A trading interruption occurred [intentionally] on Friday, June 24th, for New York-based Betterment, a robo-advisory firms with $5 billion in Assets under Management (AuM), as the firm decided to override its robots in an effort to protect clients the day after the UK’s Brexit vote.

During that Friday’s historic trading, when financial markets woke up to news of the UK’s EU referendum result, a global sell-off rippled across Asia and then into Europe before making its way to the U.S. when markets opened and later closed in the red.

Join the iFX EXPO Asia and discover your gateway to the Asian Markets

Yet, financial markets had operated efficiently, even as circuit breakers were triggered at some venues, there was no fall-out or firms going out of business, even as record volatility and price moves were reached for various securities.

Not so robo

Robo-advisors are expected to largely automate their human counterparts, advising clients on how to invest in financial markets, via the use of a programmatic rules-based approach – although the industry for robo-advisors is still relatively new and emerging in finance.

Often, human discretion is applied over the way that the robos are managed and configured, as the trading robots are occasionally steered and tweaked by people – which is a not-so robo approach.

Advanced technologies, driven by big-data and machine learning, have been employed in algorithmic trading for many years already, including in high-frequency trading (HFT) and by proprietary quantitative trading firms. However they are only starting to slowly make their way to the retail side.

Despite these existing solutions, the available retail robo-advisor products that are offered to the masses are still a far cry from being fully autonomous. Financial advisors fueled by artificial intelligence (AI) remains something to come in the perhaps not-too-distant future.

Progress is nonetheless being made as emerging challenges across both segments – including flash-crash style events for institutional firms – help set the basis for future working points, and as abnormal volatility from last week’s Brexit vote has apparently shown issues for robo advisors.

Unexpected halt

While many brokerages around the world gave advanced notice to changes made in their trading conditions, including a reduction in leverage, increased margin rates, and even some brokers enabling a close-only mode for trades, Betterment apparently made its decision with less advanced notice to clients.

The nature of Betterment’s business model as an RIA, robos aside, is focused on long-term investing, and the firm already prevents clients from trading during the first 30 minutes after the markets open each day, on purpose.

So while the decision for Betterment to halt trading as the market overreacted last Friday appears prudent, the lack of advanced notice may have caused undesirable client feedback as some customers were unexpectedly unable to trade during the brief ban. In response to feedback, the company leaned on its customer agreement noting that its actions were in line with relevant clauses.

Best execution obligations

Coverage by the Wall Street Journal noted that Betterment’s CEO Jon Stein stood by his company’s decision on the temporary halt yet acknowledged the need to communicate better with customers. Mr. Stein was quoted in an interview saying: “We are talking about different notifications we can give either in the – mobile or web – app or the moment of trading via email. There are lots of different things we’re looking at.”

Betterment has its own broker-dealer and apparently had concerns over its best-execution obligations when implementing the brief trading ban –  including how thin liquidity or wide spreads could have caused unfavorable prices for an otherwise longer-term client base that isn’t concerned with intraday timing when making longer-term investment decisions.

A securities expert, Barry Bernstein shared comments with Finance Magnates regarding the temporary trading halt: “There are circuit breakers that could halt trading at the exchange level industry-wide in an effort to prevent panic and build investor confidence but having a robo make the decision for you based on a subjective view seems somewhat reactive and can create unintended consequences.”

“The ability to turn assets into cash or cash into assets quickly is essential and by removing this ability even for a short period of time can create fear – or when done so without notice.”

Mr. Bernstein added: “Whether the fiduciary responsibility was met or not and even if the clients benefited from the stoppage the question is now raised: What if the market had continued to fall, would the robo have kept the doors closed for longer? At what point would they have opened? Hopefully, this occurrence does not damage the reputation of the robo-adviser industry but instead will drive further innovation in machine learning and artificial intelligence.”