T. M. F. Smith, 1934–2019

Terence Michael Frederick Smith, born January 18th, 1934, died December 7th, 2019.

Terence Michael Frederick Smith: Michael to his original family and Fred to the rest of the world, was born in Ilford into a family that had no history of high educational attainment. At 11 years of age he won a scholarship to the local grammar school which made him one of the first to benefit from the great Butler reform of the 1944 Education Act. He was a good scholar, in the top two or three of his year, but as a schoolboy had no ambition to go to university.

The family were not well off and if he had any idea for his future it was to become an accountant. At this time, this required being articled and there were costs which ruled this out. However, in the insurance world, would‐be actuaries were paid a salary while training and so an actuary he would become. He joined Commercial Union and for 2 years worked and took qualifying examinations. But it was not something he enjoyed and after 2 years he had had enough. He decided instead to complete his national service and joined the Royal Air Force. He became an aircraftsman and eventually reached the dizzy height of an acting corporal (paid). He had hoped that this would take him abroad as the furthest he had been to this point was a trip to Jersey. However, after basic training he was posted to Epping from which he could cycle home at nights—so much for seeing the world!

He looked back on national service as a complete waste of time and yet it seems to have had some strong influences, albeit unintended, on his future. He was dismayed by the general attitude of those in authority and the way that weak people were regularly abused. He was a keen sportsman, tall and physically fit and so not a victim of this but he did not like what he saw. Throughout his life, Fred had a healthy disregard for authority and, whatever the origins of this, it was reinforced by his experiences during national service. The other important effect of doing national service was that he met people who intended to go to university afterwards and others who were completing national service, having deferred it to complete a degree first. This seemed as good an idea as any and there was an in‐built incentive: anyone going on to university would be released from national service 2 months early to be available at the start of term. However, there were a couple of problems. His choice of pure mathematics, applied mathematics and economics for A level ruled him out from some university courses and in addition he had not studied Latin or Greek at school. A friend he knew was going to the London School of Economics (LSE) which was more relaxed about this. He wrote and received an acceptance by post, sight unseen. He had studied some statistics during his actuarial training and quite enjoyed it so statistics it would be.

The LSE was a real eye‐opener. There were people from many different countries and open meetings on all sorts of topics. Also in the first 2 years the course involved only a quarter devoted to mathematics and statistics so there were courses on economics, philosophy, politics and so on. Fred described it as a wonderful educational experience. He had always been interested in politics and on the left of centre. He remembers in the last year at school going to Gants Hill where speakers would stand up and argue for or against some issue or other—an Essex equivalent of Speakers Corner. When he left national service to go to the LSE he was placed on the reserve list for further service if the need arose. This was the time of the Suez crisis and Fred decided that if recalled he would refuse to serve as a point of principle. He was not a pacifist but felt that the British intervention in Suez was fundamentally wrong. In the event he was not recalled and his moral choice not put to the test.

Sport was always important to Fred and from the time of leaving school, through national service and until he finally left the LSE, he played football and cricket for his school's Old Boys’ teams. He recalls that, at the time of the Suez crisis, nine out of the 11 were strongly in favour of intervention whereas he and another were opposed. He was known by his team members as ’that commie bastard Smith’ but whatever the differences he maintained friendships with a number of those players for the rest of his life.

The LSE was a great experience. The staff included Jim Durbin, Alan Stuart, Morris Kendall, Maurice Quenouille and his tutor Claus Moser. Fred remembered revising for his finals from Jim Durbin's notes on regression and linear models and recalled the clarity and elegance of these. Given his future career it is ironic that the topic he liked least was sample survey theory. However, he did recall cycling to school at the time of the first post‐war election. He saw blue posters in almost every house and it seemed inconceivable that Labour could win but after the election this was the result. He remembered it as his first exposure to the bias of selection effects, although whether this was the way he reflected on it as a 12‐year‐old at the time is unclear.

Fred graduated, was awarded the Farr Medal for best statistics student and, for want of anything else, took up a postgraduate studentship. This was the time of university expansion and within months there was a vacancy for an assistant lecturer and he was appointed. He knew very little statistics, his final year being his only intense period of statistics and mathematics training, and he tended to have mathematics courses to teach since his senior colleagues wanted all the statistics courses. He had no thesis topic and worked his way through the problems in Kendall and Stuart's book which he regarded as the best statistics training he had to that time. After 3 years he had carried out no real research, had no thesis topic and no publications and was coming up for review. Fortunately he was approached by someone in accountancy with some problems which led to a couple of applied papers and was promoted to a full lectureship on the back of these. He returned to topics on statistics and accountancy several times in his career.

At this time he entered perhaps the most formative period of his professional life. Two young men, Alastair Scott from New Zealand and David Brillinger from Canada, came to the LSE on post‐doctoral fellowships and together with Fred, John Fryer and others formed a close relationship. They had the common interest of statistics but also sport and all had young families and shared values. They started working through the Box–Jenkins papers and Fred learned from the others how to examine things in detail. This gave him the real basis for his statistical development and also a long‐term interest in time series analysis. There were plenty of distinguished visitors passing through the LSE and he remembered Godambe and Ericson giving papers to the Royal Statistical Society in which he contributed to the discussion. As a result he became interested in sample survey inference.

In 1968 he followed Maurice Quenouille to Southampton and remained there for the rest of his working life. He corresponded with Alastair Scott who had returned to New Zealand. They visited each other and together published a string of important papers in various journals on questions of inference in sample surveys including ‘Estimation in multi‐stage surveys’ (in the Journal of the American Statistical Association ) which was a forerunner of subsequent approaches by others to small area estimation. This association was very productive and they also published formative papers that combined their two main areas of interest on the use of time series models to analyse repeated survey data. Again there was an influential paper for subsequent researchers. He also worked with colleagues, most notably J. A. John on non‐orthogonal experimental designs.

In 1973 I arrived in Southampton and we formed a strong association and close friendship. Indeed from the outset we decided to build up work on sample survey theory. Just as many statisticians who passed through the UK tended to visit the LSE, so our success measure would be that when distinguished survey statisticians passed through the UK they would want to visit Southampton University. We published on post‐stratification and the use of models for survey inference and together developed a research programme on the analytic uses of complex survey data. This brought many distinguished survey statisticians to Southampton including Alastair Scott, Wayne Fuller, Jon Rao, Gad Nathan and Danny Pfeffermann. Associations were forged in Southampton that would result in many important papers. These personal connections were an important outcome from the programme.

Fred was an excellent teacher who could convey the most complex ideas with clarity and enthusiasm. He enjoyed teaching, just as he enjoyed working with others—the vast majority of his published papers were jointly authored. He loved to explore ideas, in statistics or economics or public policy, and was always willing to argue his position. But he was a true academic and would freely admit a change of view if he was convinced by rational argument or analysis.

He was a member of the University Grants Council mathematics subcommittee and proud to have been a panel member for the first three research assessment exercise panels. He was awarded the Royal Statistical Society's Guy Medal in Bronze in 1979 and was President of the Society in 1991–1993. During his Presidency the delicate negotiations for the merger with the Institute of Statisticians came to fruition. The merger was smooth and successful thanks to everyone who played a central part. Fred played his part in university management as Head of Department and Dean. He enjoyed it but never wanted to move on fully to university management—a true academic who always wanted the stimulus of colleagues and the joy of exploring ideas in his chosen field.

While at the LSE he married Jean, who sadly died in 1982 after a prolonged illness. He subsequently married Ruth and they enjoyed a very happy marriage for over 36 years.

Fred's abiding characteristic was an unfailing positive approach to the people around him. He was always supportive, willing to listen to ideas and to contribute generously. This was true of any doctoral students, both those he supervised and others who were around. It was also true of his colleagues. He contributed enthusiastically to seminars and continued to be present in the Statistical Sciences Research Institute at the University until ill health finally prevented his attendance. His enthusiastic support found a new outlet in later life as Ruth completed a Master's degree and developed her own work. He was proud to change his role from academic leader to bag carrier and supporter as Ruth travelled to give presentations and to establish collaborative work in China.

He is succeeded by his three children with Jean: Cathy, Susan and Ian, and Ruth's two children Geraint and Kate and in total eight grandchildren. The family have been a constant source of support to Ruth and to each other in their loss.

Tim Holt