Over the past two weeks week the key agency forecasts were published. Last Tuesday, the EIA published STEO, OPEC released its MOMR yesterday and the IEA published its OMR this morning.
Key agency forecasts for this year’s balances vary widely. The EIA forecasts a draw of 0.3 mb/d, while we expect bigger draws of 0.6 mb/d due to our more benign forecasts of non-OPEC supply growth (1.7 mb/d versus 2.3 mb/d for the EIA). The OPEC Secretariat and the IEA don’t forecast stock changes, but if OPEC production were to hold at April levels as published by the IEA, then IEA’s projected balances would show no stock changes for the year (large draws in Q2 19 and Q3 19 being exactly offset by builds in Q1 19 and Q4 19). Using a similar methodology for OPEC Secretariat shows a 0.4 mb/d annual draw. So, in general, the key agencies all show tighter balances m/m but on the proviso that OPEC production holds around current levels.
Overall, the IEA is heavily underestimating historical demand (and revised down its figures this month), while both the EIA and OPEC overestimate non-OPEC supplies. On the supply-side, the EIA’s Q4 18 non-OPEC supply estimate is roughly 1 mb/d too large. This year, we expect growth of 1.7 mb/d while both the EIA and OPEC predict 2.2 mb/d.
The problem with demand largely stems from the IEA getting the recent history wrong, an issue that is compounded across the industry as many other organisations use IEA demand numbers as their baseline. Normally this is fine, but as mistakes creep into the IEA’s reporting then balances based on these numbers fail to capture the reality. The IEA’s Q4 18 demand number underestimating the true course of events by at least 0.7 mb/d—with missed emerging market demand the chief culprit—an error that will likely be at least partially unwound when the IEA makes annual revisions to its historical demand numbers in August. The issue is that if one uses the IEA’s Q4 18 figures as the starting point for 2019 balances, the market feels a lot looser this year even with the draws, contrary to the physical price moves.
Preliminary data for April show commercial OECD stocks rising by 23 mb m/m to 2,872 mb, led by a large build in the US. This would be higher than the five-year average OECD build of 16 mb and would push stocks above the five-year average by 4.6 mb. But this data is subject to change. Indeed, data for March now show stocks falling m/m by 25.8 mb to 2,849 mb, in complete contrast to preliminary estimates of a 5.9 mb build. The draw was far higher than the five-year average draw of 4 mb, pushing stocks to a deficit to the five-year average by 2.2 mb. Crude stocks were 26 mb below the five-year average and products at a 6.3 mb deficit.