Benefit shortfalls of the PNG LNG Project: a response to Mark McGillivray

Part of the front cover of Jubilee Australia's 'Double or Nothing' report
Part of the front cover of Jubilee Australia's 'Double or Nothing' report

The PNG LNG project promised a profound and transformative impact on PNG, including a doubling in economic size. Four years after gas exports commenced, how has it gone?

Recent analysis by Jubilee Australia (JA) in the Double or Nothing report and PNG Economics (see here and here and here) indicates that there is a huge chasm between the early promises of the PNG LNG project and what has happened. This is not due to oil prices falling – gas export revenues are almost exactly as expected with falls in oil prices offset by higher export volumes.

The construction phase from 2010 to 2014 lifted the non-resource parts of the economy by around 8% above trend growth (more on this term later) – so about K3 billion extra in 2013. This was slightly higher than the 5% above trend growth initially expected.

Unfortunately, from about 2012, there were many poor policy decisions closely related to the PNG LNG project. These are examined is some detail in sections 5.3 to 5.6 of the Double or Nothing report – a public policy example of how technocratic advice can be overruled by entrenched resource interests and politics. Gambling on huge revenue inflows, the new O’Neill Government increased the PNG budget by over 50% in 2012 and 2013. As revenues did not flow, PNG suffered its worst budget deficits and a blow-out in government debt. Some extra expenditure was poorly spent on grandiose projects and building political support. Debt interest costs increased from K480m in 2013 to over K1,600m in 2017 – a major on-going budgetary drain. After a major appreciation of the Kina during the construction phase, the central bank intervened to stop it falling back to new market levels. This intervention has led to years of foreign exchange shortages which are seriously damaging businesses in PNG. The adverse impacts on the agriculture and manufacturing sectors were poorly handled and the government is moving back towards protectionist policies. The government also made poor investment decisions, including losing at least K750m from share trading in resource companies.

These poor policies, all associated with the PNG LNG project, mark PNG’s fall into a resource curse. Arguably, this is the third time this has occurred – there were similar poor policies around the Bougainville copper mine during the 1980s and poor policies around the major resource expansion in the early 1990s.

Economic indicators since the start of gas production have been bad for PNG. GDP did not double as predicted by the muddled modelling of an ACIL-Tasman report commissioned by ExxonMobil. The GDP gain was possibly 10% in 2016 (probably less, as PNG Treasury’s 2016 GDP estimates are probably too high). However, the poor economic policies induced by the project mean that even these GDP gains are being eliminated as the non-resource parts of the economy suffer. By 2016, non-resource GDP with the PNG LNG project was actually less than non-resource GDP without the PNG LNG project, assuming an underlying growth trend of the economy of 5%. By 2020, the same occurs for total GDP with poor economic policies swamping the K10 billion initial boost to production from LNG.

A key challenge for any analysis of the impacts of any major change in an economy is how the economy would have gone without the change. The importance of this key assumption was acknowledged as the most difficult issue in the JA report (first question in appendix 2). The JA report used an “underlying growth path” based on trend analysis of how the economy was performing prior to the analysis. This is a very common approach – whenever you see projections in Australian budget papers, this is simply the number from trend analysis. A sensitivity analysis was undertaken on the impact of using an assumed 5% underlying growth path, with key results included on page 17 of the report. Any reasonable assumptions re-enforced the message that the promises of the project were massively over-stated.

In this context, it was surprising that the recent blog by Mark McGillivray claimed that the JA report “is remarkably silent on the importance of this key assumption to its findings, with no real attempt to credibly justify or evaluate it.” This is just factually wrong and my comments on his blog set out the reasons why that is the case.

As supported by PNG’s Treasurer Charles Abel’s more balanced response to the report, there is general acknowledgement that the project has failed to deliver the level of expected benefits. The subsequent JA report On Shaky Ground also highlights the failure to deliver expected benefits at the local level. An excellent article by The Monthly explores the potential social and political risks that flow from these dashed expectations.

The resource sector in PNG is very powerful. It does some good things – such as support for employment, provision of basic services around project sites, and tax revenues. However, there is a general overestimate of its importance to the PNG economy. From 1980, the resource sector has accounted for only 1% of employment, 16% of total GDP, and 13% of total revenue. This is significant, but not overwhelming. However, there are significant adverse indirect impacts from the sector that are swamping the small-scale and local benefits from such projects – some would say the “tail is wagging the dog”.

The true significance of the JA report Double or Nothing is that hopefully it strengthens the hand of the PNG government in its future negotiations with resource companies in getting greater benefits for PNG’s people. Hopefully, it may encourage a re-examination of specific “resource curse” policies adopted since 2012. Most fundamentally, hopefully it will stimulate a broader discussion on whether PNG should move away from an excessive focus on the resource sector and towards a more inclusive development agenda, one focused on areas of PNG’s true comparative advantage related to its cultural diversity, biodiversity, the richness of its soils and the potential of its people.

image_pdfDownload PDF

Paul Flanagan

Paul Flanagan has a longstanding interest in public policy issues in Australia, PNG and more broadly. His 35-year public service career was evenly shared between Treasury/Finance and AusAID. He headed up Treasury’s International Finance and Development Division from 2008-2011 before being seconded to a senior advisor position in the PNG Treasury until August 2013. He is Director of Indo-Pacific Public Policy and Economics, a leading commentator on economic developments in PNG, and a frequent contributor to the Devpolicy Blog.


  • Hi Vailala
    Thanks for the comment and providing some interesting additional material on some of the project details. Before going into the specifics, I was left wondering whether you had any disagreements with the reports conclusions? For none of the big picture issues were discussed or commented upon. Do you agree that the benefits from the project are much less than expected/promised (and please don’t use the fallacy of the 2014 fall in oil prices)? Do you agree that the government has made poor policy decisions which seem driven by the resource curse? Do you agree that PNG should look to re-focus toward a more-inclusive approach to development?
    On some specifics, all models are inevitably based on assumptions – and to that extent “conjecture”. For example, there are many climate change models. We are not at 2050, so the different results could be dismissed as just “conjecture”. However, the vast majority of climate change scientists believe that some models are better than others –using criteria such as transparency, use of best available information on what has happened previously, providing a range of scenarios, plausibility of assumptions etc. The ACIL-Tasman analysis was a seriously flawed black box that clearly failed in its prediction that PNG’s GDP would be double what it otherwise would have been in 2016. That would have required an assumption of an underlying growth path of a negative 3 per cent for every year from 2010 to 2016. That is one conjecture – PNG otherwise having a serious recession going on year after year for 8 years. My alternative conjecture was that the economy was most likely to have grown at its pre-existing rate from 2007 to 2009 of a positive 5 per cent per year (with a range of 3 to 7 per cent tested). I’m quite relaxed to say one conjecture is better than the other.
    The detailed comments seem to be based on some mis-understandings. For example, the discussion of “net” well-head value is not about impurities to determine energy – it is about whether full capital costs should be netted off from sales revenue when determining royalties. The IMF discusses using a gross basis (not allowing such capital costs for royalties) as an example of an improved tax arrangement on any new project in Box 2 of its 2017 Article IV report on PNG p18. On tax havens, I’ll leave it to readers to see if the comments about using tax haven intermediaries is convincing (why couldn’t the ‘stitching’ be done in a country with OECD average rates of taxation rather than Netherlands Antilles with a Bermuda intermediary?). Lack of budgeted figures for UBSA and other agreements wasn’t about an MOA process, it was just that governments should consider both costs (which were very substantial) as well as revenues when considering the budget impacts of a project (including implicit equity costs). The report didn’t say the PNG-LNG agreement was secret – the report went through a known set of fiscal terms in the general PNG-LNG agreement but noted “the detailed commercial agreement between the PNG LNG joint venturers remains secret”. If this is incorrect, I apologise, and look forward to the detailed revenue model from joint venture partners being released immediately (or if you could provide a web-link?). Would also be good if the PNG-PNG Agreement could be made available through an official source – interested people in PNG may not trust a web-site that describes financing of the project as a “dodgy deal”. The diversion of funds of PNG LNG income to Kumul Petroleum Holdings is just one of the key flaws of the redesigned Sovereign Wealth Fund – the first version passed by Parliament had all of these revenues going straight to the budget without the risk of them being siphoned off. I don’t understand how this has anything to do with the discussion about poor land-holder identification – although we all agree this is a serious issue that has been mishandled. Equity participation was a well known challenge but it was under control. The equity costs were largely covered through the sale of Oil Search shares the government previously owned. Indeed, when IPIC decided to pay cash to the PNG government for the Oil Search shares the government’s full equity requirement was met – there was no residual issue about covering country risk or negative pledges (the cost over-run from $US16bn to $US19bn was covered directly through the budget). The issue then became O’Neill’s decision to purchase over K3 billion in additional Oil Search shares on top of its PNG-LNG equity (a decision which led to PNG’s Treasurer being dismissed when he questioned the legality of the arrangement). We know that this bad PM call cost at least $US245m, and this bad call helps explain the diversion of PNG-LNG funds to Kumul Petroleum Holdings rather than into the budget to fund desperately needed programs in health, education and infrastructure.


    • Thank you Paul for your comments. I found some of your comments not at all understandable until I turned to your reference to the IMF 2017 report. Unfortunately the IMF report contains a somewhat common error which may at first glance seem trivial, but it covers a key point in understanding the political economy of the LNG project. The error is to assume that the royalty benefit (and sometimes the development levy) accrue to GoPNG. These two items underpin the basic legality of the LNG project and the GoPNG scheme for natural resource development. The royalty benefit forms an essential step in creating the State’s title of ownership to uplifted petroleum and the simultaneous transmission of this title (not to the “LNG Project”, which has no legal identity, or form) to the licence concession holders or co-venturers.

      Why is this fact important? The grant of benefit takes the form of a gift and as such requires, for perfection, offer, acceptance and the passing of the property from grantor to grantee. This explains why there is a Development Forum with an MOA as its outcome or product. The MOA provides the necessary documented attestation for a dealing involving land and the identified landowners signed acceptance of the grant or gift. Much the same considerations apply to the development levy and the income due to landowners, etc, from the grant of equity. The monies that the State receives from the co-venturers (which includes the PNG entity as a co-venturer) in fulfillment of these obligations is property awaiting identified landowner, LLG and PG collection and must, accordingly, be held in trust accounts before payment to the identified beneficiaries. Arrangements for the holding and management of these trust funds (SPVs) parallel those made for the ‘stichting’. The MOAs provide some detail on these matters. This also explains why the GoPNG MOA committments are limited to ‘good efforts’ and ‘best endeavours’. These efforts and endeavours are to be made by GoPNG in support of the choices that landowners, LLGs and PGs are expected to make in respect of their local social welfare, infrastructure and development plans and budgets.

      So what is at issue here is not the State failing to ‘honour’ its MOA promises but rather the recipient parties (landowners, LLGs and PGs) failing to honour their commitment to engage in local initiatives for the expenditure of the trust fund monies. In part this failure can be attributed to the long failure to identify beneficiary landowners and the confusions engendered by genealogical reckoning. Many Southern Highlanders and Hela people understand this very well. The current political turmoil is a source of great frustration to them as they work continuously to build political solutions and trust. Especially so when they look at the achievements of some other Provinces in implementing social welfare and development initiatives through their own political processes.

      Other comments made bear on the issue of the commercialization of a gas extraction and export industry, policy settings and tax issues. These can be best viewed through the lens of the unincorporated joint venture structure and its grounding in the concessional licensing scheme legislation. To take just one example, APT. If you don’t have a project people will tell you that the settings are too high. If you do have a project the same people (often the IMF) will tell you that the settings are too low. And so it goes.

      Recently PM O’Neill and treasurer Abel have commented –

      “We love our concessional loans”

      “We need free equity”

      “We need better returns from resource development projects”

      A general conclusion is that GoPNG now reckons that the cost of 14.5% LNG Project equity (19% when the landowner equity grant is included) imposed too great a burden on the PNG economy and financial system resources. GoPNG equity was initially pegged at 22.5% in anticipation of a Project capex of US$10 billion.

      The ExxonMobil discussion papers and reports on the financing arrangements that it led for the co-venturers (but not, I believe, including PNG) note on more than one occasion PNG’s difficulties caused by the negative pledge. Very briefly, this requires that any monies raised by PNG borrowing be allocated first to the pay down of the obligations created by concessional loans.

      GoPNG participation as an equity owner and co-venturer in resource development projects is an important component, perhaps the most important component, in the management of country risk. Needless to say it is backed up by contract stability agreements with the co-venturers.

      So, what do we have here? To me it looks more like a “Capex Curse” rather than a “Resource Curse”. It illustrates well the perils that may follow when you enter into a strictly commercial relationship with a very large multinational and and engage with a world of fluctuating commodity prices and markets.

      I’ll leave the SWF discussion to another day.

      You have my sincere sympathy in regard to the clumsy and error prone analysis of the IMF. I too once regarded these reports as the epitome of icy analysis based on established facts. Too often these reports are regarded as statements of factual record. That they may be no more than the tergiversations of FIFO ‘cowboy’ economists should always be weighed in the balance, as the experience of other countries shows.


  • Thank you Paul for your post. I am unpersuaded by your elaboration of your modelling methods for the PNG economy. The model remains conjectural and comparing one set of conjectures with the conjectures of the Acil model is fraught, in my opinion, with the possibility of error.

    Your ‘Double or Nothing’ report mentions the PNG LNG Agreement, signed by the co-venturers on 22 May 2008 as secret. A copy of this agreement can be found on the web-site (with links to Jubilee Australia). The PNG LNG Agreement defined the financial terms under which the co-venturers agreed to enter the Project. This document was followed late in 2009 by the Final Investment Decision document (FID) which committed the licensee partcipants to meeting equity capital calls. Post the GFC the co-venturers met with some difficulties in raising the needed capital from private equity markets. This was largely overcome by ExxonMobil taking a lead role and involving various export credit agencies (ECAs) to both shepherd commercial banks into the deal and handle the vexing issue of country risk. See for example the contribution of JBIC and NEXI (both country risk and technical risk).

    The following are some specific comments of the ‘Double or Nothing’ report.

    Page 31 argues that “net well-head value” was a “generous financial concession” on the part of GoPNG. Not so. Net value is more or less an industry norm for gas projects. Examination of the Conocophillips process will show that it is driven by gas-derived energy as are other parts of the field operations, such as the Hides Gas Conditioning Plant. Raw gas is contaminated by impurities (including water) which must be removed before sale. Using gas as energy is both energy efficient and cost-effective. Any joule consumed by processing is a loss to the gas sales volumes that accrue to all UJV co-venturers. The costs of using gas for this purpose are subject to scrutiny and approval of the co-venturers.

    Page 32 refers to tax havens. There appear to be some serious misunderstandings here. The entity referred to here is a ‘stichting’ created under Netherlands law and probably domiciled in the Netherlands Antilles (Bermuda may be merely an intermediary). It is not a commercial entity or a trading company and has no members or share capital. It’s purpose may be narrowly defined as the pay-down or amortisation of LNG project debt. As such its existence may be part of pledging arrangements. Its broad purpose is to shield the project debtors from some of the consequences of project failure, either partial or complete. For that reason it is always located in a jurisdiction in which none of the parties hold any assets, operations or bank accounts. In order to maintain the necessary legal distance from the debtors (beyond the ‘guiding hand’) it is necessary that ExxonMobil and other debtor company directors have no knowledge of or contact with the stichting directors, whose function is merely that of ‘trustees’. Creation of SPVs is a component part of risk management for both creditors and debtors.

    Accounting for the discrepant rates of tax paid by ExxonMobil as against co-venturers may find its explanation in the role of ExxonMobil as LNG Project co-lender. PNG LNG Agreement “Exhibit K Capital Uplift Provisions” may be of relevance.

    Page 22 – 24 refers to the absence of budgeted figures for UBSA and the not-mentioned LBBSA. There is again a serious misunderstanding here. The MOAs LBBSA and UBSA are just that – MOAs. In these agreements GoPNG gives an undertaking to support landowner development project wish-lists by either using its “best endeavours” or making “good efforts” to bring the proposed projects before the EIC and/or the relevant stutory authorities (National Planning, Roads, etc, etc). The MOAs also include and tie-in both LLGs and Provincial Governments. The policy intention is to push project initiation, planning and budgeting onto local authorities whenever appropriate. This general policy has been in place for decades and has produced some very good results in some provinces and marginal results in others. The creation of the Hides Gas Development Corporation is partly an attempt to replicate the success of organisations such as the Gazelle Restoration Authority. The epic failure of the anthropologist driven social mapping and landowner identification studies scheme is the root cause as to why there are no budgetary entries pertaining to the project MOAs (except for business development grants to landowner companies). Both the other Jubilee Australia report and the “The Monthly Report” that you mention are in need of a re-think of their basic assumptions on this matter.

    Page 39 mentions the SWF and the “diversion” of LNG income to the SOE KPHL. To understand the background to this decision we have to go back to the 2008 PNG LNG Agreement and the disastrous inclusion into this agreement of the ExxonMobil/OSL “sweetheart deal” with the Australian-based contract anthropologists. The 2008 PNG LNG Agreement Clause 15 covers Social Mapping and Landowner Identification issues. Clause 15 (a) (ii) and (iii) require the State to agree that existing and future SMLI studies have met the requirements of the Oil and Gas Act. Further details as to what is to be determinative here are provided in Exhibit J. This latter provides, inter alia, that the persons engaged for the SMLI studies shall be “scholars” and that their principal task shall be to “generally report” under the headings of:

    “Social-Cultural Context – detail who the ethnic groups are in the Licence Area (e.g. Huli, Febi, etc)”

    “Social Organisation – to provide a description of the clan structure, migration history, … principles of group formation (kinship and descent) … (s)ample genealogies should be collected.

    “Mapping Results – Findings – To provide a preliminary distribution map showing the relative positions of major groupings such as clans in the area”.

    An uncharitable view would be that the resulting consultants SMLIS reports never advanced beyond what could be accomplished by desk-based studies using decades-old, anthropological fieldwork reports. Actual landowner identification was simply elided. PNG law and the possibility of court action were air-brushed out of existence. Arguably ExxonMobil and Oil Search were in breach of their obligation to comply with PNG law. GoPNG was in the embarassing and impossible position of appearing to deny the relevance and applicability of its own statutory law. It was this edifice that came crashing down in the P’nyang Case in 2016.

    The effect of this colossal blunder was to eliminate any possibility that PNG equity might be financed by concessional finance obtained from development banks (IBRD, ADB, etc). These organisations not only offer concessional rates but they also supply loans that are effectively without a country risk component. Instead the risk component of these loans is met by a PNG legislative provision which is sometimes sometimes over-looked – Loans and Assistance (International Agencies Act (1971), S4 (1) and (2). A form of negative pledge.

    In 2005 the PNG equity in the Gas to Australia project was estimated at USD$800 million. By 2007 the PNG equity in the new LNG project was estimated USD$1.8 billion (total project USD$10 billion). By start of LNG production PNG equity share had grown to an estimated USD$3 billion plus. Standardly risk is assessed on an actuarial basis, except for country risk, which is logarithmic while rates are linear. It is doubtful that a country can insure itself and its creditors against country risk except by pledging assets, which brings into play the negative pledge. Accordingly, GoPNG faced many difficulties in meeting its LNG Project equity capital calls. The diversion of SWF funds to KPHL is a consequence.


  • Hi Adam. I think McGillivray was arguing with (or at least dismissing) the conclusions when he states “its (the JA report) findings must be taken with a pinch of salt.” Fortunately, there does seem to be agreement that any new projects should be on better fiscal terms – the report has helped clarify why that is necessary. Since the report, Treasurer Abel has also indicated (on 8 June) “Low Tax Revenue from PNG LNG ‘Being Addressed'”. However, there is less agreement on the need to reverse the “resource curse” policies discussed in the report. There is a wealth of ‘lessons from experience’ from other countries, and PNG’s own history, that could be better tapped into – PNG’s current economic performance indicates mistakes are still being made.

  • I don’t think McGillivray was arguing with your policy conclusions – and neither was Abel supporting your modelling. McGillivray’s critique was of your methodology – which I agree with – and Abel’s comments related to the need to negotiate things differently to create better policy outcomes – something that everyone, including the Prime Minister, agrees with. The latter are lessons you only learn from experience, and those lessons were learnt long before this report was released.

Leave a Comment