Labour productivity: what output per worker measures and what it fails to see

Labour productivity is the ratio of output to the labour spent on it: how much value added falls on each hour worked or on each person employed. It is an accounting fraction, not a verdict on how hard employees try. It rises when a person gets a more expensive machine, a better organised process or access to a larger market, and it falls when output shrinks faster than the business manages to cut its headcount. For a reader who looks at macro data in order to make portfolio decisions, one thing matters here: productivity is the limit beyond which an economy cannot grow without inflation, and a company cannot raise wages without squeezing its margin.

What is divided by what

The numerator is usually gross value added, whether for the economy as a whole, for an industry or for a single enterprise. The denominator is either the number of people employed or the number of hours worked. Both versions are called labour productivity, and they diverge more than it might seem.

Output per person employed blends two different stories: how efficient an hour of work is, and how many hours a person works on average. A country with a short working week and a high hourly return will look modest on a per-person basis. A country where people work long hours and are poorly equipped will look the opposite. That is why international comparisons are built on hours, while domestic industry breakdowns are often calculated per person employed: hours data are simply scarcer and are collected less often.

Why this is not about effort

Labour productivity is a property of the job, not of the worker. An operator on a modern production line produces more not because they try harder than their predecessor did, but because next to them stands capital that has paid for itself thanks to someone's investment. Hence the main channel of growth: the accumulation of capital per person employed, plus technological and organisational improvements.

Another indicator helps to separate these sources: the residual that is explained neither by labour nor by capital. It has an article of its own: total factor productivity answers the question of what in growth remains unexplained once all measurable inputs have been accounted for. Labour productivity is the cruder and more accessible measure; TFP is more careful and far more contested in its calculation.

There is also a composition effect. If the workforce flows from industries with low value added into industries with high value added, overall productivity across the country rises, even though nothing has changed inside each individual industry. This is a real gain in prosperity, but its mechanism is different: reallocation, not improvement.

The link to wages, inflation and the policy rate

Unit labour costs are wages divided by productivity. As long as pay grows more slowly than output per worker, the cost per unit of production declines: companies keep their margins, and there is no inflationary pressure coming from labour. When wages outpace productivity, business has to choose between a squeeze on profits and passing the costs on into prices. The central bank watches exactly this gap, and it is this gap that connects employment statistics with rate decisions, and through the rate with the repricing of the entire OFZ curve and with valuation multiples on the stock market.

That is why strong employment data are not always good news, and this is a subject in its own right: why the labour market is read the other way round. A tight labour market combined with stagnating productivity is the combination that keeps the rate high for longer than the market expects.

How to read it at the company level

At the level of an issuer, the counterpart of productivity is revenue or value added per employee, together with the trend in the payroll relative to revenue. The numerator and the denominator are taken from financial statements: headcount is not disclosed by every company and not every time, so comparisons are valid mainly within one industry and under the same methodology. The source documents are in the issuer reports section, and the publication dates are in the events calendar.

Labour-intensive businesses, such as retail, transport and construction, are more sensitive to wages running ahead: the share of pay in their costs is higher, and the gap between wages and output per worker hits the margin faster. Capital-intensive and commodity businesses are more sensitive to prices and the exchange rate than to the cost of a working hour.

What these data do not say

Labour productivity says nothing about distribution: a rise in output per worker does not have to reach wages; it can settle entirely in profits or go into lower prices. It is poorly measured in services and in the public sector, where value added is often estimated from costs, and then the fraction barely moves by construction. And it gets revised: estimates of output and employment are refined after the fact, so turning points become reliably visible only with time. Macro news on this subject is published in the news feed, and definitions of related terms are in the glossary.