The 105 page paper, 'Valuing Broadband Benefits: a selective report on issues and options" https://broadband.unimelb.edu.au/resources/ValuingBroadbandBenefits.pdf was prepared for IBES by Richard Hayes of the Melbourne Business School at the University. It looks at some of the methods that could be used to value the benefits of high speed broadband, with a particular focus on the NBN and in the context of any social cost benefit analysis of the NBN.
The author says that cost benefit analyses typically take two forms: a social cost benefit analysis (sometimes just termed a cost benefit analysis) and a private cost benefit analysis (also called a financial evaluation).
"The fundamental concept underlying the measurement of social benefits is the aggregate willingness to pay of the population for the effects of the policy or project," he says.
"In other words, social benefits are valued by summing individual's willingness to pay where willingness to pay refers to the amount a person would have to pay under the project or policy to be indifferent between the project or policy and the status quo."
However: "The basic issue is that we do not have much revealed information about private willingness to pay for broadband under the future NBN setup. This is because the product essentially does not currently exist in Australia, and so market price information is limited."
He says that some broadband benefits, (eg reductions in health care costs via the use of electronic health records) are likely to be estimable by careful analysis of cost savings. "However working out how much value end users place on, say, high speed broadband over basic broadband, requires different methods."
In the report he examines a number of possible ways of doing this:
- Discrete choice modelling, which uses "choice experiments" where people are asked to choose between hypothetical bundles of broadband options with different characteristics (such as speed and reliability etc), and then whether they would choose these hypothetical options over their existing (actual) broadband service. The willingness to pay for different characteristics of broadband is then econometrically estimated from all observed choices.
- Hedonic pricing, in which an estimate of value for a non-marketed item can be inferred from prices. The report looks at two possible hedonic pricing avenues for assessing the benefits: the effect of broadband on property prices and the effect of broadband attributes on broadband pricing.
Estimating willingness to pay presents particular challenges; "'¦methodologies for estimating WTP are not straightforward in application. They require specialist survey and statistical expertise. They cannot perfectly allow for the fact that users may not appreciate the benefits of higher speed broadband until they have experienced it. This is especially the case for applications that are not available at the consumer level yet."
An approach using hedonic pricing and property values also appears to hold little promise. "Using property pricing hedonic regression to derive value estimates for broadband is an intriguing conceptual possibility. However it does not seem likely that the data requirements can be met at this stage," Hayes concludes.
Overall he concludes "The hedonic pricing approaches discussed [in the report] have theoretical appeal'¦Unfortunately in this case the hedonic regression approaches are unlikely to be adequate for a prospective cost benefit analysis. The data requirements appear likely to be too onerous for these approaches, and even if Australian data was available it is unlikely to have the extent and kind of variation needed for useful results. This is particularly the case for moving to very fast broadband, which is essentially beyond the current Australian household experience."
He says that a social cost benefit analysis of the NBN would raise questions that may be impossible to answer analytically.
How much value will come from the development of future services and applications for very fast broadband, value like the then unanticipated value that accompanied the historical moves to basic broadband and fast broadband?
How likely is the move to very fast broadband to have the same scale of unforeseen applications and benefits that accompanied the shift to basic broadband and fast broadband?
What further unforeseen value or services might be released by having an entire population on fast or very fast broadband?
Hayes does, however, hold out some hope that some sort of meaningful cost benefit analysis could be undertaken. "Ultimately an analyst undertaking a social CBA might'¦examine the question - given estimated costs and given estimated values for calculable benefits, how much value is needed to come from various difficult to estimate benefits for the project to have net benefits? The answer to this question would still be subject to uncertainty, but even with uncertainty it would be informative to decision makers."