by base money creation. Why would they work? Because the private sector’s net financial wealth would rise. And he never mentions “permanence” – nor does Haberler, to whom he attributes this policy insight.
To Haberler and Friedman, and anyone who thinks about it, or looks at any empirical evidence, it is obvious that transferring cash from the central bank to households will raise spending. The only debate is by how much.
Curiously, when Krugman analyses the liquidity trap he assumes that base money is expanded through open-market operations, which have no effect at zero interest rates, because base money and bonds become perfect substitutes. But Friedman explicitly pre-empts this. He says open-market operations have no effect at the zero bound, but money-financed transfers do – because they raise net wealth (a point Adair Turner correctly emphasises).
This omission has sent Krugman, and I think many others, down a blind alley. By ignoring the wealth effect, monetary policy can only work by shifting inflation expectations and reducing real interest rates (it is of course simply assumed that ever lower real interest rates work – so abolishing the lower bound, by assumption, is an equivalent solution).
Krugman continues along these lines by effectively implying that the central bank must ‘permanently’ raise base money to a level inconsistent with future stability of inflation. This becomes the only way out of the trap. Michael Woodford seems to take the same position.
Krugman, Woodford, and others, have, I think, confused policy makers into thinking that helicopter drops by necessity require permanent changes in base money – which for obvious reasons creates resistance.
Helicopter drops, in the form of cash transfers to households, are not the same as tax cuts, and households don’t know what the monetary base is – nor do they care about its future path, for very good reasons.
The modelled world of infinitely-lived, forward-looking consumers reacting to tax cuts does not map intelligently onto the real world policy example of a cash transfer from central banks to households. No one is suggesting that the central bank is going to make a cash transfer and a week later levy a lump sum tax. In stark contrast, the central bank is going to say something like this:
‘We are going to transfer $250 every three months to all adult citizens for the foreseeable future until certain macroeconomic conditions are met. We expect this to last for 12-18 months.’
All the empirical evidence that we have on how individuals respond to transfers analogous to this suggests that demand will rise. We know it will. Individual household circumstances matter – low income, credit-constrained families are likely to spend more, some families will save it, some will repay debt, etc. But the net effect is an increase in demand, and the other effects – higher savings, debt repayment, are all helpful.
These effects are known without any mention of “permanence”.
Now what if a clever(-dick) journalist asks the central bank governor announcing this policy, if it is ‘permanent’. It would be entirely reasonable for the governor to ask, ‘in what sense?’
Clearly, central banks don’t have the power to levy lump sum taxes, nor would they want to – they are trying to raise spending.
But the journalist persists: ‘is the increase in the monetary base permanent?’
To which the central banker should reply:
‘Well, we don’t know. The future path of the monetary base depends on lots of factors – the demand for base money, asset growth in banking sector, which in turn is influenced by the demand for credit, the regulatory framework etc. Furthermore, we don’t know how large the output gap is, nor how responsive the supply-side will be in response to an increase in demand. We expect to see a strong recovery in investment spending and corporates exploiting many new technologies – all of which will raise productivity. We expect inflation to stay low, but if the output gap is larger than we estimate, inflation may even fall and we can keep doing lots of helicopter drops until a relatively high sustained rate of growth and incomes is occurring. Which, by the way, means that fiscal positions are likely to improve dramatically. For exactly these reasons, households shouldn’t really prioritise changes in the monetary base – and there’s lots of ways we can alter the quantity or the demand for the monetary base without impacting households – so it’s really a separate question. From a household’s perspective, all that has changed is that their financial wealth and economic prospects have improved.’
It is precisely for these reason that I have argued there is a ‘constant reserves multiplier‘, analogous to the balanced budget multiplier.
I think Krugman gets closest to this position in his “double-super-special-wonkish” discussion of the Pigou effect – although, again, he forgets Milton Friedman’s 1968 discussion of Pigou.
Krugman seems to dismiss this line of argument as “fiscal policy”. That is an arbitrary distinction, and I think fundamentally incorrect. Again, I side with Friedman on this (he’s the only economist I’m aware of that holds a rigorous distinction) – monetary policy is policy implemented by changes in base money. What is done with base money may determine who should make the decisions, but money is money, and taxes and bonds are not money. So Ricardian Irrelevance is of course an even more irrelevant consideration when considering the efficacy of helicopter drops. Or should I say, it increases the probability of success. A sensible household receiving a check each quarter from the central bank should expect improved fiscal positions prospectively – after all fiscal policy-making in much of the developed world is so perverse that taxes are most likely to go up in depressions!
A brief comment on Tony Yates’s recent take on these issues. This is the crux:
“The first, instantaneous effect of either policy [helicopter drops or negative interest rates] is the liquidity effect. More liquidity lowers its price. The second effect, if either policy is ultimately successful in raising expected and actual inflation, is the ‘Fisher effect’. This will, when the economy returns back to the inflation target which otherwise would have been missed, lead to higher nominal interest rates than if the Fed had settled for lower inflation.”
My divergence from Tony is very clear. I don’t buy either liquidity or Fisher effects. The only effect I take for granted from cash transfers is a rise in nominal spending of uncertain magnitude. This may or may not raise inflation. If supply is raised by demand – plausible in a near-depression coinciding with rapid innovation – inflation might even decline. The Fisher effect requires a model of inflation determination of implausible character and certainty. As regards, the liquidity effect, I think the effects of cash transfers could be radically different to that of negative interest rates. At the moment negative “interest rates” are coinciding and perhaps causing risk premia to rise – i.e. market interest rates on risks assets are rising not falling. Cash transfers would likely cause equity risk premia and credit risk premia to fall – because growth expectations would rise and private sector balance sheets are strengthened.
Tony’s perspective is also vulnerable to the likelihood that, when it comes to money, supply creates its own demand.
So there is little doubt that helicopter drops would raise nominal demand. The only uncertainty is by how much. All the empirical evidence shows some effect from analogous policies, although the size of effect varies considerably. A couple of additional points are worth bearing in mind. As economists we sometime unthinkingly focus on the size of the multiplier, assuming that the targeting the highest marginal propensity to consume is optimal. But this is only one consideration. The fact that cash transfers might be saved is no bad thing – we have a debt overhang after all. Similarly, there is no necessary welfare to be gained from higher inflation. If demand alone can be raised it would be even better. Of course, it would be optimal to use cash transfers to mitigate some the of the effects of income inequality – at a minimum, equal cash transfers would result in a more equal distribution of utility.
Empirical papers on tax rebates