IMF Headquarters, Washington DC
By Michael Burke
It is unusual for ‘academic’ research published by the IMF to find its way into popular media. But this has happened to the latest World Economic Outlook where the IMF deals briefly with the issue of ‘multipliers’ that is, the economic impact of changes in government spending.
The short article has caused an usually high level of commentary among economists and commentators because the data suggests that the multipliers are perhaps more than double the level generally implied by official research and forecasts. Nobel Laureate Paul Krugman has commented that the research shows that, ‘the reason for the worsening outlook is that policy makers have gotten the basic economics wrong’. In Britain Chris Giles economic editor of the Financial Times has led a counter-attack by questioning the validity of the research. A string of other commentators have joined the debate on both sides, including a Greek finance minister.
The key point in the IMF research is that the multipliers are much higher than previously thought by leading bodies such as the IMF, OECD and others. ‘The main finding, based on data for 28 economies, is that the multipliers used in generating [IMF] growth forecasts have been systematically too low since the start of the Great Recession, by 0.4 to 1.2….’ Whereas the IMF’s (and others) own forecasts implied a multiplier of 0.5, the actual multipliers may be in the range of 0.9 to 1.7.
It is useful to assess why this seemingly arcane debate has created such controversy and why that is taking place currently.
Importance of the controversy
Because of the division of labour all changes in production have a wider impact on the economy. Increased output of one sector necessarily requires increased inputs either of labour, or of capital, or of production goods or raw materials, or some combination of all factors. This is not necessarily true of changes in incomes (which might be saved) or even expenditure (which might be met from existing stocks).
There can be no single multiplier effect. The size of the impact of changes in government spending must depend on firstly on the type of change in government spending. At the same time, even where an increase or decrease in government spending has a very large impact in terms of altering output in other sectors, the impact is not infinite. The size of the impact is itself constrained by the existing capacity of the economy. Therefore the largest multiplier effect can be found where government spending requires the greatest degree of inputs from other sectors (that is, where the division of labour is at its highest level) and which increases the productive capacity of the economy as a whole.
As a result, the overwhelmingly majority of research finds that the greatest multiplier is attached to direct increases in government investment. These are usually held to be much higher than inducements to private sector investment (which may simply be saved and from which profits must be deducted). They are also higher than the multipliers attached to consumption (which does not increase the productive capacity of the economy).
In Table 1 below the Office for Budget Responsibility’s (OBR) own estimates of the different multipliers are set out. These were first published in June 2010 and have not been altered. They are characteristic of thinking among most policymakers.
The OBR has not altered these estimates even though in its latest publication its estimates show GDP will have grown by 3.6 per cent over 3 years compared to their forecast of 7.8 per cent. In addition, the OBR concedes that an assumed average multiplier of 1.3 would fully explain the shortfall in growth compared to its own forecasts. 1.3 is in the middle of the IMF’s 0.9 to 1.7 range.
It is the insistence on unchanged estimates of the multipliers which is most significant, rather than the OBR’s own very poor forecasting record. In particular there is virtually a religious Golden Rule in (semi-) official literature which places the upper limit on all multipliers at 1. A multiplier lower than 1 implies that GDP will be reduced by less than the total change in government spending. Implicitly, the private sector will always respond in the opposite direction, increasing its spending when government reduces its spending, and vice versa.
This is the crux to the whole debate on multipliers. If government spending ‘crowds out’ private spending then it should be avoided as detrimental to total economic activity. At the same time, it is claimed that ‘austerity’ measures will not prove damaging as they will be offset by increased private sector activity. This false logic explains why the OBR forecast a 20.3% rise in business in the last two years while the actual increase has been 2.5%.
It is also extremely rare that the literature acknowledges the multi-year impact of the multipliers. Very little private investment achieves its return within 12 month. The same is true of government investment. The full yield on investment in transport takes place over the life of the railway or bridge, investment in education is returned over the working life of the pupil, and so on.
Essentially, the insistence on 1 as the ceiling for fiscal multipliers and confining them to a single year is to minimise the role of the state in the economy and its capacity to determine the trajectory for the economy as a whole, in both directions.
Timing of the controversy
This is not the first time IMF research has estimated the multipliers at very high levels. Daniel Leigh, one of the co-authors of the latest piece has previously written in a chapter of the IMF World Economic Outlook (‘Will it Hurt?’) that the 5-year multiplier of a cut in government investment when interest rates are close to zero and main trading partners are also cutting is 6.
This finding was almost exactly mirrored in an IMF Working Paper (‘Effects of Fiscal Stimulus in Structural Models’ ) which used its own econometric model and those of six other institutions including the OECD, the US Federal Reserve Bank, the EU Commission, and others. The key finding was that the 5-year impact of an increase in government investment has a multiplier of 5 or more.
Since the latest IMF research is shorter, no more prominent and modestly focuses on 1-year outcomes compared to previously published research, the unusual controversy it has generated must be due to other factors. Almost simultaneously, writing in the Financial Times, ex-US Treasury Secretary and current Harvard Professor Larry Summers highlighted ‘deep differences of opinion…..between the ‘orthodox view’ (which supports austerity)….and the ‘demand support view’ (which pushes for steps to increase demand in the short run]’ (‘The world is stuck in a vicious cycle’).
The deep split within mainstream economics and the controversy over the IMF research are associated with the same trends. Countries which have adopted the ‘orthodox view’ have generally experienced sharp slowdowns and renewed deterioration in government finances. Other countries, such as the US and Germany, which have ‘supported demand’ have experienced mild but slowing recoveries. In the US this has required the maintenance of large budget deficits. While the latter policy has clearly been more effective in restoring growth it is not sustainable over the medium-term. There is little official enthusiasm in either Germany or the US for further measures to support demand. In effect both ‘austerity’ and ‘demand support’ are running out of road.
Across the OECD government spending rose during the crisis in 2009 and 2010. But this was already being reversed by 2011. Belgium, Denmark, Greece and Slovenia were the only counties where total government spending rose as a proportion of GDP in 2011. In every other OECD economy where it was recorded, government spending fell, as was the case in the OECD as a whole. Despite government investment producing the biggest impact on the economy, all the OECD economies cut investment in 2011. The solitary exception was Denmark, where government investment rose fractionally.
The well-known efficacy of increased government investment has not prevented it from being cut in all the major economies. SEB has previously shown that the cut in government investment entirely accounts for the latest recession in Britain. Evidently, even though government investment would restore growth or even improve government finances it cannot be countenanced as it would interfere with the prerogatives of the private sector.
Conclusion
The debate on the size of the multipliers is effectively debate about the role of the state in determining economic activity. The official literature tends to minimise both the scope and the timescale of the impact of changes in government spending. The latest controversy arises because in their different ways both ‘austerity’ and ‘demand-support’ policies are failing.
The alternative is government-led investment. But this has been cut in all the major OECD economies in order to facilitate an increase in the profits of the private sector. The opposite policy is required, one which increases government investment in order to boost growth, jobs and the productive capacity of the economy.
First published on the Socialist Economic Bulletin site on October 23rd and re-published with permission.
