This CT post on Stiglitz and the cost of the Iraq war reminded me to get going on one I’ve had planned for some time, as a follow-up to this one where I pointed out that the $500 billion in aid given to Africa over the past fifty years or so is not, as is usually implied, a very large sum, but rather a pitifully small one, when considered in relation to the number of people involved, and the time over which the aggregate is taken.
What are the sums of money worth paying attention to in terms of economic magnitude. I’d say the relevant order of magnitude is around 1 per cent of national income, say from 0.5 per cent to 5 per cent. Smaller amounts are important if you’re directly concerned with the issue at hand, but are impossible detect amid the general background noise of fluctuations in income and expenditure. Anything larger than 5 per cent will force itself on our attention, whether we will it or not.
To get an idea of the amounts we’re talking about, US national income is currently about 12 trillion a year, so 1 per cent is $120 billion a year. A permanent flow of $120 billion a year can service around $6 trillion in debt at an interest rate of 4 per cent, so a permanent 1 per cent loss in income is equivalent to a reduction in wealth by $6 trillion.
For the world as a whole, income is around $50 trillion, so the corresponding figures are $500 billion and $25 trillion.
What kinds of policies and events fit into this scale?
* A typical “small war”, like Iraq or the French war in Algeria, big enough to be a central policy issue for the duration, but not to impinge on the day to day life of most people (in the US/France that is, not in the country where the war is being fought) . The direct costs of the Iraq war have been running at over 1 per cent of GDP, and costs (such as looking after wounded veterans) will continue at this level for many years after the war is over. So the Stiglitz estimate of $3 trillion is about the right order of magnitude, but will probably turn out to be on the conservative side
* A typical recession reduces GDP (the right measure in this case) by around 3 per cent relative to its trend value. In most macroeconomic models, this loss is made up over a few years in the recovery phase, but I doubt that this is correct. I believe that some of the income loss in a recession is permanent, or nearly so, and countries that experience a string of bad macroeconomic episodes (for example, New Zealand) tend to lag others (like Australia) for a long time afterwards
* The amount promised by rich countries in development aid is 0.7 per cent of their income, which is in the range being discussed here. The amount actually delivered by most of them (around 0.3 per cent) is not. The amount required to achieve the UN Millennium goals is a bit over 1 per cent of rich country income. To give everyone in the world an income of at least $2 a day would cost around 2 per cent of the income of the rich countries (assuming 2 billion poor people with an average income of $1/day, doubling this would cost around $700 billion a year).
* Chronic national budget problems like the long-term deficit in the US Social Security fund typically involve a gap between revenue and expenditure of 1-5 per cent of national income
* The losses from the current financial crisis seem certain to be at least $500 billion and maybe more than $1 trillion. Allocated over, say, five years, this is comparable to a modest recession
 People mostly refer to Gross Domestic Product, which is much the same magnitude, but not a useful concept except in measuring shortrun economic activity