International REDD+ funding is now largely coming from development aid budgets, but has maintained the “performance based” element of the REDD+ idea.

REDD+ was a plan to create a system of Payment for Environmental Services, with the bulk of the funding coming from carbon markets. The failure to create a strong international climate agreement and a foundation for a global carbon market has forced policy makers to revise this plan.

International REDD+ funding is now largely coming from development aid budgets, but has maintained the “performance based” element of the REDD+ idea, at least in theory. This ‘aidification’ of REDD+ has made it similar to previous efforts of conditional, result-based, or performance-based aid.

The science is clear:

Forest loss behind Brazil’s drought

The idea of performance-based aid is simple: make payments to countries and projects based on performance or results. The performance can be measures in the form of policy reforms that will conserve forests and achieve other stated objectives, or measured more directly in the form of actual reductions of greenhouse gas emissions.

In a just-published paper I review the lessons and challenges of performance-based aid: “REDD+ as performance-based aid: General lessons and bilateral agreements of Norway”

A major conclusion from earlier research is that aid rarely can buy policy reforms. Yet this remains a major idea in current REDD+ discourses, as the basic principle of ‘no-cure-no-pay’ is very appealing. Relatedly, valuable lessons learned from performance based aid in other sectors have hardly been brought into the REDD+ debate.

In my recently published paper, I identified five challenges for the implementation of performance based aid:

Challenge 1: Donors are too eager to spend

In theory, payment should be made only if the results are achieved. But in practice there is immense pressure for donors to spend allocated budgets.

Underspending is viewed by the public as poor planning and performance. Within organizations, bureaucrats are promoted based on disbursement and overall spending. Under-spending carries a high risk of cuts in future budgets.

Donor reforms can reduce spending pressure. Bureaus and bureaucrats should be measured based on results, not on spending rates. Disbursement can be untightened from the annual budget process, and multi-year funds would reduce spending pressure. Competition among recipients for scarce REDD+ funds will also reduce the spending pressure.

In short, the donor costs of spending should increase by creating alternative uses of the funds, either on other recipients or in the future.

Donors also try to increase recipients’ willingness to undertake policy reforms, often referred to as governments assuming ‘ownership’ of the policy reforms. While efforts to create country ownership to REDD+ reforms may go hand in hand with performance based aid, the underlying idea of is, nevertheless, that the external incentives will provide the necessary impetus for reforms.

And, Paul Collier has earlier noted that: “… the ‘aid for reform’ approach takes a hopelessly naïve view of the reform process and of the game being played.”

But, the experiences are mixed, and assessment of earlier conditional World Bank lending to the forestry sector has demonstrated that, under the right conditions, funding can catalyse key forest policy reforms.

Challenge 2: Establishing performance criteria and measuring them are difficult

A performance based system must, quite obviously, establish performance criteria and measure them.

Input- or activity-based measures are generally poor indicators of the final impact. For example, a good policy approved by the parliament (an outcome) will have no or limited impact, if not properly implemented.

There is, however, a problem: Moving towards outcome and impact indicators is often more demanding and costly. Here is an example. Whether a forest has been legally designated as a protected area (an output) is easy to verify. Measuring the area of deforestation (outcome) over a specific time period is more demanding, but doable with time series of satellite images. To measure emissions (impact) one also needs emissions factors for deforestation (emissions per ha following the change in land use/cover), which requires field measurements from sample plots. But it does not stop there, we also need to benchmark the emissions.

Challenge 3: It’s hard to set realistic benchmarks or reference levels

Any performance-based payments require a benchmark – a yardstick against which performance is measured.

The simple equation is this:

Establishing that benchmark is the critical issue in all forms of impact analysis. This can also be viewed as an attribution problem, that is, to determine whether an output/outcome/impact is a result of the intervention or other external factors.

In general, such external factors play a larger role, as one moves along the causal chain. Setting benchmarks therefore becomes increasingly more difficult.

Reference levels are important both for phase two (policy reforms) and phase three (reduced emissions) of REDD+. For phase two, the typical assumption underlying performance based aid is a baseline of ‘no policy reforms’.

The Center for International Forestry Research has done much work on reference levels  related to phase three of REDD+, and let me highlight a couple of critical points from this work.

The REDD+ debate is often confusing because reference levels are used in two different meanings.

First, it can refer to the projected business as usual scenario (the counterfactual). The business as usual baseline is the benchmark for estimating the impact of the REDD+ measures implemented (and ensuring additionality).

Second, RL can refer to the crediting baseline, which is the benchmark for rewarding the country (or project) if emissions are below that level, or not giving any reward or possibly invoking debits if emissions are higher (depending on liability).

Two questions therefore arise:

1. How do we predict emissions from deforestation and forest degradation in a business as usual scenario
2. How do we set the crediting baseline?

Linked to the difficulties of predicting deforestation and the political issues involved in setting crediting baselines, there is a lot of scope for manipulating the numbers. (One of my favourite sayings is: “A reference level is a benchmark set so low that success is guaranteed.”)

Both donors and recipients have an interest in demonstrating success, and this can bias reference levels upward. In the agreement between Norway and Brazil (2008), the formula for setting reference level was historical deforestation for past 10 years (1996-2005). By the time of signing the agreement, deforestation rates had already dropped considerably, and this was an unrealistically high business as usual scenario.

Challenge 4: How to handle uncertainty and share risks

Result based systems that are based on outcome and impact indicators shift the risk to the REDD+ countries, because the results are influenced by factors beyond the control of the countries.

The business as usual baseline has several inherent uncertainties: the future values of drivers of deforestation and degradation are not known, the government may not either know the costs and effectiveness of the REDD+ policies implemented.

There are ways to deal with these uncertainties and to share risks (such as ex-post adjustments of reference levels and renegotiations), but none are quick and easy solutions.

Challenge 5: Putting money behind the promise

A final challenge―in contrast to the budget-pressure (challenge 1) is for the donor to put sufficient money behind the contract, such that the recipient country believes it will be rewarded fully for walking an extra mile.

If the contract is not fully backed by financial resources, the payment will be a lump-sum, without the marginal incentives it intends to create. For recipient countries this represents a major uncertainty, a point frequently made by REDD+ countries in the UNFCCC debates.

Again, the Norway-Brazil contract can serve as an example. In the paper I show Brazil has earned approximately USD 10 billion for the five year period 2009-13 of the contract. This is 10 times the amount stipulated in the contract.

The agreement is not credible from that perspective, and begs the question whether the agreement is result-based in practice, given that there is no way Norway can pay in full for the results.

Are the challenges so insurmountable that we should give up the idea of performance based support for REDD+?

That would be a hasty conclusion.

Five lessons need be brought into the discussion:

1. REDD+ as performance based aid can learn from lessons learned in other sectors. These lessons, e.g. from the health sector, are rarely brought into the REDD+ debate. They should be.
2. Performance based is hard―don’t be naïve. Policymakers need to have a realistic picture of the challenges and what performance based aid can achieve.
3. Don’t make promises you can’t keep. A credible agreement needs to be backed with sufficient money. The Norway-Brazil agreement is result-based on paper, but is in practice “receipt based”, as it is the spending of the Amazon Fund that determines the level of the transfers.
4. Make space for alternative uses of funds. Donors should create mechanisms to make it costly for themselves to spend. This could include creating multi-year funds, arranging competition (‘aid tournaments’), or handing over disbursements to third parties with clear instructions.
5. Don’t make all REDD+ aid performance-based. This is perhaps surprising, but nonetheless an important lesson. Some types of support do not easily lend themselves to clear and easy performance indicators. A minimum of non-performance-based support will also create higher predictability for recipients. Further, it will keep the door open for continued policy dialogue in situations or poor performance. As donors are unlikely to completely cut aid after agreements has been made, this would also increase the credibility of the performance-based part of the agreement. Donors should start small, with a well-designed performance-based mechanism (clear performance criteria, a credible reference level, and backed with sufficient funding), rather than putting all the eggs in the performance based aid basket.

Arild Angelsen is a professor of economics at the Norwegian University of Life Sciences and Senior Associate at CIFOR. Follow him on Twitter @aangelsen.