MAS taps unconventional data to monitor crypto market

Singapore

CRYPTOCURRENCIES do not currently pose a material risk to financial stability in Singapore but bear close monitoring, said the Monetary Authority of Singapore (MAS), which has been using non-traditional data sources such as blockchain ledgers in its surveillance of crypto markets.

Monitoring developments in digital token markets is hard as data is not readily available via conventional channels, noted MAS in its Financial Stability Report on Friday. To strengthen its surveillance, the central bank has turned to data sources such as digital token trading platform Application Programming Interfaces (APIs), which allow monitoring of global shifts in trading activity.

"Substantial declines in the share of trading activity of a particular jurisdiction can serve as an early warning signal that the trading activity may flow to other jurisdictions, which has both financial stability and regulatory implications," said the MAS.

It has also been keeping an eye on potential wealth effects that could occur if a decline in households' holdings of cryptocurrencies cause a drop in aggregate spending.

MAS used trading activity and inflows of the Singapore dollar into Bitcoin-SGD trades on major Singapore-based exchanges as proxies for the potential wealth effects and associated risks. It found that the wealth effects are currently small.

The lower-bound on the total volume of Bitcoin-SGD traded since 2013 is estimated to be around S$9 billion, implying monthly trading volumes of less than 1 per cent of those on the Singapore Exchange. Investor interest has also dipped substantially since January 2018.

As for Bitcoin-SGD volumes, cumulative net inflows on trading platforms that disclose their trades' directionality is estimated at S$400 million, "relatively small compared to Singaporean household assets of S$2 trillion at end 2017".

As not all major trading platforms and broker-dealers provide APIs, the MAS has gone further to analyse the Bitcoin distributed ledger. As a proof of concept, the MAS ran a clustering algorithm to find some 10,000 addresses controlled by a major wallet provider and broker-dealer in Singapore - suggesting potential custodial holdings of Bitcoin that peaked at US$20 million in January 2018.

Besides exploring digital token-related data, the MAS' recent digital efforts include the use of machine learning to identify indicators that can predict corporate distress. Using data from 700 public firms and 60,000 private firms in Singapore from 2007 to 2017, the study found that firm-level characteristics are more important than industry-specific and macroeconomic indicators.

In particular, the cash asset ratio, asset turnover and return on assets are relatively more important. The first measures a firm's liquidity or ability to meet short-term obligations, while the latter two measure a firm's efficiency in generating revenue and profits using its assets.

A revised model looking at larger firms found that similar indicators of profitability and leverage, such as equity-to-asset and equity-to-fixed assets ratios, were relatively more important in predicting distress.

READ MORE: