Monthly Archives: February 2014

Economists and our responsibilities to society

Posted on by

Do social scientists (including economists) have a responsibility to engage in public debate? Yes, said Nicholas Kristof in the New York Times. Academics have become too marginalised and – with honourable exceptions – he said: “Ph.D. programs have fostered a culture that glorifies arcane unintelligibility while disdaining impact and audience.”

Paul Krugman weighed in specifically on economics: “In my field there is indeed a problem with abstruseness, with the many academics who never even try to put their thoughts in plain language.” He concluded that economists who can’t explain what they’re doing probably don’t understand what they’re doing. Krugman supports the use of mathematical models but – echoing many great economists of the past like Marshall and F.Y. Edgeworth – argues that what we always call ‘intuition’, or non-mathematical explanation, matters too.

I like Edgeworth’s analogy best: the maths is like the scaffolding essential for putting up a building. You have to have it, and you have to take it down at the end.

Of course I agree with the general point that academics – especially social scientists! – have to engage with society. Not just engage with as in discuss one’s research in intelligible language, but understand how academic research and teaching interact with society, influence it and are influenced by it. (This was the theme of my 2012 Tanner Lectures, The Public Responsibilities of the Economist (link near the bottom of the page), which looked at the role of economics in the formation of modern financial markets, among other things.)

In the UK, this issue is now described by the shorthand of ‘impact’, with universities required to demonstrate their impact as part of the 2014 Research Excellence Framework, which will determine the allocation of funding. I’m playing a small role in the assessment exercise.

What impact do economists and other social scientists actually have? Is it by being ‘public intellectuals’? Or are there more ‘impactful’ channels? I’ve been reading a very interesting new book, [amazon_link id=”1446275108″ target=”_blank” ]The Impact of the Social Sciences: How academics and their research make a difference[/amazon_link], by Simon Bastow, Patrick Dunleavy and Jane Tinkler.

[amazon_image id=”1446275108″ link=”true” target=”_blank” size=”medium” ]The Impact of the Social Sciences: How Academics and their Research Make a Difference[/amazon_image]

It analyses in some detail evidence from data collected on hundreds of UK-based social scientists and researchers in the ‘STEM’ subjects (science, technology, engineering and maths). The first surprise is how many more students and researchers (and funding) there are in STEM subjects than the social sciences, which in turn dwarf the humanities and the creative arts and design: in the UK, the STEM subjects get 80% of total research funding, social sciences 14%, humanities 4% and creative arts 2%. Student and staff numbers are slightly less skewed, but the figures still make one doubt the often-made (by scientists) claim that STEM needs more.

On the impact question, the book argues that the tide has begun to turn on the inward-looking shift in academia, and on disciplinary silos which , the books says, were ‘concreted in’ by the late 1960s. This must be a good thing. What’s the point of fabulous climate science research if social scientists are not fully engaged in analysing how society might or might not change? Surely scientists – and research funders –  do appreciate that people do not necessarily believe what they’re told by academics? Trust in scientists and researchers is high but that isn’t the same as accepting them as legitimate decision-takers.

Using the data set on academics’ channels of influence built for the research, the book assesses the character of the ‘outputs’ of the academics (which varies quite a lot between disciplines) and looks at two arguments about academics’ impact. One is that there are ‘popularisers’ who can communicate but do little valuable research, and academics of course tend to sneer at this group. Another is that there are superstars who do the best research and are brilliant communicators. The truth is in between – most of the group are middling at both research output and public communication.

In general, though, social science research has a big influence on government, think tanks, civil society and business. The book traces many channels of engagement. The area where there is perhaps less influence – certainly less than the natural sciences – is in the media and social media. If Nicholas Kristof thinks the US lacks public intellectuals, he should feel sorry for the UK: the book suggests Stephen Fry as our leading public intellectual at least in terms of number of Twitter followers – a brilliant polymath but not a social science academic. Academics, in my experience, are often either reluctant to engage with the media (for several valid reasons such as time, deadlines, simplification…); or lack understanding of the conventions and constraints.

One example of highly effective collaboration cited here is the Great British Class Survey, between an academic team of sociologists and the BBC. There was huge public engagement and interest internationally, and it was a successful ‘co-production’ of social science between academics and the public. But no doubt there are some academics who see this as excessive popularisation.

The book ends with a discussion of what the authors call a ‘dynamic knowledge inventory’, a constantly updating repository of our current understanding of society. They commend ‘broad front’ social science and integrating social science with the STEM disciplines. I like the quotation from [amazon_link id=”0571270123″ target=”_blank” ]Tom Stoppard[/amazon_link] in the final chapter: “I don’t think writers are sacred but words are. They deserve respect. If you get the right ones in the right order, you might nudge the world a little.”

[amazon_image id=”0571270123″ link=”true” target=”_blank” size=”medium” ]The Real Thing[/amazon_image]

The LSE website link to the podcast of the launch event is broken but there are some slides available here.

Update: podcast link is here.

Robots, floods and the Kobayashi Maru

Posted on by

The discussion at the core of [amazon_link id=”0393239357″ target=”_blank” ]The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies [/amazon_link]by Erik Brynjlofsson and Andrew McAfee is whether or not the benefits of the current wave of digital and digitally-enabled technologies will outweigh the costs – or not.

[amazon_image id=”0393239357″ link=”true” target=”_blank” size=”medium” ]The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies[/amazon_image]

The book has three parts. The first describes digital technology, explains why it is only now that its consequences are becoming dramatic, and describes some of the impending technological advances. There is a useful survey of (mainly) robotics and AI advances, and a very clear explanation of ‘why now’ when computers and the Internet and even the web have now been around quite a while. The timing is partly due to the nature of exponential doubling. Moore’s law says computing power roughly doubles roughly every two years. If you start with one grain of rice on one square of a chess board and double it for every successive square, it isn’t until the second half of the chess board that the number of grains soars to an incomprehensible degree. It is also due to the need for organisational and infrastructure investments alongside technology investment. This section draws on the work of economic historian Paul David as well as Brynjolfsson’s own excellent work on how firms use the technology to increase productivity, by reshaping processes and work. In short, human institutions move far more slowly than technology progresses.

The second part of the book divides the consequences of technology into two types, which the authors call bounty and spread. Bounty is the potential for human progress, the almost science-fiction fruits of technology that are just starting to emerge in fields from medicine to driverless cars. By ‘spread’ they mean inequality or dispersion of outcomes. (I found this a counter-productive term as ‘spread’ to me carries the automatic word association of ‘evenly’, as in a million recipes I’ve used over the years: ‘Spread the icing evenly over the top of the cake.’) This section of the book has attracted much attention in reviews and articles. It discusses the well-known underlying economics – skill-biased technical change, Sherwin Rosen’s ‘superstar’ effect, and winner-take-all dynamics when there are high fixed costs and increasing returns.

There is an inevitable tension between progress for many humans and costs for a few humans. For example, IBM is getting its Watson computer to absorb all the world’s published medical information and use it to diagnose and recommend treatments. It would apparently take a human 160 hours of reading time every week just to keep up with the literature. The computer will produce far, far better patient outcomes. But what will happen to doctors’ careers? This is just one of many examples underlining the current angst about the ‘race’ between humans and robots. This is of course also the theme of Tyler Cowen’s excellent recent book, [amazon_link id=”0525953736″ target=”_blank” ]Average Is Over[/amazon_link].

The Second Machine Age has a chapter on measurement about which I want to get a bit picky, having just published my own book [amazon_link id=”0691156794″ target=”_blank” ]GDP: A Brief But Affectionate History[/amazon_link]. Brynjolfsson and McAfee rightly point out that GDP does not measure adequately the welfare benefits of innovation, and specifically the digital goods freely available. After all, if there’s no price, it isn’t part of economic activity at market prices. But they are unclear about the distinction between GDP and consumer surplus, the latter never having been measured. For example, to cite a Tim O’Reilly example, if you dry your clothes in a dryer, you’re contributing to GDP and if you hang them on the line in the sunshine (remember that?), you’re not. It is probably true that the wedge between GDP and consumer welfare has been increased significantly. The authors are also astonishingly uncritical of alternatives to GDP, even describing Bhutan’s cynical PR effort to emphasise Gross National Happiness in place of GDP as ‘promising’.

The third section of the book turns to policies that might minimise the costs of ‘spread’ while retaining the ‘bounty’. The recommendations could be summed up as more and better education, to ensure that humans and robots are complements rather than substitutes, and a minimum income or – better in their view – a negative income tax.

These are obviously valid and interesting suggestions. Do they really address the massive technology-driven structural change described in the first two thirds of the book? As so often, my mind turned to the Star Trek Kobayashi Maru example – a Star Fleet training exercise designed to test the character of a trainee captain in a situation where failure and death are inevitable. But Kirk reprograms the test in order to save the crew. This is where we’re at. We need to reprogram society’s institutions. To use a more topical metaphor, if it rains torrentially for two months, there’s going to be flood water – but the damage it does will depend on the landscape it washes over, the kinds of structures that have been built, river-management institutions.

As [amazon_link id=”0393239357″ target=”_blank” ]The Second Machine Age[/amazon_link] makes clear, rising inequality has occurred everywhere. It is indeed driven by the tidal waves of technology and globalisation. But these are interacting with economic and social institutions. Policy prescriptions will need to look at the minimum wage; at the ownership of assets especially the intangible kind – what are the intellectual property rights in a drug that has been tested on thousands of humans and draws on taxpayer-funded basic research?; at public service provision and the tax base to pay for it; and so on. The institutional revolution that meant the Industrial Revolution ultimately benefited everyone is little-studied ((Douglas Allen’s [amazon_link id=”0226014746″ target=”_blank” ]The institutional Revolution[/amazon_link] is an exception) but just as important.

Anybody interested in these questions will want to read [amazon_link id=”0393239357″ target=”_blank” ]The Second Machine Age[/amazon_link]. I wholly agree with the conclusion: “We need to think much more deeply about what it is we really want and what we value, both as individuals and as a society. Our generation has inherited more opportunities to transform the world than any other. That’s a cause for optimism but only if we’re mindful of our choices. Technology is not destiny. We shape our destiny.”

Minimum wages and robots

Posted on by

This morning I read a good article in The New Republic about the Silicon Valley jobs market and the exercise of power in the labour market. It describes documents showing that the tech giants in 2005 colluded to keep wage rates down. This was news to me and seems pretty scandalous.

The article goes on to discuss in general why labour markets are not like goods markets, although most economics courses, and many grown-up economists, often speak as if they are. The fact that searching for a job is costly gives all employers a bit (or a lot) of monopsony power (or buyer power). And the prevalence of monopsony power needs to be taken into account in analysing the effect of an increase in the minimum wage: it will slightly decrease the employers’ power and reduce both job turnover and vacancies in low-wage jobs. There is some evidence to support this. The article also cites Alan Manning’s excellent book on this subject, [amazon_link id=”0691123284″ target=”_blank” ]Monopsony in Motion[/amazon_link].

[amazon_image id=”0691123284″ link=”true” target=”_blank” size=”medium” ]Monopsony in Motion: Imperfect Competition in Labor Markets[/amazon_image]

The US and UK minimum wages are middling by OECD standards (this chart from the OECD database deflates the statutory minimum by the national CPI and uses PPP for private consumption exchange rates to convert to US dollars). There is no obvious correlation with unemployment rates at this headline level.

OECD real hourly minimum wages

Labour market economics is one of the areas of the subject where economics most needs input from the other social sciences. Jobs and pay can’t really be understood without thinking about factors such as institutions, power, psychology and social norms. Never forget this when next reading about the way technology means inequality is inevitable.

The value at the margins

Posted on by

I’m enjoying [amazon_link id=”0393239357″ target=”_blank” ]The Second Machine Age[/amazon_link] by Erik Brynjolfsson and Andrew McAfee, and will review it soon. Meanwhile, though, I was struck by their discussion of the effectiveness of crowdsourcing for problem-solving or innovation. They give examples such as InnoCentive and Kaggle. I was particularly struck by this comment:

“Another interesting fact is that the majority of Kaggle contests are won by people who are marginal to the domain of the challenge.”

[amazon_image id=”0393239357″ link=”true” target=”_blank” size=”medium” ]The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies[/amazon_image]

For instance, the best forecasts of hospital readmission rates came from people not involved in healthcare, while a contest to develop a computerised system for grading student essays was won by someone whose only expertise was a free online AI course offered by Stanford.

This reminded me a bit of [amazon_link id=”0691128715″ target=”_blank” ]Philip Tetlock’s debunking of professional expertise[/amazon_link], but more of Scott Page’s completely brilliant book [amazon_link id=”0691138540″ target=”_blank” ]The Difference[/amazon_link] (or its follow-up, [amazon_link id=”0691137676″ target=”_blank” ]Diversity and Complexity[/amazon_link]), about the value of diversity. Combining ideas that are ‘far apart’ in some way creates new ideas and insights that are more creative and innovative. Outsiders, the people in the margins, are exactly those who bring new ways of thinking.

Crowds are wise when they are composed of people with different experiences and views. If they are all the same, you get group-think and herding.

[amazon_image id=”0691138540″ link=”true” target=”_blank” size=”medium” ]The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies[/amazon_image]