MEASURING WHAT WE REALLY, REALLY WANT - THE POWER OF A SINGLE NUMBER

Australia’s productivity growth languishes. For the last two decades, growth in productivity has sat well below its 60-year average. In the 2010s, growth was half that of the 1990s.

Australia is in good company. More than 80 per cent of advanced economies have been recording productivity growth below long run averages. For emerging market and developing economies, the figure is less than 50 per cent. For low-income countries, it is roughly 30 per cent. 

Productivity has become so last century - for advanced economies anyway.

These dynamics have an upside. Incomes in advanced economies are much higher than in the rest of the world. Better productivity performance is helping poorer nations catch up, lowering global poverty along the way. But even here, there is a cloud. Productivity in emerging economies may also be slowing.

For economists, falling productivity growth is a massive concern. Long term, less productivity means lower income growth. This is leading some to warn future generations not to expect the same growth in living standards as their parents or grandparents.

The Australian Government seems to agree. It has reduced its expected long-term rate of productivity growth from 1.5 to 1.2 per cent a year. If realised, this will see incomes 20 per cent lower in 40 years than they would otherwise be. Good luck affording tickets to Tay Tay’s farewell tour.

Productivity wouldn’t matter, at least not as much, but for an in-built human desire for more and better. This desire conflicts with our growing understanding of the environmental limitations of our planet. Logic suggests that doing more with less – the essence of productivity – should be becoming more valuable, not less. Yet ideas for doing better sit unloved on the policy cutting room floor.

Explanations compete for why this is happening occurring. Many contain at least a kernel of truth. But none, even when added together, seem provide a complete answer. Something is missing.

One possibility is that our measures of productivity do not align with the version of more and better we (really, really) want. For the measure to have policy meaning, the alignment does not need to be perfect. But it does need to be pretty good. It also needs to hold, with a reasonable degree of stability, over time.

Given the importance placed on productivity by policy makers and other decision makers, alignment is not simply a matter of abstract academic interest. It is, or at least it should be, a matter of national concern.

Measuring productivity

The formula for productivity is simple – outputs divided by the inputs used to create those outputs. The higher the number, the more productive you are. Productivity growth records how this ratio changes over time. It feels like measuring productivity should be easy. Only it isn’t.

Even for an individual organisation, translating the idea of productivity into meaningful data can be difficult. W Bruce Chew, for example, examines how to assess productivity at a manufacturing organisation. As he does, the conceptually simple becomes startlingly complicated.

What is challenging at a firm level is an order of magnitude more difficult at societal level. It is little wonder that a recent Productivity Commission report noted difficulties measuring productivity without pursuing them further.

The PC’s decision was both pragmatic and somewhat reasonable. Its most used measure of productivity does not seek to capture all inputs, just one – labour. Its preferred measure of output, Gross Domestic Product, is used world-wide. One advantage of these measures is that they allow for a consistent time series. Another is that they allow comparisons with other nations. Both advantages facilitate meaningful empirical analysis.

Most of the technical debate around measuring productivity tends to focus on the inputs side. Labour productivity (partly because of its connection to personal income) is used commonly. But so are other measures, like total factor productivity, which seek to capture a fuller range of inputs.

Less debate tends to occur on the output side. GDP stands pretty much alone as a measure of societal output. The assumption is that GDP provides a good sense of what we (really, really) want. But does it?

A brief history of GDP

The history of GDP is more interesting than it sounds. Honestly.  

The story begins with William Petty in 17th Century England at a time of ongoing civil war. Petty had what we would today call multiple careers. He was a cabin boy who was marooned after breaking his leg, private secretary to Thomas Hobbes, an optician’s assistant, an anatomist who apparently revived a corpse, and physician-general for Oliver Cromwell’s army. Finally, after a bite to eat and a lie down, he founded political arithmetic (now known as statistics).

Petty’s work included combining a mix of real and estimated data to value the lands under Cromwell’s control. This provided the empirical base for Cromwell to distribute land to his army in lieu of pay. As it happens, many did not want the land and the ever-helpful Petty bought it from them. He made a fortune. It seems that Sir William was, in addition to being the ‘most rational’ man in England, also a bit of a bounder. But his idea of measuring things to inform political decisions has caught on.

A hundred or so years later, Adam Smith entered the stage. Smith, aside from some stunning bouts of absent mindedness, led a much less exciting life than Petty. But it was infinitely more influential. One of Smith’s views was that humans had a biological desire to ‘better’ their condition and, as a consequence, governments should focus on producing ‘more’. It is a view that drives governments worldwide.

Smith’s conception of delivering ‘better’ focussed primarily on the increasing the production of goods and wealth. Later thinkers, like Alfred Marshall and Arthur Pigou, argued for a broader concept – welfare (more commonly understand today as wellbeing). The problem was that only some elements of welfare could be measured. This resulted in what Philipp Lepenies describes as paradoxes.

One key paradox, which remains today, is that the value of unpaid work was excluded from measurement efforts. Too hard apparently. As Arthur Pigou observed “if a man marries his housekeeper or his cook, the national dividend would be diminished”.  His example feels archaic today. But it was a different time, and the mountain climbing Pigou was a cloistered academic, so let’s forgive him.

Paradoxes aside, technical work on building a measure of ‘national income’ continued. However, in Britain it was not until 1940s that things really started to come together. War was again a driver and production, as opposed to welfare, was the key focus.

Across the Atlantic, Simon Kuznets had been taking knowledge gained from his home nation of Russia (then a leader in political arithmetic) and applying it to the US. Kuznets sought to include a subjective element into national income to capture life satisfaction. Quantification eluded him however, leaving a gap (in his mind at least) between what should be measured and what could be measured.

Kuznets’ calculations of US national income for 1929 to 1932 were nonetheless a major hit. They were referred to extensively in the 1936 US Presidential campaign by Franklin Delano Roosevelt. The subtext was clear: government performance could now be assessed via a single number.

Despite Kuznets’ efforts, notions of welfare were ultimately overtaken in the US by a stricter focus on national production. This, once again, partly reflected wartime priorities in the US. But measurement issues and a drive to produce standardised statistics that could be compared across nations also played a role. The end result was the birth of gross domestic product (GDP).

By the way, don’t feel too sorry for Kuznets. He won a Nobel prize and had a university named after him. You win some, you lose some.

The power of a single number

GDP is arguably the single most important piece of political arithmetic in the world today. Estimates of GDP growth headline government budgets across the globe. Thanks partly to FDR, the election prospects of democratic governments can rest on whether GDP growth is meeting public expectations. It is a number with real power.

GDP is also what gives measures of productivity meaning. If GDP does not align with what we really, really want, nor will productivity. It is here where some questions arise. The focus of GDP was influenced by national priorities at times of war. Empirical pragmatism has resulted in a preferencing of what can be measured over what cannot. Some things are measured more accurately than others.

Psychologist Daniel Kahneman might describe GDP as a heuristic. It is a short cut which allows us to make fast-thinking decisions about something which is, in fact, uncertain and complicated. Like all heuristic short-cuts, it is not free from bias and not immutably fit for purpose.

It is unreasonable to expect GDP to align perfectly to what we really, really want. But to deserve its power, the relationship needs to be strong, and it also needs to be reasonably stable over time. The big question is – is it?