Counting barrels is always tough to do, as Ross McCracken discusses in this month’s excerpt from Platts Energy Economist.
How much oil does the world consume? You’d think this would be a fairly straightforward question, given its economic importance to the world economy, and indeed answers are not hard to find. The problem is those answers differ significantly.
Even with the development of the Joint Organisations Data Initiative–an evolving beast designed to bring improved transparency to oil markets–oil market data remains messy, inaccurate and opaque.
For 2013, a year which by now should be transitioning from estimate to a matter of historical record, OPEC puts world demand at 90.00 million b/d, the International Energy Agency at 91.40 million b/d and the US Energy Information Administration at 90.49 million b/d. The difference is large in absolute terms — 1.4 million b/d between the IEA and OPEC — but small if viewed in percentage terms, about 1.5% of the total market.
The differences, according to OPEC, arise from the agencies’ use of “different historical baselines.” The organization is unwilling to enlarge on this explanation, nor explain why there has been no convergence of baselines over time, but in effect it means that the agencies do not count barrels supplied and consumed, but changes in oil balances, focusing on oil inventories. For its part, the IEA says simply that it has better numbers and that a large part of the difference arises from demand estimates for South East and East Asia.
The developing world is indeed where the problem lies on the demand side. For 2013, there is consistency between the three agencies for demand data in the FSU, OECD and non-OECD Europe. Almost the entire difference comes in the estimates for developing countries.
Blog entry continues below…
|Request a complimentary issue of: Energy Economist|
|Every month Energy Economist combines incisive judgments and detailed data sets that deliver ideas you can profit by. It includes coverage of:
- Political and economic risk
- Supply and demand
- Policy and regulation
- Liberalization and market design
- Supply security
Reliable inventory data outside the OECD is often absent. Each agency develops its own methodology and makes its own judgements. This is worrying because it is the non-OECD that currently provides almost all demand growth globally. The data is worst where it is needed most.
However, all three agencies saw oil supply and demand broadly balance in 2013, which means there must be differences in the supply figures equivalent to those on the demand side. In other words, the IEA has to find from somewhere 1.6 million b/d more oil supply than OPEC and 1.4 million b/d more than the EIA. And it does — mainly in FSU and OPEC production.
OPEC has always had a difficult relationship with data. This is most manifest in its publication of one set of official output figures, derived from ‘direct communication’ with its members, and its own use of those collected by secondary sources. Both are published, but it is only the secondary source numbers that have any credibility.
JODI has not improved the situation, taking its data from OPEC for OPEC members and using the official numbers. As a result, JODI boosts both Iranian and Venezuelan production figures by about 500,000 b/d for April 2014 compared to the secondary source estimates, despite the OPEC secretariat itself not trusting these numbers. As of mid-July, the JODI database did not have complete supply data for OPEC members even for April.
However, there’s no real agreement on the secondary source numbers either. Again in percentage terms this difference may be relatively small, but these differences have a big impact on the perceived supply/demand balance. This does matter because as the absolute numbers mean little, it is the relative change in those numbers that counts.
For example, OPEC predicts demand growth in 2014 of 1.1 million b/d, the IEA 1.3 million b/d and the EIA 1.13 million b/d. The more than 1 million b/d difference in the absolute numbers disappears and a more meaningful picture emerges.
But how good are even these? Forecasts for 2014 made in July of 2014 might be expected to be fairly robust. However, while they do incorporate some harder data, they are still open both to significant revision and future events. According to the IEA, actual data for the OECD is generally two months behind, so the forecast made in July relies on data available in June, with ‘hard’ data available up to April. But even this hard data is open to revision as it becomes more complete. The non-OECD data is much further behind, with many non-OECD national statistics agencies not even producing monthly data.
2014 demand and supply data will remain an estimate rather than a matter of historical record well into 2015. Late changes as data becomes more complete can materially affect the expected supply/demand estimates for the forthcoming year.
With all these uncertainties, it is no surprise that the forecasts made by the agencies are often wrong. The time lag in receiving complete data means they are slow to adjust to fast-changing events. Based on an average of 2007-2013, the oil demand forecasts made by the three agencies in July of each year for the year ahead were on average out by more than 1.5 million b/d. The supply/demand balances were out on average by around 600-700,000 b/d.
But while forecasts will always be inherently uncertain and difficult to produce, the fact that current year data remains a forecast well beyond the point when it should be much firmer, and that there is no agreement about the oil balances of developing countries on the one hand and the supply provided by major producing nations on the other, should be of concern.
Even OECD data collection has not been given the priority it deserves in recent years, instead falling victim to the budget cuts that followed the financial crisis and ensuing recession. Budgets in some cases have been restored and the IEA continues to foster relations with national statistical offices in developing countries.
The lack of institutional capacity is not the forecasters’ fault. That the resources are not provided by governments reflects either a lack of political will or opposition to greater transparency. There remains a big gap between governments protestations that they want greater transparency in oil markets and their actual willingness to deliver it.