The best laid plans of New Zealand aid budgets

New Zealand currency (Flickr/Patrick Lauke CC BY-NC-ND 2.0)
Written by Terence Wood, Jo Spratt

New Zealand currency (Flickr/Patrick Lauke CC BY-NC-ND 2.0)Does the New Zealand Aid Programme have a problem with budgets and spending, or doesn’t it? In an earlier blog post we argued that the dramatic increase in nominal New Zealand government aid this financial year suggested the aid programme was struggling with spending predictability under a difficult minister. Vinny Nagaraj has responded disputing this. His response is well argued. On reading it, the casual Kiwi reader could be forgiven for breathing a sigh of relief and assuming all is well. Alas, this is not the case.

In short, Vinny–who is the aid programme’s chief economic advisor–argued that, because New Zealand aid is appropriated across a three year triennium, year-on-year changes in aid spending are of little concern, and last financial year’s under-spend was small anyhow. He also argued that the amount rolled over from the previous triennium (which ended in June 2015) to this one was also small compared to total budgeted spend for that triennium. He then claimed that the aid programme works well with its partners, and the predictability its planning documents brought is appreciated by them (as evidenced by input from Samoa and Kiribati to the 2015 Forum Compact Review [pdf]). He also pointed out that the OECD has commended the forward spending plans New Zealand provides its aid partners.

We are very grateful he has chosen to engage with us. We’ve written about aid spending issues and roll-overs of money at budget time in 2012, 2013, and 2014. The issue is perennial, but this is the first time we have had any official engagement. We’re thrilled.

In polite New Zealand fashion, we want to start our response by emphasising one area where we are in full agreement with Vinny: having triennial aid appropriations is great! It makes sense given the unpredictable world of aid, and it does indeed reduce “wash-ups” (the hurried last minute giving of unspent aid to multilaterals and the like). Like Vinny we think this is the way aid should be budgeted for, and we commend the previous government for introducing this process in 2003.

Next, we want to clarify a point that we probably did not make clear enough in our original blog post: we are not concerned that $20 million NZD had to be rolled over from the last triennium. We are concerned that when the 2015/16 budget was being prepared in the closing months of the 2014/15 financial year, the aid programme thought it was going to spend $588 million in 2014/15 (the final year of that triennium). Instead, it actually spent $568 million. The issue here is not an under-spend over three years, it is the fact that the aid programme could be so far off in an annual estimate just a few months before the end of the financial year. $20 million is not pocket change: it is nearly as much as the amount of bilateral aid we gave Cook Islands, our largest bilateral aid recipient in the year in question. A $20 million drift over three years: fine. A $20 million surprise over a few months: an issue.

Our other differences with Vinny are as follows.

First, while aid is allocated triennially, annual aid budgets are still important. Annual budgets help with predictability for organisations and countries that receive New Zealand aid (more on predictability in a second), but they also give a sense, within a triennial allocation, of how the aid programme is spending the money allocated to it. If the aid programme does not want to end up frantically “washing up” at the end of the triennium, it needs to come close to hitting its annual budgets. As Vinny says in his post, last financial year’s under-spend was quite small, but when it was added to the previous year’s last minute shock the numbers started to add up, and all of a sudden the aid budget needed to increase by approximately 12 per cent to give the aid programme any chance of spending the money budgeted to it this triennium. (12 per cent was our original estimate, the number will be slightly less if the numbers Vinny used, which were not previously publicly available, are correct; the difference won’t be much though.) A 12 per cent year-on-year increase is not easy to pull off. We checked using OECD data [.xlsx]. Even including the turbulent 1970s, a 12 per cent increase is within the top quartile of annual increases to the New Zealand aid budget. Manageable as a one off? Perhaps. But as part of a persistent problem? We’re not so sure. This unusually rapid increase was the central point of our previous post, and one Vinny did not address.

Second, Vinny refers to the 2015 Forum Compact Review [pdf], which states (p.19) that:

New Zealand’s Forward Aid Plans which provide four year forward projections to countries and to regional and multilateral agencies, was viewed as a concrete example of good practice in ensuring predictability of assistance to partner countries. The predictability provided by the forward estimates is valued by both Kiribati and Samoa as it enables them to forward plan and programme more effectively.

We don’t dispute the value of forward projections. Our concern is with stability of actual spending. Here charts of annual New Zealand aid spending in Samoa and Kiribati (based on OECD data [.xlsx]) are instructive. The red line is annual aid volumes under the previous Labour government. The blue line is the same information for the current National government.





As the careful observer will note, New Zealand aid to Samoa has been more volatile under National. And New Zealand aid to Kiribati has been much more volatile. We’re sure staff from the New Zealand aid programme work hard with their partner country counterparts to produce aid plans. We doubt, however, that the current funding environment in New Zealand has helped turned these plans into reality.[1]

You can see the same problem in the OECD information Vinny appeals to. It is true that in one OECD survey [pdf, p.13] New Zealand was lauded for its country-level forward planning documents. That’s great. But plans are at their best if they can be relied upon. In 2010 and 2011 the OECD published information on how good donors were at sticking to their forward spending plans (irksomely, they then stopped releasing the information). We think the OECD’s methodology could be improved (the method is explained on page 22 here [pdf]). Nevertheless, the information is the best available empirical test of how well donors put plans into action. And unfortunately, in all the various measures bar one, New Zealand scored worse than the median participating donor in both years.[2]

We’re sure New Zealand aid programme staff work hard in aid planning; it’s spending that is the problem. This was our original point.

Terence Wood is a Research Fellow at the Development Policy Centre. He undertakes research on Australian and New Zealand aid, and Melanesian politics.

Jo Spratt is an ANU PhD candidate studying NZ aid policy. Before this, Jo was an NGO advocate, manager and capacity-developer working in the Pacific. Jo is interested in global health, NGOs, and the practice of development and aid’s role in this. She is also a Registered Nurse.


[1] For those with an interest in these things, the coefficient of variation for both countries is higher under National. Also, note that while the Samoa earthquake of 2009 likely explains some of the aid increase in 2009 and 2010, it does not explain subsequent variation, nor the changes in Kiribati. Likewise, the 2011 Christchurch earthquake struggles to explain the changes: both globally and to the Pacific New Zealand ODA levels increased in the 2011 and 2012 calendar years. Also note that OECD CRS data (which are an imperfect measure but all the public has access to) suggests that much of the variation was in flows to the governments of Samoa and Kiribati, meaning that the fluctuation does not appear to be driven by one-off large infrastructure projects.

[2] Email us if you want the data and calculations behind these claims. If the aid programme is willing to provide us the data for more recent years we would be happy to test empirically if performance has improved on this front. Indeed, we think it would be great if New Zealand included a recipient-level measure of planned spending versus actual spending in its aid performance measures. This would be a very innovative initiative. If done we would happily applaud it on this blog. We would also be very happy to discuss possible methods.

Terence Wood

Terence Wood is a Research Fellow at the Development Policy Centre. He heads our program of research into Australian and New Zealand aid. Terence’s research interests include aid policy, the politics of aid, and governance in developing countries. He has recently finished his PhD, studying voter behaviour in the Solomon Islands elections. Prior to commencing PhD study Terence worked for the New Zealand government aid program.

Jo Spratt

Jo Spratt is an ANU PhD candidate studying NZ aid policy. Before this, Jo was an NGO advocate, manager and capacity-developer working in the Pacific. Jo is interested in global health, NGOs, and the practice of development and aid’s role in this. She is also a Registered Nurse.


  • Hi Jo and Terence,

    Thanks for the interesting response. There are obviously a bunch of things for me to think about, and I will try my best to make some time to do that quickly.

    Again, some (strictly) personal thoughts that in no way reflect the views of my employer. These comments are purposely terse and technical for any avoidance of doubt that this isn’t an official plain-English response 🙂

    1. The process of appropriating, allocating, expending and evaluating aid well requires optimization across multiple dimensions and constraints. These include a number of complex political economy considerations. This means that the solution set of “approaches” to how to deliver aid is vast, non-linear, and not entirely dependent on any one of those dimensions or constraints.

    2. In other words, if lower volatility was the only thing worth caring about, aid could take the form of a pre-committed agreement to simply “transfer” resources across to aid partners (whether they were governments, organizations or multilateral institutions).

    3. I have to think about this some more, but I am in the formative stages of wanting to say that volatile aid can still be predictable if there is sufficient up-front clarity about the nature of that volatility, and strong, continuous early communication of any changes. In other words, volatility might create first-order predictability problems, but managing that volatility might offer predictability.

    4. This is important because the change from red to blue was accompanied by a structural shift that involved a move within the solution space of aid approaches. That move may have coincided with a political adjustment, but ultimately involved a large tradeoff between being close to the “just make a transfer” set of solutions and somewhere else within that set which may lead to either (a) a temporary; or (b) a permanent increase in volatility. It may be temporary because this new approach involved a learning curve and fixed costs that required time for us to “settle in”. It may be permanent because the types of investments are naturally bulky, discrete, and involve large degrees of decision-making under uncertainty. My suspicion is that it may be both.

    5. Given that structural shift, forcibly smoothing volatility (e.g. by “washing-up” as described in my official response), may be sub-optimal in the short-medium term. Instead, it is more important to communicate this change of approach well and manage the volatility as described in my third point.

    6. If aid investments are more discrete and lumpy in this new approach, and if those discrete and lumpy investments are occuring in a challenging, uncertain, and unpredictable environment, then it might well be that we encounter deviations that appear “large” even over a shorter planning horizon.

    7. It also means that the 3-year horizon becomes a more relevant base to measure spending predictability.

    8. Also there is an important policy question that is missing from your commentary: while the structural shift caused increased volatility, there was a (large) levels shift which meant more aid was flowing overall to the country. How should a recipient country assess the overall impact of those two opposing forces?

    I’m sure there is an easy, pithy, 50-word way to say all this, but who doesn’t enjoy a verbose and inefficiently long-form comment? It’s so much more fun.

    Again, all unofficial, formative thoughts. But I will continue to have a think with my official hat on.


    • Thanks Vinny,

      That’s a great comment. I’ll be brief here because I’d already tried writing a longer reply. And my computer keeled over as I was doing it.

      In general I agree: reducing volatility is not the only thing we should aim for in our aid giving. If volatility is the price we pay for delivering certain types of aid, assuming the aid we’re giving is the right type of aid, then it is a price worth paying if we believe the benefits of the particular type of aid we are giving outweigh the costs of volatility. I also agree we can minimise the negative effects of volatility through careful planning with recipient countries.

      That said, I think we would agree that unnecessary volatility is still bad. As such our central difference is, I think, that you believe the volatility seen at present is an unavoidable part of the job, while I think it could be reduced (more staff, an easier domestic climate).

      Also, I haven’t seen evidence that, in the NZ case, the cost of reducing volatility is being reduced by good planning. (As far as I can tell the partner country planning documents are not available publicly. If they were and planning matched spending I would happily concede the point on this.)

      What is more, I don’t think all the volatility in spending is simply a function of the fact that we’re now doing more dams, roads and runways. If you look at the two charts below, you will see (via CRS, imperfect but all I have) NZ aid (current USD, calendar year) to Samoa and Kiribati that *went via the public sector*. Still a lot of variation. Obviously, one can funnel aid for large projects via partner governments, but I’m not confident this is what’s going on here.


      You’ve done a great job of arguing that volatility is part of life for an aid programme. And I am happy to concede aid programmes should be given leeway in their annual budgets etc. Nevertheless, I still think that when an aid budget as to increase year on year by 12% if an aid programme is to have any chance of meeting its triennial budget there are probably additional issues afoot.

      Nevertheless, I’ve learnt a lot from the exchange and am very grateful for it.

  • This is really cool to read at two levels.

    First, for the content by Jo/Terence and by Vinny reply, and the reply to that from the people-who-cant-dance. Insightful, and I enjoy trying to get my head around this.

    Second, so great to see constructive engagement. Almost a text book example of what dialogue can do for the *public* policy-making process. No, not almost. it actually is a textbook example. I can say that because I’m going to use this back-and-forth in the teaching of 41 postgraduate Development Studies students.


    • Thanks Gerard,

      To be fair, Jo can dance; I’m the only one with the issue there.

      Great to know you’ll use it as a teaching example. I’ll be very interested to see what the students make of it. Not just in a who won sense, but also in terms of what the students think the impediments are to informed public v govt department debate about aid, and what might remove these.


      • Great to know that you’re going to use this in teaching, Gerard. It has been a fun and interesting exchange, and got me thinking a lot about the challenges of improving aid predictability (and reducing volatility). I think there is the potential to do much more to stimulate informed discussion about aid policy, particularly in NZ. Maybe you could also assign students to make comments on these blogs to join in the discussion?

        For many issues in aid policy, and public policy in general, I think the winner has to be the policy at the end of the day: the discussion is only as good as it informs efforts to improve policymaking and implementation, with the ultimate goal of improved development outcomes. These are complex issues and there are layers to their discussion. In discussions with others, I think there is still more to say on the predictability issue, such as the need for careful partner country analysis (including public financial management and absorptive capacity analysis), and how to translate this into realistic planning. It also raises the question regarding whether more money should go to multilaterals rather than bilateral relationships, etc. And it raises issues about who makes decisions about aid policy and expenditure, and how a donor country can create quality aid policy. But I’ve been forbidden to write any more blogs (or blog comments for that matter – oops!) until my PhD is finished, so hopefully somebody else will in the meantime.

  • Very good analysis and response to Vinny’s blog. Twenty million is definitely not small change.

    We put a link to Vinnys blog in the July edition of our e-newsletter Update. The August edition could include your reply.

Leave a Comment