Some time after Arieh Iserles, Richard Thomas and I organized the letter to the Prime Minister signed by 25 prominent UK mathematicians, criticizing EPSRC’s July 2011 decision to take applications for postdoctoral fellowships only in statistics and applied probability, Philippa Hemmings of EPSRC wrote to us to say that she hoped she would meet us some time in the course of EPSRC’s efforts to engage with the mathematical sciences community.
We decided to take the opportunity to propose a scheduled meeting at which we’d have a genuinely open conversation about what would be best for the mathematical sciences, and what would constitute the best way to use public money. We promised to be as helpful and constructive as possible.
Philippa took this up without hesitation, and we all met at Imperial late in December 2011. Arieh, Richard and I felt the meeting went well. Philippa was interested in our views and candid about how things work at EPSRC. I don’t know whether we accomplished much, but we did learn things we wanted to know — in particular, who made the November 2011 decision that hardly broadened the subjects eligible for fellowships at all. (See paragraph 3 of section 1 below.)
Episode #3: Meeting of Philippa Hemmings (EPSRC, Theme Leader, Mathematical Sciences) with Arieh Iserles, Richard Thomas, and Burt Totaro, at Imperial College London on 20 December 2011
Notes by Burt Totaro, with clarifications/corrections from Philippa Hemmings [P.H.]
1. Shaping Capability and EPSRC Fellowships in the mathematical sciences
The EPSRC Postdoctoral Fellowship is the topic which has roused the greatest anger among the UK mathematical community. In July 2011, EPSRC announced that mathematical scientists from all areas other than statistics and applied probability were ineligible to apply for EPSRC Fellowships, until further notice. EPSRC’s decision was criticized in public letters by all five UK learned societies in the mathematical sciences, by the International Review of the Mathematical Sciences panel (commissioned by EPSRC), by 25 leading mathematical scientists, and by over 300 young mathematical scientists. These groups argued that postdoctoral fellowships play a more important role in mathematics compared to other subjects, and that shutting down EPSRC fellowships in most of the mathematical sciences cuts the pipeline which allows the very best young mathematicians to develop their research between the PhD and a permanent position in the UK.
In October 2011, EPSRC’s Strategic Advisory Team (SAT) argued repeatedly that EPSRC fellowships are particularly important for young researchers in pure mathematics. There was unanimous support for opening EPSRC fellowships much more widely at least at the postdoctoral level. Some possibilities considered were to open postdoctoral fellowships to all of pure mathematics, or to all of the mathematical sciences.
Mathematicians were therefore surprised by EPSRC’s November 2011 announcement that EPSRC Postdoctoral Fellowships would still be restricted to statistics and applied probability, whereas EPSRC Early Career and Established Fellowships would allow certain other topics such as Intradisciplinary Research. Philippa Hemmings said that this decision was taken directly by the EPSRC Executive, meaning David Delpy, John Armitt, and Lesley Thompson. Hemmings said that she had sent the SAT’s recommendations to the EPSRC Executive, but that she also told the Executive that the SAT’s recommendations were not compatible with what she described as EPSRC Council’s policy of Shaping Capability. [It would be more accurate to state that the SAT's most preferred options for fellowship support in the mathematical sciences were not compatible with Council's new Fellowship Framework. The SAT also discussed and endorsed other options that did fit with this framework. (The above is influenced by shaping capability but there are other reasons for the changes to the more focused approach). – P.H.]
Totaro asked whether Hemmings could send the minutes of the EPSRC Executive meeting which explained how they went from the SAT’s advice to their decision. She said she would try to send those minutes, if she could get permission to do so.
In an effort to do something about the pipeline for young mathematicians, EPSRC is expanding its PhD Plus program, now renamed the EPSRC Doctoral Prize, which will allow mathematics departments to keep the best of their EPSRC Doctoral Training Grant students for up to 2 years after the PhD. Hemmings understands that the funding of DTGs is being increased by 20 percent in order to pay for this.
Iserles, Thomas, and Totaro argued that although the EPSRC Doctoral Prize has some value, it cannot replace the EPSRC Postdoctoral Fellowship, which attracted some of the very best young mathematicians from the UK and around the world to the UK rather than to the US or other countries. In the same way, Thomas explained that EPSRC Programme Grants allow a researcher to hire some young mathematicians for a specific project, which has some value, but makes it impossible to hire the very best young mathematicians. The very best need to develop their own ideas, and they have the option of doing so by going to other countries, at a huge loss to the UK’s science base.
David Delpy defended EPSRC’s fellowship decision to the House of Lords science and technology committee by saying that “EPSRC funds 10 postdoctoral fellowships for mathematics each year”, whereas there are “at least 370 fellowships available to the mathematics community. We are not a major player in that space anyway, so I would not say that we are excluding a substantial fraction.” (29 November 2011).
Hemmings sent the 1-page document (with two pie charts) which EPSRC sent to the Council for Mathematical Sciences (CMS) to explain these figures. Iserles, Thomas, and Totaro explained that EPSRC’s document shows the comparison between 10 and 370 to be completely disingenous. To start with, EPSRC has told Penny Davies that the number 370 was a mistake for 310. Next, 39 percent of the supposed 370 fellowships are ERC Starter Grants, which are simply not postdoctoral grants (they can only go to researchers 2 to 12 years past the PhD).
Most important, the supposed 370 “Generic UK individual post-doctoral schemes open to mathematics” include all sorts of postdoctoral fellowships for which mathematicians can apply, even if they are open to people from any subject, so that only a tiny fraction in fact go to mathematicians. For example, EPSRC’s document lists the Royal Society’s Newton Fellowships per year as 16 percent of the supposed 370 places, which would be 59 people; but in fact only 40 Newton Fellowships are awarded each year, and only 1 of the 40 awarded in 2011 was in a mathematical science department.
The realistic comparison, from EPSRC’s document, is with “Individual postdoctoral awards in mathematics”, listed as 30 places. With the ERC Starter Grants removed from those 30, that means that EPSRC has been funding about 40 percent of the UK’s postdoctoral fellowships in the mathematical sciences. That is something EPSRC should be proud of, not dismiss as unimportant. Hemmings agreed to communicate these points to Delpy.
Thomas proposed that EPSRC introduce a new program of EPSRC Fellows in the mathematical sciences, which would be aimed specifically at attracting and keeping the very best postdoctoral mathematicians in the UK. It should match the simplicity of the application procedure for the top US postdocs: just a 1-page application form, a research statement, a CV, and three recommendation letters. It should fund 10 people per year. Hemmings saw the appeal of the idea, and agreed to explore the idea of how EPSRC might best support excellent individuals.
Hemmings explained that the fellowships in Intradisciplinary Research, which EPSRC introduced in November 2011 for Early Career and Established researchers (though not for postdocs), would be interpreted broadly. Any mathematician of the right career stage whose work combines two different areas of mathematics in an exciting way can apply for these fellowships. EPSRC deliberately did not specify in advance what counts as different areas of mathematics for this purpose.
2. One size fits all: is EPSRC willing to treat mathematics differently from other sciences?
Iserles, Thomas, and Totaro argued that EPSRC had never given any explanation of why Shaping Capability was necessary for mathematics. Shaping Capability means increasing funding for some parts of a science while cutting it for others, with funding the best research across the whole science not an option. Since mathematics does not require big, expensive laboratories, Shaping Capability seems to be purely damaging for mathematics, with no compensating benefit. The only justification has been that EPSRC is applying this policy to other sciences, and so it must do the same in the mathematical sciences. Is EPSRC capable of treating mathematics differently from other subjects when appropriate?
As evidence that EPSRC is willing to acknowledge the differences between mathematics and other subjects, Philippa Hemmings mentioned that workshops, networks, and institutes such as the Isaac Newton Institute play a different role in EPSRC’s mathematical sciences program than in other parts of EPSRC. In addition, EPSRC’s Doctoral Training Centres in the mathematical sciences must be in one university, whereas DTCs in other subjects can be in more than one university. The reason for this policy was not clear to anyone at the meeting, however. Perhaps DTCs in the mathematical sciences should also be allowed to be in more than one university.
Iserles mentioned that David Harman and Vivienne Blackstone, formerly of EPSRC’s Mathematical Sciences program, had told him that a mathematician could not be a PI on a proposal to a different part of EPSRC such as Engineering. Hemmings said that that was not true. She agreed to write on EPSRC’s web page for the Mathematical Sciences program that mathematicians are also free to apply as PIs to other parts of EPSRC.
Hemmings explained that the 2010/11 EPSRC Annual Report gives an erroneous figure for the funding of “Mathematical Sciences and Public Engagement”. She promised to send Totaro the correct figures for the Mathematical Sciences Programme over 2010/11 and the past few years. Following previous EPSRC Annual Reports, there are two relevant figures, “Total gross expenditure on research grants shown by programme” and “Research grant proposals considered and funded”. The latter figure for the Mathematical Sciences Programme was cut in half in recent years (2006/07: 21.5m pounds, 2007/07: 24.2m, 2008/09: 15.3m, 2009/10: 12.0m). Hemmings said she believed that EPSRC’s funding of the Mathematical Sciences Programme would go up soon. [I explained that the budget for maths is larger than the maths programme as maths is funded by other parts of EPSRC, including in recent years, a number of Science & Innovation awards. Funding for mathematical sciences in the coming year looks set to increase. – P.H.]
3. EPSRC organization and strategic planning for the mathematical sciences
Philippa Hemmings said that EPSRC Council is the statutory body that runs EPSRC, and that Council has the choice of how much to delegate to the EPSRC Chief Executive.
Asked how EPSRC’s Stategic Advisory Team (SAT) in the Mathematical Sciences is chosen, Hemmings said that normally EPSRC changes one third of the SAT each year, that is, 4 of the 12 members. She will ask the Council for Mathematical Sciences (CMS) for nominations. She acknowledged that it may have fallen out of practice for EPSRC to consult the CMS for nominations to the SAT in recent years.
Hemmings announced a series of activities by which EPSRC hopes to reconnect with the mathematical community. First, EPSRC will invite some mathematical scientists to Bath in January to discuss EPSRC’s recent policies, Shaping Capability and National Importance. [The purpose of the Bath workshop was to discuss strategy for pure mathematics, and national importance and the mathematical sciences. – P.H.]
Hemmings said that the CMS and IOP had nominated a small group of senior mathematical scientists (how many?) to give EPSRC strategic advice. EPSRC intends to approach this group for input in the next phase of decisions about Shaping Capability, until March. That seems narrower than the CMS’s intention of suggesting a group that could give EPSRC strategic advice more broadly. Totaro asked how EPSRC could convince this group that it would be listened to, given that EPSRC rejected the advice of its own SAT as well as the rest of the mathematical community in its November decision not to open postdoctoral fellowships. There was no clear answer. [I will post more about this. I was one of the people CMS nominated. -- B.T.]
Hemmings said that EPSRC and the CMS have started to plan a document which would show the importance of mathematics for the economy. Totaro pointed out the figures from Sir Adrian Smith’s BIS report One Step Beyond (p. 94), showing that postgraduates in the mathematical sciences have the highest starting salary among all subjects.
This is something that EPSRC’s web page and Annual Reports should be celebrating. Instead, EPSRC’s Landscape Document for the Mathematical Sciences sounds strangely condescending about the value of mathematics. Hemmings agreed that the Landscape Document should be improved.
Hemmings mentioned that EPSRC was setting up a steering committee with the CMS to discuss EPSRC’s new criterion of National Importance for all grant applications. She said that John Toland of the Isaac Newton Institute had agreed to take part in that committee. [CMS are represented on the Steering committee for a project we are about to embark on to understand the economic impact of mathematics and John Toland is also a member. I don't believe I mentioned a steering committee to talk about national importance. – P.H.]
Totaro pointed out that the businesses involved in EPSRC’s Council and other advisory bodies are almost all manufacturing companies, even though manufacturing makes up only 20 percent of the UK economy. Shouldn’t EPSRC aim to attract people from financial services, software, and retail companies? They make up a huge part of the economy, they hire a huge number of mathematicians and statisticians, and so they should have a strong connection with EPSRC. Hemmings said only that EPSRC had tried to attract more people from financial services companies in the past, and that BIS (the Department for Business, Innovation and Skills) had made support for manufacturing a priority.
4. Actions to take
Philippa Hemmings agreed to send the minutes for the EPSRC Executive’s meeting which made EPSRC’s November 2011 decision on fellowships, if she could get permission.
Hemmings agreed to tell Delpy the criticisms of his statement to the House of Lords on “10 vs 370” fellowships in the mathematical sciences.
to propose a new scheme of EPSRC Fellows to explore how EPSRC might best support excellent individuals in the mathematical sciences with her colleagues at EPSRC, and if there is support there, also with the SAT.
She agreed to write on EPSRC’s web page for the Mathematical Sciences program that mathematicians were also free to apply as PIs to other parts of EPSRC.
Hemmings promised to send Totaro the figures on EPSRC’s funding of the Mathematical Sciences Programme over 2010/11 and the past few years, correcting the errors in the 2010/11 EPSRC Annual Report. Following previous EPSRC Annual Reports, there are two relevant figures, “Total gross expenditure on research grants shown by programme” and “Research grant proposals considered and funded”.
She agreed that EPSRC’s Landscape Document for the Mathematical Sciences should be improved.