It is common to set economic policy targets as minimum percentages of GDP for the corresponding expenditure. Examples: Trump demands 5% of GDP for defense from Europe, the EU 3% for R&D investment, the OECD 1% for environmental spending, Spanish university laws 1% for universities, and a proposed popular legislative initiative (ILP) 6% for education. There are also targets, less tied to specific policies, that demand maximum levels. For example, the EU Stability and Growth Pact sets a maximum fiscal deficit of 3% of GDP.
The methodology is attractive. It is very specific, in principle, verifiable, and typically an easy way to demand more resources for good causes. However, its effectiveness is questionable because it suffers from intrinsic problems that, in practice, render it ineffective. I will describe several of these.
1. Efficiency. Spending more doesn't mean spending better, and there are many situations where the only realistic possibility of improvement is to gain efficiency. Making increased spending the central demand is often a way of sidestepping the underlying problem, with the result that, since everyone understands what's happening, the failure doesn't prompt a call to action. Consider Trump's demand for a 5% increase in defense spending. Europe only needs the increase if it maintains a non-integrated policy, with 27 separate militaries, each with its own weapons systems, tanks, and perhaps a submarine or two. The height of inefficiency. Since what matters most to Trump are orders from American companies, and since we'll be getting those, things won't get out of hand. But there will be a cost in terms of efficiency.
2. Interpretation of the content. It can be maximalist or minimalist. Minimalism is associated with the term spent with a subtype of spending that is of particular interest to a certain group of proponents. For example, spending on education can be public or private, but it's easy to detect a sense of disappointment among many proponents if the commitment is met through an increase in private spending. The maximalist approach wants to include enough spending components to make fulfilling the commitment easier and to showcase it, which may be reasonable. On the issue of European defense, it's expected that Europeans will argue that, in the age of AI and drones, the distinction between internal and external defense has blurred, and that it's necessary to include all types of cybersecurity investment in the calculation of defense spending. When the Clínic is attacked, the strategic security of Europe is attacked.
3. Asymmetry. If the purpose of the target is to stimulate spending, then the effect will be present when the actual figure falls below the target. If it falls above it, the effect can be discouraging for public action: when there are many demands, it may be preferable to meet the maximum requirements rather than exceed some. This is a relevant observation when interpreting the 3% target for R&D spending in the EU as a requirement in each member state, since some are exceeding it.
4. The denominator. A measure expressed as a percentage of GDP has a denominator, GDP, which varies and even fluctuates over time. But many of the spending components (education, maintaining an army) fluctuate less. This means that the requirement can be met to varying degrees in two years when spending is identical. One year we'll congratulate ourselves, and the next we'll be worried.
5. Disaggregation. Let's consider a hypothetical federal state where the principle is that public spending per student on compulsory education should be the same. Let's also assume that educational authority rests with the federated states, and that the overall public spending on compulsory education should be 6% of GDP. Since the demographics and GDP of the states differ, their spending as a percentage of their GDP will vary. The more prosperous states (or those with relatively fewer students) will spend less than 6%, and the less prosperous states will spend more. For example, consider two states, A and B, with the same population and 100 students each: in state A, GDP is 100, and in the more prosperous state B, it is 120. The total GDP is 220, and therefore, the overall spending on education is 13.2. The spending per student would be 13.2/200 = 0.066. Therefore, spending in A and B is 6.6%. This represents 6.6% and 5.5% of their respective GDPs. The difference could be lessened if educational costs were higher in B or if the characteristics of its economy generated greater demand for non-compulsory education. In this case, the private component to be added would be greater in B than in A.
