Bob, I accept that the two are intertwined and when assessing risks most of us will inevitably be looking ahead to how we might manage that risk. Arguably though you can’t effectively manage a risk without first assessing and genuinely understanding the extent and nature the hazard (although I am convinced that this is what some people do). I would have thought that for the most part most (maybe all) risks can be assessed without any consideration of how that risk might subsequently be managed.

Although in the main the two are considered as one (see above) I suggest that assessing risks is one thing, managing them (effectively) is another. The important consideration, for me at least, is the longer term effectiveness of controls and the ‘cost’ of maintain those controls. Although tackling hazards at source, either through elimination or reduction, MAY have a high initial cost (actually many are also ‘cheap’, ‘obvious’ and easy) there may well be longer term savings either in the cost of maintaining the controls or the cost of picking up the pieces when they have failed.

Any number of times I have seen people list controls that don’t (control) or are unlikely to be effectively in practice, and that they have missed the trick of reducing the hazard.

The essence of the original post questions whether the severity of a hazard can or should be reduced. I believe that this should be a primary consideration when considering your choice of management of the risk. Many here adopt or at least quote ERIC. Eliminate and Reduce are directly associated with changing the severity of a hazard.

“Is there any other way of bringing down the Severity ratings other than with the use of PPEs or ERPs?”. Yes, as already observed, reducing voltage, reducing the distance of a fall, using less hazardous substances etc.

“Isn’t it logic to understand that if Likelihood is reduced, the Severity automatically reduces”? I suggest that this argument is fatally flawed. While the overall risk will have reduced, I suggest that the severity almost certainly hasn’t changed (although I am open minded about other examples where this might happen) e.g. unguarded bench saw. While fitting a guard and giving someone some instruction and training will reduce the likelihood of them cutting their finger(s) off, trust me it doesn’t eliminate it; and the hazard severity remains the same. Or to use David MacCollum’s example don’t teach people to walk around the hole, either cover it or fill it; should they forget to walk around it and fall in, the severity is just that same!

Is it me are are a lot of people here just playing with words?

For me I don't care what you call it. Risk, hazard, severity, likelihood etc, it's just a word. What I care about is reducing injuries and ill health.

I do this very simply. For my risk assessments I look and see what can hurt people and write that down, and then I work down the hierarchy of controls to do whatever is reasonably practicable to reduce the chance of that person getting hurt.

Severity and likelihood are only important if you use a LxS grid for your risk assessments and for me they are not worth the paper they are written on. For likelihood there are only three possibilities. "Certain" "Impossible" and "somewhere in-between". the "in-between" bit is very subjective and to put a number on it makes it objective.

My first manager before I was professional H&S showed me how to do a risk assessment with a LxS 5x5 grid. He said anything over a score of 15 meant the risk was severe so work had to stop. I asked him if that had ever happened and he said "No. Well we did do one that was 15, but the MD wouldn't stop the process so we had another look and decided we could reduce the likelihood from a 3 to a 2 which meant it was safe enough to continue". Note, they didn't put in any control measures to reduce the score, but just decided that the outcome was less likely than they originally thought so work could continue. Pointless!

Remember a risk assessment is just a piece of paper. It does not save lives or reduce injuries or ill health. Many of my control measures have never made it to a piece of paper as I'd rather spend time implementing then recording (although of course I have some written notes so as to satisfy MoHaSaW 1992, and for significant risks a detailed narrative RA is produced).

It continually bugs me when after a successful prosecution you get an HSE Inspector quoted with something like "Had they carried out a risk assessment, this accident would not have happened", when what they mean is "Had they assessed the risks and implemented suitable control measures, this accident would not have happened". The risk assessment in itself is just that - an assessment of the risks. And that is all.

And of course don't forget ALARP and risk retention.
The 'score' (for those using) may be high(er) in comparison to others but in light of current technology and/or invention is as low as its going to be for now...

Risk retention is way down the list when you/Company decide to accept the risk is high(er) than it could be but go with it anway (or perhaps pay an increased insurance premium as an example) - of course you do so at your peril should it bite you later...

Argghhhh - you read what I wrote, not what I meant! Sorry, I wasn't very clear in my opening paragraph.

What I meant was that IMO it doesn't matter whether you assess something as needing further controls on the basis of "likelihood" or "severity", as long as the controls are implemented.

If I walk round and see a crusher unguarded, then I lock off the crusher until a guard is fitted. I just have visions of people with clipboards agonising over whether or not the likelihood is a 2 or a 3, and if the severity is a 4 or a 5. My point is that it doesn't matter at all, what matters is to make the thing safe.

To get back to the OP though, yes of course controls can reduce the severity.

Scenario: JCB telehandlers belting around the yard at 40mph whilst pedestrians use the yard for access and egress. Controls such as driver training, hi-vis for pedestrians and safe walkways reduce the likelihood of anything happening. Fitting the JCB's with electronic speed limiters to 5mph, and fitting side skirts so people can't be dragged underneath them will reduce the severity of any impact from a likely outcome of "death" to a likely outcome of "minor injury".

Canopener, I wasn't aiming my comments at anyone and I agree with a lot of what you say. There are of course different ways to go about the RA process, I just think it's important to not get tied up by the process itself, but to focus on the outcome - i.e. robust controls. FWIW neither our insurers, nor the HSE inspectors we have seen recognise a LxS matrix as being suitable and sufficient for anything other than very minor hazards, which in most cases do not require a written RA anyway.

See pages 14 and 15 of this link posted on another topic.

http://www.hse.gov.uk/enforce/emm.pdf