Where is the data on listed buildings?

Ever since the 1940s, the missing data in Historic England’s list of buildings has represented a considerable missed opportunity in terms of historical research and policy analysis.

Bath's Cavendish Crescent and Royal Crescent.jpg

Terraced housing can be covered by single entries, such as Bath’s Cavendish Crescent and Royal Crescent, but not always.

The total number of listed buildings in England is a mystery. Estimates vary from around 400,000 to upwards of 800,000. There are many reasons behind this gap in our knowledge, but, whatever the explanation, this is a major hole. Richard Griffith pointed to the problem in an article in 2010. ‘If a health service was unable to identify the number of patients it was handling, and could only suggest that it appeared to be “somewhere between 500,000 and 895,000”, then a reasonable person might conclude that its management was negligent,’ he wrote. ‘It is appropriate to ask why similar criticisms should not be levelled in the case of listed building control’. [1]

The total number of listed buildings is just one data point. It may not be all that interesting on its own, but – let’s use an architectural metaphor here – it is a keystone of data. This is because statistical analysis of the list cannot be done with any accuracy until we know the total number of buildings that are at stake. Let’s say we want to know how well represented 17thcentury buildings are in the list at large. Even if we can extract a number of 17th-century buildings from the list, we have no idea what total to divide the number by. If we want to know in what decade after 1947 the biggest chunk of the list was recorded, we will also draw a blank, lacking an adequate total.

We can trace this problem back to the origins of the list when the Town and Country Planning Act was drawn up in 1946–1947. The act happened in the wake of the second world war and fundamentally changed the relationship between building work and government. It was a landmark of legislation, and the resources of the country were beyond stretched when it was first carried out. In this context, it is understandable that the nation did not want to spend its money and time philosophising about what exactly a ‘building’ entails.

But this is something that should have been done. To have legally protected buildings put on a list, one needs to be clear on what a ‘building’ really is. This is another point made by Griffith. In some cases, the list considers an individual unit within a terrace to be a distinctive building, but in many cases it does not. In other cases, buildings with completely different functions and external walls (such as a detached water mill) are listed under one entry. The lack of consistency here is the seed of the problem we still find ourselves saddled with. Clusters of individual buildings, sometimes 20 individual units, sometimes more, are housed under individual list entries. When multiplied over a data set in the region of half a million points long, this generates the huge margins of error discussed above.

The harsh criticism quoted above was written over a decade ago, just before the National Heritage List for England was made available online. The computerisation of the list and its public availability has changed matters for the better because it made the list searchable. The advanced search function of the list enables one to select for individual criteria. And although the search engine limits results to 1,000 entries, it provides a total figure for the number of hits. This is important because it allows us to quantify things, such as buildings listed within a certain time period, or buildings of a certain date.

Realising this, I decided to take Historic England’s advanced search engine to task. In my experience as a heritage consultant in East Anglia, and then Cornwall, I had a hunch that a particular period in architectural history is very well represented: the first few decades of the 19th century. My hypothesis was that there was a statistical sweet spot achieved by buildings of this date, which satisfy the criteria for rarity and interest more than structures a couple decades later, yet are not so old that their numbers have dwindled. Any Jacobean structure that survives in recognisable form will certainly be listed: but it had to survive for over 300 years first. Moreover, far fewer Jacobean structures were built compared to Georgian ones, in the dawn of speculative development.

Setting about testing my hypothesis using manual search queries, I immediately ran into a new problem of precision. For fairness, I would need to search the list using time periods of equal length. I could not settle for the reigns of monarchs, because the Georgian period was far longer than, say, the Jacobean period. Either I had to search through every decade from 0 CE to 2022 CE, or I had to chunk things in terms of centuries or half centuries.

This search could return results with very rough numbers, after a few hours of spadework. I got some rough data for 100-year categories, which indeed showed the 19th century to be the best represented. In the end, though, I checked my hunch using a totally different data set compiled only for London, in 1988. This was the first and last time that a list was compiled in a tabular fashion, with precise dates associated with individual buildings. It turns out that there is a remarkable bunching of listed buildings around the first and second quarters of the 19th century. Who knows to what degree this distribution would survive if the survey included the whole of England, though?

A worse problem begins to show itself if we try to use the list to interrogate less-well-represented structures. With rarer types of structure, glitches begin to appear. For example, let’s say I want to check if anything built within the last 10 years has been listed. I search between the date range of 2013–2023. The search counter returns 655 results. Who knew the list contained so many neophyte structures! It turns out, however, that the correct number of listed buildings built between 2013 and 2023 is zero. The youngest listed building in all of England dates from 1997.

Other anomalies abound when we try to take a look at 20th-century heritage. One huge issue is that any building which mentions ‘C20’ due to retrofit, restoration and extension, runs the risk of returning a false positive, even if it is a 13th-century building. The search mentioned above returned nothing but false positives.

Our fundamental problem with data remains. Historic England’s search function is a wonderful tool, but it was not designed for statistical analysis. Its function is really to identify buildings, and because of flaws in the underlying data set, it cannot be used to generate reliable numbers. It cannot be used to carve up the data set in an academically rigorous way. I have done some data collection for my own interest, but I would struggle to get any of my bucket-chemistry findings through the scrutiny of peer review. And so it goes for anyone who wants to take a statistical look at our built heritage using the list.

This is a huge missed opportunity. First, the list of legally protected buildings represents a potential mine of data of great interest for architectural historians looking to inject some numeracy into their analysis. It represents a huge number of structures that are worthy of study, compiled together in a centralised format. Second, keeping track of the number and distribution of listed buildings should be a priority of governance. The concept of listed buildings is ultimately a government policy, and policies should be monitored.

Historic England cannot really be blamed for this situation, and I am not sure how change could realistically be brought about. The statistical sifting, correction and honing of the existing list would be an enormous undertaking, requiring vast amounts of time and money. It might be just as onerous as the initial digitisation of the list in the 1990s – indeed, it could be even more difficult.

On the other hand, as conservation professionals we are used to preserving artefacts and places for their own sake. Our practice could be described as idealist, and rightly so. We believe old buildings are worth our sweat, for the study and delight of future generations. If the buildings themselves are valuable in this sense, then so is the data about them. Data about the list should be among our highest priorities if the conservation profession is to keep pace: not only with technology, but with our own principles.