Teacher Supply Model: a technical description

This week the DfE issued its response to the Select Committee request for an explanation of how the Teacher Supply Modelling (TSM) process works. It took the DfE just 20 pages of lightly argued text to explain the principles to those unfamiliar with the process. This is the third such Report in response to Select Committee inquiries into teacher supply over the past 25 years. The first, issued in 1990, and entitled Projecting the Supply and Demand of Teacher – A Technical description, ran to some 78 pages in length. The second, issued in 1998, and entitled Teacher Supply and Demand Modelling – A technical description – was even longer, at 85 pages. Both received Ministerial endorsements. The first was endorsed by The Secretary of State at the time, Kenneth Clarke, and the second by Estelle Morris, the PUS of the day. The new document is seemingly devoid of any ministerial endorsement or support.

What is clear after looking at the three documents is they manner in which the TSM process has been pared down and simplified over the years. The fact that the TSM is now only run for five secondary subjects and primary (page 23) plus a catch-all is used for other secondary subjects where the numbers are then ‘divided between other subjects proportionally according to data from recent years’ must be worthy of debate by the Select Committee. What data sources are used to establish the distribution? Is it the number of teachers in the subjects as measured by the School Workforce Survey or the amount of curriculum time allotted to each subject? The absence any overall modelling for the sciences, and a concentration on just Physics and Chemistry, is also worrying.

However, the most concerning part of the document is the single paragraph on page 20 entitled Ensuring the robustness of the TSM. The paragraph is worth quoting in full.

‘The estimates in the TSM are based closely on data trends from recent years with adjustments made from known policy changes. The robustness of the TSM is assured by sensitivity testing the model against variations in all the assumptions.’

Now the earlier documents did at least identify what some of the assumptions might be. In the new document we are told of completion rates for ITT routes in the past, but not the assumption used for School Direct that has replaced the former employment based routes into teaching. We are also told of pupil growth, and of retirements, and the outcome assumption for wastage rates of teachers leaving the profession, and for those joining both from ITT and from outside the state-funded sector. However, the comments about the success of these teachers in returning as contained in section 2.3 are somewhat opaque to say the least. Here, as elsewhere, worked examples might have added to the understanding of the process.

The modelling of wastage really identifies the whole issue with the methodology: it is essentially backward looking for its inputs. This may not matter when economic and other societal trends are relatively unchanged from year to year but it risks failing to capture major shifts in the labour market until well after they have occurred. This is why the failure to discuss the outcomes of the TSM and a range of options with the wider education community always puts the government at risk of catastrophic failure in teacher supply. The situation hasn’t been helped by the lack of a desire on the part of the wider community to systematically try to replicate the TSM for its own benefits.

The section on page 19 of the new paper dealing with stability in the ITT market and the calculation of the optimum number of ITT places seems at odds with the reality of the 2014 allocation where if Table 5 is correct the estimate of places required was 34,890, but the number allocated was  some 41,000. Now either this means that the government believed that only allocating the estimated number wouldn’t produce enough trainees or it was prepared to put the Treasury in hoc for extra fee for some 6,000 students at £9,000 a throw. What happens between now and August will be of great interest, not least to the 130 history graduates likely to be recruited above estimated need.

Missing from the document are a number of areas of importance: the policy assumptions about school budgets and the effects of the minimum funding guarantee, the  consequences of the Pupil Premium; and the possible new funding formula; the presence of Teach First, and any likely increase in the use of teachers not put through the training process informed by the TSM; the effects of shortfall in recruitment into training from year to year .This last point is covered on page 1, where it is stated, ‘undersupply is double-weighted to reflect that a future shortage of state-funded teachers would be less desirable than a future surplus’: a sensible policy option. However, it is not clear how this works in practice.

Novices to the TSM process may find the document helpful at a basic level, but, by ignoring any debate about how effective the past is as a guide to the future, and also avoiding discussions about whether the TSM is part of the process in defining ITT numbers that Ministers can then change on the advice of others, the document provides little insight to the decisions taken during the past two years about how many teachers to train. Possibly, we will learn more when Mr Taylor speaks at the North of England Education Conference next week, but a lingering doubt must remain that as he said at the same conference last year:

In the future I would like to see local areas deciding on the numbers of teachers they will need each year rather than a fairly arbitrary figure passed down from the Department for Education. … I don’t think Whitehall should be deciding that nationally we need 843 geography teachers, when a more accurate figure can be worked out locally.


3 thoughts on “Teacher Supply Model: a technical description

  1. It would be interesting to see the separate calculations for the 5 secondary subjects plus MfL and ancient languages, and as you say the grouped other secondary subjects. I’m still intrigued with the student-teacher ratio assumptions and the capping process for this based on historical patterns – this ignores quite large differences between the state and independent sectors according to OECD data.

    • There may be enough data in the School workforce survey to develop a proxy model. But, the key issue is how to make the TSM more forward looking during periods of change and not just identifying what happened in the past and using that as the basis for the future. Oh, and some clearer explanation of how policy assumptions are built into the modelling

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s