Apr 27, 2021

Social impact bonds are a wolf in sheep’s clothing


Earlier this year, Vancouver city council considered a motion encouraging the city to look into the issue of “social impact bonds” (and more broadly “social impact investing”), which is a warm and fuzzy sounding term that actually refers to a type of privatization. I spoke to city council to help demystify the rhetoric around social impact bonds and to recommend they reject this model.

The bad news, at least for Vancouverites interested in protecting and expanding public services, is that the motion did pass in a narrow vote. The good news is it was watered down by amendments and ultimately only ended up asking city staff to look further into the issue and report back to council.

What are social impact bonds, and why should Vancouverites—and Canadians—be concerned about them? Below is a lightly-edited version of my remarks to city council tackling these questions.

CCPA has previously published research on social impact bonds, and I have published analysis on a related and similarly inefficient model of privatization called public-private partnerships, or P3s, which are a good analogy for understanding social impact bonds and may be more familiar to some.

A similarly inefficient model of privatization called public-private partnerships.

First introduced in Canada by the Harper government federally, and then continued under the Trudeau government, social impact bonds (SIBs) are often promoted as an innovative way to provide funding for needed social services. In reality, SIBs are simply an additional set of financial and administrative arrangements layered into public spending towards some social or environmental goal. These financial arrangements shouldn’t be confused with whatever substantive social program is provided.

At their core, SIBs are a more expensive, privatized financing model for services that could be financed publicly at a lower cost. I am going to speak to why the social impact bond model is more expensive.

Let’s think about an example. A non-profit may have an idea or proposal to provide some type of needed and perhaps innovative service to a population in the community. Under traditional public financing, that program could be pitched to government, funded directly and ultimately evaluated.

In the social impact bond model, a private intermediary (often a consultancy) is added between government and the non-profit, arranging private investors to finance the project at the outset. The investor financing is used by the non-profit to provide the program and then the private investors are repaid by government, with an additional profit on completion of the program.

This arrangement adds costs in two ways. First, the rate of return required by private investors is substantially higher than the public sector’s own cost of borrowing. This higher cost of private capital adds to the overall cost of running the program. Second, additional costs come from the need to compensate the new intermediaries who coordinate between the non-profit and government in SIB contracts: consultants, lawyers, and so on. These are called transaction costs in economics research.

SIBs are a more expensive, privatized financing model for services that could be financed publicly at a lower cost.

The upshot is that using social impact bonds as a financing mechanism wastes taxpayer dollars to fund private profit. A given social program can be provided more efficiently by using direct public financing, and those cost savings can be reinvested in additional programs and services to meet the underlying social policy goal. This is the conclusion reached in detailed case studies examined by John Loxley on the use of SIBs in Canada and internationally: “keep the programs, get rid of the social impact bonds.”

According to proponents, there is one scenario when government is supposed to be able to save money from a SIB model relative to traditional financing. This is when the underlying social program fails. In SIB contracts, it may be agreed that if a program fails to meet certain metrics, the government payments to the private investors are reduced.

In practice, however, investors are risk averse and negotiate guarantees to protect their investment, often by setting low-risk outcome metrics that are easy to meet. Additionally, tying payment to simple-to-measure metrics creates incentives that lead to additional problems. These include cherry picking participants and generally incentivizing a focus on narrower measures rather than providing the highest-quality overall service based on a more comprehensive, long-term understanding of outcomes.

If there are instances where investors actually agree to contract terms with a genuine risk of failure, that’s reflected in what’s called a risk premium, where investors require an even higher profit on their investment as compensation for the risk, meaning higher cost to government.

Investors are risk averse and negotiate guarantees to protect their investment.

Whatever progressive language is used to dress up SIBs, the reality is that using higher cost private capital and adding a new layer of intermediaries wastes public dollars compared to direct public financing.

This is closely analogous to the P3 model for building infrastructure, which is well studied. The Ontario auditor general, for example, found that Ontario had wasted $8 billion by using P3s rather than direct public financing. I have reviewed this and other key research on P3s in previous analysis. Social impact bonds are a newer repackaging of this approach in a different area of policy.

As a final note, in addition to being wasteful, the SIB approach takes us further down a road of undermining, rather than building up, the capacity of our public sector and public service to address complex social and environmental challenges. City staff efforts should be directed towards building up a robust public sector capacity to meet these challenges, not to engage in a further outsourcing and fragmentation of that capacity to meet the needs of private investors, at additional public cost.

Topics: , ,