Often we are asked if a 2-3mph change in average speeds is worthwhile when you reduce the limit from 30mph to 20mph. To understand this and how it can produce reductions in casualties of 20% or more we have to understand that these are 20mph limit implemented across a wide area. These range from slow and quiet roads where previous speeds already are less than 20mph, through roads with speeds of 20 to 24mph and even some where speeds are between 25 and 30mph.

Besides being a wide range of speeds beforehand there is also a wide range of casualty history. Many of the lower speed roads will have minimal casualties whilst faster roads may well have significant numbers of casualties..

It is this mix of speeds and hence danger which means that average road speeds across the whole network may only reduce by as little as 1mph, but the spread of speed reductions can be very beneficial for casualty reduction.

Lets take the there categories of road in terms of previous speed :-

• Low speed roads with a previous average of less than 20mph

It is unlikely that speeds will change on such roads. They are already slow due to obstacles, parking and other features. But because of the low speeds there are few casualties as well. They are included in the 20mph roll-out for consistency. It would be inconsistent to set faster roads with a 20mph limit whilst leaving these at 30mph.

• Medium speed roads with a previous average of between 20 and 24mph

DfT guidance suggests that from experience such roads will have a reduction in speed that will bring the average to 20mph. Hence such roads will "on average" show a reduction of 2mph. Such roads will probably previously have had more casualties than the roads previously under 20mph. With research showing that each 1mph reduction in an urban environment reduces casualties by 6% then this would result in a 12% reduction.

• Higher speed roads with a previous average of between 25mph and 30mph

Research from wide-area 20mph implementations on such roads will show an average reduction of between 4mph and 6mph. Enough to be a significant reduction in speed and danger but not enough to bring the average down to 20mph. But these are usually the roads with the most casualties. Hence an average 5mph reduction on the roads with the most danger is expected to reduce casualties by 30% on the 6% per 1mph principle.

There are therefore two factors at play in the interaction of previous speeds and casualties when you implement wide-area 20mph limits :

• The average speed reduction is diluted by the minimal changes in those roads which are already slow. These have been included for consistency but mask the much larger reductions in speed on faster roads.

• The effect of the larger reduction in speed on the faster roads is levered up by the fact that these are where most danger and casualties were previously reported.

These are therefore the reasons why an overall small reduction in average speeds across a road network can result in substantial reductions in casualties. In Calderdale there was a 1.9mph reduction in average speeds and a 30% reduction in casualties. In Bristol a 2.7mph reduction in average speed was accompanied by 4 fewer fatalities, 11 fewer serious injuries and 159 gewer slight injuries per annum.