Fiscal Policy - Comparing Apples with Apples

In today's The Australian, Federal Oppostion leader Malcolm Turbull writes

The gross domestic product numbers for the December quarter showed growth was negative. It was perfectly plain that the $10 billion December cash splash had been almost entirely saved: household savings were higher than they had been for many years....

So the cash splash had failed as an economic stimulus. .....

Rudd has chosen to borrow tens of billions of dollars to spend in ways that simply will not deliver an effective boost to the economy: too little bang for too much buck....

We have offered the Government the opportunity to agree on measures that would cost less and be more effective because, like tax cuts and investment incentives, they would benefit every business across the board.

Because, whether it be fiscal measures or investments in infrastructure, the priority must be to spend money to improve the productivity and efficiency of the whole economy and every business.

Shaun Carney argues in today's Age that such comments are "at best unproved and at worst nonsense".

The Opposition's claim, for example, that the December quarter national accounts prove that the $10 billion stimulus package — which it supported — had made a bad situation worse is at best unproved and at worst nonsense.

The money from the package did not reach taxpayers and consumers until the final three weeks of the quarter. And the Government expected, based on Treasury modelling, only about 10 per cent of the money to be spent around Christmas time. That is equivalent to $1 billion.

At this stage it looks like something close to that amount or more was spent. It's true that the December quarter accounts did show a substantial rise in savings, suggesting that a fair whack of the stimulus package went into paying off mortgages and credit cards. But the Government expected all along that about $3 billion would go into savings during the six months in which the package was to have effect. And savings are, in reality, deferred spending anyway, so it's too early to tell if the package was a dud.

Few in the Government made much of a fist of making these points this week.

And just as well. While it is correct to point out

• the complete absence of a counterfactual in Turnbull's argument (ie what would have the December retail sales figures have looked like in the absence of the lump sum payments?), and
• as payments commenced in the week of December 8th, one would not expect a significant portion of those payments to be reflected in the December retail sales figures, and
• even though only a small proportion of the payments were immediately spent, the increase in retail sales appear to have been sustained into January, and
• it was the decline in inventories in the December quarter that led to the negative growth result

Carney's claim that "it's too early to tell if the package was a dud" misses the point. Effective fiscal stimulus policies share two features: they are implemented in a timely manner (ie when the stimulus is most needed to employ idle resources), and; they increase economic activity as much as possible for a given budgetary cost. While Turnbull emphasises the latter point, he confuses long term productivity growth with short term aggregate demand expansion.

The fact that fiscal policy is subject to lags (both in terms of recognition and implementation) is well known. Even if the recognition lag can be overcome, policy makers are likely to find the implementation of effective short-term fiscal stimulus to be particularly challenging. As a tool of expansionary fiscal policy, an advantage of immediate lump sum payments is that they are associated with shorter implementation lags than, for example, infrastructure spending, and therefore have the potential to increase aggregate demand, and hence output and employment, in the short run. David Gruen, executive director of the Macroeconomic Group in the Australian Treasury argued in February that

we think without this package the greatest period of weakness for the economy is the whole of 2009 and that there are benefits in spending some significant proportion of money earlier

The benefits referred to here are clearly the potential short run increases in output and employment.

Where does this leave Turnbull's 'bang for the buck' or cost-effectiveness concept. Let us assume that Turnbull is referring to the value of the fiscal policy multiplier, ie the change in real GDP over a given period divided by its lifetime budget cost (ie reduction in tax revenue or increase in spending).

It is relatively uncontroversial, based on consistent empirical evidence as well as economy theory, that the multipliers associated with lump sum payments or rebates are smaller than government spending multipliers. The size of the lump sum payment multiplier will depend on factors such as the proporption of households whose consumption depends on perceived permanent income, and the expected cyclical position of the economy. Indeed, last October, following the limited expansionary impact of mid-year tax rebates in the United States, Paul Krugman called for a major fiscal stimulus that would

... take the form of actual government spending rather than rebate checks that consumers probably wouldn’t spend.

However, when we are discussing a 'bang for a buck' we need to compare apples with apples. Only the cost of fiscal stimuli that have the same intended temporal outcomes should be compared, ie we should compare the impact of lump sum payments to those policies that are also designed to have a short run impact on income and output. If policy timing matters (which it does) it is inappropriate to compare the multiplier associated with lump sum payments to the multiplier associated with infrastructure projects.

The US Congressional Budget Office published a paper in January 2008, Options for Responding to Short-Term Economic Weakness which attempted to compare apples with apples, by considering the cost-effectiveness of policies with similar enactment-stimulus lags. Only two policies with short enactment-stimulus lags were identified: across-the-board tax cuts and direct transfer payments to households. The conclusion? The cost-effectiveness of the tax cuts was estimated to be 'small', while the cost-effectivess of the transfer payments was estimated to be 'large'. As the range of transfer payments to households considered by the CBO appear to be more narrowly targetted than those in December stimulus, this points to the limited number of fiscal policy options available to generate both rapid and substantial increases in output and employment. Coupled with the fact that other components of spending in the economy are likely to be relatively weak over the next few months, all available fiscal policy tools, particularly direct government spending, will be required.