Weekly Top 5 Papers – June 19th 2017

1. Is it Ethical to Teach that Beta and CAPM Explain Something? by Pablo Fernandez (University of Navarra – IESE Business School)

My answer to the question in the title is NO. It is crystal clear that CAPM and its Betas do not explain anything about expected or required returns. There are mountains of evidence to support my stance.

If, for any reason, a person teaches that Beta and CAPM explain something and he knows that they do not explain anything, such a person is lying. To lie is not ethical. If the person “believes” that Beta and CAPM explain something, his “belief” is due to ignorance (he has not studied enough, he has not done enough calculations, he just repeats what he heard to others…). For a professor, it is not ethical to teach about a subject that he does not know enough about. – Pablo Fernández

2. Lazy Prices by Lauren Cohen (Harvard Business School) and Christopher Malloy (Harvard Business School) and Quoc Nguyen (University of Illinois at Chicago – Department of Finance)

Big Picture

Taking a step back, we provide evidence on the power of defaults and the implications of inertia in decision-making.  We show how these touch even the most watched and poured-over markets in the world. Taking a twist on monotony and boilerplates – by focusing on deviations from default behavior – we can see how powerful these “breaks” or “deviations” from the norm can be for conveying valuable future information. In an entirely non-experimental setting, across thousands of firms and almost 20 years of data, breaks from default behavior have large implications for corporations, and asset prices more generally. Given the pervasiveness of inertia in human behavior across settings, the implications of breaks from default behaviors in the corporate, financial, and political settings provide a critical, yet underappreciated area, in both finance and economics. –  Lauren Cohen

3. CDS Rate Construction Methods by Machine Learning Techniques by Raymond Brummelhuis (University of Reims Champagne-Ardenne) and Zhongmin Luo (Standard Chartered Bank, London)

If you are intrigued by AlphaGo, driverless cars, robots, you might want to know a bit more about a rapidly evolving area called Machine Learning, an area destined to change our lives in many areas. If you have interest in finance industry, you might want to know one major challenge (referred to as Shortage of Liquidity problem in the paper) faced by banks today in their efforts to improve their Counterparty Credit Risk measurement and management, a problematic area highlighted by the downfall of Lehman during the financial crisis in 2008. Our paper presents 8 most popular classifiers in Machine Learning, each illustrated by real-world example, can be a great guide for you. Meanwhile, we take an interdisciplinary approach by applying Machine Learning techniques to solve the Shortage of Liquidity challenge faced by finance industry. In derivative pricing and risk management, it is often necessary to measure the default risk of counterparties based on CDS quotes for the counterparties. Frustratingly, you will not find liquid CDS quotes for most of the counterparties in the CDS market. According to European Banking Authority’s survey for major European banks, over 75% of their counterparties do not have liquid quotes for CDS; therefore, banks have to come up with so-called CDS proxies in derivative pricing and risk management. Numerous literature published on derivative pricing, risk management and XVA (a collective term for Value Adjustments such as CVA, FVA, etc.) have conveniently assumed the existence of liquid CDS quotes and few have directly addressed the real-world predicament that confronts the finance industry in this area.

The paper, firstly, shows that two existing widely used CDS Proxy methods ignore counterparty specific default risk and potentially introduce arbitrage opportunity. Secondly, we introduce and apply well-established Machine Learning Techniques, specifically, Classification Method to provide arbitrage-free solution to the predicament discussed above. Thirdly, we systematically compare the performances of 156 classifiers out of 8 most popular classification families exclusively based on financial market data to confirm and contrast some of the findings of existing literature; when properly parameterized, we show top three classifiers (SVM, Ensemble and Neural Network) produce highly accurate classification results based on k-fold Cross Validation. Fourthly, we investigate the impacts of high correlations between financial feature variables on classification accuracies. As future direction for research, we highlight in the paper that the Machine Learning techniques introduced in the paper can be applied to proxy other financial market variables in derivative pricing and risk management as well as private equity investments. Being excited about the interests shown by our readers, we plan to continue with the research by applying Machine Learning techniques to solve practical problems in real financial world.  Zhongmin Luo

4. A Brief Introduction to the Basics of Game Theory by Matthew O. Jackson (Stanford University – Department of Economics)

5. Historical Returns of the Market Portfolio by Ronald Doeswijk (Independent) and Trevin Lam (Rabobank) and Laurens Swinkels (Erasmus University Rotterdam (EUR))

This study is the first that documents returns for the global market portfolio for a decades long period that contains various market conditions. To illustrate, our sample from 1960 through 2015 covers an inflationary and a disinflationary environment, seven recessions as well as several crises. Our global market portfolio basically contains all assets in which financial investors have invested. Composing this unique dataset also involved hard core data collection from dusty, now yellow, OECD books in the basement of the library of Erasmus University in Rotterdam. These books are invaluable since the data have not been processed digitally. Only by flipping through the pages of the books, shooting pictures of the relevant pages and processing these data manually, we were able to produce the results in this paper.

We believe this study to be valuable for several reasons. First, from a theoretical point of view, by combining the risk-free asset with the optimal portfolio on the efficient frontier we are able to construct the Capital Allocation Line. Every possible portfolio on this line has an ex-ante superior risk-adjusted expected return compared to other portfolios on the efficient frontier. Second, we aim to estimate the historical returns of the invested market portfolio as closely as possible, and therefore one could see our return series as a benchmark return for the multi-asset investor. Third, this new data enables an extensive analysis of return and risk characteristics of the global market portfolio and the asset categories over the period 1960 to 2015.

Some key findings in our study are that the global market portfolio realizes a compounded real return of 4.38% with a standard deviation of 11.6% in our sample period. Next, the reward for the average investor is a compounded return of 3.24%-points above the saver’s. Finally, we show that with simple alternative heuristic allocation schemes it is possible to achieve a better reward for risk, in particular for downward risk. We conclude that all investors together did a reasonable job in determining the global market portfolio, but there is room for improvement.

We are looking forward to conducting more research on strategic allocation in the future to add valuable studies to the academic literature! Trevin Lam