How do robo-advisors perform?

Discussion about the rise of robo-advisors has been theoretical – whether they will support or compete with wealth management or more generally disrupt the industry. Views differ, but private banks, asset managers, regulators and investors are all taking robo-advisors seriously.

However, what is missing from the debate is performance. Do the algorithm-driven portfolios actually deliver what they promise?

Last month, we completed the first round of robo-advisor participants, with the final results here.

FSA now starts its second year of the feature, with participants Hong Kong-based Aqumon, Smartly, based in Singapore and Algebra, based in Malaysia.

On July 1, 2018, we made a hypothetical investment of $1m in each of these three robo-advisors. The results in today’s first article show what that $1m is now worth. Return results will be published monthly until August 2019.

The purpose is to highlight the practical angle – how robo-advisors allocate and how they perform, particularly when there is a downturn.

Additionally, self-directed individual investors can make a comparison with their own results, wealth managers can check robo-advisor performance to client portfolio results and the rest of the industry may find the data useful in helping to form a judgment about the robo-advisor phenomenon.

FSA Robo-Advisor Showcase

Performance on 1 August 2018

Note: Three portfolios for each robo-advisor are presented – cautious, balanced and aggressive. However, because the firms operate in different markets and offer different products, the robo-advisors are not competing against each other but against their own benchmarks.

Benchmarks: 20% MSCI AC World Index + 80% US Aggregate Bond Index; 60% MSCI AC World Index + 40% US Aggregate Index; 80% MSCI AC World Index + 20% US Aggregate Bond Index

Hong Kong-based Aqumon, which launched last year, uses machine learning, a type of artificial intelligence, as part of the process to screen investment products and form allocation views to rebalance portfolios.

Aqumon can provide five-to-10 risk profiles for each client, co-founder Don Huang told FSA previously. Based on individual risk appetite, the service then creates a portfolio, which will have eight-to-10 ETFs for Hong Kong-based investors. In China, the robo-advisor makes use of mutual funds, as the costs of using ETFs are much higher than in Hong Kong and the US.

The service does not have a subscription fee and has a 0.8% annual investment advisory fee.

Benchmarks: Morningstar Conservative Target Risk Index; Morningstar Moderate Target Risk Index; Morningstar Aggressive Target Risk Index. Note: Smartly did not provide any benchmark indices for its portfolios

Singapore-based Smartly, which was launched in 2016, provides 10 risks profiles with one being the most conservative and 10 the most aggressive.

Looking at various quantitative and qualitative criteria, the platform has around 20 ETFs on its shelf to choose from and may focus on equities, bonds, commodities, real estate or  cash. It only selects physical ETFs, which means there are no derivatives involved.

When creating portfolios, Smartly uses the “Black Litterman” model, which is applied for its algorithms to measure the limitations of how much can be allocated to a particular ETF or a particular asset class or region in general. Portfolios are rebalanced regularly, as often as once a month.

The annual fee is between 0.5%-1% and there are no additional fees for rebalancing, withdrawals or trading.

Benchmarks: 50% S&P 500 + 50% US Aggregate Bond Index; 70% S&P 500 + 30% US Aggregate Bond Index; 90% S&P 500 + 10% US Aggregate Bond Index

Algebra is a robo-advisor offered by Malaysia-based Farringdon Group. It was launched in July 2017. It offers sharia-compliant and conventional portfolios. FSA features three non-sharia portfolios.

The basis of Algebra’s portfolios is a smart-beta stock-picking strategy developed by Singapore-based Farringdon Asset Management. The portfolio consists of around 50 US stocks from the S&P 500 universe. They are selected based on the analysis of portfolios of ten highly rated active US equity fund managers. From each manager the algorithm chooses five stocks in which their fund is overweight, to include in the Algebra portfolio. The three model portfolios presented here contain a different allocation of fixed income to manage the risk profile. The annual fee is 0.85%.