In the intense and public discussion of so-called algo trading over the past few years, stocks have always been in the foreground.
On the other hand, less attention was paid to the fact that during this time automated trading systems, as it were unobserved, took over a large part of the pricing and trading in bonds of better quality (investment grade), i.e. the lion's share of the volume of government and corporate bonds.
Until recently, a spread to other parts of the market prevented its lower liquidity and thus a certain degree of randomness in price developments.
In the meantime, however, these trading robots are arguably so mature that they are increasingly penetrating into less liquid parts of the bond market. In the first half of 2021, Bank of America handled an international trading volume 2.7 times the same period last year, says Sonali Theisen, who heads the e-trading and market structure for fixed-income securities at the bank. A large part of the recent growth comes from trading in low-quality bonds (high-yield bonds) and from emerging countries, i.e. from areas that are considered to be rather less liquid. Several thousand different bonds from the USA, Europe, the Middle East and Africa are now automatically traded at JP Morgan Chase. At the beginning of 2020 there were only around 250.At the British bank Barclays, algorithms are now pricing more than half of the rate requests internationally.
Covid crisis has helped developers
It was above all the upheavals during the pandemic that contributed to the fact that the algorithms could be refined accordingly.
"The Covid crisis has provided the industry with a whole range of new data on how the markets react to a crisis," says Kevin McPartland, head of research for market structure at the financial data analyst Coalition Greenwich, which is part of Standard & Poor's, which above all is known for its rating agency.
This new data was of great use to the programmers.
Large market fluctuations enabled developers to better understand how the prices of securities behave in difficult market conditions.
On the one hand, automation can handle a larger volume of business.
Second, human traders can focus on more complicated transactions.
"When you stop staring at a screen of inquiries and just typing in numbers, you have more time to do the things that add more value," says Drew Mogavero, deputy head of credit products at Barclays.
Most algo transactions are responses to automated price queries from electronic trading platforms that are experiencing immense growth, especially in the high-yield bond sector.
The proportion of automated trading in American high-yield bonds at Tradeweb Markets was 38 percent in the third quarter - three years ago it was just 5 percent.
Smaller inquiries simply ignored
The growth in bond markets is one reason robots are needed to assist.
The total value of bonds in the Bloomberg US Investment Grade Corporate Bond Index alone has risen from $ 4.3 trillion in 2015 to its current level of $ 6.2 trillion.
Thanks to the improved technology, traders can now answer a much higher number of price inquiries.
Before the Algo development, inquiries from smaller traders or unusual volumes were often simply ignored.
That affected up to three quarters of the electronic rate inquiries, according to the trading room of an unnamed major bank.
In the meantime, this only applies to a quarter.