Baumol's Cost Disease and Defence: Why 3 per cent of GDP may not be enough

Whilst there may well be valid reasons why the organisation chose a fixed percentage of GDP as the goal, it is not immediately clear whether the previous 2 per cent of GDP—or indeed 3 per cent as the UK has recently set as an ambition to be reached during the next parliament—is currently sufficient to maintain a credible level of military effectiveness and even more importantly, whether a fixed percentage of GDP is an appropriate long term metric.

GDP is measured as the market value of all final goods and services produced within an economy over a given period, excluding intermediate goods to avoid double counting. Year-on-year changes in GDP may be attributed to two factors: inflation, reflecting the average rise in prices, and real growth, driven by an increase in the volume of goods and services produced.[1]

Thus, for activities where the overall costs—e.g., wages of the participants and the prices of the goods and services required—increase at below-average rates, a fixed percentage of GDP may well represent a very fair approach to budget setting, as the inflationary increase in GDP will more than cover inflation in the cost of the activity, leaving the balance of that increase and all of the volume driven increase as real budget growth. However, in activities where overall costs consistently rise faster than average, the inflationary component—and potentially all of any volume-driven growth—may be absorbed, resulting in little or no real-terms budget growth.

The economist William J Baumol has written at length on the relative price changes of different economic goods and services. He explains that ‘if the prices of all commodities are not rising at the same pace, then some must be increasing at a rate above average’, which means ‘their inflation-adjusted—or real—prices must be rising’. ‘The list of those items whose real costs are rising remains roughly constant, decade after decade…the items in the rising-cost group generally have a handicraft element—a human element not readily replaceable by machines...which makes it difficult to reduce their labour content.’[2]

His first study on this topic, conducted jointly with Dr Bill Bowen in the 1960s, explored the apparent and persistent rise in the cost of live performing arts—namely theatre, music, and dance—in the United States. To illustrate their argument, the authors envisaged a simplified economy in which workers either manufacture motor vehicles or perform in string trios. Due to increasing productivity, it is possible to reduce the labour effort per motor vehicle by 4 per cent per annum. This means that the auto workers can receive an annual, real terms pay increase of 4 per cent without increasing the labour cost per vehicle. However, if the string trio perform at the same speed, the labour effort involved in a performance will remain constant. In this economy, if the wages of the string players rise at the same rate as those of the auto workers, then the cost per performance will also increase by 4 per cent per annum, as there is no opportunity for productivity improvements.[3]

Let us imagine that these two sectors—automobile manufacturing and live musical performance—are funded from a budget that remains a fixed proportion of GDP. If GDP grows by 2 per cent annually in real terms, the automobile industry, with constant real costs, would receive a 2 per cent annual increase. In contrast, musicians would face a 2 per cent real-terms cut each year, as their costs do not fall in line with productivity gains.

In these two simplified situations, the content of the tasks undertaken by the workers remains static, although the car workers become able to produce the content more quickly. In many industries, technological innovations allow higher-performing products or services to be produced, albeit involving increased complexity, labour content, and cost.

Since the founding of the UK’s National Health Service (NHS) in 1948, there have been major technological advances in diagnostic tests, pharmaceuticals, and medical interventions—contributing, for example, to a doubling of cancer survival rates over the past 50 years. At the same time, the cost of developing new drugs has risen sharply, and the NHS workforce has grown substantially, with ten times as many doctors in 2018 as in 1948.

These cost increases are illustrated in Figure 1. In 1948-49, the NHS consumed 2.4 per cent of GDP, rising to just over 7 per cent before the COVID-19 pandemic.

Despite rising expenditure, there is always scope to spend more on health—though often with diminishing returns. One of the key roles of the National Institute for Health and Care Excellence (NICE) is to assess the cost-effectiveness of different health interventions to ensure value for money for taxpayers. Central to this is the concept of the quality-adjusted life year (QALY), which estimates the additional quantity and quality of life an intervention is likely to deliver, weighed against its cost. This approach offers reassurance that effective and efficient treatment decisions are being made. By contrast, no equivalent framework currently exists to guide defence spending.

Figure 2 is a remarkable contrast to Figure 1. Over a similar period, defence spending declined from just over 7 per cent in 1955 to about 2 per cent. Part of this reduction reflects a narrowing of defence commitments—such as the abolition of National Service and the withdrawal from east of Suez. However, it has also been argued that the UK’s armed forces have been ‘hollowed out’ over the past 25 years as a cost-saving measure. This is illustrated by the reduction in the full-time strength of the UK regular forces, falling from 207,610 on 1 April 2000 to 138,121 on 1 April 2024. Whilst it is difficult to measure the capabilities of armed forces objectively, this reduction in manpower strongly suggests that maintaining defence spending at 2 per cent of GDP has reduced the UK’s capacity to undertake sustained military activities.[4]

Whilst continual efforts are made to improve NHS efficiency, it is widely accepted that the rising share of GDP allocated to health has been necessary to meet national aspirations for a high-quality healthcare system that remains free at the point of delivery. These increases have been driven primarily by rising staffing costs and the development of new drugs and therapies.

In defence, as in other industries, technological progress has significantly reduced the resources needed to perform administrative functions. In some cases, similar efficiencies have also been achieved in operational capabilities. For example, in 1943, generating the ‘Recognised Air Picture’—a comprehensive overview of all aircraft in UK airspace—required around 170 sites. By the 1950s, technological advancements had reduced this to approximately 66, and today, only eight sites remain, seven of which are staffed solely for maintenance purposes.[5]

In other cases—particularly where there is direct competition with potential adversaries—the development of new technologies has demanded significantly greater resources. For example, producing a modern Typhoon aircraft requires roughly 33 times more labour than was needed to build a Hawker Hunter in 1956. However, the reduced scope of the RAF’s operations means the Typhoon fleet is more than three times smaller than the Hunter fleet, so the total labour invested in Typhoon production is only about ten times greater. Moreover, as GDP growth has outpaced rises in average earnings, the cost of the Typhoon fleet, as a proportion of GDP, is ultimately only six times higher than that of the Hunter fleet. Similar affordability challenges may be observed in other types of high-technology defence equipment.

From these limited examples, we observe that while the administration of defence and a few operational outcomes are amenable to efficiency gains, there are significant areas where the resources required to maintain an acceptable level of relative operational effectiveness and deterrence increase significantly faster than GDP grows. This suggests that a fixed percentage of GDP is unlikely to be an appropriate way to set a defence budget in the long term. Therefore it appears likely that, to maintain effective defence forces, the UK will need to increase defence spending at a faster rate than the current proposed fixed percentage of GDP would provide.

---

1] Naturally if a recession occurs and the volume of goods and services decreases, then there will actually be a reduction.

[2] William J. Baumol, The Cost Disease: Why Computers Get Cheaper and Health Care Doesn’t (New Haven: Yale University Press, 2012), 19.

[3] Assuming that the musicians do not undertake additional performances for no extra pay.

[4] Similar arguments can be made about the number of aircraft that the RAF can afford to own and operate.

[5] These are the Remote Radar Heads.