Economic development and the Welsh language?

by Sep 10, 2020Arfor, Featured

By Dr Dyfan Powell, Wavehill

The idea that economic development creates jobs and allows Welsh speakers to work and live in the area[1] is one that is at the forefront of economic and language policies. It perhaps stems from the long-standing narrative that economic development and linguistic growth or viability go hand in hand. But is this true? If so, a positive relationship should be expected between economic data and the number of Welsh speakers i.e. as economic indicators increase and strengthen, speaker numbers should also increase. As part of contextual research for the Arfor programme, data was scrutinised to try and understand the statistical relationship between the economy and the language.[2]

During the course of the project, employing econometrics techniques, we sought to explore this narrative, and in particular, to explore the relationship between economic development and growth or sustainability in number of Welsh speakers in the Arfor area. GVA growth of was lagged for 1 and 2 years to scrutinise any delayed effects, and a wide range of additional economic variables were also examined. The results are shown below.

A score of +1 in Table 1 below shows a perfect relationship between variables, i.e. that the number of speakers increases as GVA does. A score close to zero indicates that there is no meaningful relationship between GVA and speakers’ numbers, while a score of -1 indicates a perfect negative relationship i.e. as the GVA grows, the number of speakers decreases.

As shown in the table below, having lagged GVA growth by a year, we only see a positive relationship in Gwynedd, although it is not particularly strong. Overall however, there is very little positive relationship between growth in GVA and the numbers of Welsh speakers.

Table 1: Correlation Coefficient (GVA and Welsh speakers)

Baseline Lagged after 1 year Lagged after 2 years
Gwynedd 0.029 0.145 -0.201
Anglesey -0.522 -0.607 -0.661
Ceredigion -0.325 -0.240 -0.129
Carmarthenshire -0.132 -0.089 0.059
Arfor -0.360 -0.264 -0.319

Data: ONS

Indeed, it is generally a negative relationship i.e. as the economy strengthens, the number of speakers declines. In the case of Ceredigion and Carmarthenshire, this negative relationship tends to weaken further over time, but in Anglesey this negative relationship is strong, and consolidates over time. Similar or slightly less negative results are reflected when employment is included as a variable in the regression analysis.[3]

Despite these results, it should be emphasised that there were time, resource and data constraints to the work. The experimental nature of the research must also be noted, and we should not look to develop far-reaching or sweeping conclusions on its basis. These results should be considered as preliminary findings only and as a starting point for further research.

However, the data does inform the discussion around the relationship between the local economy and the viability of the Welsh language. The data does not support the rationale and narrative that economic development, and in particular the creation of more and better employment, is linked to an increase in the number of Welsh speakers.[4] In fact, the data suggest the opposite; that the economic development seen in Anglesey, and to an extent in Ceredigion and Carmarthenshire since 2005, is related to a decline in the numbers of Welsh speakers.

It must be emphasised that the evidence does not support the suggestion that economic development is causing a fall in the number of speakers. But there is a statistical relationship between the economic development in places such as Anglesey and a reduction in the number of speakers The nature of that relationship is one that needs to be further explored and understood if policy makers wish to avoid economic development that is detrimental to the language.



[1] Arfor area comprises of Ceredigion, Carmarthenshire, Gwynedd and Anglesey

[2] This statistical work was done by John Pritchard and Wavehill

[3] A logistical regression of ‘Understanding Society’ data for 2009-11 suggested that there was no meaningful relationship between economic activity and the Welsh language.

[4] In fact, the ‘R’ number was not particularly high in all cases – only around 0.5.

Share This