# 8. Step 8: Estimate safety benefits

## Steps

8.1 Estimate crash rates for each severity level or crash type for each year in the Base and Project Cases.

8.2 Multiply crash numbers by unit costs.

8.3 The benefit is Base Case crash costs minus Project Case crash costs.

## 8.1 Estimate crash rates for each severity level or crash type for each year in the Base and Project Cases

A ‘crash rate’ is a number of crashes per period of time. Crashes at a location usually vary with ‘exposure’ (the number of opportunities at which crashes can occur). Exposure at a location is usually measured by the amount of traffic passing through the site per period of time. Crash numbers are usually assumed to be proportional to exposure or traffic. Hence, the relevant crash rate for forecasting future crash numbers in the Base and Project Cases is a number of crashes per year to unit of traffic (vehicles, trains, cyclists, pedestrians) or traffic-kilometre. Typically for roads, crash rates are expressed per 100 million vehicle kilometres travelled (VKT).

Depending on data availability, crashes may be considered at different categorisations and levels of aggregation:

• Severity level: fatal, serious injury, other injury, property damage only, or ‘casualty’ which groups together fatal and all injury crashes
• Location: rural, urban, urban freeway.

There may be a choice of using default crash rates for the type of infrastructure and the actual crash rate for the base case infrastructure estimated over a period of time. As crashes tend to be infrequent for a given piece of infrastructure, data over a number of years is needed to obtain a statistically reliable estimate for the crash rate. Statistical significance can be improved by estimating a more aggregated crash rate, such as, the rate for casualty crashes instead of fatal and different injury severity levels.

Numbers of crashes in each year for the Base Case can be forecast by multiplying crash rates by forecast traffic.

When forecasting the crash rate for the Project Case, be wary of using default values for the Project Case infrastructure when the Base Case forecasts are based on actual crash rates from past data. The Base Case infrastructure may have particular characteristics that differ from that assumed for the default values, or the estimated crash rate might be unduly affected by random fluctuations. It may be better to use default values for both Base and Project Cases or to calculate a ‘crash reduction factor’ (proportional reduction in crashes) from the default values and apply it to the forecast Base Case crash numbers. Crash reduction factors are available for a range of blackspot treatment types. Note that these factors apply to casualty crashes, not total crashes, which would include property damage only crashes.

## 8.2 Multiply crash numbers by unit costs

Unit crash costs need to be obtained for the particular crash severity levels or types and locations distinguished in the forecasts of crash numbers. The ATAP Guidelines publishes default values for both the willingness-to-pay and hybrid human capital approaches.

## 8.3 The benefit is Base Case crash costs minus Project Case crash costs.

No further explanation is required.