An analytics approach to debiasing asset- management decisions
Wealth & Asset Management December 2017
Nick Hoffman Martin Huber Magdalena Smith An analytics approach to debiasing asset-management decisions
Asset managers can find a competitive edge in debiasing techniques— accelerated with advanced analytics.
Investment managers are pursuing analytics- Richard H. Thaler, one of the leaders in this field aided improvements in many different areas of (see sidebar, “The debiasing nudge”). the business, from customer and asset acquisition to operations and overhead costs. The change Asset management: An industry facing area we focus on in this discussion is investment challenges performance improvement, specifically the Investment managers are facing significant debiasing of investment decisions. With the challenges to profitability. Demonstrating the help of more advanced analytics than they are value of active management has become more already using, funds have been able to measure difficult in a market where returns are narrowing. the role played by bias in suboptimal trading Dissatisfaction with active performance is causing decisions, connecting particular biases customers to migrate toward cheaper passive to particular decisions. Such discoveries funds (Exhibit 1). Actions active funds are taking in provide the necessary foundation for effective response include reducing fees, which has created countermeasures—the debiasing methods that can a competitive cycle that is compressing margins. bring significant performance improvements to a Some funds (such as Allianz Global Investors pressured industry. and Fidelity) are changing their fee and pricing structures to make them more dependent on Business leaders are increasingly recognizing the outperformance through the use of fulcrum-type risk of bias in business decision making. Insights fee structures in their funds. from the fields of behavioral economics and cognitive psychology continually emerge to reveal To push back on these trends, some investment that individuals and institutions do not base managers are broaching the topic of bias in financial and other decisions on purely rational investment decisions. They are seeking a considerations. The contours of the irrational competitive edge and—perhaps more important— biases have become increasingly known, and the to improve the value proposition of active ways bias operates in our thought processes can management. Evidence from within and outside often be predicted. Most important of all, the the industry strongly suggests that even the leading methods and means to counteract biases are asset managers in top-performing funds could becoming more sophisticated and effective, at least improve their investment performance by applying for those executives willing to inquire into their debiasing techniques. Our recent experience debiasing needs. working with investment managers has shown that enhancing these techniques with analytics can Advances in the use of debiasing techniques to improve performance significantly. improve decision making have been inspired by the research of many pioneering scholars. Bias and debiasing in action Their innovations have had numerous practical Bias is a risk in all business decision making— confirmations in the public sector and business the more significant the decision, the greater the settings as well as in the decisions of individuals. risk. Particular biases have been behind many As if in recognition of the deepening relevance costly missteps by companies and institutions in of debiasing in economic life, the Nobel Prize every sector. Consequently, behavioral scientists in Economic Sciences was awarded this year to and business leaders have developed methods for
2 An analytics approach to debiasing asset-management decisions The debiasing nudge
In October 2017, the Nobel Committee awarded its decision making is undermined by biases. Among prize in Economic Sciences to Richard H. Thaler of the biases the authors analyzed are the stability the University of Chicago, who has thought deeply biases to which investment decision making is highly about rational and irrational behaviors in economic susceptible: anchoring, for example—the tying decision making. He and his colleagues in the field of of actions to an initial value and failure to adjust to behavioral economics—including Dan Ariely, Amos take into account new information. Loss aversion Tversky, and Daniel Kahneman (a previous Nobel is another such bias—the familiar fear that makes winner)—analyze the psychological dimension of us more risk averse than logic would allow. Where economics, partly leaning on insights from the field of human decision making is more prone to bias than cognitive psychology. Their research reveals patterns to reasonable deliberation, Thaler and Sunstein that did not align with the rational assumptions recommend the debiasing “nudge”—a benign, often embedded in prevailing descriptions of economic small adjustment that counters irrational impulses. systems and individual financial actions. They The authors discuss many examples of successful found that especially under conditions of risk and nudges; reviews of Nudge most often cited the opt- uncertainty, individuals as well as institutions fail to in default for employee retirement savings plans. In behave as might be expected when making financial this example, many more employees will save for decisions. Assumed universal principles, such as retirement when they are automatically enrolled in actual self-interest and a sure grasp of dynamic a plan. That is, the nudge of requiring employees inputs including time and probability, often give way to to save unless they opt out of the plan gets better irrational and unpredictable actions, based on narrow results than an approach that requires them to opt in. or flawed data and personal experiences. Everyone recognizes the need to save for retirement, but not everyone acts on it. The opt-in nudge is in fact In their book, Nudge: Improving Decisions About a form of debiasing for those irrationally ignoring their Health, Wealth, and Happiness (2008), Thaler and own future financial well-being. coauthor Cass Sunstein reveal ways that rational
debiasing decision making. Many of the companies top performers were those that least anchored that have adopted these methods can attest to their their capital-allocation decisions to decisions effectiveness in improving outcomes: made the previous year. “Dynamic reallocators” achieved a median return that was nearly McKinsey cross-sector research has suggested four percentage points higher than that of that one of the most prevalent irrational biases, “Dormant reallocators.” the form of stability bias known as anchoring, undermines optimal capital reallocation. The An executive at the German electric utility research compared capital allocation and total RWE recently discussed his experience leading shareholder returns over a 20-year period a debiasing effort.1 Like most of its peers, the (1990 to 2010) at more than 1,500 publicly company had until recently based capital- traded companies in the United States. The investment decisions on ever-rising commodity
An analytics approach to debiasing asset-management decisions 3 CDP 2017 An analytics approach to debiasing asset-management decisions Exhibit 1 of 4
Exhibit 1 Customers are migrating to passive equity funds, which have lower margins.
North America net ow growth and revenue margin by asset class1
Passive Active Alternatives
Bubble size 2016 AUM
150 145 Public market alternatives 60 Active specialty 55 equity Active specialty 50 Multi-asset fixed income basis points 45 40 35 Active core 30 Active core equity fixed income 25 Passive 20 Passive fixed 15 equity Passive Passive income 10 Money other
2016 revenue margin, multi-asset 5 market 0 –13 –12 –11 –10 –9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 2016 net ows,
1 Active core equity includes US large cap and yield income equity; active core fixed income includes core, core plus, and municipal bonds; active specialty equity includes foreign, global, EM and US small mid-cap; active specialty fixed income includes global, EM, high yield, TIPS, and unconstrained. Source: 2016 McKinsey Performance Lens Global Asset Management Survey and Global Growth Cube
and power prices. RWE leaders realized that A global investment bank reviewed its traders’ debiasing efforts could have challenged their decisions on equity position weighting. The earlier assumptions, enabling them to hedge bank recognized overcommitment to positions potential adverse effects. For future decisions, and suboptimal execution due to endowment they implemented a farsighted cultural change effects and confirmation bias. The endowment program to identify and counter cognitive effect, discussed in the research of Richard biases throughout the organization. Care was Thaler and other behavioral economists, refers taken to ensure that dissenting and outside to the psychological effects that prejudice analyses were fully articulated. For important owners in favor of retaining their assets despite moves, a “devil’s advocate” was appointed; for changing conditions. Confirmation bias is a the company’s most pressing strategic problem, form of pattern-recognition bias, causing us separate internal and external teams—a to see nonexistent patterns in information: “red team–blue team” approach—were assigned evidence supporting a favored belief is to develop solutions independently of each overvalued, while evidence to the contrary is other. (The devil’s advocate and the red team– discounted. To counteract these biases, the blue team approach are defined, along with bank implemented “premortem” evaluations other debiasing techniques, toward the end of before each trading stage, with checklists to this article.) ensure that the stages were investigated fully.2
4 An analytics approach to debiasing asset-management decisions A global asset management company likewise investment managers could now select and identified confirmation bias in the decisions of apply debiasing methods to greater effect in their its investment managers. The company used investment decisions. devil’s advocates to develop opposing views on investment decisions.3 The analytics edge The approach requires the creation of an A leading pharmaceutical company tackled integrated data set covering the investment overconfidence bias in the proposals of its portfolio. Included are all stock decisions since researchers by using premortems to challenge a fund’s inception. Security selection, security research proposals and renewals before weighting, and selling timing are captured, as investing in any project.4 are the activities and actors that developed and maintained the portfolio. Data is thus compiled for In the healthcare industry in the United States, the processes and decisions made by the individual research revealed that anchoring and stability investment managers and their teams. This history biases significantly undermined diagnostic includes team communications and, insofar as accuracy and effective treatment in hospitals. possible, the behavior, reasoning, and emotions According to one widely cited study, errors associated with individual decisions. occurred in 6.5 percent of hospital admissions in 2000–10. NIH and other research further The funds deploy machine learning, guided by suggests that diagnostic and other errors were hypotheses developed jointly by fund managers causing as many as 200,000 avoidable deaths and experts about the biases that might have annually.5 In response, leading hospitals negatively affected their investment decisions. implemented a strong debiasing culture, Typical biases afflicting asset management including requisite consideration of “must not performance are overconfidence, loss aversion, miss” diagnoses, checklists for ICU, surgery, or the false analogy—the logical fallacy in which and diagnostic procedures, the presentation of inductive reasoning is simulated through invalid evidence in support of alternative diagnoses, comparisons. The hypothetical biases are tested and group discussions to encourage others’ by building an exploratory model to understand opinions in complex situations. emotions and processes associated with trading decisions. Emotions behind biased decisions are Debiasing in asset management diagnosed in the data through K-means cluster In the asset management sector, investment analysis—an iterative vector analysis enabled by decisions are being analyzed in light of debiasing machine-learning algorithms, which can isolate experiences in other industries. A few leading patterns in highly complex data sets (Exhibit 2). funds have employed analytics in this effort, to improve effectiveness in diagnosing bias and The use of analytics in this way, to discover biases its drivers. Working with analytics experts and and their sources, is new. The biases themselves behavioral scientists, they applied machine- are familiar and susceptible to the debiasing learning algorithms to their historical investment methods elaborated in the scholarly literature and data. They discovered clusters of suboptimal practical approaches. investment decisions that showed potential biases. By looking more closely at these suboptimal Initial performance decomposition analysis decisions, the funds identified consistent bias in The approach starts with performance the processes by which the decisions were reached decomposition analysis, the relatively simple and the accompanying emotions experienced diagnostic tool long used in the industry. This by the decision makers. Having exposed the initial step is followed by further analyses, aimed patterns of bias with the help of analytics, these at discovering the processes and emotions surrounding specific investment decisions as well
An analytics approach to debiasing asset-management decisions 5 CDP 2017 An analytics approach to debiasing asset-management decisions Exhibit 2 of 4
Exhibit 2 To detect and correct for biases in investment decisions, some funds have successfully applied analytics to large sets of historical data.