Gold could be the next benchmark to lose its lustre. British and German probes into the market may not have yet found any evidence that the price of the precious metal is being manipulated, but the twice daily gold "fixings" deserve the same scrutiny as other financial benchmarks that have come under fire. They are archaic, laborious and methodologically weak.
Fixings establish a reference price for gold which is used by investors, banks, miners, industry and the jewellery trade. They are conducted via private conference calls at 1030 GMT and 1500 GMT held by five banks: Barclays, Deutsche Bank, HSBC, Scotiabank and Societe Generale. While several other institutions trade gold, their activity does not influence the gold fixings unless they make contact with the five setting banks. There are no published records, let alone transcripts, of these conference calls. Nor is there any independent oversight, meaning the fixings are self-policed.
Elemental impurities
The lack of automation makes the gold fix heavy going. At the beginning of the fixing conference call, a senior trader at each of the five banks declares how much gold they can buy or sell at a starting price chosen by the firm chairing the meeting. If demand outstrips supply, or vice versa, by more than 50 gold bars (about 620 kg) then the price is moved until that rough equilibrium is reached. During the fixing process, gold continues to trade, both for immediate and future delivery.
This process does not restrict influence of the fix to the banks in the group, outsiders can have some sway: the final price is derived from orders for physical gold rather than completed trades or even estimates of current market price. Yet, neither the setting banks nor their clients have any obligation to honour those orders. So, a seller (or buyer) involved in the price-setting process could potentially put in a large, false buy (or sell) order to artificially move the market up (or down) temporarily.
Metal detectives
A further flaw lies in the time it takes to set the gold benchmark - typically between four and 15 minutes, but ultimately as long as it takes. All the while, information from the call seems to seep out, according to research by academics Andrew Caminschi and Richard Heaney. They found that volatility and trading volumes in exchange-traded gold fund contracts and gold futures increase during the call. They also suggest that the first market movement after the process starts is a solid indicator of where the benchmark will end up.
Other analysis has also turned up unusual trading patterns. Research consultancy Fideres looked at gold benchmarks between January 2010 and December 2013 and found several instances of sharp movements in the price of the metal during the conference call, which subsequently corrected once the benchmark was set. And, other studies have pointed out that for certain time periods the afternoon fixing price has been statistically lower than the morning equivalent.
These anomalies may have legitimate causes. The lower prices discernible at the afternoon fixes could be a product of heavy central bank selling in recent years - the afternoon fix is used more commonly because the US market is open, making it more liquid. Spikes in volumes and prices, and even subsequent corrections, could simply be indicative of traders betting on or hedging against the final benchmark price, and then being forced to close out that trade once the price is fixed. At the very least, though, trading during gold benchmark-setting looks atypical.
Gold-plating the fix
A US lawsuit filed on March 3 accuses the five gold-fixing banks of manipulating prices over the last 10 years. Deutsche Bank said the case was without merit; Societe Generale said the claims were unsubstantiated; Scotiabank has said it will defend itself; Barclays and HSBC have declined to comment.
But Scotiabank Chief Executive Brian Porter has acknowledged the gold-fixing process needs updating. The International Organisation of Securities Commissions' "global principles" would be a good start: they require the involvement of a third party to provide independent oversight for the process. Widening the fixing to include all 11 of the London Bullion Market Association's market-markers would lessen perceptions of a secret club. Automating orders would provide more accountability and deter the inputting of spoof orders.
The temptation would be to wait for investigations to conclude and then see what reforms need enacting. But there is already enough of a concern to justify modernising the gold benchmark sooner rather than later.
Fixings establish a reference price for gold which is used by investors, banks, miners, industry and the jewellery trade. They are conducted via private conference calls at 1030 GMT and 1500 GMT held by five banks: Barclays, Deutsche Bank, HSBC, Scotiabank and Societe Generale. While several other institutions trade gold, their activity does not influence the gold fixings unless they make contact with the five setting banks. There are no published records, let alone transcripts, of these conference calls. Nor is there any independent oversight, meaning the fixings are self-policed.
Elemental impurities
The lack of automation makes the gold fix heavy going. At the beginning of the fixing conference call, a senior trader at each of the five banks declares how much gold they can buy or sell at a starting price chosen by the firm chairing the meeting. If demand outstrips supply, or vice versa, by more than 50 gold bars (about 620 kg) then the price is moved until that rough equilibrium is reached. During the fixing process, gold continues to trade, both for immediate and future delivery.
Also Read
This process does not restrict influence of the fix to the banks in the group, outsiders can have some sway: the final price is derived from orders for physical gold rather than completed trades or even estimates of current market price. Yet, neither the setting banks nor their clients have any obligation to honour those orders. So, a seller (or buyer) involved in the price-setting process could potentially put in a large, false buy (or sell) order to artificially move the market up (or down) temporarily.
Metal detectives
A further flaw lies in the time it takes to set the gold benchmark - typically between four and 15 minutes, but ultimately as long as it takes. All the while, information from the call seems to seep out, according to research by academics Andrew Caminschi and Richard Heaney. They found that volatility and trading volumes in exchange-traded gold fund contracts and gold futures increase during the call. They also suggest that the first market movement after the process starts is a solid indicator of where the benchmark will end up.
Other analysis has also turned up unusual trading patterns. Research consultancy Fideres looked at gold benchmarks between January 2010 and December 2013 and found several instances of sharp movements in the price of the metal during the conference call, which subsequently corrected once the benchmark was set. And, other studies have pointed out that for certain time periods the afternoon fixing price has been statistically lower than the morning equivalent.
These anomalies may have legitimate causes. The lower prices discernible at the afternoon fixes could be a product of heavy central bank selling in recent years - the afternoon fix is used more commonly because the US market is open, making it more liquid. Spikes in volumes and prices, and even subsequent corrections, could simply be indicative of traders betting on or hedging against the final benchmark price, and then being forced to close out that trade once the price is fixed. At the very least, though, trading during gold benchmark-setting looks atypical.
Gold-plating the fix
A US lawsuit filed on March 3 accuses the five gold-fixing banks of manipulating prices over the last 10 years. Deutsche Bank said the case was without merit; Societe Generale said the claims were unsubstantiated; Scotiabank has said it will defend itself; Barclays and HSBC have declined to comment.
But Scotiabank Chief Executive Brian Porter has acknowledged the gold-fixing process needs updating. The International Organisation of Securities Commissions' "global principles" would be a good start: they require the involvement of a third party to provide independent oversight for the process. Widening the fixing to include all 11 of the London Bullion Market Association's market-markers would lessen perceptions of a secret club. Automating orders would provide more accountability and deter the inputting of spoof orders.
The temptation would be to wait for investigations to conclude and then see what reforms need enacting. But there is already enough of a concern to justify modernising the gold benchmark sooner rather than later.