Total Factor Productivity: The Residual That Explains Growth

Total factor productivity (TFP) is the part of output growth that cannot be explained either by an increase in the number of people employed or by the build-up of capital. An economist takes a production function, feeds labour and capital into it, sees how much output they should have delivered, and compares that with what actually happened. The difference is TFP. It is not a separate resource that can be bought or counted in a warehouse: it is a residual, "what was left unexplained". Hence its other name, the Solow residual, after the author of the growth model for which he received the Nobel prize. For an investor, TFP matters not as a figure in its own right but as an indicator of what exactly is driving the economy's growth: an injection of resources, or the fact that more is now being obtained from the same resources.

How the residual is calculated

The mechanics are simple in design and temperamental in execution. You take output (usually GDP or an industry's gross value added), you take labour input, preferably in hours worked rather than in headcount, and the stock of fixed capital. Labour and capital are weighted by their shares in income: how much of the value created goes to wages, and how much to profit and depreciation. The resulting "explained" growth is subtracted from actual growth, and the remainder is attributed to TFP.

Every step here is a source of dispute. Nobody observes the capital stock directly: it is accumulated from investment and written down at an assumed rate of depreciation, and the rate of depreciation is an assumption. Capacity utilisation changes over the cycle: in a crisis factories stand idle, the capital is formally still on the balance sheet, output falls, and TFP "falls" along with it, although the technology has not gone anywhere. The quality of labour changes too: a worker with a university degree and a worker without one are often identical in the statistics. That is why a proper TFP calculation always comes with a long list of caveats, and different research groups arrive at noticeably different series for the same country.

What actually ends up in the residual

The name is misleading: "factor productivity" sounds like a synonym for technology, but everything that is not accounted for explicitly flows into the residual. It includes technological progress and the diffusion of technologies that are already known; the quality of management; the efficiency with which resources are reallocated between firms, because when labour and capital move from weak companies to strong ones, output grows without any growth in inputs. It also includes the quality of institutions, transaction costs, the state of competition, access to foreign markets and to imported equipment. And it includes measurement errors as well.

This gives a practical rule for reading it: TFP growth explains nothing by itself; it merely tells you that an explanation has to be found. What is useful is not the level of the residual but its breakdown by industry and over time. When the residual is growing in manufacturing and flat in the commodities sector, that is one conversation; when the picture is reversed, it is another.

Why this concerns the investor, not only the economist

The link to the market runs through several channels, and all of them are long-term.

The first is corporate profit. A sustained rise in margins across the economy is possible either through prices or because more is produced per unit of input. The former is limited by demand and competition; the latter is precisely TFP. It is more convenient to check this not in the macro series but in the financial statements themselves: the trend in revenue per employee and the return on invested capital in issuers' reports are an industry-level projection of the same question.

The second is the interest rate and the yield on long bonds. Over the long run the real rate gravitates towards the pace of productivity growth: if the economy is able to generate a higher return on capital, capital costs more. This is exactly why the debate about productivity always sits next to the debate about the fair level of yields; it is worth following on the OFZ curve and more broadly across the debt market.

The third is the ability to pay shareholders. A company that grows only by expanding its assets finances that growth out of the same profit from which it pays dividends. A company that grows through efficiency avoids that choice. The practical test is the stability of payouts over time, not one-off generosity: the dividend history and payout calendar show who keeps to a steady regime and who pays when the occasion arises.

How to read news and announcements about productivity

TFP is a number that gets revised. The first estimates come out with a lag and are then refined together with the revisions to GDP and capital. Sharp conclusions drawn from a fresh publication are therefore almost always premature: a change in trend is confirmed not by a single data release but by the revised series.

A useful habit is to check the macro story against the corporate calendar: claims of a technological leap are tested through capital expenditure and the return on it, and those are disclosed in the financial statements. The dates of publications and corporate events are collected in the events calendar, and the flow of announcements in the news feed.

Limitations worth keeping in mind

TFP does not measure welfare, says nothing about the distribution of income and is poor at capturing the quality of new goods and services. It is especially unreliable in industries where output is hard to measure: services, healthcare, public administration. It is sensitive to the phase of the cycle and almost useless over a horizon of one quarter.

And an honest caveat: the platform's data contain no ready-made series of total factor productivity for Russia. What is published here is market and corporate data, not macroeconomic statistics. TFP estimates have to be taken from their primary sources (national statistics, the central bank, international organisations), while the platform's data should be used to check how the macro picture shows up in specific securities: in stocks, funds and financial statements. Definitions of related concepts, such as the production function, return on capital and the real rate, are collected in the glossary.