Service boom to be wary of

It is too soon to say what the shift to a service dominated economy will mean, but the implications could be profound

THE growing importance of services is a familiar trend across most advanced economies. For better or worse, Britain is in the vanguard of this trend, with services now accounting for two-thirds of total output (GDP) and three-quarters of all jobs. Only the US economy is more dominated by services.

The growth of Britain's service sector is mainly a success story, not, as some have claimed, just a passive reflection of the relative decline of manufacturing.

Since 1970 the production of marketed services (ie, excluding government) has grown at 3 per cent per year, compared with less than 1 percent per year growth in manufacturing output. The UK's international trade in services has consistently been in surplus, compared to a significant deficit in the trade in goods. In fact, last year despite the strong pound and robust domestic demand, the UK recorded its first overall current account surplus for more than a decade due in no small part to the record surplus in services. There are now five million more jobs in services than in 1970, while the number of jobs in manufacturing has shrunk.

But while the shift towards services is clear, its implications for economic policy are not. It may be, of course, that the economic characteristics of services are similar enough to those of goods that conventional macro measures (such as capacity utilisation and the output gap) and the policy rules derived from them are fully adequate. Even then, if there are greater measurement problems in services that make it hard to capture improvements in quality or productivity then economy-wide measures of growth and inflation may become increasingly distorted.

A further worry is that the effect of a policy decision (say, to change interest rates) on the economy may be different when the great bulk of producers are service companies. The export orientation and capital intensity of a firm will affect how sensitive it is to changes in exchange rates and interest rates. To the extent that service companies differ, on average, on such measures from other sectors of the economy, their growing share may affect the optimal policy response.

Consider capital intensity. On average, manufacturing firms have a higher capital/output ratio than service firms. The interest rate is a key determinant of investment because it is an important part of the cost of capital for firms. If manufacturing firms use more capital than service firms to produce the same amount of output, then a rise in interest rates will raise manufacturers' costs proportionately more and be a stronger brake to their expansion. As the share of services in total output grows, it might require a bigger rise in interest rates to achieve the same dampening of demand in the overall economy. Of course, interest rates operate through many channels other than investment so the story is not so simple. But to the extent that investment is an important part of the transmission mechanism for monetary policy, the rising share of services in output could mean that interest rates would have to move more over the course of the economic cycle. This would make manufacturing industry even more cyclical as it bore the brunt of steeper interest rate rises and reaped the benefit of sharper cuts. Such a roller coaster of interest rates would be no fun for mortgage holders either.

Fortunately, at least in terms of capital intensity and investment, there appears to be a convergence underway between the services and manufacturing, sectors. Since the end of the early 1980s recession, the share of investment in total output (ie, the investment intensity) of services has been rising while that in manufacturing has been falling. Over the last decade service firms have actually invested more, as a share of output, than have manufacturing firms. This has shrunk the gap between their capital/output ratios. Whereas in 1982, the ratio in manufacturing was roughly double that in services, by 1997 it was just 55 per cent higher. This removes some of the concern for monetary policy, but it would be even better news if the additional investment by service firms had brought about big gains in productivity. Over the last 10 to 15 years only the transport and communication part of total services has recorded average annual productivity growth anywhere near that of manufacturing, and that was a sector of net job loss. Mismeasurement may be part of the problem, but this productivity gap remains a puzzle.

A key question for anti-inflation policy is why services prices tend to rise faster than goods prices. This appears to be true across countries and during periods of both rising and falling inflation. The chart shows that in Britain over the past decade service price inflation has been higher than goods inflation in every year except 1996 when there were big reductions in utilities prices. In fact, since their privatisation, the utilities component of service price inflation has been an important offset to higher price rises in other services. The biggest contributors to high services inflation over this decade have been restaurants and leisure services.

In the year since the Government gave the Bank of England responsibility for setting interest rates to meet its inflation target of 2.5 per cent, goods price inflation has been consistently below that level (at an average of 2.1 per cent) while service inflation has been above it (3.0 per cent). Household spending on services makes up 35 per cent of the inflation index, much less than services share of output or employment. This is partly because the important services of health and education are paid for through taxes rather than user charges. In the US, where most healthcare and higher education is paid for directly, services account for 57 per cent of the inflation index for households.

There are at least four possible reasons for the consistently higher rates of service price inflation.

First, we could blame the statisticians. There could be greater measurement bias for services than goods in terms of underestimating quality improvements for which the consumer is willing to pay, thereby overestimating the price increase. Many visitors to London have remarked on the vast improvement in restaurant food and service since the 1970s. It must be difficult for the statistician to decide how much of the higher restaurant bill reflects better service and how much is simple inflation.

Second, the lower exposure of services to international trade may result in less competitive pressure on service prices at home.

Third, even if competitive pressures are strong, there may be certain characteristics of some services that allow a greater degree of price complexity and therefore market power by producers. Customisation (versus commoditisation) and direct interaction between the producer and consumer (versus arm's-length sales through intermediaries) make price comparison more difficult for consumers. Component services may be embedded in packages (eg, mobile telephones) that can obscure the price actually paid or the service actually bought.

Finally, if average productivity growth is intrinsically faster in manufacturing than services, then a fixed inflation target for the economy and competitive conditions in both sectors implies that service prices will continue to rise relative to goods. However trend productivity growth would decline, so a 2.5 per cent inflation target would be associated with lower average wages and GDP growth.

It is too soon to say what the shift to a service dominated economy will mean for future growth, productivity, employment, trade, competition policy or monetary policy. The implications could be profound. Research is under way to address such questions at the Bank of England and elsewhere, but more needs to be done. Until we understand better some of the fundamental economic characteristics of services, there is a risk that using old rules and old tools could trap the economy in its old tendencies to underperformance.

DeAnne Julius is a member of the Monetary Policy Committee, Bank of England.