Why Is There So Little Contractually Contingent Financing?

Most foreign aid comes in one of two forms: either we pay a person or an institution today in exchange for delivering some beneficial activity in the future, or we observe something bad happen to them and then give them support to recover from it. This kind of aid is simple to design and deliver, but in the former case has limits in how sharply it incentivizes success and effort from a range of actors and in the latter case leads to the inefficient and undignified “begging bowl” approach to humanitarian financing. In what follows, I identify a broad family of alternative approaches, which can loosely be grouped together as “contractually contingent financing,” and explain why they are still relatively underused.

What is contingent financing?

There are various kinds of contractually contingent financing, but what they all have in common is the basic premise: if X, then Y. For example: if you can design and produce 1 million units of a safe, effective vaccine for disease X, I will pay you $10 per dose; or, if you suffer damaging and widespread flooding, we will provide you with immediate financial relief, calculated according to a pre-agreed formula. This general structure underlies pull financing and advance market commitments; some anticipatory social protection (which may be but is not always contractually contingent on an expected natural event); insurance; disaster, pandemic and catastrophe bonds; payment-by-results; and cash-on-delivery.

There are two characteristics that define what I consider “true” contractually contingent financing. The first is that the contract pre-specifies under what circumstances payment is made, and if those circumstances never materialize, payment is not made. And the second is that the circumstances chosen are not wholly predictable: they depend on the materialization of uncertain external conditions or the achievement of some uncertain outcome or impact by the implementer. Some traditional aid may be “contingent” in the sense that payment installments may be made only the achievement of some input milestones like the appointment of a project team or the submission of a final report, but this kind of instalment structure is not uncertain in the way the delivery of a new vaccine is, and I do not consider it true contractually contingent financing in what follows (a more substantial edge case is IMF conditional lending, where financing is disbursed only on the achievement of specific conditions, which are largely, but not entirely, under the control of the borrower).

Proponents of the rather different forms of contractually contingent financing point to a number of benefits: in the social protection space it should be faster and more dignified as well as more impactful, because the need to fundraise is eliminated in situations where speed of action has an outsize importance. It reduces uncertainty in financing disaster response and may facilitate more effective action. Its use as pull financing or various forms of payment-by-results creates a different set of incentives for effort and may facilitate the engagement of a different, wider set of potential actors. It also transfers risk between parties: in disaster response, it reduces the risk borne by recipients and increases risk borne by funders; in payment by results and pull financing it increases the risk borne by innovators and delivery agents and reduces risk borne by funders (transferring risk is not always a good thing, but in some cases will be).

Even if you think these arguments are overblown and the reality is much more complicated (and it is absolutely the case that the benefits can be overstated, and such approaches should only be used when they are appropriate to the actual constraints to better outcomes), it’s hard to look at the balance of international financing and not conclude that we should at least be trying more contractually contingent financing. Comprehensive numbers are hard to find, but in the humanitarian sector, for example, “pre-arranged financing” amounts to less than 5 percent of total crisis finance, despite growing evidence that it can deliver much better results. And even after the announcement of a new £185 million pull financing programme by the Department of Energy Security and Net Zero, only a small portion of UK aid, for example, is provided using such mechanisms. Other donors, including the World Bank, and have used contractually contingent financing in various guises, but only sparingly. You need to have very strong views on the uselessness of contractually contingent financing to think our current state is the right balance.

So what accounts for the relative dearth of financing that takes this if X, then Y structure? Sometimes the barriers to a better mix of funding relate to real and meaningful concerns about the appropriateness of contractually contingent funding in specific cases. But often, constraints are largely institutional and organizational. Most of these are surmountable (or at least possible are possible to mitigate), with creativity and effort—which, sadly, are also scarce resources. My colleague Rachel Bonnifield has looked at the constraints for two specific donors, the European Commission and US government; here, I set out four general constraints that affect all donors.

Institutional capabilities

First, contractually contingent financing requires specific organizational and institutional capabilities which must be invested in. Writing contracts that specify payment in the event of an outcome being achieved or a state of the world being realised requires that we can fully (or fully enough) specify that outcome or state of the world. This is not trivial: defining a successful innovation means defining what we take as success (lab solutions? Field solutions? Field solutions that are purchased? Field solutions that are bought and used to some minimum level?); defining a disaster requires a full-enough list of what constitutes a disaster, and how we identify them, both conceptually and in a specific time and place. These contracts require specialist contracting skills. We may outsource the resolution of the ambiguities that are inevitable in such contracts to institutions with authority to resolve them either through adjudication or by taking on responsibility for declaring success (or disaster) in the first place, but there are relatively few such institutions, and it takes time to acquire authority. And once contracts are agreed, contingent financing only really works if all parties believe the contract will be upheld or, if need be, enforced through legal means—all of which requires further institutional capability.

Administrative complexity

Secondly, contingent financing is practically more complicated to manage for most donors. It requires that funders keep resources to one side to pay out in the event of the trigger being met. They might achieve this in many ways. They could literally put the payout aside, perhaps in some fairly liquid investment, ready to extract when necessary. They could hold the liability on their books, but “double-allocate,” that is spend most of the money elsewhere, with some flexible reserve being used to finance a payout and future spending reductions making it good if the payout is ever made. They could use some form of insurance to cover unexpected payouts, either formal or informal (for example, an arrangement with another organization or number of organizations to “risk share” on their respective contingent financing projects). These arrangements are always unusual: they are not the normal, routine way in which budgets are allocated and spending managed in most funding organizations. They require some special plan to be made, executed and monitored. As such they also run down another scarce resource in organizations: attention. The more dedicated thinking time or active management an arrangement needs, the less attractive it is to a stretched organization. As a result, as my colleague Rachel Bonnifield has written, these financing mechanisms are more attractive at large scale, where the effort and admin costs of managing them are effectively amortized or spread over a number of projects; yet larger projects are also seen as riskier and donors are less likely to act as first movers into them when they are less familiar.

Distribution of risk

Third, donors often have an asymmetric attitude to risk that influences the mechanisms they deliver aid through, potentially at the cost of effectiveness. Specifically, donors often worry more about “false positives,” paying out when they shouldn’t, than “false negatives,” not paying out when they should. The desire to mitigate such risks can lead donors to undermine the very rationale for contingent financing by either including excessively onerous checks on payment or by using incredibly cumbersome contracting arrangements—a classic example of which was the 386-page prospectus and 12 week wait-and-see period for the World Bank’s pandemic bonds, an instrument whose whole rationale was to remove uncertainty and increase speed (to be clear, a long prospectus is itself no barrier to a well-functioning contingent financing instrument, and there were many other problems with the PEF).

Ironically, the desire to avoid “false positives” is also sometimes the true motivation for using contractually contingent financing in non-disaster settings, with donors effectively arguing that by paying out only on the completion of specific results they reduce “waste.” This transfers risk from the donor to the implementer, but doing so is not always a good thing. Risk transfer is only one of the possible motivations for using pre-arranged financing, and if donors use the mechanism when it is the only criteria satisfied (and, for example, the true constraint to achievement of good outcomes is not effort or the range of delivery agents incentivized towards a target but instead structural factors beyond the control of a delivery agent), then the mechanism tends not to work any better than other forms of financing and is usually more expensive as delivery agents usually require compensation for taking on additional risk. What’s more, many implementers do not have the cash flow or capacity to take on all of the payment risk in this fashion; a structure that is partly contingent and partly up-front (or low risk) may make for a more effective mechanism, as the potential universe of implementers would expand.

Underprogramming of resources

Finally, a related constraint: all contractually contingent financing means running the risk that a payout is never made. For insurance and crisis financing, we hope the payouts are never made: we want disasters not to strike, and will invest in making countries more resilient to them and less susceptible to their human cost. For payment-by-results and pull mechanisms, we hope they pay out because that means a successful investment that has achieved the results we set out at the beginning of the project, but not paying out means that the risk transfer has worked: we are not paying for failure (or paying less). In each case there is a moderate chance that the “disaster” or “results” trigger never gets pulled. Most organizations are highly averse to underspending for two reasons: first, it can complicate their financial management, particularly so with those that use spending targets for aid (notably the UK); and secondly because it can make life more difficult when advocating for funding replenishments. It is difficult to argue that one needs more generous allocations when one has unspent balances on the year.

A dedicated institution can resolve these challenges

Many—but not all—of these constraints can be mitigated with a simple institutional solution: set up a separate contractually contingent financing organization, or establish a dedicated multilateral mechanism within an existing organization (something Bernat Camps Adrogue and I have argued for elsewhere). There is precedent to such a step: donors have in the past set up institutions to run the kind of projects they find difficult to manage through their normal channels, such as the Global Innovation Fund, and the Global Concessional Financing Facility, which provides concessional development finance to middle income countries with large refugee populations. Similarly, the Government of Colombia has created Fondo LOGRA, specifically to implement various results-based financing mechanisms.

A dedicated organization that administers many such financing arrangements at once, and agrees new ones regularly will quickly develop the capacity to write such contracts well, and accumulate a number of reliable institutional partners to which verification can be outsourced. Donors can disburse their full allocation for the pre-arranged financing mechanisms they wish to fund the pull mechanism to the institutions in one go (or in large and predictable chunks), thus removing the difficulty of how to manage uncertain disbursements. Of course, this difficulty is only outsourced to the dedicated institution, but by managing a portfolio of pre-agreed mechanisms with a range of likelihoods of payout and time horizons mitigates this somewhat, and it can develop specialist insurance of financial arrangements to deal with the uncertainty. What’s more, since this approach will be it’s “business as usual,” it will require less management attention (though it is likely that the admin costs of such an institution will be higher than usual; this is the trade-off for the potential benefits of contingent finance). And finally, by taking the risk of underspending off the books of the ultimate funder, it mitigates the political/institutional cost of underspending. It doesn’t, in itself, mitigate the issue or asymmetric risk attitudes, but this is a problem of institutional culture (as well as political culture, which is more difficult to tackle). Institutionally, a more “rational” approach to payout risks is possible to build, starting from leadership. Political risks can be managed by relying on a range of donors to fund the institution.

Note that this institution does not need to originate and decide on policy across all of the disparate policy areas it might run contractually contingent financing. A single organization with a policy function across health, humanitarian, education and other sectors would effectively be a new major donor. Instead, it could operate as a platform to which execution is outsourced, with origination of programme ideas and policy retained by the ultimate funders (either individually or collectively).

Getting something like this done is not trivial. Setting up new institutions is time consuming and difficult; most donors would like to avoid it. And doing so requires commitment to attempt this financing at scale, which is already one of the barriers to overcome. Nevertheless, it is doable. It requires a concrete plan for how it would be structured (two, in fact—one for a standalone institution and one for an institution housed in an existing international body); a sufficiently large pipeline of actionable projects to form its opening portfolio; and a first mover or champion around whom other actors can coalesce. The first requires effort and consultation, but nothing out of the ordinary. The second is almost achievable now, through the work of CGD, the University of Chicago Market Shaping Accelerator and organizations like Instiglio, the Center for Disaster Protection and others. The third is the most difficult, and most unpredictable. But it can also go from 0 to 100 quickly. Difficult things often happen.

Contingent financing is not a panacea. It will not work everywhere. And it can’t solve all the problems we face in development financing. But it is underused, and many of the reasons for this underuse can be addressed. What we need now is a coalition of donors who recognize this and who can take the first necessary steps.

I am grateful to Parth Ahya, Rachel Bonnifield, Sebastian Chaskel, Mark Plant, and Benjamin Stephens for excellent comments that improved this piece. Any remain errors are my own.