How difficult can it be to trade contracts based on the cash value of wheat, crude oil, unroasted coffee beans and other raw materials? Sure, there are future risks, physical and financial, but they can be usually managed with the help of trading exchanges, standardized data and lightning-fast computer technology. Political news, such as President Trump’s threat to increase U.S. tariffs on Chinese imports from 10 to 25 percent, can drive down the price of equity shares on Wall Street. But the lean hog futures contract fell just three cents a pound, the Chicago exchange’s daily price movement limit on May 6, the day of the threatened tariff increase.
Commodity markets already have their own rules. What business does the government have in regulating agricultural commodity futures markets? IATP went to Kansas City to hear what was on the mind of market participants and the regulators.
IATP attended AgCon 2019, a conference on agricultural futures trading co-organized by the Commodity Futures Trading Commission (CFTC) and Kansas State University, and a meeting of the CFTC’s Agricultural Advisory Committee (AAC). Of the many topics addressed at the April 11-12 meetings, three deserve further discussion: The failure of wheat futures contracts to perform their historical role of serving as price benchmarks for forward contracting; the increase of automated trading, particularly trading by computer algorithms, of agricultural futures contracts; and position limits on the trading of commodity contracts by financial entities with no commercial interests in the selling, buying, transporting or processing of those commodities.
An audio recording of the AAC meeting will be posted on the CFTC website and slides of some of the presentations can be found through the links above. But some of the presentations of most interest to IATP had no slides.
IATP began work on commodity market regulation in November 2008, in the midst of the Wall Street-synchronized boom and bust in commodity prices driven by commodity index traders (CITs). CITs, mostly trading desks in such investment banks as Goldman Sachs and Morgan Stanley, packaged economically unrelated contracts (e.g. crude oil, copper, soybeans and live cattle) into financial investment vehicles sold to large institutional investors, such as pension funds.
Beginning in 2009 and as a member of the Commodity Markets Oversight Coalition, IATP advocated for position limits (i.e., limits on the quantity of futures contracts a financial speculator can hold) on agricultural and non-agricultural commodity contracts and for position limits on CIT investment vehicles. The absence of well-calibrated and enforced position limits, including new limits applied to CITs, contributed to excessive speculation and extreme price volatility that undermined the ability of processors and producers to manage their price risks.
Although the Dodd-Frank Wall Street Reform and Consumer Financial Protection Act of 2010 authorized the CFTC to set new position limits for agricultural and non-agricultural commodities, by 2016, successful Wall Street opposition lead IATP to write “Death of the position limit rule?” The current CFTC Chairman Christopher Giancarlo has not advanced the position limit rule since it was re-proposed, for the fourth time, in December 2016.
One recent instance of the failure of exchanges to implement pre-Dodd-Frank position limits was implied in a CFTC lawsuit against the agribusiness giant Kraft Mondelez for market manipulation of the Chicago Mercantile Exchange’s (CME) soft red winter wheat contract. The settlement of the lawsuit, as reported by the Financial Times, said that terms had been agreed. The CFTC delegates its authority to the CME and other exchanges (Designated Contract Markets) to monitor trader positions and report to the CFTC positions which may indicate market manipulation. The CFTC accused Kraft Mondelez in 2015 of market manipulation for pocketing $5.4 million in illegal profits by buying about $90 million in CME wheat contracts beyond their commercial hedging needs and position limits.
The CFTC did not issue a press release about the lawsuit victory, which awaits a May 28 court hearing to finalize the settlement. The settlement terms will likely include a fine and could include a “bad actor” designation, which would require more record keeping and reporting requirements of Kraft Mondelez. CFTC’s current enforcement policy, announced in September 2017, is for market participants to self-report rule violations and to cooperate with the agency to avoid litigation and more severe penalties. In any event, the CME will not be disciplined for having failed to detect Kraft Mondelez violation of CFTC-delegated CME position accountability rules. (Position accountability is the exchange’s self-regulation of position limits.)
Grain traders at AgCon 2019 debated why the CME wheat futures contract price failed to converge with the cash price towards the expiration date of the futures contract (nearby futures price). Futures prices historically have served as benchmark prices for forward contracting, e.g. by farmers selling local grain elevators of a portion of their crops at an agreed delivery date, quantity and quality.
Professor Mykal Taylor of Kansas State University presented research that she, Randy Fortenbery and Glynn Tonsor had done on “wheat basis,” i.e. the difference between cash prices for wheat at county elevators and the “nearby futures price” for wheat at the contract expiration date on the Kansas City Board of Trade (KCBT). For example, with the KCBT wheat futures price at $5.03 a bushel on July 5, 2018, the wheat basis in Kansas averaged 45 cents a bushel but ranged from 8 cents to 83 cents depending on the county location of the elevator. If the cash price of the physical wheat and the futures price do not converge during the delivery month of the futures contract, the futures price does not serve as a reliable benchmark for forward contracting. If farmers do not see the futures and cash prices converging, they may forgo forward contracting and hope for a better price for the grain they have stored in their own bins, assuming both storage and price risks that forward contracting transfers to the grain elevator.
Data gathered from eighteen Kansas grain elevators from 2008-2012 showed a non-converging average basis risk of 30 cents a bushel, spiking to an average of 83 cents a bushel in 2011, a time of extreme wheat price volatility. The elevators responded to the price convergence failure by doubling the cost of forward contracting. Farmers either pay the high basis risk premium or they avoid forward contracting and are exposed to the risk of receiving lower seasonal cash prices for their wheat.
At the CFTC/KSU conference in 2018, the CME explained price convergence failure as due to fixed storage rates for grain in its contract specifications. By revising the CME wheat contract to include a Variable Storage Rate (VSR) for farmers, the exchange proposed to fix convergence price failure. In 2019, Fred Seamon, the executive director of agricultural markets for the CME, again touted the VSR as a solution to convergence failure, which he characterized as a “rare” occurrence. Professor Taylor said that the VSR built in further cash price uncertainty for the forward contracting farmer. One audience member said that adding a VSR to the terms of the wheat contract was “putting a band-aid on a hemorrhage,” adding that with “tons of money sloshing around the market,” financial speculators “could care less about convergence.”
Mr. Seamon proposed a definition of “cash price” that would not concern the point of sale at the local grain elevator, but the price of the grain shipment sales from local elevators to the terminal elevators owned by such mega-processors and traders as Cargill, ADM and Kraft Mondelez. This redefinition of “cash price” extended to the commodity’s delivery point—no longer to be confined to a specific delivery point, but to any delivery point within a geographic area defined by the futures contract. This flexible definition of “cash price” would structurally narrow the convergence between the cash price and futures contract price, but not to the benefit of the forward contracting farmer or farmer cooperative.
CFTC economist Christa Lachenmayr, not speaking on behalf of the Commission, said that one of the CFTC’s Core Principles requires contract design to enable price convergence. A regulation based on that Core Principle required a “good faith effort” by the exchanges to ensure convergence. However, the CFTC would be reluctant to react too quickly to a specific convergence failure and change a rule, in case market manipulation had caused the failure. The AAC voted on April 11 to study contract design and make recommendations to amend the existing contract design rule, if needed.
Both the AAC and AgCon meetings focused on the automated trading of agricultural contracts, including by High Frequency Traders (HFT). Many HFT firms have relocated to K Street (traditionally the home of lobbyists, not traders) in Washington, D.C., to be electronically milliseconds closer to the electronic release of official government data, including the U.S. Department of Agriculture’s World Agriculture Supply and Demand Estimates (WASDE). Seth Meyer, chair of the board that releases the monthly WASDE, said that he was “agnostic” about whether HFT enabled or disrupted price formation in futures contracts. He said that HFT “was not going anywhere so we have to adjust” to release WASDE without a ninety minute “time out” for agricultural journalists and data reporters to interpret the WASDE data before its release to the public.
Much of the academic research presented at AgCon 2019 about the impact of automated trading concerned Treasury bills, interest rates and other data underlying financial commodity derivatives contracts. Matthew Edelstein said that he started as a manual “point and click” trader in 2003 but found that trading method prone to errors, such as doubling clicking and inadvertently buying or selling more than intended and increasing price volatility. Edelstein said that algorithmic trading took (“scraped”) data packet inputs (e.g. from WASDE) to trigger trading instruction outputs, following an if/then binary logic of computer software fundamentals. “To a computer,” he said, “even a millisecond is an eternity” when trading is transacted near the speed of light. In response to a question from IATP about how computer algorithms processed non-standardized data, Edelstein replied that question was a subject for the next panel, “Unknown Unknowns: Futures Markets Responses to Significant News Events.”
During the next panel, Ed Prosser, who has traded commodities contracts for more than thirty years, pled with the CFTC and the audience to ensure that automated trading, with its lower transaction costs, would not wipe out manual “point and click” traders like him. His baptism by fire was managing price risks in the live cattle markets during the outbreak of Mad Cow disease in the late 1980s, when the extent and causes of the disease were not known to market participants. No amount of automated trading strategies or exchange “kill switches” to stop trading could substitute for the commodity specific knowledge of manual traders, he argued.
A late April research note by Goldman Sachs, perhaps the most influential commodity index trader, indirectly illustrated part of Prosser’s argument. The note announced that Goldman had cancelled ten commodity trading recommendations, covering energy, metals and agricultural price risk management contracts. A Financial Times article about the research note reported that commodities trading was Goldman Sachs most profitable asset class during the first quarter of 2019. Why abandon recommendations made for trading the contracts of your most profitable asset class?
Goldman Sachs was withdrawing its recommendations because of what the firm termed “idiosyncratic risks” in commodities production that impacted CIT prices and the profitability of the asset class for Goldman and its clients. The FT stated, “For example, Goldman noted the near 50 per cent rise in lean hog prices this spring, thanks to an African Swine Fever outbreak spreading to China, was missed by all momentum strategies.” Momentum strategies attempt to manage investor risks by formulaically trading some contracts in anticipation of price increases while trading other contracts in anticipation of price decreases. Momentum investors did not anticipate the price jump in lean hog futures contracts predicated on the expectation that China would be importing U.S. pork to compensate for the loss of pork from its own hogs.
The trading algorithms that incorporate standardized market data to execute momentum strategies do not respond effectively to price and contract transaction volume volatility data generated by such events as the outbreak of African Swine Fever. The algorithms are limited by the binary logic of computer software, which responds to an “if” data input, with a “then” trading instruction. Trading algorithms respond to other algorithms, often without human intervention, except in the algorithm's design and changes to it. The much-criticized herd behavior of algorithmic trading in stock market equities likewise afflicts the commodity derivatives market.
For Goldman and other CITs, abandoning one trading strategy for another because of alleged “idiosyncratic” risks is just another opportunity to profit from new trading recommendations. But for the commodity producers and processors who are hedging price risks critical to their business, managing risks under adverse circumstances of commodity production cannot be accomplished by algorithmic trading in response to standardized data. For regulated commodity traders hedging price risks to compete against unregulated automated trading systems, automated traders must be regulated.
However, more than five years after the CFTC first released a concept paper about the regulation of automated trading (Reg AT), it remains unregulated. Automated trading now accounts for more than 60 percent of all U.S. live cattle futures contract transactions and more than 52 percent of corn futures according to a CFTC staff report, which was discussed at AgCon 2019. (The report does not include data on Over the Counter agricultural contract trading not transacted on regulated exchanges. Inclusion of such data would show yet a higher percentage of automated trading.)
IATP was among the public interest groups that criticized a CFTC-proposed Reg AT in November 2015 as insufficient to prevent market disruptions, particularly by traders (then about 35 percent of the total) unregistered with the agency. The proposal was revised in November 2016. CFTC Chairman Giancarlo advocates industry self-regulation of automated trading and has not permitted further work to finalize a Reg AT.
However, Giancarlo’s term as Chairman ends in 2019. It will be up to the incoming Chairman (Treasury official Heath Tarbert has been nominated but not confirmed by the full Senate) to decide if the market disruptions resulting from herd behaviors in algorithmic trading require re-proposal and finalization of Reg AT. Likewise, the incoming Chairman will have to finalize a position limit rule required by the Dodd-Frank Act and determine also whether a change in the CFTC’s rule on contract design can reduce trends in commodity price convergence failure.