This is the third of a four part series covering the corruption of PNG’s economic statistics. The government has failed badly in trying to explain why they were using false nominal GDP numbers. They blame a change in a “Price Index”. But this excuse is farcical because you don’t even use a price index in determining nominal GDP. A price index is used in measuring real GDP – but even then the NSO/ABS/IMF new price index was accepted for all the years 2007 to 2014. The price index was only questioned when the facts didn’t suit the economic narrative of the O’Neill/Abel government.
This article sets out in full the government’s attempted explanation for the mismatch in the PNG’s GDP numbers – I didn’t want to be accused of not fully representing the government’s excuse. Assessments are then made of the key arguments. This makes for a long article but some detail is required to justify why it is the final straw for judging that PNG’s economic statistics unquestionably have been corrupted.
The 2019 Budget included a very interesting box on the GDP figures (Bottom of Box 2, 2019 Budget Volume 1 pp 27-28). The following paragraphs provide the full government explanation in italics followed by a response.
“The National Statistical Office (NSO) is the agency responsible for the collection and compilation of statistics including economic data and the compilation of PNG’s National Accounts. The office is currently in the midst of a reform period (2015-2019) with assistance from the Australian Bureau of Statistics (ABS) providing the Technical Assistance following NEC Dec 162/2014 for the overhaul of NSO. Several milestones have been achieved from this reform program, including the timely release and update of the National Accounts 2007-2014 in 2017 which updated the outcomes of GDP for that period which had remained outdated for more than a decade. While this demonstrated much of the progress emanating from the reform activities, more needs to be done before NSO can fully function on its own. In this regard, Treasury and the BPNG have been continuously supportive of the reform process by providing timely assistance in terms of data validation based on economic data provided by NSO through constant collaborations since the start of the reform process.”
This background information includes a dangerous half-truth. The latest IMF Article IV report released in December 2018 indicates that the ABS is no longer providing technical support: “however, the ABS recently suspended its program due to increasingly uncertain NSO leadership and management, which threaten its current operations” (IMF 2018 Article IV p10). The ABS has walked away.
According to the 2019 Budget, what now is the problem with the NSO update?
“In 2018, NSO, Department of Treasury and BPNG, continued this process during the preparation of the 2015 National Accounts. During this stage, Treasury and BPNG observed a major change or shift in the methodology of compiling the national accounts. This included the use of a Composite Price Index as the new deflator replacing the Consumer Price Index (CPI). This caused a significant variation in the 2015 National Accounts prepared by the NSO compared with the official estimates produced by the Treasury (see below):
Because the change in methodology was something new and since it caused some anomalies in the accounts, the Treasury and BPNG advised NSO to refrain from publishing the 2015 National Accounts until it was comprehensively understood and all associated anomalies from it were reconciled. Unfortunately, NSO proceeded to publish the 2015 National Accounts during the year without addressing the outstanding issues.
This is an absolutely extraordinary explanation.
First, yes, the table above does show the difference of K5,021 million between the “NSO Composite Price Index” and the “Treasury Consumer Price Index”. But a price index has no impact on nominal Gross Domestic Product – it is a total non-explanation dressed up in technical jargon. In basic economics, one learns the clear difference between “nominal” and “real” GDP. Nominal GDP does not take into account price changes through time while real GDP does. So a difference between price indexes (composite vs consumer) can affect measures of real GDP, and especially real growth rates. However, by definition, they have no impact on nominal GDP figures. The K5,021 billion gap is a nominal GDP difference. Price indexes do not change nominal GDP calculations. This attempted explanation is farcical.
Second, any concerns about the price indexes did not apply to the NSO 2017 release of updated information. The complete set of those tables, from 2007 to 2014, were immediately incorporated into Treasury’s GDP tables (although with some technical glitches, but there was no discussion of problems with any of the new methodology). So there was quick action in accepting the first eight years of updated information (2007 to 2014). The issue only occurred for the ninth year (2015) and probably 2016.
Third, going beyond the vital but simple technical issue that nominal GDP is not affected by price indexes, could there be a significant difference between a composite and a consumer price index? Short answer is that the PNG Treasury doesn’t actually use a “consumer price index”, and the NSO change is unlikely to have a significant impact on the deflators used in PNG’s national accounts. More specifically, PNG Treasury had been using for many years several deflators depending on the sector of the economy (which is what should happen). So the mining sector had its own price index to reflect changes in international minerals prices as did the petroleum and gas sector. The agriculture sector also had its own price index to reflect changes in international agriculture prices such as coffee and oil palm and forestry. All three of these price indexes are dominated by changes in international prices, not changes in the consumer price index. As these sectors together accounted for over 40% of GDP, it is difficult to describe the Treasury index as just a “consumer price index”. It clearly is a composite index of the CPI and international price movements. There is not much information on the other “composite price index” produced by the NSO. The original NSO update talked of using “constant purchaser prices”. Unfortunately, there is no public information on this index including on the NSO website. However, as mentioned above, this NSO index has been used by the PNG government for all of its GDP figures over the previous eight years and is the basis for the PNG Treasury numbers in the 2019 Budget from 2008 to 2014 – the deflators for each of the sub-sectors in Table 1 of Appendix 3 of Volume 1 of the 2019 budget are identical to those of Table 15 in the NSO publication PNG National Accounts 2007 to 2014. And as stated at the beginning, the key issue is that these price indexes do not affect nominal GDP anyway. They could affect real GDP growth calculations – but generally only by a small amount. The baby was thrown out with the bathwater in a very convenient outcome for the government.
The final element of the official explanation is below.
“In view of this, Treasury and the BPNG, including the NSO, established a working committee in September to evaluate the 2015 National Accounts and address all the anomalies before releasing the final accounts again. This work is in progress and will also assist to establish the foundations for the publication of the 2016 National Accounts and the forward years’ national accounts projections.
It seems extraordinary that it took six months to decide to form a committee to try and resolve this massive difference in GDP estimates (9 March 2019 for the initial NSO release and September before the committee was formed). The PNG Opposition was querying the GDP figures by this time, along with the UPNG/ANU economics update (see here and here), so presumably there was some demand for action.
Sir Humphrey Appleby from “Yes Minister” fame would be pleased that the resolution to this political issue was to call for a committee!
This article goes through in considerable detail the full government’s explanation of the GDP figures mis-match. However, as assessed above, this attempted explanation fails miserably. Indeed, it is so filled with half-truths and blatant errors that it is the final straw in the formation of the view that PNG’s economic statistics have been corrupted. It is one thing to make an extraordinary technical error in the employment figures in the 2019 Budget (see here). But there is such a clear pattern of convenient erroneous numbers that one can only conclude PNG’s statistics are now being manipulated for political purposes. This pattern of corruption is likely to continue to slowly seep through other statistics, especially as there do not seem to be any good governance checks on what is published. My next article will explore in greater detail the possible motives for the corruption of PNG’s economic statistics.