San Francisco Fed researchers linked bitcoin’s reversal to futures

The Federal Reserve Bank of San Francisco on May 7, 2018 published an Economic Letter arguing that the arrival of regulated bitcoin futures helped turn a one-sided speculative market into one in which pessimists could express a negative view. The researchers said bitcoin’s late-2017 peak and subsequent decline were consistent with a familiar futures-market mechanism, while stopping short of a formal causal estimate.

That distinction mattered. Bitcoin had already fallen far below its December record, and the paper offered an institutional explanation for why the reversal began when it did: derivatives changed who could participate in price discovery. It did not establish that futures were the sole cause of the decline.

The researchers’ case

The letter, written by Galina Hale, Arvind Krishnamurthy, Marianna Kudlyak and Patrick Shultz, used Bloomberg prices and the authors’ calculations. It identified a bitcoin peak of $19,511 on December 17, 2017, the date CME bitcoin futures began trading. Cboe had launched its contract on December 10, but the authors said trading remained thin until CME entered.

Before exchange-listed futures, buying bitcoin was an easy way to express optimism, while taking a negative position was substantially harder. The authors’ mechanism was that cash-settled futures let pessimists sell exposure, diverted some demand from the spot market and introduced downward information into prices. They compared that sequence with research on financial innovation and the U.S. housing boom and bust.

The comparison was an interpretation, not proof that the same forces operated with the same strength. The letter did not present a counterfactual model of what bitcoin would have done without futures, isolate other news, or estimate a causal coefficient.

What the market measurements showed

For its principal price comparison, the letter indexed three 2017 bitcoin peaks to 100 and followed the days around each peak. Bloomberg supplied the prices; the authors calculated the indexed paths. The decline after December 17 was larger than the reversals following the June 6 and September 1 peaks, and bitcoin had not regained its pre-futures high by late April 2018.

The volume comparison covered the first week and first month of the new derivatives market. The authors estimated that CME contracts represented about 12,000 bitcoins of average trading volume during the first week, against estimated spot turnover of 200,000 bitcoins. They also said average daily volume in the month after CME’s launch was approximately six times the volume observed when only Cboe offered futures.

Those figures are not event-day measurements for May 7. They combine contract-equivalent quantities and an estimated spot-market total in a fragmented, continuously traded market. They should be read as scale indicators within the paper’s methodology, not as a complete census of global bitcoin activity.

Why the conclusion was consequential

The publication moved the futures debate beyond a simple chronology. The CFTC’s December 1, 2017 record confirms that CME and Cboe self-certified their contracts and that the agency warned the underlying cash markets were largely unregulated, volatile and vulnerable to problematic trading practices. Self-certification was not CFTC approval or endorsement.

By May 7, the question was no longer only whether regulated derivatives could coexist with bitcoin. It was whether those instruments had changed the spot market’s price-discovery process. The San Francisco Fed researchers’ answer was qualified but clear: the observed rise and reversal were consistent with optimists bidding up an asset before a sufficiently deep shorting mechanism arrived, followed by pressure from pessimists once futures became available.

The surviving evidence supports describing the letter as a consequential market interpretation, not a verdict. CoinDesk’s contemporaneous report accurately framed the finding as futures having played a role. No synchronized May 7 closing-price reaction is asserted here because the letter did not supply one, crypto trades continuously, and the publication itself was analysis of the December-to-April period rather than a new market transaction.