How could physics teacher supply be best improved?

Despite recent improvement in teacher recruitment generally, and physics teacher supply specifically, there remains justified concern about the availability of physics specialist teachers across the school system in England.

This is compounded by the Government’s ambition for all pupils to have the opportunity to study triple science at GCSE, which would depend on greater availability of physics specialists to be implemented effectively. Further policy action is needed to improve physics specialist teacher supply.

In this blog, supported by the Gatsby Charitable Foundation, we summarise analysis from a newly enhanced teacher supply simulation and forecasting model to look at the prospects for physics teacher supply under current policy and in the current economic environment. We also assess the potential impact and costs of new policy measures that could be brought forward to further support physics teacher recruitment and retention.

The availability of physics specialist teachers remains challenging, despite recent improvements in recruitment and retention

Data from the initial teacher training (ITT) census, published in December 2025, showed that physics recruitment in 2025 was much improved on previous years. Trainee numbers increased by 36 per cent on the previous year and represented 77 per cent of the Department for Education’s (DfE) target of how many the system needs, up from 30 per cent in 2024.

This improvement was in part due to a general increase in trainee numbers that has affected all subjects, likely driven by weakening job opportunities available in the wider labour UK market. Physics ITT has also seen increasing numbers of international trainees: non-UK nationals represented 63 per cent of trainees in 2025, up from 12 per cent in 2021. This has itself been driven by the expansion of bursary eligibility to non-UK nationals in 2022.

However, physics recruitment remains below target and sustained under-recruitment over many years means there is a dearth of subject specialists in schools. NFER’s analysis for the Institute of Physics revealed that around a quarter of secondary schools have no specialist physics teacher and a further 30 per cent have only one physics specialist (for context, a typical science department has around ten teachers).

This means the quality of physics education, and progression to further study of the subject, may not be as good as it could be with more specialist teachers available. NFER’s data dashboard demonstrates that schools in disadvantaged areas struggle most with attracting and retaining physics specialists.

Following the recent Curriculum and Assessment Review, the Government aims to expand access to triple science at GCSE by offering the opportunity to all. There is no estimate yet of how many additional teachers would be required to deliver this, but a lot more physics specialists would be needed for this aspiration to be implemented effectively.

Further policy action is therefore needed to continue improving the supply of physics specialists and gradually improve the availability of specialists across the system. The School Teachers’ Review Body (STRB) is currently drafting its recommendations to Government for teachers’ pay over the next three years, amid a wider labour market that is faltering, although projected to improve again.

Meanwhile, the DfE is drafting the delivery plan for its pledge to recruit 6,500 teachers within a severely constrained fiscal environment. What policy measures the Government should take to spend scarce public resource on improving physics teacher supply is therefore a crucial policy question of the moment.

Using our enhanced teacher supply simulation and forecast model, we can analyse the impacts and costs of different policy scenarios

NFER, funded by the Gatsby Charitable Foundation, has developed a new teacher supply forecast and simulation model. The model is an enhanced version of a simulation model we previously used for policy analysis: see here for an example.

A key methodological development is that the model is now agent-based, which models the impact of policy changes on the career decisions of individual teachers. These are not real teachers, but what is known as a synthetic population of teachers. Using data from the School Workforce Census we constructed a simulated population of teachers with characteristics (subject, age, sex, experience etc.) that reflect the characteristics in the real population of teachers.

The model uses the latest evidence on how teachers’ recruitment and retention decisions are influenced by policy changes – particularly related to pay and financial incentives, as this is where the evidence is strongest – to predict how patterns of recruitment and retention behaviour might evolve under different policy scenarios.

Key improvements to the model now enable us to: 

• analyse the likely impacts of policies on teacher supply in terms of teacher numbers
• better assess the impacts on the distribution of teacher characteristics
• better capture the range of plausible outcomes given the substantial amount of uncertainty that is involved in making a forecast.

As with the previous model, the enhanced model is closely aligned to the DfE’s Teacher Workforce Model. More details about the model’s methodology are available here.

We use the model to analyse the likely path of teacher supply over the next four years under current policy proposals and forecasts of pupil numbers and the economy. We then use the model to assess the likely impact and additional costs of different policy measures that could improve physics teacher supply.

We compare and contrast these policy options to identify approaches that are likely to be most impactful and cost effective.

Under current policy, we expect short term improvements in physics teacher supply but continued challenges in the medium term

Our modelling approach begins with a baseline scenario, which forecasts what may happen to teacher supply over the next four years under existing policy. Even though it represents only a suggestion to the independent pay review body at this stage, we include within current policy the DfE’s proposal for teacher pay awards over the next three years.

DfE’s evidence to STRB suggested that ‘a 6.5% pay award over 2026/27, 2027/28 and 2028/29 would be appropriate, with the level of awards weighted towards the latter part of the remit’. We assume a profile of pay increases of 1.5 per cent in 2026/27 and 2.5 per cent in 2027/28 and 2028/29.

We assume in our baseline that the current targeted retention incentive (TRI) policy of retention payments for early-career physics, chemistry, maths and computing teachers is retained long term, even though this has not been officially confirmed by DfE. We also assume that bursaries are maintained long term at their current level, although in the past these have varied in response to changes in the economic and policy context.

Figure 1 shows the forecasted path of physics teacher supply. This is a measure of teacher availability from recruitment and retention flows and does not necessarily predict how many teachers will be employed to teach physics, as this will also depend on school’s vacancies, budget constraints and ability to deploy non-specialists to fill any supply gaps.

The data suggests the under-recruitment in 2022 and 2023 is very likely to contribute to a reduction in the availability of physics teachers in the market until this academic year, but that physics teacher supply is likely to improve over the next two years as the influx of trainees enter the labour market. However, this improvement is very modest compared to the under-supply of specialist physics teachers across schools.

This also assumes the latest cohort of trainees complete their training and enter the state sector at the same rate as historically. Questions have been raised (most recently by the Association of Science Education) about whether this will happen, given that many more are international trainees who tend to experience barriers to entry such as visa issues.

In this baseline scenario, physics teacher supply is expected to plateau after 2027/28, due to an expected recovery in the wider labour market and the long-term effects of a series of pay awards that are not expected to keep pace with average earnings.

As shown by the shaded areas, which represent the range of outcomes from the simulation model, there is also a lot of uncertainty about the longer-term as economic forecasts and teacher behaviour become more challenging to predict accurately. Nonetheless, the central forecast is for physics teacher supply to increase by around 230 between 2024/25 and 2028/29 under current policy and then perhaps fall slightly.

Figure 2 shows a similar picture for recruitment to ITT against the estimated target. Recruitment is forecast to remain high under current policy due to continued weakness in the wider labour market and maintained levels of international trainees, but fall slightly as the labour market recovers through to 2028/29.

While the central forecast is for physics recruitment to remain below target – albeit considerably higher than it has been in the last decade – there is considerable uncertainty associated with this forecast further into the future.

The simulation model also provides further insights on how the physics workforce may change over time under current policy. For example, due to the recent influx of trainees and its continuation into future years, the proportion of physics teachers who are in the first five years of their careers is forecast to increase, from 24 per cent in 2023/24 to 26 per cent in 2028/29.

Given that physics recruitment is unlikely to reach the target under existing policy, this underscores the need to consider what further policy measures might be required to increase physics teacher supply, and thereby increase specialist teaching and the prospects of the Government’s curriculum ambition of triple science for all being feasible to deliver effectively.

Improvements to pay and financial incentives could all drive greater improvement in physics teacher supply, with targeted measures offering the greatest value for money

Using the same simulation and forecast model, we next assess the potential impacts and costs of some potential policy changes available to Government.

We first consider four key policy options in isolation:

1. Pay increase: increase pay for all teachers by one percentage point per year more than in the baseline scenario for the next three years, i.e. 2.5 per cent in 2026/27 and 3.5 per cent in 2027/28 and 2028/29.
2. Bursary increase: raise training bursaries for physics, chemistry, maths and computing from the current £29,000 to £33,000 for the 2027/28 training cohorts onwards.
3. TRI extension: extend eligibility for TRI payments to all physics, chemistry, maths and computing teachers in the first five years of their careers, from 2026/27 onwards. This would extend the current policy to make teachers in all schools eligible for at least the £3,000 payment .
4. TRI expansion: extend eligibility for TRI payments to all physics, chemistry, maths and computing teachers (as above), and apply it to teachers in the first ten years of their careers, from 2026/27 onwards.

We also assess the impacts and costs of two combination scenarios: 

5. Pay and incentives combination: pay increase as in scenario 1 above, bursary increase as in scenario 2 and TRI extension as in scenario 3.
6. Incentives expansion combination: bursary increase as in scenario 2 and TRI expansion as in scenario 4.

Table 1 summarises the key findings from the simulation analyses.

Table 1 Summary impacts and costs from policy simulations

	Number of physics teachers, 2029/30	Physics recruitment to ITT vs target 2028/29	Leaving rate of physics teachers, 2028/29	Total cost above baseline, 2028/29
Baseline	9,550	68%	11.1%	0
Pay increase	+165	86% (+18pp)	-0.5pp	+£1,600m
Bursary increase	+119	81% (+13pp)	0pp	+£90m
TRI extension	+41	72% (+4pp)	-0.2pp	+£40m
TRI expansion	+92	76% (+6pp)	-0.3pp	+£100m
Pay and incentives combination	+331	105% (+36pp)	-0.6pp	+£1,690m
Incentives expansion combination	+209	88% (+20pp)	-0.3pp	+£108m

Note: Total cost includes salary, national insurance and pension costs for teachers of all subjects, bursaries, retention payments and central Government portion of teacher training costs.

The individual policy lever with the largest impact in terms of physics teacher numbers is increasing pay. This change could be expected to lead to more physics recruits and a higher retention rate.

The additional physics teachers retained have a range of experience levels, meaning the policy change would be likely to retain teachers with higher levels of experience as well as more inexperienced teachers.

While the scenario also comes with wider benefits, such as increased recruitment and retention of non-physics teachers, it also comes with the largest additional cost: £1.6bn more than the baseline in 2028/29.

Bursary increases and changes to the TRI policy come with more modest impacts on physics teacher supply but much lower cost because they are targeted towards shortage subjects. Bursary increases could be expected to lead to more recruits and more teachers by 2029/30 at a relatively low additional cost of around £90m.

A TRI extension to all early career teachers in their first five years could be expected to lead to higher retention among shortage subject teachers, resulting in more physics teachers by 2029/30 than in the baseline scenario and a lower-than-otherwise recruitment target.

A further expansion of the TRI policy to include teachers in their first ten years of teaching could be expected to yield a still larger impact of 92 physics teachers by 2029/30, particularly by retaining more teachers with at five to ten years of experience. This could have added benefits of bolstering the availability of experienced physics specialists in science departments.

Combining the impacts of separate policy measures could lead to more teachers being recruited and retained. For example, combining a pay increase with bursary increases and a TRI policy extension could be expected to add more than 300 physics teachers by 2029/30, although the cost remains similarly high to the pay increase scenario.

Combining multiple financial incentives could also lead to more teachers at a relatively modest cost compared to pay increases.

Figure 3 below summarises the trade-offs between the supply impact and cost. While scenarios involving a pay increase tend to achieve the greater scale of difference to physics teacher supply overall, the cost of doing so is much higher than financial incentives that are targeted by subject. Indeed, the incentives expansion combination scenario could be expected to yield more additional physics teachers than the pay increase scenario, but at a fraction of the cost. 

Conclusion

There is no one right answer to the question of how best to combine pay increases and financial incentives into a most effective or cost-effective package. The financial incentive scenarios we have assessed appear to have relatively good cost effectiveness compared to pay increases because they are targeted towards shortage subjects.

However, wider implications such as fairness and the impact on other outcomes such as the number of teachers generally and experience levels are also factors that policymakers need to consider.

Policymakers should therefore focus on developing approaches with the evidence in mind and with regard to the trade-offs involved. Our analysis indicates that combinations of policy changes to pay, bursaries and early career payments can be powerful for developing an impactful and cost-effective long-term strategy.

However, given the relatively healthy state of teacher supply for primary and many secondary subjects and the tight fiscal environment, now seems a very appropriate time for policymakers to consider policy measures that are tightly focussed on subjects in greatest need of targeted support, such as physics.

Acknowledgements: Many thanks to Dawson McLean and Sarah Tang for developing and building the simulation and forecast model and conducting the scenario analysis and to the Gatsby Charitable Foundation for funding this research project.