A private business calculating the return from an investment project through a presented discounted cash flow analysis takes into account gross revenues minus operating costs (NE). The social benefits of transport projects go wider than this assessment of the returns captured in the farebox. Users gain if the investment lowers the cost of transport to them. The consumer surplus gain (CS) is the incremental difference between what they would have been willing to pay and what they now have to pay for each unit of transport - in the jargon of economic textbooks it is the additional area under the demand curve. The numerator of a social rate of return calculation of transport benefits will include both NE and CS.
The social savings calculation made by Gary Hawke is an (upper bound) estimate of CS made by estimating the difference between the cost of rail and the cost of the best non-rail alternative transport and multiplying this amount by the volume of rail transport in each year. The most difficult data to obtain are the costs of non-rail transport. Hawke estimated that in 1865 for freight rail costs were 1.21d per ton-mile and non-rail costs would have been 2.3d while for passenger traffic the estimates were 1.35d per rail passenger-mile and 6.5d for non-rail. In addition, rail freight saved some supplementary wagon haulage and cut inventories. Virtually all of the total freight cost savings of £25 million accrued from minerals traffic.
Willingness to pay for transport is a derived demand stemming the ultimate demand for inputs by firms and final products by households. Assuming that the estimated transport demand curve is well-specified to allow for rearrangement of productive activities as transport costs change, transport benefits are acceptable as an estimate of overall economic benefits in some circumstances but not when there are significant externalities or where there is imperfect competition in the transport using sector. Hawke's calculation is innocent of these complications and must be seen as no more than a rough and ready approximation.