Aid and the Maintenance of Infrastructure in the Pacific

Why is it that the public infrastructure of developing countries is often decaying? Roads are in many cases littered with pot holes, wharfs are sometimes no longer operational. Infrastructure maintenance in developed countries is certainly not perfect, but there appears to be a stark difference in the maintenance and consequently the economic life of infrastructure projects in developed and developing countries. Because such infrastructure is generally the responsibility of government, this begs the question: why do governments in developing countries find it so difficult to maintain existing infrastructure?

The obvious answer is lack of funding. Governments in developing countries commonly have less financial resources, and have pressing expenditure priorities such as extension of infrastructure into both rural areas and urban squatter settlements.

But governments in developing countries do find the money for new infrastructure projects. In Fiji for example, the government recently upgraded the last unpaved section of the King’s Highway, linking Fiji’s two biggest cities along the northern coastal road (this was done in several stages, with the final stage funded using a loan from Exim Bank Malaysia). In the meantime, the much used Queens Highway, which was originally constructed with donor funds and links the east and west of Viti Levu along the southern coast, is decaying. Many roads in Suva are in worse condition, with potholes everywhere. Why is the government funding new highways while failing to maintain existing roads? The answer cannot solely be about funding.

A more convincing set of arguments relate to the incentives of those in government. This was the argument of Elinor Ostrom (the 2009 Nobel Prize winner), Larry Schroeder and Susan Wynne, in a book they co-authored in 1993. Governments in developing countries, they argued, have political incentives to build new infrastructure; new infrastructure is widely publicised and benefits particular constituencies. The same is not true for maintenance. Maintenance of existing infrastructure – generally built by previous governments – simply does not attract the attention of the media, nor does it win votes.

The same incentives exist for politicians in developed countries. However, due to independent bureaucracies, strict procurement rules, and the less salient nature of patron-client relations; these incentives are less pronounced. Governments in developed countries are also not recipients of aid.

The construction of infrastructure using donor funds strengthens the incentive bias against provision of maintenance. According to Ostrom, Schroeder and Wynne, governments in developing countries will not invest in maintenance if they believe that a donor will replace infrastructure once it is no longer operational. Donor provision of infrastructure thereby creates perverse incentives for partner governments not to invest in infrastructure maintenance. These incentives flow through to different levels of government, with no incentives established for public servants to ensure that infrastructure maintenance is adequate. On the demand side, communities are less likely to lobby for maintenance of infrastructure when they perceive that infrastructure as being a gift from a foreign donor.

Similar arguments are made by the World Bank in relation to the Pacific. In the 2006 Pacific Infrastructure Challenge report, the Bank noted that: “In the Pacific, governments have often focused on building new infrastructure, rather than investing in sustainable infrastructure operations and maintenance” (p36). The report argued that donor aid is partially responsible:

“Aid dependence … weakens accountability and risk management. When an infrastructure asset is created with aid money, the normal disciplines of valuing the asset and providing for its replacement in the future can be avoided … making capital free may have contributed to poor governance and poor management over time by removing the pressure to have to pay for valuable assets or provide for their eventual replacement in the absence of donor aid” (p25).

The implication is that aid dependence weakens governance.

The problem is also the result of the way that donors operate. Donors have traditionally funded new infrastructure projects instead of maintenance of existing infrastructure. In part this is due to the view that aid is not permanent. Donor aid for infrastructure is instead provided in order to spur economic growth, often in poor regions, with the idea being that this growth will subsequently be self-sustaining (as will maintenance of infrastructure). In the Pacific, this view of aid has recently been challenged in a discussion note by the World Bank (see this blog for a summary).

Donors also face incentives to prioritise spending on new infrastructure over ensuring maintenance of existing infrastructure. The construction of new infrastructure is an “announcable”. It is a tangible asset that can be photographed and reported by the media, in a way that maintenance cannot. New infrastructure is also expensive, and provides donors with a means of spending money (the amount of which can also be reported). Maintenance does not provide the same benefit, given its low but recurrent cost over many years.

How can such problems be addressed?

Some analysts argue that involving the private sector in infrastructure provision can improve maintenance (see for example this ADB report). In the Pacific, private sector provision of services has achieved mixed results. In some cases, private sector involvement has been a success, and has led to provision of better maintenance, such as with road maintenance in Samoa. In other cases, private sector involvement has been a failure, as in the case of private sector maintenance of solar rural electrification schemes in Fiji (see this paper of mine on that scheme). Corporatisation of government entities is an alternative, although again the results have been mixed.

In all cases, the key to success involves a combination of adequate funding and establishing appropriate incentives for maintenance. Importantly, for private sector involvement to be successful, the government must have the capacity to establish appropriate incentives for private sector stakeholders. This capacity is sometimes lacking in line agencies in the Pacific – as it was in the case of rural electrification in Fiji.

Maintenance can also be incorporated into donor-funded infrastructure projects. This is being done by some donors in the region, and is consistent with the Paris Declaration’s emphasis on improving the capacity of partner government bureaucracies and ensuring that aid generates sustainable results.

The Pacific Regional Infrastructure Facility (PRIF), which is co-funded by AusAID, New Zealand, the World Bank, and ADB, deserves special commendation. The PRIF provides both funding for physical infrastructure projects, as well as support for partner governments to ensure that these infrastructure works receive adequate maintenance. In some cases, this involves incentive payments to local governments that provide adequate maintenance to infrastructure (see for example this article by AusAID). The PRIF is an excellent example of how donors can ensure the infrastructure in which they invest fulfils its economic potential.

Another example of successful integration of maintenance into infrastructure programs is the transport infrastructure schedule to the Papua New Guinea – Australia Partnership for Development, which draws on PNG’s Medium Term Development Plan (2011-15). The schedule states:

“The Governments of Papua New Guinea and Australia acknowledge that, as a general rule, the returns from maintenance and rehabilitation of roads are higher than those from the construction of new roads” (p4).

It subsequently commits the Australia and PNG Governments to rehabilitating key roads in PNG and putting in place systems that will ensure their continued maintenance. No new roads are included in the document, although it does provide scope for “assessment of future transport needs”.

Unfortunately, similar approaches have not yet been adopted by all donors in the region. In one of the least sustainable infrastructure projects in recent years in the Pacific, the Chinese Government provided funding for construction of the US$12.9 million Samoa Aquatic Centre for the Pacific Games – one of only three certified Olympic swimming complexes in the region (the other two are in Australia, New Zealand has none). The centre originally included a statue of a white elephant, although this has since been re-painted yellow. An article in the Samoa Observer two years after the Games laments the fact that no international competitions have since been held in the complex, and observes that the pool is beginning to rust.

Donors clearly need to ensure that maintenance is provided to the physical infrastructure that they finance.

Matthew Dornan is a postdoctoral research fellow at the Development Policy Centre.

image_pdfDownload PDF

Matthew Dornan

Matthew Dornan was formerly Deputy Director at the Development Policy Centre and is currently a senior economist at the World Bank.


  • Excellent points on this important issue.

    As a further comment, I understand that reasons for the failure to maintain roads in PNG and other developing countries also include financial and institutional. At the national level, the PNG DOW relies on its annual budget submission and allocation to fund routine and periodic road maintenance. In practice, the DOW’s budget request is cut back and even the agreed amount is only partly received. Why? Does this reflect a more substantive issue related to the politics of budgeting and the management of public finances?

    Is there also an underlying problem in the PNG Government’s tendering system for infrastructure projects? While the public procurement environment in PNG has improved in recent years a recent Central Supply and Tender Board (CSTB) PNG Procurement Assessment stated that ‘whilst steps have contributed to improving PNGs procurement system, many challenges still remain …. The report paints a picture of a procurement system in dire need for a strong push towards reform, in order to bridge the inherent weaknesses in the system’. The assessment identified a number measures that could be taken to strengthen procurement, which included the following:
    • Strengthening existing procurement laws
    • Extending the legal framework to cover all procurement using public funds
    • Increasing compliance with existing laws
    • Improving the quality of technical specifications
    • Improving transparency on bid award decisions
    • Improving the complaints process
    • Ensuring completeness of procurement records
    • Providing better support the development of competitive procurement markets
    • Improving the procurement control mechanism
    • Increasing independence of CSTB
    • Improving legal provisions for targeting corruption.

    E. John Blunt is an Institutional and Public Procurement Expert. Much of his recent work has focused on increasing the use of partner government procurement systems. He has worked on infrastructure projects in Asia and the Pacific. He is currently on assignment with the Southern African Development Community Secretariat in Botswana.

  • Ami, your point is a good one and of course, this information is crucial form a technical point of view but what I think Matthew’s blog highlights is that in this arena (as in so many others), no amount of technical savvy will assist with pro-development outcomes if the political environment is disregarded. In fact not only does it need to be regarded, it needs to be well understood and engaged with in a nuanced way

  • The discussion of this important issue would be much better informed by contributions from the many professional engineers advising governments and donors or who are working on government or donor funded infrastructure projects. Comment by the various engineering professional associations including the Institution of Engineers and FIDC would also be useful as would comment from various donors.

    Are engineers and their representative bodies advising governments and donors to combine future maintenance contracts with construction/reconstruction activities? If not, why not? If such advice is being proffered, whay are governments and donors not accepting that advice?

  • Excellent post Matt. In Timor-Leste, there are certainly significant infrastructure maintenance issues (district roads in particular). But driving around Dili, it is also clear to see maintenance nightmares in the making as large government offices and other facilities are being built (many funded/donated by China) using methods that are non-traditional and materials that may be hard to replace or source locally. A number of these buildings seem to be aiming for grandeur rather than practicality, which will only create a bigger maintenance burden for the national government down the track — just like that Samoa swimming pool. I’ve heard a number of stories around Dili about how these new buildings are already suffering from problems.

    There’s also the challenges of the climate and weather. For example, in the wet season in Timor, many roads are washed away or damaged by landslides. Some materials are more prone to rust and wear in humid climates. Donors need to consider this at the beginning of projects as well to make sure that they minimise the maintenance burden in the medium to long term by choosing appropriate materials and methods for construction. For example, why build a sparkling new road alongside a steep area prone to landslides without some kind of barrier or protection from falling rocks, or drainage for the water flowing downhill to run underneath the road? Without addressing these considerations, the road is clearly going to be badly damaged and need significant repair every wet season, reducing its usefulness.

  • I agree with Robert Cannon that this is an important contribution to an area of focus in donor aid programs.

    A number of interesting points are made by both Cannon and ManRuz and are worthy of further discussion, perhaps in a wider forum.

    The importance of maintenance has been often forgotten when designing aid funded infrastructure projects.

    I remember being on Nauru when AusAID was again contemplating further work on the power station which in 2006 was falling into disrepair. Why? As I understand, the AusAID funded project was for the purchase, installation and commissioning of the power station. Maintainence? In a culture that was essentially ‘non-maintenance’, it should have been expected that the maintenance of the facility was going to be a problem.

    In PNG as well as in many other developing countries, the problem is not only obtaining funds for road maintenance, but of more significance, obtaining funds for more complex infrastructure activities including the maintenance of bridges. I wonder how many of the current donor funded bridges construction/reconstruction activities either underway or planned for PNG have a maintenance component. If not, should they?

    I note and agree ManRuz’s suggestion that combining future maintenance contracts with construction/reconstruction activities for infrastructure is perhaps a start in the right direction.

    E. John Blunt is an Institutional and Public Procurement Expert. Much of his recent work has focused on increasing the use of partner government procurement systems. he has worked on infrastructure projects in Asia and the pacific. He is currently on assignment with the Southern African Development Community Secretariat in Botswana.

  • While not suggesting that aid dependence is not an issue the suggested primary link to “the normal disciplines of valuing the asset and providing for its replacement in the future …” is I think invalid. While it is aid dependence is linked I suggest the underlying drivers of decision making need to be explored in more detail and the cultural component (outside normal political posturing) needs to be developed.

    The position was best summarised for me by the national of a developing country when asked about the issue – his response was that it is fundamentally cultural. In a day-to-day living environment where the standard practice has never included the need to put away for the future notions such as preventative maintenance and replacement provisions have no place. Personal property in the form of homes are built and not worked on until they fail – why should the attitude to public infrastructure be any different.

    I would suggest that this is further exacerbated when examining a subsistence environment at the national level of a developing country. It is pretty hard to focus on setting aside funds for the future when you are preoccupied with finding enough money for school fees.

    How to address it? I suggest that some of the strategies being employed in PNG of combining future maintenance contracts with construction/reconstruction activities for road infrastructure is a start in the right direction. However, there are a host of associated issues to be addressed and the only certain thing is that if the answer was easy it would have been done a long time ago.

    • Thanks for the interesting comments. I wasn’t aware of the Indonesian school maintenance example.

      Ostrom et al.’s argument is clearly a simplification/abstraction of a much more complex reality. I’m of the opinion that culture and government systems will tend to reflect underlying political economy incentives. At the very least, cultural norms/govt systems cannot be against the material interests of those in power; if they are, these norms/systems will not be followed and will be changed. I see the cultural norm of maintenance neglect in the Pacific as occuring precisely because there are limited incentives for infrastructure maintenance. Were incentives to provide maintenance established, I strongly suspect “cultural” factors would quickly change. The same applies for govt systems – if leaders and senior bureaucrats had incentives to establish systems of the kind that Robert describes, I think this would occur – although I agree it isn’t a simple process, and does warrant donor focus.

      I recently read a discussion paper on these sorts of issues, which also explores the link between material incentives and culture. I don’t agree with all the arguments in the paper, but the link is here if you are interested:

  • This is an important contribution to an area of focus in donor aid programs. The image Matthew creates of the white elephant in Samoa will stay with me for ever!

    A similar story with maintenance exists in Indonesia with school infrastructure projects. As part of an aid effectiveness review in basic education I led for the World Bank in 2008, we visited schools that had received ADB project funds and technical support for substantial school refurbishment in both Lombok and in Bali. In both locations we observed frequent examples of schools that had been refurbished in the period 2003 – 2005.

    Yet by late 2008, the condition of many school buildings showed heavy wear-and-tear including broken tiles, filthy windows, damaged furniture, worn paintwork and decaying ceilings. Children’s toilets were in poor condition and in some cases totally unusable. In one primary school in Bali the building rehabilitated in 2004 was being torn down and replaced by an entirely new structure to provide more suitable accommodation for children. To be fair, there were other schools that were, on the whole, neat, clean and pleasant places in which learning could occur but there was almost no visible evidence in classrooms of changes in learning and teaching practices, mainly because this was not a part of the project design at all – astonishingly, a school education project with no attention to teaching in the design, but that is another story!

    During an evaluation for the USAID Decentralized Basic Education Project, Package 1 (DBE1), in late 2011, some of the reasons for lack of school maintenance emerged which seem more fundamental and solvable than simply blaming such outcomes on lack of funds or on laziness, which of course can be important reasons as well.

    One reason is lack of technical knowledge of how to judge the condition of schools and to plan for maintenance at the level of schools and District governments. In many cases, local governments simply do not have accurate data on the stock of schools in the District nor do they know their physical condition. Accordingly, planning for school maintenance, when done at all, has been at best haphazard. This helps explain the depressing frequency with which one still reads in the Indonesian press of school roofs and walls collapsing on children leading to death and injury.

    However, when DBE1 developed appropriate school data gathering tools, and analytical strategies, in consultation with schools and Districts, accurate numbers, distribution, size and condition of all schools in participating District were known for the first time. Planning for maintenance could therefore be based on increasingly accurate data as data collectors and data users began to realise the importance of accurate data and its use in planning. Further, some Districts that were planning to build new schools realised their stock of schools indicated further construction was completely unnecessary.

    If Districts lack accurate school and school-age population data (as many and possibly most do), this raises the very real possibility that school construction projects may be adding unnecessary stock when the immediate need is for school mergers, re-locations or even closures, and for resources and technical skill in school and equipment maintenance.

    By way of conclusion, and to answer Matthew’s question “How can such problems be addressed?”, one step in addition to those he describes is to help local stakeholders collect and to use accurate data about the condition of their physical assets, to help in planning maintenance programs, and to provide access to reliable technical support for maintenance.

    I hope Matthew’s excellent article will attract further contributions on this important theme.

Leave a Comment