Measuring the journey to a million Welsh speakers

Darllenwch y dudalen hon yn Gymraeg

Cymraeg Logo

The new Welsh Language Strategy Cymraeg 2050: a million Welsh speakers was published recently and contains an ambitious target of a million Welsh speakers by 2050. Measuring how we might meet this target has provided an interesting challenge for statisticians in Welsh Government, and we’ve been working with policy teams to consider how their policy aims might translate to achieving this aspiration, and how many Welsh speakers we might have based on recent demographic trends.


So, how do we estimate future numbers of Welsh speakers?

To start with we revisited the demographic projections of Welsh speakers produced by the Welsh Language Board in 2012. Using the same approach, we modelled trends based on how the language is passed on within a household (language transmission) and a cohort model of Welsh language prevalence. We did this using data from the 2011 Census, mid-year population estimates and the national population projections produced by the Office for National Statistics (ONS). This allowed us to project how many Welsh speakers there might be if current trends continued until 2050. This work estimated that by 2050 there would be around 666,000 Welsh speakers in Wales – 334,000 short of the aspiration.

Using this model as a basis we then considered various scenarios of how you might move from 666,000 to a million in 2050. We were able to model differences in the numbers of adults learning Welsh, different transmission rates within a household, the impact of different Welsh-medium education provision, and the impact of seeing more children retaining the language after leaving school.

This allowed us to set a ‘trajectory’ for reaching a million based on the ambitions of the strategy, that is to: increase Welsh-medium education; aim for 70 per cent of school leavers to be speaking Welsh; and aspirations around the number of adults acquiring the language annually. This trajectory was included in the strategy and provides a picture of the path to a million that will be reviewed continuously. Alongside this work, we modelled the number of teachers that might be needed to achieve aspirations around Welsh-medium education and used the trajectory to set targets on Welsh language use.

Graph of trajectory to one million Welsh speakers


We’ve published a more detailed technical note about this process on our website.

Just to be clear this is only one possible trajectory. There are a range of factors, be they demographic, social or the impact of policy that could well influence the journey to a million. As stated in the strategy, modelling demographic change in the future is complex; modelling the impact of policy changes on specific characteristics of the population is even harder.


What else are we doing to support this work?

Of course, using the Census as a basis relies on having a consistent time series in the future. We are working closely with the ONS on their future plans of the Census and one priority for us is to ensure that there is a robust set of data available on the Welsh language that continues the historical time series. We’re also aware that surveys such as the National Survey for Wales and the Annual Population Survey provide different, usually higher, estimates of the proportion of Welsh speakers. Particularly in recent years we have seen an increase in the percentage of people who speak a little Welsh. To help us understand how people differ in their responses to the Census and surveys we are undertaking some innovative work with the Administrative Data Research Centre-Wales and the ONS to research more into these differences.

As stated above projecting the future is not easy. We will continuously review the trajectory and demographic projections and provide regular updates. We will also share the findings of the work looking at the Census and surveys when it is complete. If you are interested in any of this work please contact

Post by Glyn Jones and Martin Parry, Welsh Government

6 thoughts on “Measuring the journey to a million Welsh speakers

  1. Truly a flawed policy with huge and negative consequences for the democracy and the future of Wales unless this madness is stopped and stopped now.

    We must have an open and a frank debate on this and not spin and deceit that emanates from the Welsh Government in abundance.

    For starters, there are no 0.5 million Welsh speakers as your graphs suggest – Plese see what I wrote on this very subject when the WESP measures were first announced:


  2. There are problems here aplenty. Take the requirement for an increase in the number of teachers. This year, 2016/17, there are 66 fewer teachers working in Welsh medium schools but the number of children in those schools has increased by 886. The number of WM schools has decreased by 7 and the percentage of pupils in WM schooling has increased from 22.49% last year to 22.69%. Your aim is to have 40% of all pupils in WM schooling by 2050.
    There are ways to predict how many people will go into WM teaching in the future; the number applying for WM ITT places last year was 5% of that cohort that entered Welsh first language GCSE. Similarly the previous year’s entry for WM ITT was 5% of the cohort that entered GCSE Welsh L1. Both these years under supplied against the aim of WM teacher recruitment.
    The two previous years, 6% of the Welsh L1 GCSE cohort applied to become WM teachers, the year before that, 2011, 5%.
    The point that I am trying to make is that the teachers that you need to go into WM teacher training are already in the school system. On GCSE entries you can predict that next year you will have 286 WM ITT entries then 278, 276, 297, 295, 290, 281.
    In other words, using the average percentage figures of GCSE Welsh L1 entries who may become WM teachers you already know that you will be below aim right up to 2022. Worse, you can also use KS2 assessments in Welsh L1 in the same way to predict even further into the future.

    Why might that percentage entering WM teaching go down?
    Firstly it is only FIRST LANGUAGE Welsh speakers that are going into WM teaching and the percentage of those pupils in school is a steady 10% of all pupils, 52% of Welsh L1 GCSE entrants. These Welsh first language speakers are in increasingly high demand as Carmarthenshire moves to have Welsh as the language of its LA and the WG pushes to expand the provision of Welsh language services in both the public and private sector.
    Is teaching more attractive as a career than other public service employment? NO.
    Too much competition for a scarce resource.


  3. There is another difficulty that regularly crops up when it comes to prediction. Hywel Jones, the statistician at the Welsh Language Board, made a longitudinal study of Welsh speakers which compared numbers in Wales in 1991 and 2001. The theory was that Welsh speakers in a particular age group would replicate in the next census amongst the people counted as 10 years older. He was checking the cohort as it aged.
    So, in theory, if 40% of 15 year old pupils were able to speak Welsh in 1991 then 40% of people who were 25 in 2001 would be able to speak Welsh. Hywel Jones supposed that any deviation from that 40% would be as a result of migration out of Wales and migration of non Welsh speakers aged 25 into Wales.
    It didn’t work for two reasons, one, picked up by a quality audit of the 2005 Welsh use survey, was that parents overestimated the Welsh language ability of pupils. In 2011 parents may have been more realistic but the suspicion remains that when parents in Blaenau Gwent, Newport and Monmouthshire claim that their child speaks Welsh, although that child goes to an English medium school, those parents are using a very weak definition of “able to speak Welsh”.
    The second reason is one that is very evident where I live in Anglesey; English L1 pupils who go through WM primary school and then Bilingual secondary school speak Welsh when they HAVE to; that is, in school. By GCSE they are planning to enter all English medium papers and writing essays in English if that is their first/ most competent language. By “A” level they are working exclusively in English in order to get their best grades, if English is their first language. After university, particularly if they have travelled to an English university, they have been using a diminishing amount of Welsh since the age of 14 and have grown used to communicating exclusively in English.
    If those people return to Wales and fill in a census they might quite rightly say that they are no longer “able to speak Welsh”. They have not used it and have lost it.
    This is where the prediction data falls down…Welsh attained in school, even WM school, is not necessarily for life.
    What we do know from Hywel Jones longitudinal study is that, in the Fro Cymraeg particularly, non Welsh speaking young people are much more likely to leave Wales than Welsh speakers. I would go further and speculate that young people who have learned Welsh but whose first language is English will also be disproportionately amongst the migrants out of Wales.


  4. What is very serious and has not been recognised about the intention to use Welsh medium schooling to drive up the numbers of Welsh speakers is the effect that this will have on Wales’ ability to compete economically in the modern world.

    You can see what I mean when you look at the specialist degree subjects of students entering teacher training post initial degree. I’m looking at the tables here:-

    The first, very noticeable thing, that strikes you is that 37.6% of students who enter training to become Welsh medium teachers have a degree in…Welsh. Only 2% have a degree in English.
    For students training to teach in EM schools only, 0.3% have taken a degree in Welsh. From this we know that the EM schools will be largely recruiting Welsh medium trained teachers to their Welsh departments…in other words, even fewer WM trained teachers are going to enter WM schools than is apparent at first sight.

    For EM teachers, the degree specialism most often seen is in Maths; 12.8% of ITT entrants have a maths degree against 5.7% amongst WM ITT students. This imbalance exists throughout STEM subjects…WM schools will not have enough teachers who specialise in the “wealth making” subjects necessary in a modern economy.

    The drive for One Million Welsh speakers is going to be a death knell for the Welsh economy and may even be a death knell for the Welsh language as the technologists that Welsh industry requires are recruited from England and the wider world and the Welsh born percentage of the population falls.


Comments are closed.