Every year the RES invites an independent journalist to attend the RES conference and write a report for the Newsletter. This year the RES asked Richard Reeves to do this and his personal view is printed below. Richard Reeves is a former research associate for the Work Foundation and columnist for Management Today and the New Statesman. He has also been the economics, and Washington, correspondent for the Guardian.

Bertrand Russell, perhaps not surprisingly, was not a sunny adolescent: ‘I hated life and was constantly on the verge of suicide,’ he admitted in his brilliant essay The Conquest of Happiness. What kept him from hurling himself off a cliff, he added, was ‘a desire to know more mathematics’.

Russell would have been happy for much of the three days of the RES 2003 Annual Conference. Each presentation contained hundreds if not thousands of numbers, and several plenty of equations, too. On the second night, after too much wine at the conference dinner, I dreamt that I was living in a kind of Matrix-world, with every object consisting of row after row of tiny, flickering numbers…..

This is not a complaint. Economics is necessarily a numbers game. (It may also have been Russell who said that the mark of civilized person was the ability to weep over a column of numbers.) And one of the distinguishing skills of an economist is the ability to look at a page of numbers and see a pattern — see quickly where the ‘action’ is. Just occasionally, though, it is good to remember that the numbers only matter in so much as they tell us something about the world. The numbers are a means, not an end.

How best to summarise a three-day numberfest? It is impossible to do justice to the mountain of work which was represented, tip-only, at the conference. Some presentations or sessions were important enough, or remarkable enough, to be worth summarizing here. And some overall impressions were strong. But so as not to bury the lead, my overwhelming sense was that despite some great research, some startling findings and some top-notch presentations, that the conference as a whole represented a significant missed opportunity for the economics community.

The annual conference is a chance to do a number of things which are difficult to manage on a day-to-day basis. First, to think about the broad state and role of economics in intellectual and public life, and get a sense of the key challenges lying ahead. Second, to hear and contribute to topical debates or controversies within the economics research community. Third, to allow researchers the opportunity to learn from others while their research is in progress. Fourth, to pick up emerging trends in areas of economics outside one’s own field. According to my non-representative sample of one, the conference only really manages the last of these.

OK. Let’s go for some good news. Alan Manning. He may look like a children’s TV presenter — I kept wondering where Bungle was — but he was undoubtedly the conference’s star performer. Giving the Review of Economic Studies lecture, which is for a distinguished young economist, he wins top prize not only for the best one-liner — ‘I now realise that being a young economist is not the same thing as being young’ — but also for providing a model for other presenters. He took a theoretical innovation, applied it to a live policy issue, explained his terminology and held his audience.

The innovation was to look at the labour market from a ‘search model’ perspective, a departure from the traditional approach inspired by Mincer, which looks at human capital development as the key factor explaining wage differentials in a presumed competitive labour market. Manning’s view is the monopsony is a better way of understanding labour market behaviour. As he put it: ‘What happens if an employer cuts everyone’s wages by one pence?’ Not, as the competitive Mincerian approach suggests, a mass stampede for the exit. ‘Stayer bias’ is a hugely important variable in employment behaviour — not one to simply be controlled for, but which has explanatory power. While traditional Mincerian approaches treat job ‘quits’ and ‘lay-offs’ as identical, Manning argued they needed to be treated as distinct — and that additional consideration needs to be given to returns to experience and job tenure; costs of job loss; returns to job mobility; and the evolution of wages within jobs. ‘Progress through life is like snakes and ladders,’ he said. ‘So we need to understand both the snakes and the ladders.’

Manning then turned his theory towards the issue of the gender pay gap, and in particular to the fact that younger women (eg those entering the labour market between 1975 and 1979) are doing no better than slightly older women (eg the 1968-72 cohort). The conclusion was that job mobility, differential promotion chances, training or career breaks for children have little to do with it — this last finding meaning that ‘there is not much action, to be got, from a gender pay gap perspective, from increases in family rights, maternity leave etc’. The principal cause of the pay gap in the early years of labour market experience turns out to be within-job wage growth, with no obviously observable cause.

Why then this difference? For someone who believes that social factors are now impacting economic ones more than the other way around, Manning’s suggested explanation was heartening. Having shown during the course of a theoretical and empirical tour de force, that the principal cause of the pay gap is not E(DW1|X,Z,Q=1)- E(DW0|X,Z,Q=1) but E(DW0|X,Z), he concluded that the reason for the difference may be the fact that there is an economic return to making a ‘song and dance’ about things, and that men are more ‘moaning’ and ‘bolshie’ than women.

One last point about Alan Manning, before we move onto some of the other couple of hundred presenters. He recognized that his audience were from different fields of economics and so took the trouble to explain his terms. At one point he quickly said that, for those who were not labour market economists, ‘years of experience’ meant simply the number of years in the labour market minus age. It was a nice touch: other presenters might take note.

That said, there did seem to be an awful lot of labour market economists about. Indeed, a crude summary of the two species of principal presenter on show would be i) terrifically laid-back, casually-dressed British labour market specialist, likely habitat the LSE or ii) earnest, tie-wearing and slightly anorak-ish American number-wizard, likely habitat MIT.

Of this latter genus, Joshua Angrist was the most perfect specimen. His Sargan lecture, on ‘Treatment Effect Heterogeneity in Theory and Practice’ was dazzling. It was like watching Mozart composing on speed, without being able to read a note of music yourself. I know that Angrist’s work is important and innovative. And I know that the Sargan lecture encourages world-class economists to stretch their theoretical muscles. But it wasn’t the easiest start to the conference — perhaps the Sargan lecture could be moved? One doctoral candidate, who is working herself with some of the techniques of instrumental variables, admitted to understanding ‘about 10 per cent’. One of the UK’s leaders in the field said they got ‘about 63 per cent’. (Only an economist could approximate to 63 per cent.) Was the lecture a success? Not in a conference of fairly young economists from a variety of research branches, if success is related to new understanding. But maybe in this case it isn’t. Anyway, the one thing I learned was that parents who decide to have a third child thereby reduce their chance of still being married some years later by a (significant) 5 per cent, a research finding I passed on unvarnished to my partner, with whom I have two children, aged two and one.

James Poterba (MIT), who spoke on ‘Portfolio Risk and Retirement Saving’ for his Frank Hahn lecture, touched on a huge policy issue here in the UK, the move from defined benefit to defined contribution pension plans. And he threw some light on the attractiveness of Employee Share Ownership Schemes (ESOPs), which Gordon Brown is keen on. In particular, he looked at the investment by workers in their own company’s stock and whether this represented a good risk strategy for individuals. He showed that over time, large-cap stocks and individual company stocks have the same real mean rates of return (9.4 per cent), but significantly different levels of variation (with annual standard deviations of 20.2 per cent and 40.4 per cent, respectively). Simulating the retirement incomes for 300,000 imaginary workers, he concluded that own-company investment could be worth half as much as a diversified holding. Poterba correctly stated that the real issue about investing for retirement is the classic liberal dilemma: should people be free to make bad choices? More specifically, should annuities be required?

A special session on income and wealth in old age added a different twist to the debate. IFS research showed that people are now, on average, more wealthy at the point of death than at the point of retirement. James Banks, one of the researchers, suggested that the UK may now be ‘over-annuitised’ — another fascinating viewpoint from the perspective of public policy. The problem with pension planning is that, as Tibetan writer Sogyal Rinpoche points out, the only two certainties in life are that you will die, but that you don’t know when.

Here, though, is an example of how the conference failed to meet its potential. The area of pension provision is highly topical, controversial and personal. Why not organize a debate between advocates of DC and DB? Or for and against annuities? The conference is instead built on a ‘download/upload’ model: people turn up, download their own research, upload other people’s, and then depart. It is not obvious how this differs, in outcome terms at least, from reading and writing for the right journals. What is surely needed is some debate, some conversation, some interaction. Even where sessions did contain time for Q&A, it was usually limited and limp. The best moments were when one person added some value to the work of another by offering their own insights or experience, or when a conversation led to new questions being posed. But these moments were rare.

This is partly the fault of the incentive structure underlying the conference, and the resulting structure. What seems to happen is that doctoral or post-doctoral researchers apply to present, knowing that if they are accepted, a) they will get funding to attend and b) they can put it on their cv. This means that the conference organizers have an incentive to accept lots of papers, in order to guarantee sufficient attendance (the vast majority of attendees were presenting a paper). The result is that most of the parallel sessions were sparsely attended, with researchers all too often presenting their research in order to get the funding to present their research and then to be able to say they had presented their research. To rub salt in the wounds, the organizers then run special sessions on sexy, topical issues, which seem specially designed to ensure that only the hardiest souls are ever tempted to one of the dozens of parallel sessions

Here, in an empirical spirit, are some numbers. In the last session of the conference, there were 23 people in the special session on Competition Policy and 24 in the special session on Productivity. There was, meanwhile, a mean of 5 people in each of the eight parallel sessions taking place at the same time — which, given that there was a mean of 3 presenters for each, means that the mean ‘audience’ was just two people. Now I’m not an economist, but this doesn’t look very efficient to me.

It is a shame that parallel sessions are so clearly the very poorest conference cousins, given that they can provide a good space for expert gatherings. I will not easily forget the warnings about ‘fat tails’ and ‘improper posteriors’ in the paper from Rodney Strachan on ‘Bayesian Analysis of Stochastic and Deterministic Processes in The Error Correction model’. To explain: ‘there is no more fatal result for a Bayesian analysis than an improper posterior’. In some of these sessions, the specialists did seem to be swapping notes and updating their knowledge. Usually though, the feel was more undergraduate lecture than lively intellectual endeavour (of course the rooms don’t help).

If you were to attend the conference as a policy-maker, or indeed a journalist, the best approach would be to seek out the papers based on government-funded research, which at the request of the relevant department have not been press-released and/or made available on the conference website. Two examples stand out. The work of Jonathan Haskel, Denise Hawkes and Sonia Pereira at the Centre for Research into Business Activity (CeRiBA) represented a significant empirical step forwards in productivity research. For the first time in the UK, two official data sets — the Employer Skills Survey and the Annual Business Inquiry — were matched together, allowing Haskel et al to test the relationship between individual-level skills and firm-level performance. They found that the most productive firms do indeed have significantly higher proportions of skilled labour; the top decile of firms, ranked by productivity, hire workers with, on average, two years more education compared to the bottom tenth. Now why would the DTI and Treasury not want to shout this from the rooftops?

Perhaps because of the next stage of the research. Digging deeper, the CeRiBA team probed the factors underlying Total Factor Productivity (TFP), and found that the difference in employee skill levels accounted for only 7-8 per cent of the TFP gap between firms at the top and those at the bottom. Haskel describes this as ‘puzzling’ — and it certainly flies in the face of the high-skill, high-productivity orthodoxy driving public policy.

The second ‘stealth’ research project, funded by the Department for Education and Skills, was the first assessment of Excellence in Cities, one of the Government’s flagship policies for raising school performance in depriving areas. The basic message of the research, conducted by Sandra McNally and colleagues at the LSE, is that EiC is working. One effect of the policy – which costs an average of £120 per pupil — is to lift 4 per cent of boys up a whole level in the Key Stage Three tests. EiC areas are also the only ones in which there has been a reduction in the number of unauthorized absences. The research also suggests that in EiC areas where more money was spent, results improved more dramatically. Why the DfES were so reluctant to share this knowledge (the researchers were not even allowed to post their paper on the website) is anyone’s guess.

I also got a kick out of the parallel session on central banks. This is probably because I need to get out more, but also because the moment when Gordon Brown announced independence for the Bank of England is one of my most treasured journalistic memories. Sitting between the economics editors of two national newspapers, I heard two simultaneous expletives when Brown dropped his bomb in 1997.

Two interesting papers at the session also provided evidence of the ever-present need to look beneath the surface of research. (Academics are sometimes the very antithesis of journalists, seeking to bury the really interesting stuff in the dense thicket of the paper) Maria Demertzis’s paper showed that central bank ‘transparency’ has no impact on the level of inflation but does reduce variability in inflation (by about 50 per cent), which has clear benefits. The most interesting finding, however, is that among the various elements which comprise transparency, constitutional independence and the presence of an inflation target has no impact at all. Other factors, such as the publication of data and forecasts, clarity about the central bank’s economic model, are more important than institutional arrangements. The top line: transparency is not synonymous with independence, and is more important.

James Talbot, from the Bank of England, wins top prize for research chutzpah. His research compared group decisions on interest rates to individual ones, using a series of simulations with undergraduate and postgraduate economists (and no, there was no difference in the performance of the two groups!). And he did find what he wanted: while the average individual rate-setting ‘score’ was 41, the average group score was 68. Hurrah! Committees are better than monetary dictatorships. But the performance of the best individuals within each group averaged 65 — statistically different from the group score at only the 10 per cent level, as James rather quietly admitted (0.10 being shorthand for ‘Desperately Seeking Correlation’). What this tells me is that an exceptional individual is likely to do at least as well as a normal group. So unless your committee is full of stars, just let Mervyn King set the rates (after all, a similar approach has worked in the US for years).

I feel I should try to resist the temptation to write about Andrew Oswald – it being difficult to imagine an economist in less need of extra publicity. But I can’t. In part, this is because his paper was so straightforwardly interesting, but also because he delivered it in a way that others should emulate. His opening line was, ‘it is very important to get married, especially if you are a man and especially if you are a smoker.’ Using BHPS data, he showed that being married adds 5-7 years to the life expectancy of a man, which is almost exactly the loss of life expectancy associated with being a smoker. Rather than giving up smoking after getting married, as so many of us do, this research suggests that opposite course would be more rational — enjoying a first, post-nuptial, fag.

Oswald’s session was far and away the most interactive, engaged and conversational of any that I saw over three days. Rather than going laboriously through each stage of his methodology, he cut to the chase and then encouraged people to ask questions — some of which were to clarify methodological issues, others to suggest alternative explanations or argue about consequences. One questioner suggested that presence of grandchildren might have an effect, someone else confirmed that the BHPS collects data on this, and Oswald said he would look at it. These few minutes felt like a community of scholars sharing findings, ideas and suggestions, adding value to each other’s work. The lesson — which was also demonstrated by the excellent Chiara Criscuolo in the productivity session — was to cut to the chase at the beginning of the presentation, encourage interruptions and enjoy the conversation.

And now for a delicate issue. How, in a report on the RES Annual Conference, being written for the RES, to say that the address given by the RES President was a disappointment? Steve Nickell is a man whose academic reputation precedes him, and his current status as a member of the Monetary Policy Committee is formidable. The Presidential Address takes place in front of the assembled conference at the end of two long days of papers, numbers and overheads and just before the conference dinner. It is the moment for some broader thinking, some vision, some passion, and maybe even some philosophy.

Prof Nickell admitted up front that he had supplied the title for his talk — ‘Poverty and Worklessness in Britain’ before giving any thought to what he was going to say. I am afraid it showed. At the beginning he described the two central problems of the UK economy as being productivity and poverty, and for an instant it looked like we were in for an interesting overview. It was not to be. (PS to conference organizers: get some roving mikes. No one could hear the questions.) There was little new to be found; the narrative line was shaky and the concluding policy recommendation — ‘bribe’ good teachers into poor areas — was neither original nor fleshed out with any modeling, projections or theory.

The RES Annual Conference could be so much more than it is. Great stuff happens — but, as in so many institutional contexts, this happens despite the structures and incentive systems rather than because of them. Rethink the parallel sessions; rehearse the main speakers; run debates, question times and panels; give more time and better spaces for networking. Otherwise the conference risks becoming one of those events that takes place because it has always taken place. And given the importance of economics not only to intellectual but also to public life, this would be a great shame.