Johan Fourie's blog

I'd rather be a comma than a fullstop

Archive for July 2018

What universities can teach us about job incentives (or how to make South African researchers more productive)

with one comment


Let there be little doubt: academics have the best jobs. When we teach, we get to fill young, smart minds with ideas we care about and believe in. When we undertake research, we get to explore these ideas further, understanding the world and how it works a little bit better. We work in tranquil settings (most of the time), surrounded by like-minded individuals in search of (the) truth, or, for those of us who shy away from people, books that do the same thing. Sometimes we get to travel to nice places to meet more like-minded people and share our ideas. Sometimes we even take sabbaticals, a time to reflect more deeply about the world and how it works without the need to do anything else. And best of all: even if we do not do most of these things, we have job security for life.

Universities are some of the oldest institutions. Although the role of professor has changed somewhat over the centuries – we used to have to earn our income when students paid to enter our classrooms! – the system of academic tenure, where an appointment is permanent and one cannot be fired except under extraordinary circumstances, has been around for more than a century. It is a decidedly different system than the private sector, where the biggest incentive for working hard is to not get fired.

While South African academics get tenure almost immediately after their appointment (it varies, but there is usually a probation period), ‘getting tenure’ is a big thing in the US. The first five years after appointment is a race to publish in top journals. If your tenure evaluation comes up, and you have not published well enough, you won’t get it, and you will have to move somewhere else, or quit academe. Once you get tenure, though, all the incentives to publish are removed; continued research depends entirely on the goals and objectives you set for yourself.

Here are two very different systems that are perfect for analysis. In the first, the incentives are clear: publish or perish. In the second, there are no external incentive like the overt threat to job security. Which of the two systems produce the best results?

Before answering this question, it is perhaps useful to ask why the system of academic tenure was introduced in the first place. There were mainly two reasons. First, tenure provides academic and intellectual freedom to pursue new avenues of inquiry. Second, it provides a sufficient degree of economic security to make the profession attractive. It is the first of these – the unencumbered pursuit of truth – that is still upheld as the indisputable defense for tenure.

Does this defense stand up to empirical support? Three economists, Jonathan Brogaard, Joseph Engelberg and Edward van Wesep, used their own profession to find out. In a paper published in the Winter 2018 issue of the Journal of Economic Perspectives, they measure the research output of almost a thousand academic economists in the five years before tenure and the ten years after. They not only measure the quantity of output, but also the quality. They create two measures: ‘home runs’ are papers that are highly cited (in the top 10% of papers published in the same year) and ‘bombs’ are poor-performing (papers in the bottom 10% of citations that year).

Their results are emphatic: publication and home run rates rise to tenure, peaking in the year a researcher comes up for tenure and a researcher’s first year as tenured faculty, but then fall off a cliff, with publication and home run rates 15% and 35% lower in years 2 to 10 after tenure. Most surprisingly, bomb rates, publishing papers that get very few citations, increase by 35% after tenure.

The authors consider various reasons that might explain this drop in productivity and success. Perhaps this is just a ‘time since PhD’ effect, in that older people are less productive, but the authors find no evidence to support this. Perhaps it is the rise in service, teaching and other nonacademic obligations post-tenure, but that would not explain, for example, why researchers publish more bomb-papers. Perhaps tenure encourages researchers to take bigger risks and branch out into new, explored areas of research. The authors measure this by looking at where the authors publish, and find no difference in the number or uniqueness of co-authors or journals. Perhaps the averages mask elite researchers’ performances. But even if the authors only limit their analysis to the top US universities, the results hold true. Perhaps it takes time for truly novel research to gain traction. But when the authors limit the sample to papers with 20 year lags, the results stay the same.

What emerges from their analysis is that tenure is bad for research productivity. This is not necessarily to say that the tenure-system is bad: had it not been there, the number and quality of PhD students that aim for academic positions would probably have been lower. The possibility of future economic security is the incentive that really matters in drawing the sharpest minds into the field.

But it does suggest two things. On a practical level, giving tenure too early may be a bad thing. The South African system almost assumes tenure at the time of appointment; I don’t know anyone that has not received a permanent appointment for failure to publish. By extending the timing of tenure to at least five years, and making ‘not getting tenure’ a realistic threat, the South African government can get more research for their proverbial buck. At a more general level, the study clearly shows how important incentives are. A world where permanent employment is guaranteed with no performance appraisals is a world where output falls and innovation dies. Even academic economists sometimes need reminding of that.

**An edited version of this article originally appeared in the 21 June edition of finweek.


Written by Johan Fourie

July 19, 2018 at 07:30

Join me in New York and Boston!

leave a comment »


On 26 July, the Economics department and LEAP will host a Stellenbosch alumni event in New York. I’ll give a short talk on ‘The Data Revolution in African Economic History’. Four days later, on Monday, 30 July, we’ll host another event in Boston. If you are in the neighbourhood and want to drop in to hear what we’re doing at LEAP, in the department and at Stellenbosch University, please send me a mail. In New York, we’ll meet at the ING offices and in Boston, we’ll be at the Residence Inn hotel in Cambridge.

The reason I’ll be in the US is to participate in the World Economic History Congress, which is hosted by MIT this year. I’m responsible for five papers (yes, I know, this is bad planning), so it will be a busy conference. The programme can be downloaded here. I hope to share some of the results of this research on this blog over the coming three months. (Also, I’m excited about plans for a new look blog. More about that later.)

After the WEHC2018, I’ll take a two-week break before going on a seminar tour in Gauteng, delivering papers at the universities of Pretoria (27 August), Wits (29 August), Johannesburg (30 August) and North-West, Potchefstroom (31 August). I’ll post more detail about those talks closer to the time.