By Nicolas Rabener
Creating factor portfolios requires investors to make a number of decisions regarding portfolio design and can be considered as much an art as science. The sorting metrics need to be defined, the stock universe, the top and bottom cut-off levels, and the rebalancing periods.
The latter, i.e. how frequent the portfolio needs to be rebalanced, might seem trivial, but markets have gotten more efficient and quicker rebalancing might be advantageous. Academic research typically assumes quarterly or monthly portfolio rebalancing, which seems rather infrequent in the age of high-frequency trading.
In this short research note we will analyse the impact of the rebalancing frequency on a multi-factor portfolio and four single factors.
In this research report we focus on four factors namely Value, Size, Momentum and Quality. The factors are created by constructing long-short beta-neutral portfolios of the top and bottom 10% of stocks of the US stock market.
The multi-factor portfolio is comprised of the four factors and is created via the inter sectional model, which selects the stocks in the intersection of the factors. Only stocks with a market capitalization of larger than $1 billion are included and each transaction occurs costs of 10 basis points.
Multi-Factor Portfolio: Rebalancing Frequencies
The chart below shows the performance of a long-short multi-factor portfolio in the US, which ranks stocks for the Value, Size, Momentum and Quality factors. E.g. the long portfolio contains stocks that are cheap, small caps, have shown a good relative performance over the last 12 months and rank highly on quality metrics.
The portfolio is rebalanced at different intervals and we can observe that the profiles are almost identical in terms of trend. However, the portfolios with more frequent rebalancing, i.e weekly to monthly, show a consistent outperformance to portfolios that rebalance on a quarterly or semi-annual basis.
Investors might argue that a more frequent rebalancing adapts the portfolio quicker to company news and market changes, naturally at the price of more transaction costs. The multi-factor portfolio selects stocks based on the intersectional model, which means the stocks in the intersection of factors (please see this report for further model details Multi-factor Model 101).
These stocks don’t rank particular high on a single factor, but rank highly on average across the factors, which results in a portfolio of stocks that does not change often as the multi-metric ranking process provides a degree of stability. The chart below exhibits the risk-return ratios of the multi-factor portfolio with different rebalancing frequencies. More frequent rebalancing generates higher ratios, albeit with only a marginal difference between weekly and monthly rebalancing.
Value Factor: Reblancing
In addition to analysing the impact of the rebalancing frequency on the multi-factor portfolio we can also observe the changes for the individual factors. The chart below shows the risk-return ratios for the Value factor and highlights that the rebalancing frequency has a minor impact on the ratio.
The factor is defined as a combination of price-to-book and price-to-earnings multiples, which only change significantly when corporate earnings are released. Given that listed companies in the US publish quarterly earnings, stocks in the long and short portfolios of the Value factor do not change too often.
Size Factor: Rebalancing Frequencies
The impact on the Size factor, which is defined as buying small and shorting large companies, highlights a linear increase in the risk-return ratio with a decreasing rebalancing frequency. The results can be explained by portfolio construction, which requires companies to have a minimum market capitalisation of $1 billion.
More frequent rebalancing leads to more companies being beneath the threshold, which then results in a higher turnover and more transaction costs. If we exclude transaction costs from the analysis, then the risk-return ratios do not show that less frequent rebalancing is more favourable.