| 29.09.2020

Hanken Finance Day: a deeper look at factor investing

Hanken’s annual Finance Day moved online this year, with three speakers presenting their work in an area of increasing interest to finance academics and fund managers alike: factor investing.

In the wake of the 2008 global financial crisis, the Norwegian Government Pension Fund Global – the world’s largest sovereign fund – asked professors Ang, Goetzmann and Schaefe to evaluate the fund’s active management. It was in their report that the three academics first introduced the idea of what came to be known as “factor investing”.

Traditionally, investment portfolios are arranged by sector and/or country, combining securities or bonds of different risk weights to ensure diversification. Factor investing looks at portfolio management from a different perspective: instead of structuring portfolios by sector or country, they are created around certain market anomalies, also known as risk factors.

“Factor investing means aggregating securities by their exposure to predefined characteristics, which is very different to what has been done before,” says Marie Brière, Head of Investor Research Center at Amundi and Affiliate Professor at Paris Dauphine University.

“From a paradigm that was mainly about allocating across market-cap weighted indices, we move to a paradigm where we should allocate across dynamically rebalanced factors.”
“Factors are very good in the high-risk, high-return space, whereas sectors dominate in the low-risk, low-return space. Factors also don’t do as well during times of crisis,” she says. “So diversifying a portfolio made of both factors and sectors is definitely a good way to allocate.”

As risk factors need to be constantly evaluated and rebalanced, Briére says that one of the key issues is to understand if they can resist transaction costs.

“Our estimates, which are based on an exceptional database of institutional investors’ trades, show that the factors of size, value, profitability and investment can easily resist transaction costs. However, the momentum factor can involve high trading costs for large-size portfolios.”


Competing pricing models

Today, a number of the world’s largest funds are using factor modelling for part of their asset allocation. There has also been an explosion of academic interest in the subject, both to identify new market anomalies and to create new multifactor models to explain them.
Paulo Maio, Professor of Finance at Hanken, refers to the work that Kewei Hou, Chen Xue and Lu Zhang have done in the area. The three researchers identified 158 significant anomalies in the US market, splitting them into six broad categories. They also created a four-factor asset-pricing model, later extending it with an additional factor in joint work with Haitao Mo.

The group’s work competes with the multifactor models of Fama and French, who proposed extending the traditional Capital Asset Pricing Model (CAPM) with up to six risk factors.  
“According to several of my studies, the empirical performance of the Hou, Xue and Zhang model is better than the Fama French model,” says Professor Maio. “It also appears to be more consistent with Ross’s Arbitrage Pricing Theory than the other models proposed in the literature.”
“When properly implemented, factor investing is more about tailor-made investment products than about commoditized ones,” says Maio. “In this way, it is more appropriate for sophisticated, large and long-term investors.”

A new framework for factor investing
So does factor investing work in a market where both equities and bonds are suffering at the same time? This was the question that Kari Vatanen, Chief Investment Officer at Veritas Pension, set out to answer.
“There are two types of risk factors driving the returns of major asset classes: macroeconomic risk factors and strategy-style factors,” says Vatanen. “We have constructed a framework to classify asset style factors according to their nature and risk characteristics.”
Vatanen’s study collected live data from seven international investment banks to construct the most common alternative risk-factor strategies. From this data, 28 strategies across five different asset classes were identified. Vatanen has divided these strategies into three categories of risk: fundamental, behavioral and structural.


Click here to watch the seminar.


Text: Andrew Flowers