Is Cardiff the least Welsh town in Wales?

A recent poll for The Sunday Times reported that 31 per cent of Welsh voters want independence — compared to 50 per cent of Scottish voters and 51 per cent of Northern Irish voters. This could be seen to demonstrate that Wales is more integrated with England than the other two nations in the UK — an interesting premise to test.

It also raises the fascinating hypothetical of where the Welsh border might be drawn — if the two nations were ever separated on the basis of genuine differences in identity and heritage.

Looking at names is a powerful way to do this. Our Origins tool is often used to chart more recent waves of migration — helping clients to understand shifts in Nigerian or Bangladeshi named populations for example. But our work on diversity reveals just as much information on historic movements of people within the UK.

Often, decisions about education, health, housing and economic strategy need to take into account differences between people who are grouped together by national surveys as ‘white’ or ‘white British’. Where did Cornish tin miners move to when the tin industry there collapsed? Why did The Black Country recruit its industrial workforce more locally than nearby Birmingham — and what are the political implications of this? When southerners were recruited to professional jobs on Teesside did they live in the same estates as indigenous north-easterners? Each of these questions relates to the existence of historic migration within the UK among people with regional as well as national identities. These migrations have cultural and behavioural implications.

After all, it is not just recent arrivals to the UK who have distinctive names. Each region of the British Isles also has a stock of recognisably local surnames.

Nowhere is this truer than in Wales. Just compare the number of Welsh Rugby icons who are named Jones with those named Smith. As the table above shows, 5.76 per cent of the 1,164 who have played for Wales in senior Rugby Union internationals since 1881 bore the surname Jones (67 players in total). That is one seventeenth — not far off an average of one Jones per team. This is compared with three Welsh rugby players called Smith, the most common UK surname — a figure which represents just 0.26 per cent of all players.

Overall, an astonishing 466 of the 1,164 players to have ever played for Wales — i.e. just over 40 per cent — have carried one of the 21 surnames listed in the chart above. And more than 20 per cent have had one of just four names: Jones, Davies, Williams and Evans.

This illustrates a distinctive feature of Welsh culture — namely the relatively small number of different surnames you are likely to find in a typical class of Welsh children. One consequence of this is the adoption of double-barrelled names, such as Lloyd-Jones, Griffith-Jones. Another is the historic Welsh practice of referring to people by their occupation as well as by their surname, as in ‘Jones the Butcher’ (or, as Michael Foot, the former MP for Ebbw Vale was sometimes dubbed, ‘Foot the Bill’!).

Why is this?

Most European surnames derive from where ancestors were born, their occupation or their ancestry. The commonest practice in Wales, by contrast, was to take the personal name or the father and add a possessive ‘-s’. Check the 21 names in the chart above, for example, and you will see that all but five of the most common surnames among Welsh Rugby players are of this sort. The same is not true of Anglo-Saxon names.

The fact that Welsh ancestry is so accurately inferred from names can be seen in the map below. This is taken from a piece of research we did for the Welsh Assembly in 2006, but the general patterns are unlikely to have changed very much, based on more recent figures we have looked at.

The map shows the geographical distribution of Welsh surnames in England and Wales, by postcode sector. We can see that there are areas of north Wales (coloured darkest red on the map) where over 60 per cent of the population still bear a distinctively Welsh surname. This is eight times the figure in the blue-coloured areas that cover most of England.

It is also evident, however, that levels of Welshness (if we can use the distribution of residents’ surnames as a proxy for it) vary considerably within Wales. The proportion with Welsh surnames ranges from 68 per cent in Bala in rural north Wales to under 20 per cent in Cardiff. Residents in the English towns of Shrewsbury and Hereford have higher levels of Welsh ancestry than residents in the Welsh capital city of Cardiff — or, for that matter, than in nearby Newport. Were the Welsh border to be drawn purely on the basis of people’s surnames, this suggests, the nation’s capital might remain in England.

The above map also reflects the way in which the south west of Pembrokeshire (an area sometimes dubbed ‘Little England Beyond Wales’) long ago started to see significant English migration inflows. Linguists and place name specialists will be familiar with the Landsker Line, which divides those parts of Pembrokeshire which have Welsh place names from those carrying English places names.

The map on the left below shows where such scholars place the Landsker Line. It clearly matches the divide between English and Welsh surnames, shown in our map on the right (which is zoomed in, with the colour contrast increased — darker shades of red again indicating more Welsh names).

These distribution patterns reflect — and are reflected in — the performance of Plaid Cymru at Westminster elections — with the party’s electoral strongholds tending to lie in the north-western seats of of Arfon and Dwyfor Meirionnyd. Likewise the darker shades of red on our surname map correspond to the places where Welsh is most commonly spoken (mapped here). Bala, which has the highest proportion of Welsh surnames, also has the highest proportion of Welsh speakers.

One interesting thing to observe from our first map, meanwhile — returning to the initial question about integration — is that the incidence of Welsh surnames declines fairly steadily once you leave Wales — from around 25 per cent in the neighbouring counties of Herefordshire and Shropshire to around 2.5 per cent in areas along the North Sea coast.

This pattern is gradual, and, in this respect, is different from that of the Scottish border — as the map below, taken from more recent datasets, shows. The proportion of residents bearing Scottish names falls sharply as soon as you go from Scotland to England — especially if you do so to the west, around Carlisle.

The line of the Anglo-Scottish border is a much better predictor of names than the Anglo-Welsh border — where the divide is more blurry. It takes around 45 to 60 minutes to drive from the Welsh border to places where less than 14 per cent of names are Welsh. In the case of the Scottish border this figure is more like half an hour.

This reflects the fact that, historically, most English and Welsh migration has been local — people marrying someone in the next village or moving to a nearby market town. It has tended to occur irrespective of ‘national’ boundaries.

Scottish migration into England, by comparison, has typically been long distance. Those with Scottish ancestry have more often moved to London or to other major cities for professional reasons. This perhaps explains why Scottish independence has tended to be more politically contentious.

Indeed, it is easy to forget how permeable the boundary between England and Wales has been to officialdom. As recently as 1958, when the Post Office first introduced regional issues of stamps, there was uncertainty over whether Monmouthshire was in England or Wales.

This was evidenced by an exchange between the Deputy Director General of the Post Office and the Home Office, in which an official from the latter wrote that the Home Secretary “fully appreciates the difficulty of having two sets of stamps, but thinks it would be a prudent step in Monmouthshire to make both sets of stamps available.” (You can read about this episode in more detail on the Postal Museum’s online resource, by clicking here).

Wales’ more gradual migration to England raises a separate question: where did those Welsh residents who moved further afield migrate to?

Having access to files of residents’ names in other Anglophone countries, it is possible to identify the extent of Welsh ancestry in a wider context — looking at parts of the New World as well as the rest of the UK.

Clearly, by this point, it is unrealistic to be able to expect to identify people of exclusively Welsh ancestry. People may be one half, one quarter, one eighth or one sixteenth Welsh — often even less. But, if we can establish the proportion of a population with a Welsh surname, then we should have a pretty good proxy for the share of Welsh ancestry in today’s population.

The chart below shows the proportion per 1,000 residents with Welsh names in the rest of the UK — as well as in the ‘New World’ countries. It shows that New Zealand, with 47 Welsh names per 1,000 residents, is not far behind the rest of the UK on this score. (For reference, it is worth noting that in Wales itself the figure would be 349 Welsh names per 1,000). Canada has the lowest proportion of Welsh names — which may partly be due to the much higher numbers of French names there.

In all, the prevalence of surnames such as Jones, Davies, Williams or Evans in Anglophone countries is slightly over 10 per cent of its prevalence in Wales. However, given the much larger population sizes of these Anglophone populations — especially the Unites States — we can estimate a combined adult population of 16 million with Welsh names across the United States, Australia and New Zealand.

Even within this we see some interesting variations. The map above, for instance, shows the prevalence of Welsh names by US States (with a score of 100 being the average, and orange or green states being those which over and under-index, respectively). It reveals a much greater concentration of Welsh names in the southern states, especially those towards the East Coast. Tennessee, Georgia, Alabama and North and South Carolina all have around twice the proportion of Welsh names as the American average.

What relevance does this have to present-day decision-making? I think there are two lessons that can be drawn. The first is that the distribution of Welsh surnames shows clear differences that are consistent with received wisdom and common-sense interpretations. The other is that surnames can map underlying variations in cultural history, which run much deeper than modern identities acknowledge.

As the future of The Union and national independence continue to be discussed, Origins helps us to understand how historical patterns of UK migration affect contemporary debates. The comparison of the Welsh and Scottish borders, in particular, informs some of the deeper cultural questions which drive debates about sovereignty. Such patterns continue to shape the level of commitment of local communities to the identity of the larger region or nation of which, administratively, they are a part.