16 January 2012

The Apparent Paradox of Productivity Growth

On the one hand, we see that increasing productivity can kill jobs -- in today's manufacturing sector 177 American workers can produce the same output that required 1,000 workers in 1950. One the other hand, productivity is viewed as central to job creation, as this article in today's FT notes:
The rate at which workers are raising their productivity in the world’s advanced economies fell by half in 2011, and is even starting to slow in some emerging economies, according to a report that suggests that unemployment is likely to rise in the months ahead.

According to the Conference Board, the global business organisation, productivity – defined as output per worker – in the most advanced economies fell from 3.1 per cent in 2010 to 1.6 per cent in 2011.
It turns out that decreasing productivity also kills jobs:
Bart van Ark, chief economist at the Conference Board, said that in the short term, the drop in productivity suggested that employers would cut labour to match the drop in overall output. “But in the longer term, productivity gains come from technology innovation and investment,” he said.

Moreover, there are concerns that the focus on austerity by governments may exacerbate the loss of productivity because without expenditure on new technology, any gains will be limited. “With calls for austerity, you have to be cautious not to cut the investments in new technology that increase productivity,” he said.
Read that again. If labor productivity falls, then businesses will eliminate workers, but over the longer term innovation will more than compensate for the short-term losses -- or at least that is the argument.

Even leading economists are of two minds on productivity. Here is Martin Wolf writing in 2005, extolling the virtues of productivity growth:
Productivity determines the wealth of nations. The proportion of the population at work matters, too, and so does the number of hours worked by each person. But neither is as important as productivity.
And here is Martin Wolf writing last summer extolling the virtues of productivity decline:
[I]f one is going to pursue austerity, as the UK government does, it greatly helps to have poor productivity performance. With US productivity, too, the UK would have a jobless rate of over 12 per cent.

On balance, I am grateful that the UK job market has responded to this recession in this curiously continental way.
By a "continental" response to the financial crisis Wolf means "a market that adjusts to shocks via hours worked per person rather than via jobs." So from this perspective, poor productivity performance is a consequence of decisions about how to spread the pain of an economic crisis.

But labor productivity is only one element of total productivity -- other factors matter as well. An economy can weather declines in labor productivity -- even those that are self-imposed -- if total factor productivity continues to increase. Wolf's apparently contradictory statements can be reconciled if we understand that in 2005 he was referring to total productivity and in 2011 he was referring to labor productivity. On his blog I am going to strive to be very clear about what I mean by "productivity" when I use the term.

Understanding the modern economy requires making sense of the easily confusing concept of productivity and how it relates to jobs and economic growth. [Total and especially labor] productivity growth does indeed kill jobs, but it creates jobs as well. Creating a virtuous cycle of total productivity growth is a key challenge of managing the 21st century economy.


  1. The paradox between jobs and productivity (total and labor) may be explained through capital productivity. The majority of the population have limited capital and when the cost of goods and services decline through increased productivity, their limited capital becomes more productive and, by virtue of our individual nature, also more diversely invested, thereby increasing opportunities for producers and workers.

    I think the problem arises when individuals and cooperatives (e.g. government) are capable of exerting sufficient leverage and through the introduction of false signals into the market manage to distort real supply and demand. The result may be desirable (e.g. instant gratification), but it is also unsustainable (i.e. accumulation of personal, corporate, and sovereign debt). This is why people fear, and rightly so, the establishment of private monopolies; but, they fail to recognize that the government is, in fact, the largest monopoly and it also has a granted authority to enforce its will over the people it serves.

    It's worth noting that the current government introduced distortion represents a false economy of approximately 10% per annum. The current total government (i.e. federal, state, and local) spending currently accounts for nearly half of our GDP. Neither of these situations are sustainable.

  2. There are two basic ways of dealing with this phenomenon: either we can have the 177 continue making everything, and tax them so that they only get 17.7% of what they produce and spread the rest around to the 823 (let's call that mode S), or have everybody work and everybody gets 5.65 times as much as before (call that mode G).

    Seems to me the USA was in Mode G up until the Great Stagnation (circa 1970) and everybody expected to be better off than their parents, etc, etc. Then we shifted into mode S.

  3. From the perspective of a small business owner, labor(?) productivity growth means survival and profitability. If you can produce a product for less than your competition while making a profit, your competition will either improve or get out of that business. Most large manufacturers expect their suppliers to have continuous improvement programs and contracts will be written to spend less per product the longer the manufacturing run continues. So car models approaching significant design changes can still be profitable even though they are deeply discounted. The customer that pays $100 less per month for a new car with dated styling can afford to spend more money on other things. So all this is pretty obvious.

    But Roger, what about industries that seem to have precious little productivity gains such as public education and some parts of health care. How does their lack of productivity improvements affect the overall economy and competitiveness? It's another side of the coin you've tossed out but would make an interesting discussion on the impact on people's lives and finances.

  4. I'm not an economist, so maybe I have this wrong, but as I understand it, one key question which you haven't addressed is how productivity changes affect aggregate demand.

    If productivity gains lead to demand growth that meets or exceeds productivity growth, then growing demand plus growing productivity --> economic growth and steady or rising employment coupled to rapidly rising quality of life. If aggregate demand does not keep up with productivity gains, then productivity growth leads to decreased employment.

    Some of this may depend on the way the benefits of rising productivity are distributed. Back in the 1980s Richard Freeman and James Medoff wrote a book called "What Do Unions Do?" which argued that unionized workplaces had significantly higher productivity than non-unionized workplaces (partly because workers were more skilled and worked harder; partly because the higher cost of labor motivated owners to invest their capital in better equipment; as you say, labor productivity is only one factor), so there was a paradox that unionization would increase total GDP but decrease profits because the differential in labor costs was larger than the differential in productivity: Freeman and Medoff's conclusion was that unions were good for the country (GDP), good for workers (higher pay), but bad for the owners (lower profits).

    Other economists challenged Freeman and Medoff, so it's not like their data or their conclusions are obviously correct, but I've long thought the idea was worth keeping in mind as we think of problems related to productivity.

    Coming back to my original macroeconomic question, can the relation between pay and productivity be relevant here? If higher productivity means higher worker pay, and those workers spend their money on all the cool new gizmos that they're producing, can this produce a virtuous cycle of economic growth with steady or rising employment?

    Conversely, if higher productivity means reducing the labor force and using the reduced demand for workers to justify cutting pay, can the reduced consumer demand lead to a deflationary spiral with falling employment?

    The question comes down to the one people have been arguing heatedly with regard to economic stimulus over the past few years: do we increase demand most by increasing income in lower-income households (e.g., extending unemployment benefits and cutting payroll taxes) or in upper-income households (e.g., cutting estate taxes, capital gains taxes, and marginal rates at the top brackets)?

    Finally, since you and I are academics, Roger, it's worth noting that in teaching profession is notorious for failing to improve productivity. There's no way 177 professors can teach the same number of students that 1000 professors could in 1950. Nor can 177 professors accomplish anywhere near the amount of research that 1000 professors used to. The same goes for teaching at all levels.

  5. -4-Jonathan Gilligan

    Thanks for the comment, I'll look that piece up.

    On this comment, "There's no way 177 professors can teach the same number of students that 1000 professors could in 1950."

    You should visit my program at CU, administrators are sure trying (6 faculty FTE + 1,100 undergrads + 60 grad students) ... the result is not pretty and not sustainable!

  6. @Roger: The arguments you often make about access to energy in poor countries apply also to education. They need quality education (primary, secondary, and university) for billions of people, but can't afford anything like the thousands to tens of thousands of dollars per pupil per year that it costs here.

    Revkin has been commenting on this problem recently in Dot Earth.

    For university administration simply to declare that faculty productivity shall increase (like Canute, but without the irony), is as futile as nations setting carbon emissions targets without a plan for achieving them; but as with decarbonization, it's essential to come up with feasible ways to improve teaching productivity by orders of magnitude, no?

  7. -6-Jonathan Gilligan

    Thanks ...

    "it's essential to come up with feasible ways to improve teaching productivity by orders of magnitude, no?"

    Let me try the "No" -- but first be clear about what I mean by teaching productivity, measured as resources devoted to education compared with educational outcomes (which might be measured as knowledge, skills, credentials).

    It seems to me that poor countries need to be able to dramatically scale up the resources devoted to education, which is a function of gross wealth -- not productivity gains in teaching, but rather total productivity gains in their economies.

    Productivity gains in teaching are possible of course, but limited in important respects (just as they are in many services that require human to human interaction).

    What do you think? Thanks!

  8. The Wall Street Journal has an article covering this specific topic and today's business environment, (caustion, it may be behind a paywall)
    What is interesting is also how the cost of capital also figures into the equation:
    "Billy Cyr, chief executive of Sunny Delight Beverage Co., a Cincinnati-based beverage company, says he is buying new machinery partly because it is a bargain. "When the cost of capital goes up, it is harder to justify an equipment purchase and may, instead, result in higher employment using existing equipment," he says, such as by adding shifts or overtime for existing workers. Today, the opposite is happening.

  9. 7. Roger Pielke, Jr. said...

    It seems to me that poor countries need to be able to dramatically scale up the resources devoted to education, which is a function of gross wealth

    You posted a TED talk on that a while back...the answer is the 'Washing Machine'. Primary School education is for the most part a female endeavor, as is clothes washing.

    Without Washing Machines the necessary labor pool to teach young children is too busy doing laundry.

  10. "It seems to me that poor countries need to be able to dramatically scale up the resources devoted to education, which is a function of gross wealth -- not productivity gains in teaching, but rather total productivity gains in their economies."

    You have a point, except that resources devoted to education is not simply a function of gross wealth; it is also a function of decisions about what % of a country's wealth will be devoted to education.

  11. Productivity is another subject which has been poorly served by the economics discipline.

    In general, economists measure productivity via dollar based measures like GDP.

    However, this is a terrible method because it skews productivity 'enhancements' to financially based instruments.

    Equally so many anecdotes on productivity measurements are equally useless.

    How many products being made in 1950 are being made today? In 1950, the United States was the predominant manufacturer of steel. Today this is very, very far from the case.

    Certainly there have been improvements in steel production, but by and large steel production in the United States has fallen off a cliff:


    In 1950, according to the above link, US steel production was 58.5 million tons. In 2009, it was down to 19 million tons.

    Yet another failure of the economics profession is to relate productivity to consumption. Yes, productivity has gone up, but equally so has consumption. The ratio of the two is more important than the measurement of productivity alone, but dividing GDP and GNC in dollar figures is simply worthless.

  12. Here's an alternate link to the WSJ article upthread, which is paywalled:


    Man vs. Machine, a Jobless Recovery
    U.S. Companies Are Spending to Upgrade Factories but Hiring Lags; Robots Pump Out Sunny Delight

  13. Roger:

    You close by saying: "Creating a virtuous cycle of total productivity growth is a key challenge of managing the 21st century economy."

    I agree.

    As I've suggested in previous threads here, pursuing continuing economic growth is unsustainable as long as the 'free rides', exacerbating depletion of nonrenewable (capital) resources, continues virtually unabated.

    The 'real' costs of all of those resources must appear in their market prices (e.g., the costs of global warming and pollution). Such transparency also requires that people begin to recognize that "total productivity growth" is not guaranteed to translate into equitable ("virtuous?") growth in the global quality of life! And growth in quality of life, RATHER THAN "total productivity growth" should probably become a prime motivator of behavior (both economic and social).

    To increase transparency, enforced regulation of markets, in ways that are often perceived as moving Capitalism towards Socialism, are required. But that's necessary to decrease lying and cheating and GROSS inequities – and to move towards your "virtuous cycle".

    VERY CAREFUL formulation of regulation SHOULD generate MORE – rather than LESS – true equity, and should broaden the 'freedom' and 'virtue' within markets, contrary to the predictions of most LIBERTARIANS.

    All of this will require increased tolerance of others views, by ideologues of all persuasions, despite their personal beliefs/faiths about what is CERTAIN/TRUE!