One of the principal claims made to support various government projects is that they will offer a positive return on investment (RoI). From one perspective, levels of RoI are scrutinised as part of the cost/benefit analysis and consultation process on major projects and form part of the ‘business case’ for such projects (presumably the term ‘business case’ is used to distinguish such discussion from non-financial issues involved in any project). The same is true when subsidies are given.
For a full discussion Mark Pennington’s Robust Political Economy is excellent reading and is, naturally, quite correct. If governments are going to spend vast sums of taxpayer’s money on public projects it is only right that they are not loss-making; especially when the project is not a pure public good (not that this excuses the Ministry of Defence for wasting money!). Government departments’ projections and performance are scrutinised by the National Audit Office and Parliament itself.
Unsurprisingly, predictions of RoI are frequently over-ambitious and even without hindsight can be called into question. While large transport infrastructure projects are à la mode it’s worth recalling the Channel Tunnel but I’m sure everyone has their favourite; the South Sea scheme, groundnuts, Concorde, the Millennium Dome, the Edinburgh tram etc etc etc. It is to be remembered that plenty of other countries’ governments also have a strong record in this sort of thing so it is not simply British incompetence. This should all be borne in mind when governments and lobbyists put forward favourable RoI as a justification for projects as diverse as HS2, the Olympics or UK film funding.
Nonetheless, although predictions of RoI are often optimistic, or downright fraudulent, there is a bigger epistemological question at stake here. For instance, in the case of film funding: “For everyone [sic.] £1 … invested, British films are reckoned to generate £5 at the box office”. To which the response must be “How much would the £1 have earned if it were left in private hands?” Clearly this question cannot be answered, even if we approximated it using the average RoI for every £1 invested throughout the economy (also, one wonders, if films are such a profitable investment, why aren’t private investors falling over themselves to get into the industry?). Proponents of projects fail to recognise that they are making a counterfactual claim by using RoI and are assuming knowledge that they cannot possibly possess.
This ‘knowledge problem’ was most famously expressed by F.A. von Hayek in his article ‘The use of knowledge in society’ (although I think Hayek was discussing the dispersed nature of knowledge rather than the impossibility of it as in these instances). Of course, there may be other reasons for pursuing a government spending project, hence we now have the concept of ‘Social Return on Investment’, but these will likewise be subject to incentive problems, as Public Choice theory demonstrates*.
Of course, private investors have to make a similar assumption of knowledge with similar constraints on that knowledge. The key differences are, however, that private projects are undertaken at the risk to the individual(s) concerned and without the constraint of Public Choice effects. A private scheme that fails to generate sufficient RoI will be liquidated or sold on to someone who can generate a profit or – if undertaken for reasons other than profit – will be run at a loss. Government schemes that fail to make a profit simply incur greater charges on the public purse.
To most readers of the ASI blog, I expect, these points will be so commonplace as to be hardly worth mentioning. However, as we hear such claims made every day in favour of x project and y subsidy and go unquestioned it seems worth re-iterating. It is beholden on us to remember that rationalisation of government spending based on claims of positive RoI is not simply occasionally fallacious but always and entirely specious.
*For a full discussion, Mark Pennington’s Robust Political Economy. It is excellent reading.