Markets and the scientific method

By Shaun Hendy 06/07/2011 9


I went undercover last week at the New Zealand Association of Economists conference to see what they had to say about innovation.  Not so incognito was the Undercover Economist, Tim Harford, Financial Times columnist and author of several highly entertaining popular economics books, who delivered the opening keynote address at the conference.  In this post I will touch on some of the ideas that Harford covered in his talk.

Harford made his name with 2005’s Undercover Economist, an account of the effects of markets in our everyday lives.  Via a deft deconstruction of the factors that govern the price you pay for your morning espresso, he delivers an orthodox expose of the inner workings of the marketplace:  markets are good for you, except when they aren’t, in which case there are straightforward interventions that will correct most failures.

The trouble with markets

Yet the Harford of 2011 is not the Harford of 2005.  The recent global financial crisis gave many economic commentators pause for thought, and Harford’s response comes in his most recent book, Adapt: Why Success Always Starts With Failure*.

Harford’s new world view is a tad less orthodox, although it stands firmly on the shoulders of the old.  The Undercover Economist told us of three ways in which markets can fail:  externality, information asymmetry and monopoly.  To this list, argues Harford, we must now add a fourth cause of market failure:  complexity.

Harford’s original faith in markets was vested in their ability to tell the truth.  In a market economy, bad ideas will ultimately be shut down by the bankruptcy court, whereas good ideas will spread as they are copied by competitors.  In a centrally planned economy, bad ideas can become Great Leaps Forward.

Nonetheless, Harford sees the financial crisis as an example of where complexity may have overwhelmed the ability of the market to sift the good from the bad.  The bewildering variety of complex financial hedges that were in place to manage risk instead ended up concealing that risk; at least for a time, markets were unable to tell the good loans from the bad.  This was not so much an information asymmetry as an information deficit.

Beyond efficiency

Harford has not abandoned his confidence in markets altogether.  Rather, he draws a lesson from the fact that markets seem to work at all in the face of complexity, and applies this wisdom to the broader swathe of institutions that advanced economies rely on to regulate, to innovate and to govern.

Beyond the efficient allocation of goods, Harford celebrates the ability of markets to explore new ideas, experiment with them and eliminate those that fail.  In contrast, governments, bureaucracies, and even most companies are not good at taking risks or experimenting.  It is an unusual political career than can survive more than a few failures, and middle managers in a hierarchical corporate or government structure have little incentive to report failure up the chain.  This generally results in organisations that struggle to filter good from bad.

Yet some institutions have learned to flirt with failure.  The scientific method, for instance, formalises the procedure for proving ideas wrong.  Peer review, the double blind trial and the requirement for repeatability in any experiment, are all tools the scientific community use to weed out the ill-founded ideas from the sound.  Companies like Google expect 90% of their projects to fail, relying on the 10% that succeed to keep the company ahead of its competitors.

Picking winners?

So what can the rest of society learn from the way in which markets explore and scientists experiment?  Imagine a government, Harford muses, that had the confidence to experiment, that was able to run properly controlled trials of new initiatives, and that above all was ready to accept failure.  Such thoughts sit quite nicely along side those of Sir Peter Gluckman, regarding evidenced-based policy making.  A government that could properly trial and refine educational, social or correctional initiatives would get my vote.

There are also lessons for how New Zealand should spend its innovation dollars.  It is frequently argued that New Zealand is doomed to choose; our resources are too limited to fund a full portfolio of science.  We must put our resources where we think they will do best.  We must pick our winners.

Yet much of Harford’s talk was spent busting the myth that this is possible.  Fonterra today seems as good a bet to Kiwis as US Steel must have to Americans at the beginning of the twentieth century:

This was a company with everything going for it:  it was the market leader in the largest and most dynamic economy in the world; and it was in an industry that has been of tremendous importance ever since.  Yet US Steel had disappeared completely from the world’s top one hundred companies by 1995; at the time of writing, it was not even in the top five hundred.

So if we can’t pick winners, what do we do?

An entrepreneurial government

Harford argues for an entrepreneurial approach to funding science and innovation:  governments should intervene in research and development as if they were entrepreneurs rather than investment bankers.  When pressed for an example of what this might look like, Harford suggested that innovation prizes were one way in which governments were being more entrepreneurial.

Harford is not alone in holding this point of view.  I recently read a very interesting evaluation of the human genome project, which concluded:

… that reframing science policy around the notion of conducting entrepreneurial experiments — experiments that increase the diversity of technical, organizational and institutional arrangements in which scientific research is conducted — can provide policy makers with a wider repertoire of effective interventions.

…  policy makers can use the levers of entrepreneurial experimentation to transform scientific progress, much as entrepreneurs have transformed economic progress.

Huanga and Murray, Research Policy 39  567—582 (2010).

To attempt this would entail a radically different approach to funding science in New Zealand.  It would demand a commitment by government to maintaining a diverse set of scientific and technical capabilities.  It would require new methods for evaluating the effectiveness of these experiments.  It would require an acceptance of greater risk and a tolerance of failure by our policy makers.

But the pay-off could be huge. Do we have the courage to rise to the challenge?

*The talk I saw on June 29th was largely based on the first chapter of this book.


9 Responses to “Markets and the scientific method”

  • I have to add the followings as well since, there’s only 2 links per post before the spam filter is triggered, so I couldn’t have posted more than 2 in my first post.

    From Simplistic to Complex Systems in Economics

    Impact of the topology of global macroeconomic network on the spreading of economic crises

    Whatever theory that attempts to describe the complexity of the markets and its failures, all the problems can be traced back to Government interference. Some of these have been pointed out by scholars in the Austrian School of Economic thoughts such as Economics Nobel Laureate, late Prof. Hayek, but no one is listening them. No wondered that we have encountered economic crises after crises even when bureaucrats have passed many laws (based on perceived causes of past crises) in order to prevent future crises . But crises are still occurring. As Austrian economists have said, it is the government interference.

    Some econophysicists think that Hayek was the forefather of complex system theory. He lacked mathematical techniques in his days, but now, advanced mathematics have been developed to deal with system complexity as in economics.

  • Not to sound like a broken record, but even assuming we could get government to change its innovation funding models; what sort of time frame would we be looking at (I’m think 5 -10 years min)? Will New Zealand be in a position to make those difficult choices if we continue our GDP per capita decline for another 5-10 years when there seems to be insufficient prosperity even to make them now? Furthermore can something as slow to change and adapt as policy keep pace with the rapid evolution of things like innovation and entrepreneurship…it’s beginning to seem to me that innovation will always be 2 steps ahead of policy, by necessity, so why expect policy to keep pace and solve it? Perhaps a grass-roots approach would be more effective in the short term, even if it had less resources than the government?

  • Falafulu Fisi: Yes, the Austrian School doesn’t believe in market failure, or public goods like the benefits from government funded R&D for that matter, so I guess this is post is from rather un-Austrian point of view. A less extreme position is that government failure is always worse than market failure, but especially when it comes to government investment in basic research, I would suggest this would be a minority point of view.

    Elf: The concern is that the government’s RS&T investment portfolio is too risk averse and in the long run we are not best value from our RS&T system. We should certainly look for novel ways to do science with running to policy makers, but I would challenge your assumption that they have to be two steps behind: the Human Genome Project was a great example of an entrepreneurial government-led science initiative. In some fields, grass roots can produce results (e.g. open source software development), but in others you’ll need someone to pay for that synchrotron …

  • The idea of entrepreneurship doesn’t sit very well with government funding- it works because of the powerful incentives on the entrepreneur to balance risk and reward very carefully. These incentives are largely absent when someone else’s money is at stake.
    And you can’t escape the fact that taxpayer research funds have to rationed in some way- there will always be more projects than there are dollars.
    So the question is, how best to do the rationing? You could have a system of random selection ( actually not as daft as it sounds) but most players would demand a selection process based on agreed and transparent criteria. Even Harford’s ‘innovation prizes’ would have to be awarded by a committee or whatever.
    In the NZ system nearly all research priojects are investigator-conceived. The selection criteria vary across funding mechanisms and over time. Sometimes they have been too specific and this may have inhibited innovation.
    You could go down the South African route and award research funds to top scientists and let them do what they want- but that is just another rationing process which may also inhibit innovation.
    I think we should take Harford’s advice and urge other countries to experiment with new funding mechanisms. Then we could adopt the ones which work best (as indeed we have done in the past).

  • Kemo sabe: Yes most people would agree that entrepreneurship doesn’t sit well with government funding but Harford is arguing that this shouldn’t be the case. It is not just the fact that it is some else’s money – there are plenty of trusts and charitable organisations that are far more entrepreneurial with their R&D spending than government. There is an interesting study (http://www.nber.org/papers/w15466) comparing funding outcomes from the Howard Hughes Medical Institute, which which tolerates early failure and allows freedom to experiment, with the NIH, which invests much the way we do: naturally, it finds the HHMI funded researchers do better.

  • Shaun
    Thanks, this is very interesting, although I’ve only read the abstract.
    You could look at it another way; the HHMI is simply rationing its funds according to a different set of criteria than the ones used by NIH. The decision-makers are still spending someone else’s money, and both organisations are large bureaucracies.

    I have had experience of independent funding agencies (charities, trusts etc) which were unbelievably conservative, in some cases only funding work which the government was also funding.
    It probably boils down to the criteria reflecting the goals of the organisation, and government’s goals are often unique.

Site Meter