Measuring Australia’s foreign aid generosity, from Menzies to Turnbull
Menzies and Sukarno, December 1959

Conventionally, and for good reasons, the ratio of aid to Gross National Income (GNI) is the principal measure of a country’s generosity as a provider of foreign aid.

Based on the economic growth forecast contained in the government’s December 2016 Mid-Year Economic and Fiscal Outlook, Australia’s 2016-17 aid budget of $3.8 billion is estimated to amount to 0.22% of Australia’s GNI. That’s 22 cents in every $100, lower than ever before.

If we stick to recent times—going back, say, to the turn of this century—a graph of Australia’s foreign aid generosity, in aid/GNI terms, resembles an asymmetrical hill. It rises from what was at that time a historic low of 0.24% in 2000-01, reaches a 26-year high of 0.34% in 2012-13, then falls to its new historic low in 2016-17.

If we extend the same graph back several decades, though, the twenty-first-century hill seems more of a hillock. Considerably higher levels of generosity were seen during the 1960s and 1970s. How much higher, exactly?

Trust me, I’m from the OECD

Answers to questions of that kind are often drawn from Organisation for Economic Co-operation and Development (OECD) aid statistics because they are readily accessible, allow for easy comparisons between donor countries, and are presumed to be authoritative. In addition, the OECD holds information on Australian aid levels in the 1960s that is not easily found in Australian government publications.

World Vision Australia’s Chief Advocate, the Reverend Tim Costello, was relying on OECD aid statistics when he said, in December 2016:

Aid was at its highest under Menzies, at 0.5% … when per capita income was much lower.

So was the Sydney Archdiocese of the Catholic Church when it issued a statement two years earlier, condemning aid cuts and saying:

Back in the 1960s and 1970s Australia’s foreign aid budget stood at 0.65% of Gross National Income.

Ditto The Guardian a year before that, when it said:

In the 1960s and 70s, Australia was seen as one of the world’s most generous donors, spending up to 0.65% of its GNI on aid.

The OECD statistics on which these various commentators relied, it turns out, are incorrect. Based on the Australian government’s own data, the highest aid/GNI ratio seen under any Australian government since annual reporting began was actually 0.48%. That was in 1967-68 under three Prime Ministers in quick succession: Harold Holt, John McEwen and John Gorton.

Costello’s statement, unlike the other two just quoted, was only slightly inaccurate with respect to Australia’s peak aid/GNI ratio—out by no more than a couple of tenths of a percentage point and one prime minister.[1] The other two statements were quite inaccurate. All these inaccuracies, small and large, stem from problems with OECD data on Australia’s ODA/GNI ratios from the 1960s right into the 1990s.

Measuring up Menzies

As the result of a recent fact-checking exercise undertaken for The Conversation, which led me to the above conclusion, I’ve prepared a policy brief presenting an accurate series of data on Australia’s aid generosity over time, extending for the first time as far back as 1961-62, based on appropriate Australian government ODA and GNI data.

The Development Policy Centre’s interactive Aid Tracker already brings together the most readily available Australian government statistics on Australian aid flows over time, extending back as far as 1971-72, but until now it has been hard to find any information on flows during the 1960s.[2]

Fortunately, it happens that the Australian Bureau of Statistics (ABS) used to report on aid expenditures in its series (discontinued after 2012) of annual Year Books, which can be found—though not particularly easily—on the ABS web site. By burrowing around in these, you can piece together information on expenditures from 1961-62 to 1970-71 with the exception of one year, 1962-63—and the rough level of aid in that year is easy to infer ($75 million).

On the basis of the ABS data and other information already reflected in the Aid Tracker, together with the most recent ABS time-series data on Australia’s GNI, it is evident that the OECD’s aid/GNI ratios for Australia are quite inflated over the three decades or so from the early 1960s to the mid-1990s (Figure 1). On average, the ratios published by the OECD are inflated by about 20% up to 1995. In some individual years, including 1975, they are inflated by over 40%.

Figure 1: Ratio of Australian ODA to GNI: Australian government vs OECD

Figure 1: Ratio of Australian ODA to GNI: Australian government vs OECD

Click image to enlarge. See this spreadsheet for a collation of the data on which the above chart is based, with links to sources.

The best of intentions

While OECD aid data are organised on a calendar year basis and Australian data on a July-June financial year basis, that is not a significant factor in the marked variations between the green and blue graphs above. The variations are mostly due to the use of very different GNI denominators up to the mid-1990s. It is likely that the GNI figures originally reported to the OECD by the Australian government, at least for the early decades of Australian aid, were subsequently revised upward by the ABS but not by the OECD.

In fact, the 1973 ABS Year Book said,

Australia has consistently been among the first three or four of the major aid donors, with a figure for direct government aid averaging 0.56 per cent of GNP over the last four financial years.

This statement, even allowing for the minor difference between GNP and GNI, is consistent with the ratios reported by the OECD for the same period and implies that contemporary estimates of national income were lower than today’s.


The moral of the story isn’t that the decline in Australian aid generosity is fake news. Aid generosity under Menzies was twice as high as it is now, even though per capita income was less than half of its present level. Australia’s aid/GNI ratio stood at 0.44% in the final year of the Menzies era and increased to 0.48% soon after. In addition, contemporary documents indicate that governments in the late 1960s and early 1970s believed they were spending some 0.56% of national income as aid. OECD statistics still reflect this belief. They’re unreliable, up to the mid-1990s anyway, but they preserve the best of intentions.

Robin Davies is the Associate Director of the Development Policy Centre. The above post and the related policy brief draw on a fact-checking article and related methodology note first published on The Conversation.

[1] If OECD statistics were correct, Costello’s statement would actually be the least accurate of the three quoted, since those statistics assert a maximum ratio of 0.65% under Whitlam.

[2] The Aid Tracker data will shortly be updated to incorporate the additional information on which this post is based, as well as additional information on 2015 aid allocations that became available at the end of 2016.

image_pdfDownload PDF

Robin Davies

Robin Davies is an Honorary Professor at the ANU's Crawford School of Public Policy and an editor of the Devpolicy Blog. He headed the Indo-Pacific Centre for Health Security and later the Global Health Division at Australia's Department of Foreign Affairs and Trade (DFAT) from 2017 until early 2023 and worked in senior roles at AusAID until 2012, with postings in Paris and Jakarta. From 2013 to 2017, he was the Associate Director of the Development Policy Centre.


  • This is important work Robin and great that the Centre has embarked on this search for the facts on aid spending. The Policy Brief is an excellent summary.

  • It would be interesting (and challenging) to also reflect on the legitimacy and effectiveness with the spending over time too.

    Do you have comments on what is/was counted as Aid and how that has changed over time?

    • Ian — I suggest looking at the comments on the related article published on The Conversation, and the responses to some of those from myself and others from the Development Policy Centre. They touch on both the effectiveness-over-time question (challenging, as you say) and your question about what was and is counted as part of Australia’s aid effort.

  • A question – is the expenditure on the original Colombo Plan, and subsequent iterations of that program over the earlier years of your time-span, included in the calculations of Australian Aid data? Or was it, in the earlier days, part of the Education Department budget on higher education? I don’t have access data on it, but my recall from APS in the 70s-80s is that it was the responsibility of the Department of Education, and if so, would not have been in the line budget of the aid program. I am not sure when it was moved into the line budget of AIDAB/AusAID. However my recall may be faulty, so I’m wondering if that detail can be made explicit?

    I have no idea if the level of expenditure in the earlier years would be sufficient to change the proportion by much – but the proportion of the ADS/Australia Awards programs in the aid budget in recent times is fairly high, and if the expenditure level of earlier, roughly equivalent, programs is not included in these proportional calculations, might that create a distorting effect on the estimation of relative decline?

    • Gai — Yes, the ABS Year Books were based on data from across government. Colombo Plan expenditure was included in the reported aid figures. For example, in 1969-70 the total amount of aid reported was $166 million, which included about $23 million in Colombo Plan expenditure (covering 1,635 people in training, including 949 new awards), $17 million in bilateral aid for countries other than PNG, $11 million in multilateral contributions and $115 million in budget support and other costs associated with PNG. Expenditure on the ‘Commonwealth Co-operation in Education’ program (only $885,000) was reported as being administered by the Department of Education and Science, budget support for PNG ($96 million) by the Department of External Territories, and contributions to the World Bank and the Asian Development Bank ($6.5 million) by the Treasury. The implication is that all other assistance was administered by the Department of Foreign Affairs, though this is not stated.

Leave a Comment