Erwin Bolthausen became IMS President in July (he is pictured below, center, with Past-President Bin Yu, left, and President-Elect Richard Davis, right). Erwin shares his thoughts on the future of the institute:
The IMS is extremely successful as a scientific society, mainly as a scientific publisher, and also with its important conferences. The institute is also very cost effective, which results in quite low prices for its top journals. When I was editor of Probability Theory and Related Fields (PTRF), we were fighting with Springer, its publisher, about lowering the high price of this journal, by arguing that the price per page of PTRF surpassed that of the Annals of Probability by a factor of 17. This did make Springer knock it down; I think the factor is now about five.
Although at present things in IMS are running perfectly, there may well be some problems in the future, and the Institute will do well to think about and prepare for a number of potential issues ahead. I present here some personal thoughts on topics which in my view are important.
The most severe problems that IMS will probably be faced with in the coming years are to do with the rapid changes in the field of scientific publishing. Already income from subscriptions is slowly decreasing, which, for the moment, can easily be offset by slight increases in our prices. The reason for this decrease in subscriptions seems evident and unavoidable: Formerly, many universities had more than one subscription, one for the central library, and then another, or several more, for individual institutes. That is no longer necessary.
Apart from this rather minor problem, there is a lot of insecurity about the further development of mathematical and statistical publishing, and of course publishing in science, generally. One has also to be aware that our situation considerably differs from that in other fields of science, like in biology, where the relevance of a publication is decaying much faster than in our fields. Therefore, the experience of others is only of rather limited relevance for us. The one big problem we are facing is that, evidently, the journals nowadays are of much less relevance for the reader than they used to be. They are of course still of tremendous importance for young authors who need publications in highly estimated journals for getting a job.
There are many visions around how the system could change. Commercial publishers, like Springer, come up with the idea to get journals financed to a large extent by publication charges, and no longer so much by subscriptions, hoping that science foundations are strict with an open access policy. Another idea is that scientific publishing in its present form could be abandoned completely, and the refereeing process could be replaced by open electronic discussions, for instance in the arXiv. I myself don’t believe that this would work properly, but one has to admit that within the present refereeing system, many results pass through which contain serious mistakes. So, one may ask what the advantage of the present system is. If nobody reads a paper, then it is irrelevant anyway whether it’s correct or not, and if it is read, then possible mistakes would come to the surface maybe more rapidly with the proposed system than they do now.
I believe that the IMS should find a strategy how to face these problems, which has to be flexible, as nobody can foresee precisely how things will develop.
There is a different problem the IMS is faced with, namely to define in future its scientific identity. I am not very familiar with the history of the IMS, but from its name I conclude that, originally, it was thought to cover that part of statistical science which has a mathematical foundation. This foundation, to a large extent, was closely connected to probability, and in fact, big parts of probability theory were at that time closely tied to statistical problems. However, probability theory and statistics have drifted quite apart nowadays. As a probabilist, I don’t have a profound knowledge about the situation in statistics, but for probability theory, it is clear that many of the major developments of the past 20 years have now no close connections with statistics, but more so with other branches in mathematics, like complex analysis, differential equations, algebra, number theory, or mathematical physics. As a consequence, a majority of the younger generation in probability is no longer interested in a membership in the IMS, but would rather join mathematical societies. After I became president-elect of the IMS, I did ask many of my friends and younger colleagues why they are not IMS members. Most answered that it makes more sense for them to join a mathematical society.
The visibility of the IMS for theoretical probabilists is essentially reduced to the fact that we are publishing the most important journals in the field. It has the, maybe minor, practical consequence that it becomes more and more difficult to find probabilists to work in committees of the IMS, but I think there is in the long run the more important problem of the identity of our society. An important step for the IMS certainly has to be to open up to new developments which are important for statistics, like machine learning and big data, for which modern mathematical tools are important. I think, it will always be important for the IMS to stress the connections with mathematics, as for “non-mathematical” parts of statistics, there are other societies.
I don’t want to give the impression that I feel concerned about the future of the IMS. Quite to the contrary, it is a great society, and it is of utmost importance for the future of our scientific fields, and I am sure that it will find the proper answers for the challenges posed by recent developments.