Some years ago now, I started serious work in the Population Estimates Unit of Population Statistics Division 2 of the Office of Population, Censuses and Surveys, then based in Somerset House, subsequently kicked out of there (too grand for them) and eventually subsumed into what is now the Office for National Statistics. One of our tasks was the production of the annual mid-year estimates of the population of every local authority - then around 400 of them as I recall - one number each - in England and Wales. There was also the question of production and dissemination of national statistics, shared with Scotland, Northern Ireland, possibly with the Home Office and what was then the Government Actuaries Department. Quite a lot of this work went on in an out-house in Stubbington near Fareham in Hampshire.
At the time I joined the unit, the estimates were produced in manuscript, with the help of large and noisy calculators - which looked a lot like typewriters - on very large working sheets. Loss of or damage to which might well be disastrous. Most of the data came from the (usually) decennial censuses, the registration of births and deaths and the electoral register. Internal and external migration were tricky, with the result that these estimates were sometimes well adrift as they approached the next census.
By the time I left, the unit was moving into the world of computers, in the form of terminals into an ICL computer running the then advanced George III. And, I imagine, not long after that it moved into the world of PCs and the products that became the Excel we have now. It was also moving beyond one figure for each authority into attempting to estimate by age and sex.
These estimates were important to local authorities as they were an important element in the calculation of the central rate support grant, then as now a very important element of local authority finance. There were occasional meetings with town clerks who tried to persuade my boss to adjust the estimate for his town. And the drift mentioned above sometimes resulted in Questions in the House.
One of the hot topics of the time was how much easier all this would be if we moved back to the war-time scheme of a national register with identity cards - a scheme that failed then and has failed several times since. Although, over the water, the Irish have managed a useful step in the right direction, as noticed at reference 2.
One of the curiosities of the time was that this sort of statistics was much more like accounting than the statistics that one might do, for example, on an agricultural research station. I was not qualified in that sort of statistics, but a lot of my colleagues were. While population estimates were more a matter of new population equals old population minus deaths plus births plus a few twiddles for migration. Clerical errors were the issue, not standard errors and p-values.
All of which has left me with a soft spot for the estimates, so I picked up this morning at reference 1 on the publication of those for this year.
I learn that migration of the ordinary sort is still difficult and it looks as if all the movement, all the toing and froing generated by the massive growth of higher education has become difficult. And part of the response has been the development of something called a dynamic population model, which appears to sweep away the curiosity noted above and I do not get anywhere near the statistics of reference 6. No doubt it is all good stuff, but I do wonder whether the sophistication of the analysis has outstripped the quality of the data.
Difficult to the extent that we have official statistics and developing statistics produced in parallel, with the latter slated to replace the former in a few years time.
I also felt that while a lot more stuff is published than was the case in my day, a lot of it in the form of Excel workbooks, it is also rather less accessible. Rather less effort is put into the descriptive task. Nor did I care for the emphasis on change: I would have preferred a bit more on levels, more on where we have got to and less on how we got there.
PS 1: the Stubbington operation was on the first floor of one of the retail blocks around the green. But Stubbington is no longer listed by ONS as one of their sites and I am unable now to say which one it was. But a sample is snapped above. Maybe the one middle left.
PS 2: I was reminded of the existence of the National Health Service Central Register, in my day a massive clerical operation involving lots of cardboard and lots of big ledgers, then housed in a former resort hotel in Southport known as Smedley Hydro. The posh bar in Southport at that time was called the Bold, where they were too posh to serve pints, although they would allow you to buy two halves instead. They probably sold sherry too.
References
Reference 1: England and Wales record biggest annual population growth in 75 years: Office for National Statistics points to impact of immigration as births slip to two-decade low - Valentina Romei, Financial Times - 2024.
Reference 2: https://psmv5.blogspot.com/2024/04/senior-moments.html.
Reference 3: Population estimates for England and Wales: mid-2023: National and subnational mid-year population estimates for England and Wales by administrative area, age and sex - ONS - 2024.
Reference 4: Population estimates for England and Wales, mid-2023: methods guide: National and subnational mid-year population estimates for England and Wales, broken down by administrative area, age, sex and components of population change - ONS - 2024.
Reference 5: Dynamic population model, improvements to data sources and methodology: local authorities in England and Wales, mid-2011 to mid-2023: Update on the data used by the dynamic population model (DPM) to produce admin[inistrative data] based population estimates (ABPEs) - ONS - 2024.
No comments:
Post a Comment