As Australia plummets toward the ranks of the least-generous donors, policy choices of unprecedented difficulty must be made about what not to fund. Those who thought that the aid program was at least respectable at A$5 billion per annum, and also many cut-weary people inside the Department of Foreign Affairs and Trade (DFAT), might very much prefer not to contemplate the options. But the Development Policy Centre exists in part to provide gratuitous advice on Australian aid policy, and that includes aid-cutting policy. It is the unpleasant duty of this post to contemplate the options.
In what follows, the chosen problem is to identify savings of $1 billion in the 2015-16 aid budget in line with the government’s announced savings objective. The government will also exact savings totaling a further $300 million in the 2016‑17 budget. The latter problem is not addressed here, in part because it is less immediate, and in part because it will likely be achieved by across-the-board cuts. That approach certainly will not work in 2015-16.
For present purposes, the aid budget can be divided into ten categories of funding, as follows. (Brief explanatory notes are provided at the end of this post.)
Cash payments to multilateral development banks (MDBs) are a no-go area. These payments are made pursuant to legal obligations. The banks hold numerous promissory notes from the Australian government, each of which has an associated draw-down schedule. The timing of payments can be re-negotiated to a degree but the payments must be made. It is not possible to discover what the aggregate cash payment obligation will be in 2015-16 but it does tend to be fairly stable from year to year, so let’s assume it holds steady.
The budget allocation for humanitarian, emergency and refugee programs was rashly cut in the Coalition’s early-2013 revision of Labor’s 2013-14 budget, then restored in the 2014-15 budget to almost exactly the level previously envisaged by Labor. Labor’s 2013‑14 aid budget was to total $5.6 billion, so Labor was allocating six per cent of it to humanitarian, emergency and refugee programs. This year’s budget is $5 billion, so the Coalition’s restored allocation represented seven per cent of it. One could argue, therefore, that the 2015-16 allocation should be reduced to seven per cent of $4 billion, which is $280 million. However, given the state of the world, the far-reaching impact of the funding provided and the ease of administering it, there is a very strong case for not touching this line.
The allocation for UN, Commonwealth and other international organisations took quite a hit in the last budget, falling from $395 million in the estimated 2013-14 aid budget outcome to $318 million in 2014-15. While commitments to these organisations are in general not legally binding, despite their apparent firmness in so-called ‘partnership agreements’, there would now be little scope to decrease them further. In fact some painful reallocations might need to be made within this allocation to accommodate recent and substantial pledges to the initial capitalization of the Green Climate Fund and to the second replenishment of funding for the Gavi Alliance, as well as commitments [pay wall] to renew funding for medical research into malaria and tuberculosis treatments. This budget line might not be a no-go area, but neither is it a hollow log.
Messing with funding in some of the budget categories above will be fiddly, often diplomatically excruciating, and in many cases only marginally rewarding in savings terms. Nevertheless, some of the savings required will clearly need to be achieved through general shaving. So, for the sake of argument and without endorsing anything that follows, let’s assume a 10 per cent level of savings from the ‘government departments other than DFAT’ category (yielding $39 million), the ‘UN, Commonwealth and other international organisations’ category ($32 million) and the ‘centrally-funded NGO and volunteer programs category’ ($20 million). In addition, let’s assume the current five per cent cap on DFAT’s aid-related departmental expenses as a proportion of total aid is maintained, such that a fall in the aid budget from $5 billion to $4 billion will yield savings of $50 million on the administrative component of the aid budget.
The four assumptions above yield total savings of $141 million in 2015-16, leaving $859 million to be found from the first four of the budget categories above. Together these four categories currently account for 64 per cent of the aid program, or $3.2 billion. Assuming for a moment that the budget cuts could be made in one fell swoop—i.e. that there was total flexibility in all programs funded by the relevant budget lines—one scenario for achieving the required savings is as follows.
This scenario, clearly enough, lands enormous hits on cross-regional programs and Indonesia, while subjecting Papua New Guinea (PNG) and all other country and regional programs to a uniform cut of 10 per cent. (The cut for PNG would preferably be exactly 10 per cent, given the size of this program, while cuts to smaller country and regional programs need not be uniformly applied, provided they yield savings of 10 per cent in aggregate.)
The cross-regional program, it should be recalled, ballooned from $309 million in the estimated 2013‑14 budget outcome to $686 million in the 2014-15 budget. It is not known what this huge increase was intended to fund, though it is reasonable to assume that part of it was contingency funding for regional resettlement deals, as with Cambodia, and that part of it was meant to cover some of the replenishment costs listed above. Reducing the allocation to $336 million would constitute no cut at all relative to normal levels of funding for this budget line.
As for Indonesia, the size of this country program increased by around 20 per cent after the 2002 Bali bombings, to around $150 million in today’s prices, then almost quadrupled from that base since the 2004 Indian Ocean tsunami. The first increase was warranted, the second not, or at least not at that magnitude. It is very difficult to maintain that Indonesia, a country whose net aid receipts turned negative for the first time ever in 2013, needs anywhere near $543 million in Australian aid each year. This is arguably the right time to make a step-change in our aid relationship with Indonesia, based on the fact that it is an emerging middle power—a member, with Australia, of the recently formed MIKTA grouping and of the G20 grouping, which has assumed greater significance since being elevated to leaders’ level in 2008. It would be better to cut quickly and cleanly rather than year after year. There is, though, still plenty of work to be done with a budget of over $230 million in Indonesia.
A second, quite different scenario would drop the constraint that limits the impact of the cut on smaller country and regional programs to 10 per cent, while retaining this constraint for less flexible programs (budget lines 5, 8 and 10).
Under this scenario, there is no disproportionate hit on Indonesia. All country and regional programs, including the Indonesia and Papua New Guinea programs, are cut equally by 20 per cent. The cross-regional budget is reduced by $351 million as per scenario 1. The hit on PNG, while not disproportionate, is very large in absolute terms at $100 million.
Essentially the two scenarios presented above involve a choice between (i) making a big step-change in aid to Indonesia, with a cut of some 57 per cent, and (ii) cutting all country and regional programs by 20 per cent rather than 10 per cent. The difference between the shape of the aid program now and its shape under both of the two budget cut scenarios is illustrated in the figure below.
Aid budget scenarios compared ($m)
Admittedly, inflexibilities within programs might be such that cuts, wherever they are made, cannot be made in one fell swoop. A pattern of allocations like either of those presented above is very unlikely to be achievable in a single budget. However, as noted earlier, the timing of payments to MDBs is negotiable to some extent, so it may well be possible to move to one of the above patterns over two budgets by adjusting the timing of payments that will fall due in 2015-16.
Of course, the two scenarios above are drawn from an infinite universe of possibilities. Nevertheless, they illustrate the fundamental choice that the Minister for Foreign Affairs faces, given that she cannot extract large savings from programs other than country and regional programs, and in some cases should not do so anyway. So, which way should she jump?
My vote goes to scenario 1. As argued above, Indonesia can well use a substantial allocation but does not need to consume over 10 per cent of Australia’s aid, or over one-fifth of Australia’s aid through country and regional programs. And aid to PNG should not be reduced much given that country’s woeful development indicators and narrow donor base—though it would be a fine thing to see this part of the aid budget more effectively used.
Nothing in what is said above is based on pragmatic considerations, and certainly not on any assessment of diplomatic feasibility. However, it should be possible to communicate a step-down in aid to Indonesia as quite a deliberate correction now that the tsunami is far behind us and Indonesia is increasingly playing a middle-power role internationally. There is of course a large danger that a cut in aid to Indonesia this year, of any magnitude, will be perceived, or indeed presented, as a retaliation for the (now seemingly imminent) execution of two Australian citizens. But aid to Indonesia can hardly escape a substantial reduction under any scenario.
Robin Davies is the Associate Director of the Development Policy Centre. He serves on the board of an NGO which has received funding from the Australian aid program, and the Centre itself also receives funding from the aid program.
Notes to budget categories table: (a) Figures are drawn from Table 1 of the DFAT publication ‘The 2014-15 development assistance budget: a summary’, in which allocations to countries and regions reflect DFAT country program budgets, not ‘total flows’ estimates as in Table 2 of the same publication. (b) The cross-regional allocation funds programs that benefit multiple regions, such as sector-based initiatives, but also functions as a war chest in which to park funds that the government does not wish to allocate to programs at budget time. (c) The negative adjustments are not explained in the 2014-15 budget papers but are of a similar magnitude to those made in previous years, which were said to be required to exclude some expenditure not eligible to be reported as Official Development Assistance, like GST payments, and reconcile expenses calculated on an accruals basis with cash payments. (d) This total adds to more than 100 per cent owing to the application of the adjustments described in note (c).
Hi Robin. Thanks for this thought provoking piece. I was surprised that there was no mention of the framework proposed in the independent aid review (p.126) for allocating aid to countries that Stephen played a key role in crafting for the Rudd Government.
Country allocations in this framework were based on i. poverty; ii. national interest; and iii. Australia’s capacity to make a significant difference. Indonesia unsurprisingly scored the highest priority with a ‘1’ – ‘medium’ for poverty, and ‘high’ for the two other categories.
While you highlight the practicalities of having to take the aid portfolio to the chopping block and there has been a lot of water under the bridge since the independent aid review, wouldn’t these same criteria make for a worthwhile framework in considering the aid cuts now?
The allocation framework you mention was applied for the most part to building blocks defined at the regional level. The only individual countries mentioned were PNG, Indonesia, Solomon Islands and Timor-Leste (and these last two were bracketed together). In addition, the framework served only to sort the building blocks into a hierarchy of five ‘focus categories’, with category 1 comprising the four countries just listed.
Obviously it’s possible to question the current relevance of the above process. When undertaken, it was about identifying where billions of additional dollars might reasonably be absorbed over several years with the most benefits per dollar. The task now is very different in nature: to identify where $1.3 billion can be extracted over two years with the least possible damage. Perhaps country A can better absorb additional aid than country B, but it clearly does not follow that country B can better absorb the pain of aid cuts than country A.
Relevance aside, even when first propounded the aid review’s allocation framework entailed nothing much about the appropriate magnitude of aid to specific countries. The ‘scope for expansion’ table at p. 131 of the report of the aid review suggested that PNG and the Pacific shouldn’t expand much, and that Indonesia could expand quite a bit, but did qualify this judgement by saying, ‘at the same time, Australia is starting [in Indonesia] from a high base’. The aid review was elsewhere silent on this last point, effectively taking it for granted that the marginal aid dollar had more impact in Indonesia than elsewhere.
Even with an implausibly low ‘scenario 1’ allocation of around $230 million per annum, Indonesia would still be Australia’s second-largest bilateral aid recipient, by a very long way, and therefore a focus category 1 country. So the aid review’s allocation framework is consistent with both a bigger allocation to Indonesia, and a much smaller one, provided Indonesia remains among our top few recipients. That reduces the utility of the framework to aid pruners.
Thanks Robin. Interesting scenarios. Do you think that Indonesia might just stamp its sovereign foot sometime very soon and say ‘you know what, keep your aid’? And perhaps in light of the ‘koinuntukAustralia’ push it might not be that far away.
The Indonesian government could certainly play that card, as the Suharto government did with the Netherlands decades ago, but I suspect would not do so unless there were a disproportionate cut to the bilateral aid program, along the lines of scenario 1 above, which they perceived as retaliatory in intent. No matter how much offence they might have taken at the ex post conditionality implied in the Prime Minister’s recent remarks, they, like successive Australian governments, tend to take a long view of the aid relationship because they see aid as a stabiliser in what is regularly a tempestuous bilateral relationship.
Joel, Ryan, Sam and Garth,
Thanks all for your comments. Forgive this blanket reply, but it’s easier to do it this way:
1. I’m not a proponent of eliminating aid to Indonesia or reducing it to a rump. $230 million per annum, or something like it, is a lot of money.
2. If Australia’s aid allocation policy were to make country allocations broadly proportional to absolute-poverty headcounts or total population levels, a lot of country allocations would have to change dramatically.
3. I’m not a proponent of basing aid allocation on per capita GNI alone. The piece above does not base its recommendation, a step-down in aid to Indonesia, on improvements in Indonesia’s per capita GNI (which currently stands at US$3,580, in the same ball park as the Philippines, Sri Lanka, Vanuatu and Mongolia).
4. The recommendation is, rather, based on the observation that aid to Indonesia was ramped up to an extreme degree for purely political reasons post-tsunami, that it remains at an artificially elevated level ten years later, and that Indonesia has meanwhile moved on in several ways. I don’t recall anybody arguing that Indonesia was receiving too small a share of the Australian aid budget prior to 2005.
5. It’s possible in principle that Australia is getting such good development returns on this massive investment in Indonesia that we should forget about the origins of its huge aid-share increase and be happy that the money turned out to be so well spent. But it’s extremely unlikely that a dollar spent in Indonesia is so superior in impact to a dollar spent somewhere else.
6. It is also extremely unlikely that DFAT could identify a ‘bottom billion’ dollars’ worth of genuinely poor-quality aid projects to axe across the whole aid program. It might find some tens of millions by taking this approach but any further cutting would have to be justified on non-quality grounds.
7. Given the government’s decision to take the aid program down by $1 billion in 2015-16, there are actual, large trade-offs to be made here. If the suggestion is that Indonesia should suffer no greater proportional cut than other countries, it should be acknowledged that many other countries will feel their cuts much more keenly, as they are more reliant on our assistance than is Indonesia.
I agree with you Robin that poor quality programs are likely to make up much, much less than 20% of the current program, however in a constrained funding environment I think it makes it even more important to systematically identify and prioritise those activities which have the biggest poverty reduction impact per dollar invested. Of course, in reality, other criteria such as Australian political and strategic interests and contractual obligations will also apply, but I would hope that DFAT will be working to ensure that this 20% cut to aid program funding will result in a much smaller reduction in the aid program’s poverty reduction impact. This will require more than just giving the program a hair cut, but assessing each of the individual strands of hair.
While I understand, and sympathise with, your argument for larger cuts to the Indonesia program Robin I think that average income per capita is an inadequate criterion by itself to determine aid levels. This is especially so when there is still a very large number of very poor people in the country and the average per capita income at around US$3600 is still less than one third that of the upper bound for ODA eligibility. In addition there are provinces in Indonesia that are poorer and have larger populations than some of countries to which Australia gives aid. In 2011 the Independent Review of Aid Effectiveness concluded that significant increases in aid to Indonesia could be effectively used and that Australian aid had a high capacity to make a difference in the country.
Instead I would hope that DFAT would take a more results-oriented approach and determine, in co-operation with partner governments, and across multilateral programs, which activities are failing to meet a benchmark for cost effectiveness in reducing poverty.
Thanks Robin for a thought provoking and well argued piece. The rationale for cuts in Indonesia appears to be pretty similar to the winding down of the aid program in China and India, so there’s at least a precedent. The bilateral relationship is deep and broad, so hopefully any major aid cut would be seen as part of a mature evolving relationship.
While I’m not privy to the internal deliberations around the cuts, it strikes me that the whole process would be more transparent and credible (not to mention diplomatically paletable) if the analysis was undertaken and communicated to those effected before the actual announcement was made – it’s one thing to make big cuts to the aid program (that may or may not be warranted, depending on your ideological perspective and view of economics) – it’s quite another to make them on the run, with little pre planning and in a way that hurts the stakeholders involved and, potentially, Australia’s national interests.
Thanks Robin, for an interesting and clean post, on an important topic for which I have not yet seen much constructive engagement.
Its worth noting that in in 2014 Indonesia’s poverty headcount was 28.4 million: well over triple PNG’s entire population (poverty in Java alone is over twice PNG’s population). Regarding the scale-up (which if mostly disaster aid should probably be wound back when the disaster is over, I agree), many of the important Indonesian programmes have been well-evaluated and I suspect we have a much better idea of what is working better there (here), than we do in just about any of the other programmes. There is a strong case to be made that the aid to Indonesia has been some of the more effective aid delivered (even if we just focus on poverty, and governance broadly speaking, not the other non-development objectives the program no doubt serves), particularly when we contrast it against some of the other countries’ in the region.
I’d vote Option 2, or an unnamed third one that targets particular themes/types of aid instead of countries, although this would likely make the quantum harder to hit in a simple way.
Thanks Robin for a wonderful piece. As you note, the level of funding to Indonesia is the elephant in the room. Australian cash is insignificant compared to Indonesia’s GNI (US$2.3 trillion). And whether it buys us “influence” is very debatable. Australia can engage on policy issues (universal health coverage for example) without needing huge amounts of cash injections.
I think grouping all non-Indonesia and non-PNG country programs together is a bit messy as some of the countries including in that – like Indonesia – are wealthy enough to fund their own development programs. Iraq, Fiji and Tonga for example are upper middle income countries (according to World Bank) – so in theory at least should have sufficient funds to drive their own development agenda (perhaps with some small policy support from DFAT). Samoa and Timor-Leste are just on the cusp of upper-middle-income status and Indonesia and Philippines (two large recipients) are not that far behind (see here http://data.worldbank.org/data-catalog/GNI-per-capita-Atlas-and-PPP-table).
Perhaps WB classifications are a bit of a blunt instrument and there of course remain (large) pockets of poverty in those countries. But using the WB classifications as a method for reducing aid would make sense to the Australian public and would – in some ways – return the aid program to a poverty reduction emphasis rather than a model where middle-income Asian countries get the most aid.