27 June 2017

Social Policy Bonds and price signalling

One of the less obvious benefits of a Social Policy Bond regime arises from the price signalling of the market for the bonds. At flotation, the bonds would be auctioned, and difference between the sums raised at flotation and the total redemption value of the bonds would supply the market's best estimate of the cost of achieving the targeted goal at that time. This estimate would vary over time, depending on many factors including bondholders' performance in undertaking or financing goal-achieving projects. The market for Social Policy Bonds, then, as with all markets, plays a vital role not only in allocating resources but also in signalling; in this instance to policymakers, the approximate costs of achieving social goals.

A competitive market for Social Policy Bonds would minimise the total cost of achieving a specified objective, as well as signalling it. More subtly, and more technically, it would also indicate the marginal cost of achieving further improvements. Say one million crime reduction bonds issued by a local authority were to sell for $5 each. This would tell the issuing body that the present value of the expected maximum cost, including bondholders’ profits, of reducing the crime level from, say 50 to 40 units, would be $5 million. The local authority might then suppose that it could afford to be more ambitious, and aim for a further fall to 30 units. It could issue a million additional bonds redeemable when this new lower rate were reached. These would (probably) have an initial market value of less than $5, reflecting the (probably) diminishing returns involved in preventing crime. The point is that, by letting the market do the pricing of the bonds, the local authority would be getting an informed view of the marginal cost of its objectives. So if the bonds targeting the new level of 30 units were to sell for $4 each, then the maximum cost of achieving that objective would be $11 million, being equal to: $5 million (paid out when the level fell from 50 to 40 units) plus $6 million (paid out when the level fell from 40 to 30 units). The marginal cost of a 10-unit drop in crime would thus have been revealed to have risen from $5 million to $6 million. Should the local authority aim for a further fall to 20 units? Following such crime rate-targeting bond issues it would have robust information about the cost of doing so.

This is, of course, a simplified example and in fact the bond market would continuously update its pricing information. Say that new research, of the sort that might be stimulated by an initial bond issue targeting crime, suggested new ways of reforming or deterring criminals. Bondholders may, for example, have financed successful research into more effective reform programmes, or set up more appealing alternative lifestyles for especially hardened criminals. How would the market react to such developments? Once their effectiveness had been revealed, the value of all the bonds would rise. Instead of being priced at $5 and $4, the two crime reduction issues of the example might sell for $8 and $7. The total cost to the government of redeeming these bonds would not change: it would remain at $11 million (though redemption would most probably occur earlier). But the market would be generating new information as to the likely cost of future reductions in the crime rate. The market would now be expecting reductions of 10 units of crime to cost $2 million (from 50 to 40 units), and $3 million (from 40 to 30 units). The new research would have reduced the costs from $5 million and $6 million (respectively). So the cost of any further crime reductions would also fall, and by following market price movements policymakers could gauge approximately by how much. 

These figures are hypothetical, but they do indicate the role that markets for Social Policy Bonds could play in helping the government, and taxpayers, decide on their spending priorities. The market for the bonds is elegantly efficient in conveying information about the cost of achieving objectives and, crucially for policymakers, how this cost varies with time and circumstances. I discuss this in chapter 5 of my book.

No comments: