Market structure is changing. Some like the advent of machine learning could enhance investor returns, while others like fragmented clearing are adding an additional cost.

What Machine Learning Will Mean for Asset Managers (Harvard Business Review, 7 min read) Harvard asserts that machine learning (ML) will actually reverse the trend towards passive investment. ML can make more accurate forecasts by using alternative data (eg images and sound), employ non-linear modelling, reduce human biases and be quickly implemented without regulatory approvals (unlike say the auto industry) . [Bullish ML quant managers, bearish traditional passive}

The Cost of Clearing Fragmentation (BIS, 30 min read) Latest BIS research shows that an architecture based on multiple but fragmented central counterparties (CCPs) lead to daily opportunity costs of $80 million for end users. This happens mainly because global liquidity dealers cannot net trades cleared in different CCPs. [Bearish banks]