Publications: Notes at the Margin

The Data Must Really Be Wrong (May 12, 2015)


Economists at the Federal Reserve found a flaw in our calculations (see the PDF posted here). To be blunt, we screwed up.


As we wrote in the May 11 issue of Notes at the Margin, there is a problem in the data. Oil inventories are not increasing at the rate everyone has projected. (We will add that we have been a consumer of forecasts, not an originator.) Had the projections been correct, we would have seen very steep contango. Furthermore, had the projections been correct, we would have seen tanker rates rise as oil was stored at sea. We have seen neither. Instead, the incentive to store has declined. In addition, firms such as PBF have stated that they have reduced their purchases of Bakken crude. Exports to Canada have declined. Rail movements have decreased.


The strongest signal of diminishing stocks can be found in returns to storage (see Table 4 of the May 11 report). Returns to storage for WTI adjusted for storage costs peaked March 30 and have dropped sharply since. This is a clear indication that inventories are falling, not rising.


This evidence is clear. However, our error has confused the issue. We apologize.