Safety interventions should be practicable and cost-effective, but too much of an imbalance towards safety does not make economic sense for employers, argues Geoff Vaughan, who suggests ‘gross disproportion’ provides a practical limit.
The legal precedent for the phrase ‘reasonable practicability’ was set in the 1949 case of NCB v Edwards, which concerned the death of a miner. In this case, Lord Justice Asquith outlined his interpretation of the distinction between ‘practicability’ and ‘reasonable practicability’.
Referencing the 1937 case of Coltness Iron Co Ltd v Sharp, Asquith said that ‘reasonably practicable’ seemed to ‘imply that a computation must be made… in which the quantum of risk is placed in one scale and the sacrifice involved in the measure necessary for averting the risk (whether in money, time or trouble) is placed in the other, and that, if it be shown that there is a gross disproportion between them – the risk being insignificant in relation to the sacrifice – the defendants discharge the onus on them.’
However, it wasn’t until the Health and Safety at Work Act (HSWA) came into force in 1975 that employers started to seriously consider what was ‘reasonably practicable’ when they implemented safety measures. The employer’s basic statutory duties, as outlined in this ground-breaking legislation, are all qualified by the phrase, ‘So far as is reasonably practicable’ (or SFAIRP).
Although the phrase ‘As low as reasonably practicable’ (or ALARP) has become more commonly used, Great Britain’s Health and Safety Executive (HSE) has emphasised that the requirements for meeting ALARP should be the same as for SFAIRP.
Disproportion or gross disproportion?
Whenever a safety decision is made, it is not unreasonable to ‘err on the side of safety’. Justice Mance stated this explicitly in the 2011 Supreme Court case of Baker v Quantum Clothing in which he quoted Lord Justice Smith (in a previous case) noting, ‘there must be at least a “substantial disproportion” before the desirability of taking precautions can be outweighed by other considerations’.
The difficulty in enforcing the HSWA is how to apply the legal precedent of ‘gross disproportion’. The question, ‘How much disproportion’ is reasonable needs to be asked. Clearly, there must be an upper limit to the degree of disproportion. If there isn’t, employers would do everything at their means to reduce the risk, regardless of the cost.
Most decisions on what is ‘reasonably practicable’ are based on ‘relevant good practice’ which has been informed by industry experience and discussions with interested parties; it is often set out in HSE-approved codes of practice.
In novel or complex cases statistical data and advanced techniques are used to estimate a probability or frequency of an adverse event and the value of the detriments, including the value of preventing a statistical fatality (VPSF) or injuries and illnesses. These are then used in a cost-benefit analysis that monetises the actions required.
By carrying out such a computation it is possible to compare the costs of implementing a safety measure and the benefit (risk reduction) obtained. The degree of disproportion can be derived as well as a judgement on whether it is ‘gross’.
Why use gross disproportion?
The main issue with applying the SFAIRP requirement is that a numerical definition of ‘gross’ has never been decided by the courts. Nor has British law determined what factors should be included in ‘time, trouble and cost’.
At an Employment Tribunal involving London Underground in 2000, the chair thought ‘that a factor of 2.5 could not be described as “gross disproportion” based on a definition in Chambers Dictionary of “extreme, flagrant, glaring and palpable”’. However, tribunal findings cannot set legal precedents.
Nor can statements at public inquiries. The HSE’s then director-general said at the inquiry into the Sizewell B Nuclear power plant [in 2014] that a factor of up to 3 (i.e., costs three times greater than the benefits of reducing risks) would apply for risks to workers; for low risks to members of the public a factor of 2 and for high risks a factor of 10 should apply. However, its validity in a court of law has never been tested.
The NCB v Edwards judgment concerned a single worker’s death. However, accidents are not restricted to single fatalities, nor are they limited to the death of workers. Accidents also lead to injuries and can involve the public. Applying ‘gross disproportion’ to these situations does not follow directly from the NCB v Edwards judgment, nor from any other legal case.
However, the HSWA wording implies that any form of risk to people’s health and safety should be subject to the test of ‘reasonable practicability’. With this in mind, it seems correct and proper that any actions taken to prevent any adverse health and safety impacts should be subject to the test of ‘gross disproportion’. However, physical damage to property should be excluded.
The limit to the degree of disproportion is the level at which it is deemed gross. In the context of NCB v Edwards, Lord Justice Asquith’s use of the word ‘gross’ was simply to set an upper level to the degree of disproportion that is reasonable.
Unfortunately, contemporary commentators have tended to view ‘gross’ in a more negative light, claiming it distorts safety regulation and places unnecessary burdens on industry. The phrase is often misinterpreted to mean all measures up to those that would be ‘grossly disproportionate’ must be implemented.
Very little action, if any, has been taken to produce a rational argument for the value of ‘gross disproportion’. Going back to the NCB v Edwards judgment, one might question what Asquith had in mind. What did he mean by ‘quantum of risk’? How did he propose that it should be ‘shown that there is a “gross disproportion”?’ In relation to ‘computation’, was he proposing a cost-benefit analysis? Did he consider what the value of a worker’s life was? Significantly, what did he mean by ‘gross’? Could it be legalese? Could it be a function of the times when this language may not have had the same pejorative meaning?
In response to a 2006 House of Lords Select Committee Report, which queried the concept of ‘gross’ in relation to the policy of ‘reasonable practicability’, then then Labour government defended the policy’s continuation, stating that, ‘the use of apparently arbitrary “proportionality factors” may not be the best way to account for a range of factors … in particular in reference to nuclear energy’.
Professor Andrew Evans’ 2013 review of railway safety concluded the use of ‘gross’ in the NCB v Edwards judgment should be considered in the light of the compensation payments current in 1949, which were much lower than present day valuations of a life. He noted that, ‘the 2009 British official value of preventing a statistical fatality (VPSF) was £1.59m, or about 60 times greater (in real terms) than the 1949 compensation’ of £984 paid to Edwards’ widow.
In a 2006 Rail Safety and Standards Board (RSSB) report, it was suggested, ‘the net output value of preventing a fatality … would, in 1949, have resulted in a figure of … about £3,600 (in 1949) prices’ (equivalent to £120,896 in 2017 prices). These figures suggest the current valuation, based on willingness to pay (WTP) surveys is, at worst, about a factor of about seventeen greater. However, it is well-known that net output valuations are much lower than WTP ones. Applied to the net output figure above would increase the figure to £169,380 and reduce the factor of 17 to about 12.
Evans’ review agreed the HSWA intended the test to be a comparison of cost and benefit, but the use of ‘gross’ was no longer appropriate due to the higher VPSF values used in 2013. He said there had been no change in the legal phraseology since 1949. However, the RSSB noted in a 2008 report that if the government had intended a different test, it would have worded the HSWA differently.
If a revealed preference approach based on wage differentials for undertaking risky work were used to determine the VPSF, it would be higher. The NCB v Edwards case was a civil action for compensation for the widow of a single worker, so it did not consider other outcomes such as injuries to other workers.
‘Reasonable practicality’ considerations, however, have to take into account the number of workers at risk. Consequently, the argument against ‘gross disproportion’ that Evans employs is really about ‘damages’ to an individual, not failure to implement safety measures, which is how the HSWA applies the term.
A National Economic Research Associates (NERA) report published in 2017 querying the use of ‘gross disproportion’ noted that historically the HSE and the ONR have maintained that a safety measure should be implemented ‘unless its cost is grossly disproportionate to the benefit’. However, both the HSE and ONR state that, ‘Ultimately, the courts determine what is reasonably practicable in particular cases.’
The NERA report also found that the use of ‘gross disproportion’ seems to have fallen out of favour with judges who have tended to use the slightly less contentious phrase ‘disproportion’ or the simpler ‘proportionate’.
NERA has accepted that ‘erring on the side of safety’ is a reasonable policy, but with a low factor between cost and benefit and that ‘as risk increases towards an “intolerability limit”, it is not unreasonable to require a raising of the bar as that limit is approached.’
The ONR’s 2018 response ‘recognised that NERA is sensitive to the term “gross”, as it potentially indicates an undue economic imbalance, and to the length of time since the 1949 Edwards judgement’.
However, it added that ‘ONR (and HSE) use a modern interpretation based on evidence given at the Sizewell B public inquiry, together with our guidance on “tolerability of risk”. ONR also noted that NERA ‘recognise that nuclear safety sometimes demands a “substantial margin of safety” which is ‘entirely consistent with the modern interpretation of gross disproportion’. ONR also noted that it ‘is acutely aware that future health and safety court cases may lead to a change in interpretation’ which would lead to a change of policy.
If it is accepted that it is reasonable to ‘err on the side of safety’ and strict proportionality is not to be used, then there must be ‘disproportion’. Also, the scale must be determined, so it is not unreasonable to state that the disproportion should not be ‘gross’.
Whether a value will ever be promulgated through the courts remains to be seen. However, something needs to be done to avoid the possibility, raised by Prime Minister Tony Blair in 2005, that, ‘We are in danger of having a wholly disproportionate attitude to risks we should expect to run as a normal part of life. This is putting pressure on policymaking … to act to eliminate risk in a way that is way out of all proportion to the potential damage.’
‘Reasonable practicability’ suggests safety measures should be implemented that are practicable and cost-effective, accepting that there should be an imbalance towards safety.
The view is that ‘erring on the side of safety’ is sensible, but too much erring would not make economic sense, so a limit is needed. This limit is expressed in the term ‘gross disproportion’.
None of the arguments put forward above detract from that concept, though the usage may be inaccurate. Until a clear legal definition is agreed so we know what the courts’ view is, even in cases of a single fatality, regulators will apply their own judgement on the definition of ‘gross’.
Geoff Vaughan is a former HSE regulator who is currently studying for a PhD in public preferences for health and safety risk regulation.