Two weeks ago, we reported that when Goldman observed the latest gasoline demand data, it said that either something must be wrong with the data, or the US is in a recession: as the firm's commodity analyst Damien Courvalin put it, such a steep drop in in US gasoline demand "would require a US recession." He added that "implied demand data points to US gasoline demand in January declining 460 kb/d or 5.2% year-on-year. In the absence of a base effect, such a decline has only occurred in four periods since 1960 during which time PCE contracted."
Bloomberg's Liam Denning confirms that "big dips in U.S. gasoline demand, especially of 5 percent or more, are almost unheard of outside of a recession or oil crisis." Goldman then adds that "to achieve the 5.9% decline suggested by the weekly data, our model requires PCE to contract 6%, in other words, a recession."
At this point Goldman - which naturally was aghast at the possibility that there is an under the radar consumer recession taking place at a time when the bank was predicting three rate hikes - quickly pivoted and explained that a far more likely explanation is that the latest weekly report was an aberration, and that there was simply something wrong with the data.
Our analysis identifies weekly yield and exports as systematically deviating from their final values and such biases suggest that demand could be revised higher by 190 kb/d. The EIA's real-time export data still includes estimates and we see potential for the recent shifts in the Mexican gasoline market to exacerbate the overstatement of US exports by an additional 185 kb/d given (1) lower PEMEX refinery turnarounds, and seasonally lower demand exacerbated by the January 16% hike in prices. Adjusting for these lower exports points to US gasoline demand declining only 85 kb/d yoy in January, in line with our macro model.
After chosing to ignore the data, Goldman then reiterated its fudged, rosy outlook based on its own fudged data:
Looking forward, we reiterate our outlook for strong global demand growth in 2017 and view the recent US gasoline builds as reflective of transient regional shifts in gasoline supply instead. Given our outlook for strong consumer spending in 2017, we believe that US gasoline demand growth will remain resilient this year at 60 kb/d, albeit below last year's 150 kb/d growth because of higher prices. From a global perspective, these declines remain modest, especially compared to the 510 kb/d 2016 demand growth from the 40 countries we track.
The problem with this narrative is that with every passing week that the DOE fails to correct the "error" in its residual calculation, assuming there is one, the less credibility the "non-recession" thesis has.
Which is why among the numbers released in today's weekly update by the DOE, the one most anticipated was the one showing gasoline demand: what it revealed was troubling, with gasoline demand in the past 4 weeks sliding some 5.2% compared to last week, the much anticipated rebound, or data revision, remains elusive.
It also suggests that the worst-case scenario may be the right one: something is wrong with US consumer spending.
While one can debate if the gasoline demand data is accurate, there are other troubling indicators: as Bloomberg points out, the recent pace of recovery in US miles driven is slowing sharply.
"Vehicle-miles traveled increased by just 0.5%, year over year, in December. For much of the prior two years, roughly coinciding with the crash in oil prices, growth had averaged more than 3 percent, sometimes spiking above 5 percent. In addition, the slowdown is fairly broad in geographic terms. Of the five regions into which the Federal Highway Administration divides the U.S., three showed outright year-over-year declines in December, the first time that's happened since March 2014.
Those three regions represent almost 60 percent of the entire country's vehicle-miles traveled; again, it's the first time that much of country has experienced a decline in about three years:"
Among the other notable observations on American driving habits, namely that "the number of miles driven by the average American peaked more than a decade ago (see this blog post by Harry Benham of Carbury Consulting).
In addition to the recent economic improvement on the back of trillions in central bank liquidity, one can also point to the gasoline price base effect: today's average of about $2.22 per gallon is far below the levels of $3.60 or more paid back in the summer of 2014. But the rebound over the past year, highly visible on gas-station placards along every highway, must have registered with at least some drivers:"
Whatever the reasons, Bloomberg concludes that "the size of the change is acute, and does suggest that a shorter-term action may be the cause. However, there may be a combination of near-term price sensitivity intertwined with longer-range shifts in vehicle technology and driving patterns. The next few weeks will reveal more."
Alternatively, it may also indicate that the entire narrative about the "recovery" is quietly unraveling on America's roads.
Finally, as we have been pounding the table since last year, no matter what is the driver, so to say, behind the decline in gasoline demand, it will have broad implications on both the equilibrium price of crude oil, and supply: and as Bloomberg similarly concludes, any impact on US gasoline demand this year, perhaps causing trend growth of 100-200,000 b/d to transform into a reduction of say 250,000 b/d, would force the oil market into having to find that net 350-450,000 b/d demand from elsewhere.
"And that seems unlikely given the trends in Europe and China toward more conservative fuel policies, and non-gasoline technologies."
Which means that all those hedge funds who currently have record long WTI positions and who have been ignoring the troubling fundamental data from America's roads, will sooner or later be forced to close out, resulting in one of the biggest oil price drops in history, in the process perhaps unleashing the next deflationary shock.