How much does it matter whether we have labour unions or not?

In the popular imagination labour unions are a significant factor in the incomes of ordinary people and a major reason we don’t endure the relative poverty of past generations. Here I will try to argue that in a long run view labour unions, even given very generous assumptions, are pretty inconsequential, using some data from the Australian case. In the process I’ll help explain why just 6% of leading economics bloggers described unions as very important ‘for the health of the US economy‘ in a recent survey, while 70% described them as unimportant.

Recent research from the Melbourne Institute suggests that the lowest income earners in Australia receive a premium of ~12% of their wage from union activity, while the highest income earners receive a ~6% premium from union activity. They speculate that the reason for this is that high income employees are more skilled at bargaining individually and so benefit less from expert union representation. Such a premium is roughly comparable with union premiums observed across Europe and the US, and a bit higher than past estimates of the premium in Australia probably due to recent changes in industrial law.

This wage premium is entirely what we would expect from economic theory; where monopolies or cartels can push up prices and improve their earnings they will surely do so. Even where unions lack what economists term ‘market power‘, by allowing workers to pool their knowledge and skill when negotiating wages they can produce better outcomes for workers who might otherwise be convinced to take low wages. Unions can also use their power to push for better working conditions. Here I will assume that workers use unions to push for higher ‘non-wage compensation’ (safe and pleasant work environments, breaks, amenities) and wages roughly equally. In any round of bargaining workers might get larger improvements in one or the other of these but over decades they probably want them to rise proportionately.

Other than improving the hourly compensation that workers receive and helping to coordinate its form to maximise worker satisfaction, what other impacts will strong unions have on an industry and the economy? *For ease of reading I’ve stuck union costs at the end of the post.*

For the sake of discussion, let’s assume away all of these downsides and instead imagine that the 12% income premium falls magically out of the sky into workers’ wallets at no cost to anyone. Is this an impressive wage premium that we should fight hard to preserve?

Australia had an average GDP growth per person over the 20th century of 1.7% per year, resulting in per capita incomes increasing by 540% between 1900 and 2000. For a variety of historical reasons, this was among the slowest rates of growth in developed countries. This measure of wage changes fails to take into account changes in the total number of hours worked and doesn’t factor in that workers probably take home a somewhat larger share of national output today than in 1900 due to increased education.

Probably then a more accurate estimate of future wage growth can be found looking at modern data: between 1994 and 2004, hourly labour productivity rose 24%, implying annual productivity growth of around 2.2%. Let’s make the reasonable assumption that wages and other redistribution to workers grow hand in hand with productivity (this follows from the share of income received by workers being mostly constant over time). If this growth is predictive of the future, the union wage premium of 12% is then equal to the natural wage growth that occurs every 5.2 years. Amazingly, fifty years of productivity growth at this rate would triple wages!

The obvious indicator of the relative value of the union wage premium compared to productivity growth is how much we would need to increase that growth to compensate a future worker for the disappearance of their union. Those who have seen the power of exponential growth will know that the answer will fall drastically the farther into the future you look. As shown in the figure, to compensate a worker 10 year in the future we would need 1.1 percentage points higher annual growth over that 10 years (3.3% rather than 2.2%), over 20 years 0.56 more, over 30 years 0.36 more, over 40 years 0.29 more, over 50 years 0.23 more, over 60 years 0.19 more, over 70 years 0.165 more and over 100 years just 0.11 more.

The upshot is, if you care about the earnings of low-income people only in the present and immediate future, unions are an easier way of increasing that than improving productivity growth, as long as the drawbacks listed below aren’t too large in practice. However, if you care much about your or your descendants’ incomes a few decades hence, or any of the many other people who will live in the future, pretty small increases in productivity growth would compensate you for the disappearance of unions pretty quickly. Redistribution today will benefit present workers alone; increased productivity growth now will benefit everyone passing through the Australian economy on for hundred or thousands of years into the future. If you care for all those hundred of millions of future people even slightly, productivity growth should be by far your overwhelming concern.

Is there really a trade-off between union activity and productivity growth? Probably at least a small one. Union organisation and political activity beyond what is necessary to simply negotiate wages takes up resources that could otherwise be put towards producing and investing wealth, inventing new things, and learning new skills – all of which would boost productivity growth. If by pushing up wages they increase unemployment (and so push down GDP), they will also slow productivity growth. Any policies unions promote to ‘protect jobs’ from innovation or competition, can reduce the creativity and efficiency of the economy; these sorts of policies were at least partly responsible for the sluggish productivity growth in Australia prior to the 90s. Most importantly, the more voters, public servants and academics are focussed on using unions to redistribute income, the fewer people are thinking up and debating ways of increasing our long-term productivity growth.

I can think of three interesting responses to this line of argument: that the real benefit of unions is not to increase worker compensation, but rather to level out the status of different groups in society, resulting in lower stress and more respect for low earners; that relative incomes are what determine people’s happiness and satisfaction rather than absolute incomes; and finally that people in the future will be so much richer than us that we shouldn’t worry too much about how we affect their incomes. I’ll discuss these in future posts.

Union costs

The following are all possible or likely outcomes of a union pushing up wages:

Reduced total working hours (when the prices of something goes up it is usually purchased less) and so somewhat lower total output (labour is a crucial factor in producing wealth).
Increased prices of whatever they produce (if the price of an input in producing a product goes up, the price of the product will often be pushed up), lowering the effective incomes of people who buy that product.
Lower incomes for the industry’s other employees, investors who supply capital and providers of its land and natural resources (firm management can choose not to pass on all of the extra wage costs to the consumer, instead squeezing everyone else involved with the firm, including themselves, for savings).

So you can see that some of the wage increases unions achieve come out of the pockets of their customers, people with savings and through various other price changes for everyone else they trade with both directly and indirectly. Some of it also comes at the expense of total employment, either in the form of some people going unemployed or some workers getting fewer hours of work than they would have without the union. Unfortunately, unions representing the highest earners in the economy (which achieve that 6% wage premium) are almost certainly on net taking from the poor and giving to the rich, as well as reducing the total output of the economy. Without them we would have both more income equality and higher incomes for the poor.

Advertisement