By John Auers and John Mayes
Although the oil business was only in its infancy when Mark Twain coined the famous quote in the title of today’s blog, it certainly came to mind when we were looking at petroleum statistics recently. The foundation of petroleum forecasting is firmly rooted in data. Historical information regarding volumes, prices, differentials and other aspects provides the basis for virtually all future projections. If we are uncertain as to the validity of past or current data, then forecasts become even more tenuous. A similar situation arises in the historical relationships which are related to petroleum. For example, there is generally a good correlation between economic growth and petroleum demand growth, particularly for diesel fuel. But what if the economic projections (and even historical data) vary dramatically? How certain can we be of the petroleum demand forecasts? A dramatic example of the inconsistency with data is the difference in world petroleum demand growth for 2015, which is being reported by different sources. The International Energy Agency (IEA) reports that global demand increased by 1.8 million barrels per day (BPD) last year, while the U.S.’s Energy Information Administration (EIA) has demand at a full 450,000 BPD (25%) lower at 1.35 million BPD. Some other analysts have even shown demand higher than IEA’s numbers, with a recent report by Raymond James estimating that the actual petroleum consumption increased by 2.2 million BPD in 2015! Although not always this dramatic, substantive differences in data reporting exist for other key statistics. So who are we to believe? Why do the numbers differ so much, and what does all this mean for the future? We’ll dig into this as we ponder another famous quote that is applicable (this one by the 19th century British statesman Benjamin Disreali) – “There are three types of lies – lies, damn lies and statistics.”
Probably the biggest factor in global demand uncertainty comes from the difficulty assessing actual and expected demand in China. For the last three decades, Chinese economic growth has led the world, and it has increasingly been the engine supporting robust petroleum demand growth for all of East Asia. This economic boom has had its side effects, however, in producing an over-inflated stock market and a massive level of uncollectible corporate debt. These uncertainties have grown in recent years with many of the shock waves being felt in 2015. To bolster faltering exports, the Chinese government devalued the yuan by 3% in August of 2015, while the stock market experienced dramatic swings in the second half of the year and into January of 2016. In July, more than half of the Chinese companies listed on the Shanghai exchange temporarily ceased trading while all trading was stopped only two weeks ago when a sharp drop triggered the regulatory “circuit breakers.”
These recent events are now creating a great deal of uncertainty in future Chinese economic growth projections. Following a recent peak in 2007 at 14.2%, economic growth has remained strong, but has been gradually falling. By 2014, growth had declined to 7.3% and the 2015 estimate is around 6.9%. Because of the recent turmoil, growth estimates for 2016 vary from around 6.7% to below 6.0%.
Even more uncertain is the future relationship between economic growth and petroleum demand growth for China. Most analyses tie future oil demand growth with economic gains, but one of the largest state-owned petroleum companies, Sinopec, has a completely different conclusion. Sinopec has concluded that Chinese diesel demand will peak in 2017, while gasoline demand will peak in about a decade. Given the size of the Chinese economy, these variations can have a sizable impact on total global petroleum demand.
While uncertainties in developing countries are taken as a given, it is surprising how tentative the data can be even for the U.S. A case in point is the level of U.S. crude production. The EIA issues both weekly and monthly estimates of production rates. The weekly estimates are given for the prior week while the monthly volumes are issued two months later. Technically, neither are actual output volumes but simply the most current estimates of oil production. The problem is that the two sets of numbers do not agree.
The monthly figures indicate that U.S. crude production has flattened during the last five months (Figure 1). In fact, October output (the latest monthly figure currently available by the EIA) was higher than the June level by 32 MBPD. Since June, total U.S. crude production has consistently been above 9.3 million BPD, with July, August, and September above 9.4 million BPD. The EIA has weekly production estimates into January of 2016. These weekly volumes point to a December output level of slightly under 9.2 million BPD.
The weekly estimates present several problems, however. In recent months, the weekly estimates have been as high as 284 MBPD above and as low as 332 MBPD below the eventual monthly figure. This creates considerable confusion in determining current production levels. Will the November and December 2015 rates actually decline to below 9.2 million BPD as the weekly data suggests, or will the monthly volumes continue in the 9.3-9.4 million BPD range? The weekly averages for both November and December are above the weekly October estimates (by about 65 MBPD), which could indicate the final two months of 2015 could be in excess of 9.4 million BPD.
A second area of confusion with the EIA production data is the propensity to revise past output levels. In recent months, headlines have indicated that U.S. production has declined when the monthly data was released. The current month’s data was, in fact, often below the previous month’s output. What is not always reported, however, is that data for previous months is often revised. For 2015, each of the monthly levels for January through September has been revised at least once and some several times (Table 2). In seven of these months, the monthly values were raised (compared to what is currently being reported online by the EIA) with reductions in two months. For the first nine months, the average change has been an increase of 78 MBPD. This average change does not include the additional changes for these months, which are likely to be made in future revisions.
It is interesting to note that both the weekly-to-monthly differences and the revisions to the monthly values are both in the direction of higher output levels. These two factors are part of the explanation as to why output is remaining stubbornly high in spite of the perception that crude production is declining. This topic is discussed more extensively in our 2016 Crude Oil and Refined Products Outlook which is to be issued in early February.