top of page

Simon Scott: A New Low - Bond Purchases as ODA, May 2023

stevecutts
May 31, 2023
4 min read

Updated: Nov 15


In this May 2023 paper, Simon Scott examines a proposal before the OECD's Development Assistance Committee to count bond purchases as Official Development Assistance. He explains how bonds differ from concessional loans and explains why buying a market-priced security does not involve "donor effort". The proposal subsequently failed.


Click on the link below to see the original pdf. An html version follows, which fixes broken OECD links.



A new low: bond purchases as ODA


by Simon Scott


Governments and companies issue bonds to raise funds to finance their operations. They are tradable at a market price and therefore provide a market return. Each day investors can go into the market and buy bonds at various maturities, in various currencies, issued by various entities at various rates of interest. They can buy initial issues, or they can buy bonds from those who already hold them. The return they are prepared to accept depends on constantly changing judgments about risk, inflation, and the opportunity cost of foregoing investment elsewhere.


A bond is like a loan, but with one important difference. Loans are issued by lenders; bonds are issued by borrowers. The difference is especially important when considering ODA eligibility. ODA is supposed to be a measure of something given away. Lenders can decide to give something away in their loan terms. But borrowers cannot decide to give themselves anything when issuing their bonds. They can offer bonds at whatever terms they like, but they cannot force anyone to buy them, and they will only be able to sell them if they offer a return that meets market expectations.


For this reason, buying developing countries’ bonds has never been reportable as ODA. Even bonds issued by the World Bank or similar institutions, where the money will clearly go to developmental projects, have always been ineligible as ODA, because any government buying them receives a market return, and is therefore not giving anything away.


The 2016 HLM decision, which negligently opened the door to counting non-concessional transactions as ODA, did not dare mention the possibility of counting bond purchases. Nor was the idea raised in the provisional reporting rules on private sector instruments agreed in 2018. And DAC statistical reporting directives issued only last month still classify the purchase of bonds at market terms as an “other official flow”, not as ODA (see paragraphs 18 and 512 here).


Yet a recent DAC document proposes to score bond purchases as ODA (section 3.3 here). This will mean that the very epitome of a non-concessional transaction – buying a tradable, interest-bearing security – is going to have a fictitious “grant equivalent” calculated for it, and scored as ODA.


As has become customary in DAC documents, the text of the proposal uses various tricks to obscure the origin and background of this idea and hide the magnitude of the change in ODA rules that it involves.


First, neither the document’s title nor its lengthy abstract even mentions bonds or explains why they are being dealt with. The only hint is a remark in the text that “members’ deliberations on PSI and other topics have invited clarifications” on the issue. Since this wording does not claim that members have invited the clarifications, it may well be that the DAC Secretariat has decided to do this off its own bat: we simply don’t know.


Second, the text nowhere makes clear that bonds have never been reportable as ODA, and that there is no HLM mandate for changing this.


Third, it repeatedly describes the fundamental rule change it proposes as only a “clarification”, as if the aim is merely to resolve some doubt that has arisen about how to interpret an existing rule.


Fourth, it uses the present tense and the passive voice to reinforce the impression that the unheard-of change it is introducing is somehow already the correct method, thus:


Donor effort in subscribing bonds is measured using the method and parameters for the treatment of loans to the private sector

and

Upon sale if traded indeed, an ex-post adjustment is introduced to the grant equivalents reported ex ante as a correction for the shortened bond maturity.


Fifth, it uses the technical term “subscribe” for the bond purchase. This conceals from the layman the reality of purchasing a security with a market return. It also gives the impression that only the purchases of initial bond issues will be counted, but there is nothing in the proposal to stop the purchase of existing bonds on the secondary market being counted as ODA as well, which from a financial point of view amounts to the same thing.


Despite these tricks, it’s easy to see why buying a bond never has been and never should be ODA. You are not giving anything away; there is no “donor effort”. You are getting a quid pro quo. And the quid is exactly worth the quo, according to the free judgment of the buyers and sellers in the market. It makes no more sense to score ODA for buying a bond than it would to score ODA for buying developing countries’ minerals, manufactured goods, or farm produce. Which only makes me wonder, will those be next?



Additional references from the original PDFs



Comments


bottom of page