Our Q4 18 balances have significantly changed from last month and now show liquids builds of 0.4 mb/d compared with a 0.2 mb/d draw last month. The downward revision was driven by demand revisions in Q4 18 which amounted to 0.19 mb/d and that is with some pessimistic forecasts for final November 2018 and December 2018 US demand built in. Our crude balances now show a massive 1.5 mb/d build across Q4 18 compared to our initial estimates of a 0.9 mb/d draw set out after the Iranian sanctions. Crude builds were driven by a shocking decline in refinery runs, which globally fell y/y by 0.44 mb/d. Stocks rose by 2.7 mb/d y/y in October 2018 followed by 2 mb/d in November 2018 (we had expected draws of over 1.5 mb/d), while December 2018 crude draws were paltry at just 0.1 mb/d (again, here we had expected draws of around 0.7 mb/d).
A lot of post-mortem has been done on Q4 18 balances and why the huge draws predicted—particularly for crude—failed to materialise. While the easy answers have always been Iranian waivers and the outperformance of US production, the numbers never quite added up for us—until now. After all, Iranian exports were likely only 0.3 mb/d higher than what we had expected (at 1 mb/d vs 0.7 mb/d) and US tight oil production was only revised up by around 0.2 mb/d for Q4 18. We had mostly expected the surge in OPEC production in November 2018 and so, the variance with actuals was not significant. As it turns out, it was demand that fell out of bed, specifically refinery runs, which led to large crude stockbuilds. We had noted that end-user demand had started to weaken, but Chinese, Korean and European demand came in weaker than we had expected. Even so, the downward revision to end-user demand in Q4 18 only amounted to 0.5 mb/d vs predictions made in Q2 18.
Instead, it was global refinery runs that fell by a shocking 1.2 mb/d y/y in November 2018. Preliminary indications for December 2018 peg runs at flat y/y, taking Q4 18 global runs lower y/y by 0.44 mb/d, the first decline since Q4 13. This compares to our expectations of at least a 0.5 mb/d y/y increase in runs, even after accounting for turnarounds. What is more perplexing in this is the fact that refining margins had improved in November 2018. Margins weakened in December 2018, but they were still $1-2 above run cut levels last seen in 2014. Importantly, forward margins in September 2018 and October 2018, when the crude buying for November 2018 and December 2018 would have occurred, was still comfortably above run cut levels.
November 2018’s run cuts were likely driven by a lack of crude availability as refiners sought to replace all Iranian exports, fearing no waivers from the US government, and the surge in Saudi, UAE, Iraq and Russian production was not able to make up for zero Iranian exports, which is what the refiners were buying to replace. By December 2018, end-user demand had started to fall sharply in Asia, exaggerated by destocking in China, which cascaded through the entire region. Gasoline cracks were putrid through most of Q4 18 and petrochemical margins collapsed in December 2018 too. Moreover, end-user demand in Europe slowed too, with the macro backdrop deteriorating in Germany and strikes in France pushing demand sharply lower.
Going forward, we expect builds in Q1 19 driven by hefty turnarounds. Crude stocks are set to rise by 0.6 mb/d, offset partially by product draws, although the US government shutdown and the Mexican standoff creates downside risk to our demand estimates for this quarter and can lead to only moderate product draws. Crude draws commence from late Q2 19 as refineries return from maintenance and intensify through Q3 19 as runs rise, driven by new refinery ramp-ups and the need to build diesel stocks ahead of IMO 2020. So, product stocks will build, especially as the more refineries run to make diesel, the more gasoline production will also rise given the lightening of the global crude slate. Overall, once we look past the expected softness in crude balances during the turnaround period, H2 19 balances look constructive.
Our crude-only balances peg H2 19 draws at 1.2 mb/d, concentrated in Q3 19, while Q4 19 sees far shallower draws on the back of sharply higher US production. Indeed, we maintain our US crude supply growth forecast of 1.2 mb/d y/y as we believe that commitments on new Permian pipes will ensure production growth remains high. Our current balances do not assume that the OPEC deal will be extended beyond June, although we do not expect production to go to the November 2018 surge levels on a sustained basis either. Even after accounting for further declines in Iranian production as we expect waivers to be reduced by another 50% in May, we currently expect OPEC production to be 0.4 mb/d higher in H2 19 vs H1 19, with Saudi output around 0.3 mb/d higher in H2 19. So our balances, by no means based on overly aggressive bullish estimates, will still register large draws driven by a sharp increase in runs, barring a complete collapse in the global economy.