Usually I try to avoid engaging in spats on social media – they’re time consuming and invariably end up with both sides all the more convinced of their original positions. Nevertheless last weekend I found myself engaged in one with Ryan Bourne of the Centre for Policy Studies over a recent post on ConservativeHome by Andrew Lilico.
In it Lilico claimed that the very idea that government transfers could boost aggregate demand and, as a consequence, GDP is a “fallacy”. He writes (emphasis mine):
“The naive idea behind this is the intuition that if people spend more of their money, rather than saving it, that will boost growth. Except that's just wrong. In orthodox macroeconomic models, money that isn't consumed is invested. And investment boosts GDP as well as consumption. So when the economy is operating properly, even if we could costlessly transfer money from high savers to high consumers and would not distort their incentives by doing so, there would be no effect upon GDP - it's just that a bit more of that GDP would be consumption and a bit less investment.”
Now I think it’s important to address this point. Yes, at the most basic level orthodox economic models demonstrate that savings always equal investment. As Simon Wren-Lewis explains:
“In the most simple model of a closed economy without government, income (Y) = consumption (C) + saving (S), but also expenditure (Y) = consumption (C) + investment (I). So S=I by definition.”
The problem here is the so-called “paradox of thrift”. That is, if there is an external shock a reduction in consumption reduces output (and therefore incomes), offsetting any temporary rise in saving.
This fall in output can be exacerbated because people decide not to invest their excess capital but instead park it in bank accounts as money. During “normal” times this should not pose much of a problem as the bank can then lend that money out but this may not happen if, for example, the financial sector is in the process of deleveraging and de-risking.
Instead that additional capital could simply sit on the bank’s balance sheet in order to improve its loan-to-deposit ratio. Or as Wren-Lewis puts it – “the bank may just decide to hold on to the cash”.
During these periods there could be good reason for a government to seek to reallocate idle capital from those with a low marginal propensity to consume (MPC) to those with high MPC. As The Economist’s Daniel Knowles suggests, this should increase the velocity of money (eg what GDP is designed to measure).
To be fair to Lilico his central case is slightly more nuanced than the classical savings = investment model. He states that even if a government were able to transfer money to those with high MPC it may still fail to improve output:
“[It] isn't the amount of saving that counts here, but the amount of over-saving. The fact that one person saves a lot doesn't mean that she over-saves - it might be optimal for her to save a lot. And the fact that someone else hardly saves at all doesn't mean he doesn't over-save - ideally he might save even less! So transferring money from low MPC folk (say, "the rich") to high MPC folk (say, "the poor") might increase over-saving, rather than reducing it.”
Now this looks like a valid point. If we are saying that an external shock can make saving the overwhelming priority for people, then we should assume that it applies as much to those on low incomes as those on higher incomes.
But here is where this time really is different.
As the chart below shows, during almost all recent recessions prices have fallen much further than average earnings. During this latest downturn, however, the opposite has happened.
Household costs in particular have held up during the recent recession, shrugging off falling incomes to post in many cases inflation-busting price increases.
What this suggests is that, unlike during a traditional recession, those whose incomes were already hitting up against their expenditure before the crisis would have struggled to cut costs even if there was an overwhelming pressure to save. Many would have had to raid savings, sell investments or borrow against fixed assets just to meet current spending commitments.
Yes, this could also apply to those at the higher end of the income spectrum but they are likely to have been far less impacted by increases in household costs and would be in a far better position to adjust their finances towards saving (liquidating assets, cutting down on discretionary spending etc). Simply put, under current circumstances with the overwhelming incentive to save it is doubtful whether the concept of "optimal saving" is really relevant to the policy discussion.
Bourne, however, claims that people’s spending decisions are not based on short-term factors. In his recent blog post on the CPS site he writes:
“Those advocating redistribution now usually do so by advocating one-off transfers because of the current state of the economy. But economic analysis generally finds that people’s spending decisions tend to be based on permanent income, so this sort of one-time transfer is unlikely to have any significant effect even if you believe all of the assumptions above hold.”
Yet that was not what was suggested by the work of Christopher Carroll and Miles Kimball in their 1996 paper “On the Concavity of the Consumption Function”. In the paper the authors conclude that “adding income uncertainty to the standard optimization problem induces a concave consumption function in which, as Keynes suggested, the marginal propensity to consume out of wealth or transitory income declines with the level of wealth”.
Or, as Carroll later described in a NBER article:
“[In] the presence of uncertainty, households with low levels of wealth will respond more to a windfall infusion of cash than households with ample resources”
As such there is good reason to believe that precisely because of the current state of the economy transfer payments could increase aggregate demand. While during normal times savings held as cash in bank accounts may have made their way into productive investments, the dysfunction within the banking system at present makes that highly unlikely.
Now, simply because transfer payments could increase aggregate demand does not make them ideal policy. Both Bourne and Lilico rightly point out that there are significant deadweight costs to such policy moves. My preference would, instead, be for government to increase spending on infrastructure, which would have long-term benefits for the economy as well as the short-run boost to aggregate demand.
However, simply because it is not ideal policy does not make it a “fallacy”. The question Lilico, Bourne et al need to answer is whether the economic status quo is really better than the alternative that they attack?