In a recent post ^[1], Neil James made some interesting points about defence spending metrics and the political economy of defence in a democracy. With the federal budget due in three weeks’ time, I thought I might add some observations of my own on how to measure defence spending. A response to Neil’s provocative points on the politics of defence funding will have to wait for another day.

There are four measures of defence spending in common use: dollars, growth rate, percentage of GDP and percentage of government outlays, each of which gives useful and complementary information about the financial aspects of a country’s defence effort.

The most direct measure of a country’s defence spending is what it spends measured in its own currency. But it is also the measure most beset by complications. Comparisons of defence spending within a country over time are made difficult by inflation. Although it’s routine to talk about ‘real’ dollar figures adjusted for inflation, there’s actually no unique way of doing so. What’s more, the steady introduction of new products into the economy reduces the meaningfulness of comparing the ‘value of money’ over very long time periods, even within a consistently applied methodology.

Comparisons of defence spending between countries are even more fraught, buffeted as they are by ever-fluctuating exchange rates. Even within a single year, exchange rates can easily vary by 20%. Comparing defence spending between countries and over time is harder still because the already substantial vagaries of inflation and foreign exchange are compounded.

One way to avoid these complications is to calculate near-term rate of growth in defence spending. This works because inflation can be reasonably well (though still not uniquely) defined over the short to medium term. Knowing how much defence spending has risen or fallen over preceding years is useful information. Here in Australia, the figure of ‘3% real growth’ has become a talisman-like benchmark for the adequacy of defence spending since it was introduced in the 2000 Defence White Paper. If only it were that simple.

Expressing defence spending as a share of GDP (often termed ‘defence burden’) avoids a great many complications by calculating a dimensionless ratio ^[2], usually expressed as a percentage.

^[3]

Defence burden allows comparisons between countries and over time without reference to tables of exchange rates or inflation indices. But as Neil pointed out, there are still complications—most especially the ebb and flow of economic growth. In a recession, it’s entirely possible for GDP share to rise even as defence spending falls. Yet GDP share is still a valuable measure because it serves as a proxy of the importance assigned to defence as a national priority. High levels of defence burden indicate that a high priority is being given to defence, and conversely so. A significant limitation of this measure, especially as a comparator, is that it provides no indication of the absolute size of defence spending. For example, the denizens of Tuvalu could spend 100% of GDP on defence and we’d know that they were deeply worried about something. But, in the absence of an extraordinary set of events ^[4], their US$37 million wouldn’t do them much good if they took on the United States which ‘only’ spends 4.5% of GDP—around a lazy US$500 billion.

Sometimes, though not frequently, GDP per-capita and/or defence spending per-capita are raised in discussions of a country’s defence effort. Apart from reminding voters of the substantial personal opportunity cost they incur as a result of defence spending, it’s not clear what’s achieved by doing so.

The final measure in common use is the percentage of the federal budget allocated to defence—Neil’s preferred option. As a comparative measure between countries, this measure suffers from the wide international variation in both the scale of government spending in absolute terms (and as a share of GDP) and the variable demarcation between federal, state and local spending. Even within a single country over time, there can be substantial variations due to policy choices that shift revenue and expenditure from the state to federal level, or which arise due to the shifting boundary between the private and public provision of services such as health, education and retirement savings.

The impact of shifts in spending between levels of government can be avoided by aggregating spending at all levels to yield what might be called a national budget, but no such approach is possible when it comes to the private–public split, short of going back to GDP share. And, as with GDP share, changes to the denominator (the size of the federal or national budget) can alter the result irrespective of what happens to defence spending.

As illuminating as the various measures of defence spending are, they all ultimately suffer from focusing on financial inputs rather than military outputs. So while these metrics can help to capture gross trends and perform broad comparisons, they fail to say anything about what we are getting for the vast amounts being spent. Or to put it differently, they tell us a lot about the opportunity cost that we are incurring but nothing about the benefits we hope to accrue as a result.

Mark Thomson is senior analyst for defence economics at ASPI.