The Local Elections are coming up on May 7th. With candidates announced, I wanted to pull some demographic data from all candidates.
The data predicts that the 2026 Local Elections will show a roll-back for gender equality in local government. While the total percentage of women standing has a small decrease when compared to prior data. However, with the much lower percentage of women standing for the poll leaders, Reform UK, at 23%, I would suggest that following the election, we can predict there will be less women councillors after this set of elections.
This data was initially created for my POL2010 Research Methods for Politics module at Edge Hill.
Gender Balance at the 2026 Local Elections
The gender imbalance in UK local elections has been well reported. In 2024, the Fawcett and Democracy Club found that only 34% of candidates were women. However, this is projected to actually decline. In 2025, Bunting’s analysis reported that of all the Reform Party’s candidates, less than a quarter were women (the Conservatives had just under 30%, and Labour and the Greens were at 40%).
The 2026 candidates have remained stable in regard to gender inequality. With women making up only 32.4% of candidates, a slight decrease of 1.6% compared to 2024. The Reform Party, in particular, has a poor percentage with only 23% of their candidates being women in this election. Likewise, independent and other third parties have brought the overall level down also.
| Gender of Candidates | N | % |
| Male | 16,066 | 64.1 |
| Women | 8,122 | 32.4 |
| They/Non-Binary | 21 | 0.1 |
| N/A or Unknown | 850 | 3.4 |
| Total | 25,059 | 100 |
Gender balance by party
| Party | Men | Women | Total | % Women |
| Conservative and Unionist Party | 3,134 | 1,417 | 4,551 | 31.1 |
| Green Party | 2,561 | 1,780 | 4,341 | 41.0 |
| Labour Party / Labour and Co-operative Party | 2,755 | 1,979 | 4,734 | 41.8 |
| Liberal Democrats | 2,572 | 1,282 | 3,854 | 33.3 |
| Reform UK | 3,635 | 1,083 | 4,718 | 23.0 |
| Other/Independent | 1,409 | 581 | 1,990 | 29.2 |
It should also be stated that this does not necessarily mean the proportion of candidates will translate to elected councillors. Historically women have been placed in unwinnable seats, or provided less campaign support.
Most Popular Name
Given that there is a significant gender imbalance in candidate numbers in the 2026 Local Elections, it’s no surprise that the most popular name is David, followed closely by John, Paul, and James. The most popular name for a woman only appears at number 20.
It’s safe to say those with first names which are masculine, and originate from a biblical origin, will be very well represented in the upcoming elections.
| No. | Row Labels | Gender | Count | % |
| 1 | David | M | 498 | 1.99 |
| 2 | John | M | 444 | 1.77 |
| 3 | Paul | M | 376 | 1.50 |
| 4 | James | M | 344 | 1.37 |
| 5 | Mark | M | 294 | 1.17 |
| 6 | Michael | M | 287 | 1.15 |
| 7 | Andrew | M | 278 | 1.11 |
| 8 | Richard | M | 278 | 1.11 |
| 9 | Peter | M | 277 | 1.11 |
| 10 | Stephen | M | 237 | 0.95 |
| 11 | Chris | M | 203 | 0.81 |
| 12 | Simon | M | 201 | 0.80 |
| 13 | Ian | M | 198 | 0.79 |
| 14 | Robert | M | 181 | 0.72 |
| 15 | Christopher | M | 156 | 0.62 |
| 16 | Martin | M | 146 | 0.58 |
| 17 | Alex* | M | 137 | 0.55 |
| 18 | Steve | M | 135 | 0.54 |
| 19 | Daniel | M | 134 | 0.53 |
| 20 | Sarah | F | 134 | 0.53 |
| 21 | Matthew | M | 133 | 0.53 |
| 22 | Tom | M | 116 | 0.46 |
| 23 | Tony | M | 114 | 0.45 |
| 24 | Mike | M | 110 | 0.44 |
*Names which could potentially be androgenous in UK usage
Some Interesting Names
- None Of The Above X, An Independent Candidate standing in Langdon Hills and Westly Heights Division Wards. Their statement suggests they intend to act as a mechanism for voters to spoil their ballots.
- Esther Erzsebet Aiko Pimprenelle Saurigny, is the longest named candidate with 41 characters in these elections. They’re standing in Hollybush (Welwyn Hatfield) for the Green Party.
- The shortest name is Tagl (4), Who is a Green Party candidate in Caxton & Papworth (South Cambridgeshire)
Overall
It seems much of the dataset is incredibly consistent from a similar analysis conducted by the Fawcett Society back in 2024, in terms of gender equality within the parties, and even in terms of the most popular names.
With the rise of Reform UK in the polls (Reform are currently at 24%, Green 18%, Conservative 19%, Labour 17%, and Lib Dem 13%), I would overall expect the total number of women elected to local government to be reduced, and an effective roll-back of gender equality in local government.
Methods & Gender Inference
The data used for this came from Democracy Club/ Who Can I Vote For via their awesome Candidates and Results API. There’s also a CSV option available for those that want to play with this. This returned 25,059 candidates.
I also wanted to infer gender characteristics of candidates. This was a little messy as only a small percentage had their gender listed in their profiles (n=5,997, or 24%). Therefore, I opted to infer gender by their first name. This isn’t perfect, it doesn’t account for non-binary identities, gender-neutral names, and there will be edge cases. For example, Alex is often a masculine name but is also used for women too. Overall, I suspect that while most are accurate, the figures here won’t be perfect. Inferring gender by age, according to the paper by Van Helen et al. (2024) has an accuracy between 82% and 96% depending on the approach used, and origin of the name (some names, such as those originating from South Korea are less accurate than German names, for example).
My first approach Firstly, I took the biggest groups of names to manually code them. The rest I planned to undertake through the use of Genderize.io’s API & APIs (using multiple keep under the rate-limit of their free tiers). After about 20 minutes of this approach, I realised that manually coding the large names would take ages, and even when using multiple API’s, I still wouldn’t have enough requests.
My second approach is probably what I should have done from the start. I found a dataset at the UC Irvine Machine Learning Repository titled Gender By Name. This added together 4 large datasets and produced a gender categorisation of 147,270 names in total. From this, I used an Excel function (XLookup) to provide inferred genders for my total dataset of candidates. This completed most of the dataset apart from 1,537 candidates (Or 6%) which were listed as unknown.
I tested this against the gender data available from the profiles (n=5,998), and it had an accuracy of 93.25% with 5,405 being correct, and 365 incorrect. I corrected any incorrect entries, rectified any candidates’ names initially listed as unknown but contained profile data.
I then took any candidates with a middle name and ran the same Gender By Name XLookup for those. This associated a gender with an additional 345 names, leaving only 850 names unknown. For candidates who had profile information which indicated ‘they’ or ‘non-binary’ status, I left these as is.
This resulted in a mostly completed dataset with inferred gender.