How do Canadian corporate bond mutual funds meet investor redemptions? We revisit this question using decision tree and random forest algorithms. We uncover new patterns in the decisions made by fund managers: the interaction between a larger, market-wide term spread and relatively less-liquid holdings increases the probability that a fund manager will sell less-liquid assets (corporate bonds) to meet redemptions. The evidence also shows that machine learning algorithms can extract new knowledge that is not apparent using a classical linear modelling approach.
Since 2010, the liquidity of corporate bonds has improved on average, while their trading activity has remained stable. We find that the liquidity and trading activity of riskier bonds or bonds issued by firms in different sectors have been stable. However, the liquidity and trading activity of bonds issued by banks have improved. We observe short-lived episodes of deterioration in liquidity and trading activity.
In recent years, the liquidity in the secondary market for Canadian provincial bonds was a concern for many market participants. We find that a proxy for the bid-ask spread has deteriorated modestly since 2010. However, a proxy for price impact as well as measures of trade size, the number of trades and turnover have been stable or improved since 2010. This holds for bonds issued by different provinces and for bonds of different ages and sizes. Alberta bonds provide an interesting case study: After the fall in oil prices in 2014–15, the province increased its borrowing in the bond market and its credit rating was downgraded. Yet trading activity for Alberta bonds increased significantly. Overall, we interpret the evidence as a sign of resilience in the provincial bond market.