Was ‘troubles’ related death toll as high as 30,000?

How many people died in the recent conflict in Ireland? You’d think this would be a relatively easy to answer question. But depending on how you decide to define a death as conflict-related, the total, which is usually given as around 3,700, is probably at least 5,733 and may be as high as 30,000.

A quick trawl of existing databases puts a detailed death toll in the region of 3,640-3,760. This is a considerable figure and itself only a fraction of the number who received injuries or were harmed in some other way by their experience over the same period. But a comparison of some of the components of those totals shows that this is an estimation and a very conservative one and doesn’t seem to fully reflect the extent of loss of human life arising from the conflict. Initially, I was looking at this to see the methods employed to determine what might be the best way to estimate the loss of life in the north (and Belfast in particular) in 1919-23. Instead, I noticed that, depending on how you choose to define whether a death is related to the conflict here, you can argue that the actual death toll is at least 5,733. Or you could even put it as high as 30,000.

There are a number of publicly accessible databases recording deaths arising from the post-1966 conflict here. This includes the following: Lost Lives ; Malcolm Sutton’s An Index of Deaths from the Conflict in Ireland 1969-1993 (with a draft list up to the present); and, Michael McKeown’s Database of Deaths Associated with Violence in Northern Ireland, 1969-2001. The latter two databases are hosted, in various formats, along with contextual literature and other resources on the Conflict Archive on the Internet (known as CAIN).

The 2006 edition of Lost Lives records 3,720 deaths from 1966 to that date. The database compiled by Malcolm Sutton covers the period from 1969 to 2001 and records 3,532 deaths (in relative terms, this is 149 less than the total for the same period in Lost Lives). It also includes a provisional total of 88 deaths from 2001 to the end of 2017. The second database available via CAIN, prepared by Michael McKeown, records 3,649 deaths over the period from 1969 to 2005. Compared to the same period covered by Lost Lives, McKeown lists 3,622 deaths, 98 less than Lost Lives. Combining the figures in Lost Lives and those on CAIN for the period up to 2017 gives a total of 3,762 conflict-related deaths from 1966 to 2017.

This variability hints at the complications that underlie what deaths are deemed to be conflict-related and how that is defined. It is also still possible to identify some deaths that have been overlooked in all them, such as Erwin Beelitz in Berlin in 1972. Lost Lives does provide discussions of individual cases that illustrates the scope of what is considered as a conflict-related death. It generally just includes violent deaths and so would not typically include, for instance, someone whose health suffered from conflict-related stresses leading to a premature death. The latter may be much more difficult to determine and be quite subjective (although more on this point below).

What got me interested in the divergence from the general quoted figures of around 3,700 was in the officially declared military death toll. I’d thought that a relatively obvious way of checking how robust the available figures were, was to compare figures for individual groups against published records. In December 2012, the UK Ministry of Defence provided a breakdown of British military personnel who died during Operation Banner (the British army deployment to the north). For deaths as a “…result of operations in Northern Ireland or Irish Terrorism in other countries…” it gives a total of 1,441. Lost Lives gives an overall figures for the British Armed Services of 503 along with 206 UDR and RIR. The breakdown for Operation Banner, provided by the Ministry of Defence, though, is 814 regular army, 548 UDR and RIR, and a further 79 for other branches (making up that total of 1,441). This is 732 in excess of the figures provided for in any of the relevant databases. This underestimate, by some 103%, is considerable. As the term used in the letter detailing the figures very specifically says that the deaths were as a “…result of operations in Northern Ireland or Irish Terrorism in other countries…” this would be appear to be the official total.

Critically, to understand the methodology, the existing databases all list every individual who is included. The official figures for Operation Banner do not provide individual details.

It is also possible to look at the figures given for republican fatalities. A ‘Roll of Honour’ was published in An Phoblacht in 2010, identifying the deaths of those named as having being conflicted-related. This gives a total of 336 conflict-related IRA deaths (it also lists 25 Sinn Féin members). Using the tables provided by Sutton, his equivalent figure appears to be 292 (excluding Sinn Féin members) suggesting there were a further 44 deaths on top of those conventionally associated with the conflict. A further 83 republican combatant casualties are recorded by Sutton (including INLA, Official IRA etc). McKeown reports some 271 IRA deaths, of a total 350 republican dead, while Lost Lives provides a combined total of 396. If the An Phoblacht figure reflects deaths deemed to be conflict-related but not conventionally captured by the methodologies employed by Lost Lives, Sutton and McKeown, the existing method may underestimate republican conflict related deaths by around 15%. So, if extrapolated for republican casualties as a whole, a figure of around 455 may more accurately reflect the scale of loss. If it might be argued that similar factors would be at play, the Lost Lives figure for unionist paramilitaries, 167, should probably be revised upwards on the same basis to 192. Neither of these figures would necessarily capture conflict-related deaths that are due to factors connected to their experience as combatants, incarceration or self-harm.

The issue of self-harm, more particularly suicide, is one that arises in a variety of contexts. The number of recorded RUC fatalities is given as 301 by Sutton, 303 by Lost Lives and 304 in McKeown. But on a number of occasions, official statements and figures have been given for suicides among serving RUC members, with 55 recorded by 1996 and 75 recorded by 2007. Without trivialising such a complex and emotive issue as suicide, it may never be known how many of these could be directly attributed to conflict-related factors. However, as noted with the difficulty of assessing the factors in any individual death, a review of overall figures might show any increased mortality that, in the context of the north during the conflict, is likely to have arisen from factors related, at least in part, to the conflict.

Occasionally, individuals who took their own lives are included amongst the conflict-related deaths. Patrick Sheehy, an IRA volunteer who appears to have shot himself in Nenagh in County Tipperary on 1^st January 1991 is listed in Lost Lives (3170), although not in either Sutton or McKeown. It seems difficult to argue that it wasn’t related to the conflict and so it’s inclusion seems reasonable. On the same basis, it would seem that RUC fatalities should also be included. Taking Samaritans figures, it is possible to make wider comparisons with suicide rates across Ireland and Britain to see if the broader impact of the conflict on mental health is reflected in elevated suicide rates. There doesn’t appear to be any evidence for this until 2007, after which time an apparent rise by about 6 deaths per 100,000 since around 2007 is considered by some to be attributable to earlier conflict-related stresses. This would amount to around 1,080 deaths and continues to this day.

Combined with other additional deaths noted above for Operation Banner, the IRA, RUC etc, this would suggest the total number of deaths is closer to 5,733.

To follow this through, a similar comparison of general mortality for the north in the 1970s through to about 1998, makes for equally grim reading. Typically the male mortality rate was at a level significantly below that of the UK and the south of Ireland, by around 200 deaths per 100,000 per year. By the early 1980s, this had closed to around 100 per year and the mortality rates were then roughly comparable throughout the 1990s. The female mortality rate was significantly higher in the north than the UK and south of Ireland average in the early 1970s, by around 210 (per 100,000 per year), staying just under 200 higher in the early 1980s, rising to around 220 higher in the 1990s and the dropping back to just under 200 deaths higher (per 100,000 per year) by the end of the 1990s. Taking the overall differential in the mortality rates over the period from the early 1970s to 1998 suggests that maybe 24,000 more people died than would be expected, based on the rates prevalent in the UK and the south. Are there a series of complex factors underlying the increased mortality rate? Undoubtedly, but it is hard to see how it is likely that any complex factors are not, in themselves, somehow connected to systemic and structural issues related to the conflict. In that sense, it could be argued that these 24,000 should also really be regarded as conflict-related and arise from what is known as ‘structural violence’ where socio-economic places limitations on a population’s quality of life causing harm and increased mortality.

So is it plausible that we should consider the death toll from the recent conflict to be 30,000 or roughly 5,733 rather than around 3,700? The currently used conservative estimates of the quantum of deaths arising from the conflict has at least two origins. A concern of many of those who complied the databases was in painstakingly researching and detailing each individual death. Where the deaths are anonymised into collective data (such as the casualties from Operation Banner), it seems that there is no mechanism for either Lost Lives or the likes of Sutton or McKeown to include them.

Another factor, though, is that acceptance of the reduced figures is also an artefact of the same security policy that sought to minimise the nature and intensity of violence and brought us terms like ‘ulsterisation’ and ‘criminalisation’ but didn’t speculate on the capacity of those same policies to inflict structural violence on a population. This isn’t to imply that Lost Lives, Sutton or McKeown are somehow complicit in furthering the same policy. They are simply following convention and documenting instances of violence in which lives were lost. This differs significantly from British (and Irish) government strategy that sought to present violence in the conflict in the language of decontextualized, criminal acts. A logical outworking of this would be to continue to adopt a minimalist approach in assessing the human cost of the conflict even where, for instance, the official death toll of Operation Banner is way in excess of the figure normally cited for military losses. Unfortunately this attempt to promote a conservative estimate minimises the actual impact and adds to the dissonance between the official narrative and the impact of structural violence as experienced by individuals, families and communities. For them, perhaps we need to recognise that the scale of related fatalities is much greater. In that regard, a fresh consideration of how we define the death toll from the recent conflict may be worth further exploration and debate.