A Note on the 2016 aid statistics

The preliminary totals for 2016 aid (official development assistance – ODA) were published by the OECDOrganisation for Economic Cooperation and Development on 11 April. The overall headline in the OECDOrganisation for Economic Cooperation and Development press release was: ‘Development aid rises again in 2016 but flows to poorest countries dip’. The text explained that:

‘Development aid reached a new peak of USD 142.6 billion in 2016, an increase of 8.9% from 2015 after adjusting for exchange rates and inflation. A rise in aid spent on refugees in donor countries boosted the total – but even stripping out refugee costs, aid rose 7.1% . . . Despite this progress, the 2016 data show that bilateral (country to country) aid to the least-developed countries fell by 3.9% in real terms from 2015 and aid to Africa fell 0.5%, as some DACDevelopment Assistance Committee (of the OECD) members backtracked on a commitment to reverse past declines in flows to the poorest countries.’

There is a lot to unpack here, and some of that work has been done, by the Development Assistance Committee of the OECDOrganisation for Economic Cooperation and Development itself, and by independent analysts (including Eurodad, Development Initiatives, and CONCORD). An additional point to add relates to the adjustment for exchange rate movements: holding exchange rates constant, and ignoring e.g. the effect of any devaluation of a donor country currency, distorts what aid is actually worth to recipients. This was key for the UK in 2016. More on that below.

The conclusion is that the ‘narrative’ on oda in 2016 needs to be nuanced, for the world as a whole, and for the UK.

For the world as a whole, ODAOverseas Development Assistance went up, but by much less than the headline figure, especially when the impact of the UK devaluation is taken into account. Other key factors to take into account are in-donor cost of refugees and Spanish debt relief to Cuba. Taking account of all these factors, the overall increase in the value of oda to developing countries may have been less than 3% in 2016. It remains to be seen what happened to Country Programmable Aid and what level of overall funding (including bilateral and multilateral) was provided to least developed countries. The trend in humanitarian spending is also notable – it increased by 8% in real terms in 2016, less than the DACDevelopment Assistance Committee (of the OECD) headline figure of 8.4%, but more than the 3% estimate above.

For the UK, oda went up by 10% in £ terms, or by £1.2bn, because of rising GNI (45%) and the use of new accounting standards (55%). However, devaluation of 14% against the $US during the course of the year reduced the real value of UK aid, which actually fell in $US terms compared to 2015.

For the future, there are two main conclusions. First, that it remains important to disaggregate aid statistics and improve measures of real aid. And, secondly, a better understanding is needed of the impact of exchange rate movements on the value of oda to recipient countries. 

The background note from the DAC is worth reading in full. In addition to the points in the press release, it demonstrates that

• ODA rose in real terms in 22 out of 29 countries, with some remarkable increases, for example of 192% in Spain, 43% in Poland, and 36% in Germany. The figure for the UK was 8.4% (NB after adjusting for inflation and exchange rate changes).
• Despite these increases, the average ODA/GNI ratio was only 0.32 (against the UN target of 0.7), with only six countries reaching 0.7: Denmark, Luxembourg, Norway, Sweden, the United Kingdom, and, for the first time, Germany.
• Furthermore, the increases reflected a number of factors, especially the first year cost of receiving refugees in donor countries. So-called ‘in-donor refugee costs’ rose by 27.5% in current terms, from $US 12.1bn to $15.4bn, and accounted for 10.8% of total expenditure in 2016. Refugee costs accounted for over 15% of ODAOverseas Development Assistance in Austria, Belgium, Denmark, Germany, Greece, Iceland, Italy, Norway, Sweden and Switzerland. The figure for the UK was 3%.
• In the case of Spain, the large increase was the result of debt relief to Cuba.
• Humanitarian aid also rose by 8% in real terms in 2016, to a total of $US 14.4bn.
• As a consequence of these factors, and also a slight increase in the share of oda accounted for by multilateral agencies, bilateral aid to least developed countries fell by 3.9% in real terms, and bilateral aid to sub-Saharan Africa fell by 0.7%. Note the headline in the press release is misleading in claiming that aid to poorest countries dipped, unless there is evidence that the fall in bilateral spending was not matched by an increase in multilateral spending in those countries.

There are graphs and tables to illustrate all these points. The main ones are pasted in at the end for ease of reference.

As far as the UK is concerned, the headline figure is an increase of 10% in overall oda in current prices, or 8.4% in real terms, correcting for inflation and exchange rate movements. The increase results partly from an increase in GNI, and partly, as DFIDDepartment for International Development itself explained when the UK aid statistics were published at the beginning of April, from a revaluation of GNI, using the most recent EUEuropean Union standards (ESA 2010, replacing ESA 1995). A rough estimate is that about 55% of the increase was attributable to the revaluation of GNI, and 45% to the increase in GNI. The treatment of the post-Brexit devaluation is important, however. In current $US terms, UK oda actually fell, from $US 18.5 bn in 2015 to $US 18.0 bn in 2016.

The main point commentators make about the statistics relate to what is and what is not valued in aid.

The DACDevelopment Assistance Committee (of the OECD) itself uses the concept of ‘Country Programmable Aid’ as a key indicator. This is defined as

‘the portion of aid that providers can programme for individual countries or regions, and over which partner countries could have a significant say. Developed in 2007, CPA is a closer proxy of aid that goes to partner countries than the concept of official development assistance (ODA)’. Specifically, CPA is defined as gross bilateral grants and loans, minus:

• Debt relief;
• Humanitarian aid;
• Food aid;
• Administrative costs;
• Imputed student costs;
• Promotion of dev. awareness;
• Refugees in donor country;
• Core support to national and international NGOs;
• Aid from local governments;
• Equity investments;
• Aid not from main agency; and
• Other unallocated

2016 CPA data are not yet available from the DAC.

Some NGOs have used different concepts, focusing on what has been called ‘real aid’ and subtracting from the overall total what is variously called ‘phantom aid’ or ‘inflated aid’. For example, the EUEuropean Union NGO organisation, CONCORD, discounts:

• Spending on refugees in the donor country;
• Tied aid;
• Spending on students in the donor country;
• Interest repayments on concessional loans, and future interest on cancelled debts; and
• Debt relief.

CONCORD argue (Pg 12) that ‘according to aid effectiveness principles, these items do not contribute to international development, nor do they represent a genuine transfer of resources to developing countries.’ Note that, unlike in the calculation of CPA, humanitarian aid, food aid and support to NGOs, among other items, are not subtracted.

Adjustments of this kind have been much debated. See, for example, Richard Carey, then with the OECD, commenting on the early attempts to measure Real Aid; or David Roodman, then with the Centre for Global Development, defending technical assistance, administrative costs, and subsidies for students, but agreeing that aid for refugees in donor countries should be excluded, along with most debt relief. Jonathan Glennie is another who has written on this topic, inter alia criticising the counting of debt relief. The CGD’s Quality of ODAOverseas Development Assistance Index (QuODA) reflects many of the debates, with scores relating to efficiency, and reducing the burden on developing countries, among others.

As far as 2016 is concerned, the data are not yet available to permit full analysis. However, commentators have picked up some of the issues, focusing especially on the high share of first year refugee funding. Thus:

• Polly Meeks for Eurodad, wrote ‘taken in context, these new statistics should set alarm bells ringing for anyone committed to achieving the Sustainable Development Goals’.
• Concord said that ‘CONCORD, the European Confederation of Relief and Development NGOs, regrets that aid is being diverted from the countries most in need. Development aid is being instrumentalised as an incentive for governments to cooperate with the EUEuropean Union on its migration and security measures.’
• Bernd Bornhorst, Chair of the German NGONon-governmental organisation platform said: ‘The achievement of the 0.7 percent target in the election year (in Germany) is a swindle. A large amount of the total expenditure is spent on domestic refugee costs. Germany thus inflates the spending on development cooperation and remains the largest recipient of its own resources for development cooperation. We strongly discourage other EUEuropean Union countries to follow this path.’
• Amy Dodd, Director of the UK Aid Network said ‘Supporting refugees arriving in Europe is absolutely the right thing to do, and something we as Europeans should be proud of. But counting in donor refugee costs as aid – money spent in the donor country which never reaches a developing country – is of questionable development impact at best and certainly an attempt to artificially inflate countries’ aid figures.’

Not discussed in the commentary so far is the impact of controlling for both prices and exchange rates. The impact stands out most clearly in the case of the UK, where oda:

• Went up in £ terms by 10%, according to DFID, from £12.1 bn to £13.25 bn in current prices;
• Went up in real terms by 8.4%, controlling for both inflation and exchange rates, as in the table below;
• But went down in current dollar terms by 2.8%, from $US 18.5bn to $US 18.0 bn.
• And therefore went down in inflation-adjusted dollars by somewhat more, say 4%.

The main reason for the disparity is that the pound went down against the dollar during the year, as in the chart below, from around £1 = $US 1.45 at the beginning of the year to below $US 1.25 at the end of the year. This is a devaluation of nearly 14% during the course of the year as a whole, or an average of about 7%,  off-setting part of the rise in the £ value of oda during the year.

Figure 1

£:US$ exchange rate, 2016

Source: http://www.xe.com/currencycharts/?from=GBP&to=USD&view=2Y

This raises the question of why aid statistics are being adjusted, not just for inflation, which would be normal, but also for exchange rate movements. The DACDevelopment Assistance Committee (of the OECD) explains its procedure as in Box 1, and says that it is trying to standardise to the purchasing power of a US dollar in a recent year. But this is not easy to understand. If the DACDevelopment Assistance Committee (of the OECD) wishes to estimate the value to developing countries, then would it not be better use current exchange rates, not constant ones? Again, the UK example is informative: in US$ terms, the value of UK oda fell in 2016, because of the devaluation. Why ignore this? In fact, we know that the devaluation has caused problems for DFID, because commitments and expenses are often fixed in foreign currency, and have therefore cost more in £. For example, contributions to EUEuropean Union budgets are fixed in euros.

When colleagues were consulted on this, they agreed that there was an anomaly, though thought that the DACDevelopment Assistance Committee (of the OECD) methodology could be defended as giving an assessment of donor effort, rather than value to recipients. Some comments were:

• The DACDevelopment Assistance Committee (of the OECD) scoring of ODA, as you will recall, exists to enable donors to hold each other to account for the size of their respective efforts (that is, to disincentivise free-riding by one rich country on the generosity of the others). As WP-STAT and DCD have made plain on several occasions, it is no part of their mandate to provide information that might be useful to developing countries about how much money they might receive or for what purpose, or to evaluate what the money might achieve.  Any such broader benefits are incidental to DAC’s purpose in collecting and publishing this information. From this perspective of measuring donor effort, measuring changes at constant prices and exchange rates makes perfect sense.  Though as you imply, it makes no sense from any other point of view.
• Presumably, as a donor consortium, DACDevelopment Assistance Committee (of the OECD) thinks it should be measuring something like donor effort, but this is really only of interest to DACDevelopment Assistance Committee (of the OECD) and they have the oda/GDP ratio for that.  If you want to measure what is happening to the value of what developing countries receive, I think you are right to use just prices.

Box 1

Information Note on DACDevelopment Assistance Committee (of the OECD) Deflators

Source: http://www.oecd.org/dac/stats/informationnoteonthedacdeflators.htm

• What price index are they using to deflate?  If it is a single index that in some way represents what aid is spent on (not easy to design, but I used to use proxies like average OECDOrganisation for Economic Cooperation and Development export prices or average traded manufactures prices; now it might be better to find something including traded services and non-OECD donors as well), then there is no reason to adjust for exchange rates (and I am not even sure what this would mean).  From the point of view of  countries receiving aid, you want the total value of aid received each year, in any common currency (so dollars are fine) deflated by some index of what aid is spent on, also in dollars.  (If you are just playing with targets, then UK aid in sterling/UK GDP, of course.)
• I can see the challenge. The problem is that exchange rates can be so volatile – and hard to distinguish between permanent and temporary changes. in the long run a better – and feasible - approach might be the one the World Bank use for GDPGross Domestic Product per capita figures. As you may know their ATLAS methodology uses three year average exchange rate.
• Yes, there is a real cut to the value of UK aid in African (and other developing) countries due to the devaluation. On top of that (or partly as a result of that) DFID’s Africa portfolio faces an immediate cut of 30-40% with every single budget line being scrutinised by Ministers, again after similar scrutiny not so long ago (the review should have been finished ). This is a real pain for programmes that are due for renewal. I know of a project that could face closure by December if it doesn’t get some new UK funds in coming months. That would be a pity as it has good results and delivered everything it promised.

The question then is what difference it would make to the overall analysis of oda data if no adjustment were made for exchange rate movements. Would the overall increase be less than the DACDevelopment Assistance Committee (of the OECD) estimate of 8.9%. The answer seems to be ‘yes’. In current dollars, the increase was 8.4% and if this is corrected for inflation only, using the OECD inflation rate of 1.08% in 2016, then the increase was only 7.2%, significantly less than the DACDevelopment Assistance Committee (of the OECD) estimate – and allowing a smaller margin for an increase in real aid after allowing for in-donor refugee costs and other contested items.

To conclude, then, the ‘narrative’ on oda in 2016 needs to be nuanced, for the world as a whole, and for the UK.

For the world as a whole, ODAOverseas Development Assistance went up, but by much less than the headline figure, especially when the impact of the UK devaluation is taken into account. Other key factors to take into account are in-donor cost of refugees and Spanish debt relief to Cuba. Taking account of all these factors, the overall increase in the value of oda to developing countries may have been less than 3% in 2016 It remains to be seen what happened to CPA and what level of overall funding (including bilateral and multilateral) was provided to least developed countries. The trend in humanitarian spending is also notable – it increased by 8% in real terms in 2016, less than the DACDevelopment Assistance Committee (of the OECD) headline figure of 8.4%, but more than the 3% estimate above.

For the UK, oda went up by 10% in £ terms, or by £1.2bn, because of rising GNI (45%) and the use of new accounting standards (55%). However, devaluation of 14% against the $US during the course of the year reduced the real value of UK aid, which actually fell in $US terms compared to 2015.

For the future, there are two main conclusions. First, that it remains important to disaggregate aid statistics and improve measures of real aid. And, secondly, a better understanding is needed of the impact of exchange rate movements on the value of oda to recipient countries.