March 6, 2013 No comments

The Sociology of Morality

Do you see yourself as a morally strong person?  Are your actions consistent with your self-conception?  How would you feel if you did something bad?

These simple questions are never asked in employment interviews; candidates’ declarations of integrity would be inherently unreliable.[1]  Few recruiters and managers in the investment industry are trained in reading the facial micro-expressions that indicate whether someone is responding truthfully.[2]  Moreover, nobody’s actions are entirely predictable.  Yet there is empirical evidence that individuals’ internal moral standards are indirectly related to their conduct.  It seems desirable for gatekeepers—indeed, for anyone who is concerned about an organization’s culture, values, and reputation—to understand the factors that make some people more prone to misdeeds than others in like circumstances.

In a study published last year, two sociologists proposed and tested a theory of the self that explicitly links the cognitive, behavioral, and affective facets of moral choice.  Jan E. Stets and Michael J. Carter wrote, “To understand the illicit behavior of some, we need to study the moral dimension of the self and what makes some individuals more dishonest than others within and across situations.”  They cited the “practices of stock brokers, investment advisors, and mortgage lenders whose behavior facilitated the recent economic recession” as an example.[3]

Among many other issues, philosophers have investigated the status of ethical codes, the nature of personhood, the meaning of virtue, and the relationships between thought, language, and action.  Psychologists have examined motivation, the desire for external validation, and the affective consequences of inner conflict.  Now Stets and Carter have enriched the study of moral choice by framing testable hypotheses on the basis of a sociological model of the internal operations of the self.  They found that model in identity theory.

Identity theory sees people as actors, not merely reactors, and fully acknowledges the possibility of choice, but also recognizes that human action is constrained by social structure and social interaction.  The theory crucially postulates that actions are shaped by the way individuals see themselves and the environment; and the meanings people attribute to themselves and to situations are developed through their interactions with others.  In fact, people have as many identities as they have social roles—spouse, parent, child, sibling, manager, employee, colleague, and the like—and those roles are structured in a “hierarchy of salience” defined by individuals’ commitments to being a particular kind of person.  (Salience, in this context, is the probability that a given role will be activated in a specific situation.)  Sheldon Stryker, who largely developed identity theory, expresses its fundamental proposition in these terms: “Commitment impacts identity salience impacts role choice.”[4]

In applying identity theory to the domain of moral choice, Stets and Carter emphasize the mechanism of “identity verification.”  They diagram the self in its environment and explain identity theory in the terminology of control systems, with inputs, outputs, and a feedback loop.  The inputs are definitions of situations (interpretations that frame situations as calling for certain roles, behaviors, and feelings) and “reflected appraisals” (interpretations of others’ reactions to one’s behavior).  These inputs are fed into a metaphorical “comparator” which relates them to stored identity standard meanings and produces error signals when it registers variances.  If an actor thinks others perceive her as a just and caring person to the same extent that she so perceives herself—say a “moral identity score” of 7 on a scale of 1 to 10—then her identity is verified.  However, if she thinks others see her as less, or more, moral than she considers herself, her identity is not verified.[5]  Similarly, there is a discrepancy if the meaning of a situation is unaligned with the person’s salient identity.  “For example,” Stets and Carter said, “if meanings in a situation are about acquiring wealth rather than behaving morally, one will have difficulty verifying the moral identity.”[6]

It is through the internal operation of identity verification that emotions come into play.  Stets and Carter focus on two negative moral emotions that may follow from the experience of non-verification: guilt (an actor judges that he did something bad) and shame (an actor judges that he is a bad person).  These emotions may trigger a variety of responses, including cognitive strategies such as rationalizing one’s behavior, evading responsibility, minimizing one’s participation, understating the consequences, and blaming the victim.  Often, however, they result in behavioral changes intended to reduce the inconsistency.  Stets and Carter stress the social significance of these emotions.  “Ultimately,” they wrote, “guilt and shame keep people integrated into society through internal monologues with the self and feedback from others.”[7]

Stets and Carter formulated five hypotheses and designed an experiment to test them.  For instance, one hypothesis was, “The higher a person’s moral identity score, the more likely the person will behave morally.” Another was, “The more a person defines a situation as containing moral meanings, the more likely the person will behave morally.”  The authors recruited 369 undergraduate sociology students and conducted the study in two parts.  In the first phase, they administered a survey that measured the participants’ moral identity, behavior, and emotions in scenarios that were relevant to the students’ lives.  In the second phase, conducted three months later, the same population—there was no attrition— responded to a survey that measured moral meanings and feeling rules (cultural expectations about the emotions that ought to be experienced in moral situations).

Stets and Carter’s article in the American Sociological Review presents the design, results, and limitations of their study in considerable detail.  In general, they state, their hypotheses were supported.  For example, both the moral identity and situations with moral meanings are positively associated with moral behavior, and reports of moral emotions are positively associated with identity non-verification.

This brief treatment does not remotely do justice to Stets and Carter’s meticulous study.   Well written and insightful, it repays careful reading.  What are the implications of their research for investment professionals?  To open the dialogue, here are two suggestions.

  •  Senior management should set clear expectations for ethical conduct.

Because role choices guide behavior, the firm should commit itself to creating a culture that prizes integrity and to defining its employees’ role in terms that prominently include ethical behavior.  Staff should know, for instance, that they are expected without reservation to place clients’ interests ahead of the firm’s and their own interests.  Moreover, because codes of conduct are all too often seen as meaningless totems, managers throughout the organization should consistently enforce internal controls, taking care to ensure there are no mixed messages that would interfere with moral identity verification.  Establishing solid cultural and social expectations within the firm, and communicating them in practice every day, might encourage moral behavior—or at least discourage wrongdoing.

  •  Training should importantly include practice in identifying and articulating the moral meanings of situations.

People can change, even radically.  Examples of moral conversion come to mind readily: Constantine.  Augustine.  In our own times, the late Charles Colson.  But these men may be so easily recalled precisely because their stories are extraordinary.  Change is hard, and radical change betokens a profound spiritual experience.  Identity theory appears to consider the actor’s internal standards inalterable; and, realistically, an adult’s moral identity is fixed, at least in the short term.  Investment management firms may, however, be able to influence the other input to the comparator, the moral meanings of situations that are likely to arise in the course of business.  Helping staff perceive the ethical dimensions of their work might facilitate moral identity verification.

Understanding identity theory, and especially the mechanics of identity verification, is not particularly helpful in spotting likely miscreants, but it does suggest why they might make unfortunate choices.  They may have relatively low internal standards; define their professional role in a way that minimizes or excludes ethical obligations; overlook moral meanings in situations; and/or fail to experience moral emotions, at least with the same intensity as others.  Because these are hidden characteristics, it remains possible that bad actors will occasionally slip through the hiring process—all the interviews, reference calls, and background checks—and disrupt the organization or put its reputation at risk.  Let those people find themselves alone at investment management firms that take ethical conduct seriously.

[1] See “Pants on Fire,” December 16, 2011.  http://middle-office.com/2011/12/16/unclassifiable/.  Accessed March 4, 2012.

[2] Online training is available from Paul Ekman Group at www.paulekman.com.  Accessed March 4, 2012.

[3] Jan E. Stets and Michael J. Carter, “A Theory of the Self for the Sociology of Morality.” American Sociological Review, 2012, 135. http://asr.sagepub.com/content/77/1/120.  Accessed March 2, 2013.  DOI: 10.1177/0003122411433762.  Many thanks to Sharon Gould for bringing this research to my attention.

[4] Sheldon Stryker, “Identity Theory,” Encyclopedia of Sociology (Macmillan Reference USA, 2001), 1253-1258.

[5] Stets and Carter point out that non-verification of moral identity can result from surpassing as well as from falling short of one’s identity standard.  “Exceeding one’s moral standard may generate the view that one is a moral fraud, while failing to meet a moral standard may result in feeling morally inadequate.” Op. cit., 135.

[6] Op. cit., 125.

[7] Ibid.  The authors cite Jonathan H. Turner, “Natural Selection and the Evolution of Morality in Human Societies,” in S. Hitlin and S. Vaisey, eds., Handbook of the Sociology of Morality (New York: Springer, 2010), 125–145.
January 28, 2013 No comments

Smarts

One of the many things Howard Gardner has taught us is that displaying a talent for logico-mathematical thinking (or, in our line of work, quantitative modeling) is not the only way of being smart. His theory of multiple intelligences, first set out in Frames of Mind (Basic Books, 1983), holds that there are other, separate kinds of intellection, including for example linguistic, spatial, musical, and interpersonal competencies. Other ways of learning and exhibiting intelligent behavior.

Gardner has profoundly enriched our understanding of intelligence. But his influential early work comes to mind now because Adam Gopnik unexpectedly made an interesting observation about bright people and their cognitive styles in an article on the science of sound. Gopnik is a writer of exceptional range and subtlety. “Among highly intelligent people,” he said,

there are two kinds of minds, the sharp and the soft. We expect smart people to have minds like swords, made to fight and slash and slay. Soft smart minds, though, are of another, rarer sort. They absorb great quantities of data and opinion, often silently, even sluggishly, and turn them around slowly until a solution appears.[1]

Traders, at least those who survive their first year or two, have sharp minds. They are paid well to see evanescent opportunities, exploit asymmetrical information, and execute advantageous decisions swiftly. They use their minds like sabers, and, eyes on the Bloomberg, they “fight and slash and slay” in an arena where relentless self-regard and aggressive behavior are not only acceptable but mandatory. On the trading floor there are the living—the “quick”—and the dead.

Is there any place in the investment industry for Gopnik’s soft smart minds?

Contrasting “sharp” and “soft” is a curious move. The opposite of sharp is dull; of soft, hard. In ordinary language, to say that someone is “not the sharpest knife in the drawer” is no more a compliment than calling him “soft in the head.” It’s just common sense, it seems, that investment organizations competing in fast-moving, impersonal, zero-sum markets need hardheaded men and women who will spot the opening and go for the kill. And it seems equally clear that soft minds—one is tempted to say flaccid minds—belong in philosophy departments, or rather in the caring professions. In my experience, academics, armed with pen-is-mightier vocabularies and finely honed instincts for one another’s vulnerabilities, have shown themselves intensely competitive. Popper said Wittgenstein threatened him with a fireplace poker.[2]

Yet in Gopnik’s usage soft smart minds are anything but weak. They merely have another, less familiar way of ingesting and processing complex and often contradictory information: they ruminate. This is a form of high intelligence that is difficult to describe favorably; Gopnik chose the unfortunate adverb “sluggishly” to characterize how soft minds assimilate facts and opinions, and I’ve fallen into a bovine metaphor. Nonetheless, precisely because they don’t reach conclusions precipitously, soft smart minds may be more likely to see remote connections, and less likely to overlook significant but unobvious trends.

Some investment management firms appear to value unhurried deliberation, not, of course, on the trading desk, but on the part of business and portfolio strategists. In my view, many more would do well to hire a few good, soft minds—in Gopnik’s special sense—and supply them with economic and capital markets data. After a while they might come back with surprising insights.

[1] “Music to Your Ears: The Quest for 3-D Recording and Other Mysteries of Sound,” The New Yorker, January 28, 2013, p. 35.

[2] David Edmonds and John Eidinow, Wittgenstein’s Poker: The Story of a Ten-Minute Argument Between Two Great Philosophers (Faber and Faber, 2001).
December 12, 2012 No comments

Analyzing Tactical Asset Allocation Performance

Market timing has a bad reputation; all too often, it has meant crowding into the latest hot asset class or sector when it is already overvalued. But intelligently researched, ably executed tactical asset allocation (TAA) is another matter altogether. Prospective and contrarian, the TAA approach is to monitor expectations and correlations, increase investments in unpopular assets with the greatest potential for appreciation, and reduce investments in those with the greatest potential for loss. Some managers seem to have the skill and, above all, perhaps, the discipline to add value in a fairly consistent manner by thus restructuring portfolios advantageously.

Conventional buy-and-hold attribution analysis breaks out the value-added contribution from asset allocation, but it does not isolate and quantify the results that are specifically due to tactical shifts among asset classes or sectors. In a 2010 article in the Financial Analysts Journal, however, Jason C. Hsu, Vitali Kalesnik, and Brett W. Myers proposed a framework for unpacking the effects of managers’ static and dynamic asset allocations.[1] Although it is not without inconveniences, their instructive methodology fully deserves the attention of investment performance practitioners.

The Brinson model views portfolio returns as the sum of the products of component weights and returns. (The components can be any single dimension of interest, such as economic sectors, geographic regions, or characteristic ranges.)  Hsu, Kalesnik, and Myers’s fundamental insight is that portfolio returns can be expressed as the sum of two terms: E(wt)E(Rt), representing a static asset allocation effect, and cov(wt, Rt), reflecting the manager’s dynamic allocation skill. (See Exhibit 1.) If the covariance of component weights and returns is positive, the manager’s market timing has added value; if it is negative, the manager has destroyed value.

Consider a hypothetical long-only portfolio whose manager has latitude, within limits, to shift assets among financial stocks, telecommunications stocks, and cash. The benchmark is 45% S&P 500 Financials, 45% S&P 500 Telecommunication Services, and 10% cash, and the manager is permitted to invest up to 50% of assets in a single equity sector and hold up to 25% in cash. In the year ended 30 November 2012, the Financials and Telecom sector indexes had total returns of 25% and 24.1%, respectively, but they traced markedly different paths across the twelve months. (See Exhibit 2.)

The portfolio’s total return for the same period was 26.8% vs. the benchmark’s 22.6%. (See Exhibit 3.) The active return, therefore, was 4.25%. With monthly rebalancing, how much did the manager’s tactical asset allocation add to this outperformance? Conventional attribution analysis (Exhibit 4) shows that the value-added return stems entirely from asset allocation (the stock selection effect was negative) but does not tell us—or the firm and its clients—anything about the impact of the manager’s month-to-month reallocations.

Two preliminary remarks about the Hsu-Kalesnik-Myers procedure are in order. First, as is well known, arithmetic attribution methodologies don’t handle multiple periods deftly, and, if unadjusted, produce results that may leave a significant portion of the added-value return unexplained. In Exhibit 4, we addressed this problem by applying the Cariño smoothing algorithm to the monthly attribution effects. Hsu, Kalesnik, and Myers use another expedient: they calculate average effects over the measurement period.

Second, because manipulating covariances is clumsy business, Hsu and his co-authors also define the dynamic allocation effect as the difference between the total and static allocation effects. This practical step vastly eases the computational burden.

Applying the formulas shown in Exhibit 5 reveals that the manager’s dynamically adjusting sector weights from month to month accounts for almost all the average asset allocation effect. (In this analysis, it also explains most of the total added value because the average stock selection effect rounds to zero.) Tactical asset allocation decisions contributed 36 of the 37 basis points due to asset allocation in total. In other words, the firm can tell its clients that, in the average month, more than 95% of the active return resulted from market timing. This is encouraging evidence that the manager is skilled in anticipating relative sector performance and positioning the portfolio to benefit from potential market shifts.

The Hsu-Kalesnik-Myers framework is an outstanding innovation, deficient only in its treatment of the multiple periods that are an essential dimension of the TAA strategy. It is most desirable to present attribution analyses for the entire measurement period, as we did by means of the smoothing algorithm in Exhibit 4. Perhaps another theorist will resolve this problem. Nonetheless, Hsu, Kalesnik, and Myers have furnished the investment industry with a powerful means of measuring the dynamic allocation effect.

[1] Hsu, Jason C., Vitali Kalesnik, and Brett W. Myers, “Performance Attribution: Measuring Dynamic Allocation Skill,” Financial Analysts Journal, vol. 66 no. 6 (November/December 2010), 17-26.
November 19, 2012 2 Comments

The Motivation to Work

The Belgian psychologist Joseph Nuttin wrote, “it is a notorious fact that many people are not motivated to do the work they do; they work for extrinsic reasons: in order to do certain things alongside or after their work.” Nonetheless, many others, he observed, “are deeply involved and find their pleasure in the work they accomplish.”[1] What factors account for the difference?

Nuttin, who died in 1988, was an original thinker who brought creative insight as well as experimental ingenuity and analytical power to the study of general psychology. Unlike Freud, whose starting point and abiding interest was psychopathology, he primarily sought to understand well-adjusted, highly functioning people at all levels of their activity. Unlike Thorndike, Hull, and their respective epigones, he recognized and explained the rôles of cognition, intentionality, and self-concept in human behavior. Nuttin developed a dynamic, relational model of personality, one in which “it is suitable to say that personality is not simply situated in a world and open to this world, but that this world enters as an integrating element into the personality itself.”[2] And he expounded a theory of motivation that is, in my opinion, incomparable for its richness and subtlety.

There is no question of summarizing (let alone criticizing) Nuttin’s theory of personality in this short piece. All I can do right now is to mention a theme that is germane to an understanding of the motivation to work. Nuttin objected to the thesis that people are driven primarily by physiological needs—sex, hunger, thirst, self-defense—and its corollary that higher-order needs are at best secondary or derivative. In his view, for example, the organism’s fundamental need to maintain its existence against the external environment is equally present in the hunt for food and the endeavor to be someone in the social world.  “We can easily recognize the psychic energy this ‘drive’ contains by imagining, just for an instant, all the effort and tedious work that the education and preliminary training for one career or another, precisely due to which a young human being will become ‘someone’ among the others, requires of the one who submits to it.”[3]

Nor, of course, can Nuttin’s theory of motivation be adequately presented in this space. Given that he places personality, behavior, and motivation in a unified conceptual framework,[4] I can at most indicate some key elements. Building upon his insight that there is no personality without a world and no world without the personality that constructs it, Nuttin states that a person is “a being in situation doing something,”[5] and he defines behavior as “a meaningful response to a situation that also has a sense.”[6] A situated subject acts on a perceived state of affairs in view of a conceived state—a goal—which is more or less ‘realized’ or achieved in the result of the action.[7] The fact that behavior has a goal is central to Nuttin’s understanding of motivation.

With this overlong yet sketchy introduction, let’s turn to Nuttin’s account of the motivation to work. One of the most striking aspects of human behavior, he observes, is that people can’t leave things in the condition they find them. If they don’t undertake to destroy what disturbs them, people are tempted “to intervene, to change, to restructure, to improve.” Human beings also try to intervene in their own development, to become who they want to be, and this tendency is “personalized and made concrete—at least in our culture—in a vocational project.” People generally seem to be motivated to achieve or create something which would not be accomplished or produced without their action. Often they identify with the things they succeed in making real; their opus is an extension of themselves.

In the ideal case, Nuttin says, individuals’ intervention in producing things—their work—is incorporated in the project of their personal development and the realization of their self-concept. In most cases, however, one observes the contrary: “the work one must in fact execute consists in collaborating in the realization of other persons’ projects without the least effort toward a certain integration of projects having been undertaken.” Such extrinsically motivated work remains connected with the need for self-development in the sense that collaborating in an impersonal, collective project provides the feeling of being useful, a feeling which affects one’s self-esteem, as the unemployed will attest. “Nonetheless,” Nuttin writes, “intrinsically motivated work is accompanied by profound satisfaction, and the quality of the work supplied is generally superior.”

Knowledge workers and those who are their own bosses seem to have less difficulty than others in finding tasks that really occupy them. Nuttin realistically acknowledges, however, that, independently of financial problems, few people appear capable of giving themselves work that “counts.”[8] Moreover, an exaggerated fear of taking risks and an immoderate desire for security are factors that, in our society, impede the motivation for accomplishment and the tendency to pursue constructive personal projects. Insofar as there is a solution, the theoretically coherent remedy Nuttin suggests consists in “establishing lines of communication by which the person can find a certain personal development through and in his work.” He cites well-known measures to engage employees, such as involving them in enterprise planning and assigning them responsibility for the results obtained, in short, instituting arrangements that favor the workers’ identifying with their work. “But,” he writes, “psychology still seems a very feeble instrument in this struggle against differently powerful social and economic forces.”[9]

In my view, Nuttin’s pessimism on this point may be unjustified. He might have underestimated how much a strong manager in a progressive corporate culture can do on behalf of employees who are open to the possibility of integrating work and personal growth. And creating the conditions for rewarding work that is, incidentally, of superior quality seems well worth the struggle.

[1] Théorie de la motivation humaine: Du besoin au projet d’action, 5e édition (Presses Universitaires de France, 2000), 195.

[2] La Structure de la personnalité, 6e édition (Presses Universitaires de France, 1985), 214.

[3] Psychanalyse et conception spiritualiste de l’homme, 3e édition (Publications Universitaires de Louvain, 1962), 261. Coming full circle, Nuttin wrote, “the different forms of needs pervade one another like the physiological and psycho-social activities themselves. Thus the need for nourishment is concretely manifested as a need to earn one’s living and, even more concretely, as a need to maintain and develop one’s social standing.” Tâche réussite et échec: Théorie de la conduite humaine (Publications Universitaires de Louvain, 1953), 428. “Social standing” is in English in the original.

[4] Théorie de la motivation, op. cit., 122.

[5] Ibid., 104.

[6] Ibid., 38.

[7] Ibid., 78.

[8] Nuttin found the widespread incapacity for constructive self-direction worrisome because, when he was writing, leisure time was expected to increase substantially. That prediction certainly hasn’t come true. Many workers are compelled to sacrifice almost all their free time and hold more than one job just to make ends meet.

[9] Ibid., 194-199.
November 6, 2012 1 Comment

Risk-Adjusted Performance Measurement

Investment results can be calculated, reported, and to a considerable extent analyzed without recourse to ex post risk measures. For example, the arithmetic difference between portfolio and benchmark rates of return conveys some information about the worth of active management; attribution analysis on the Brinson model isolates the value added by the manager’s sector weighting and security selection decisions; triangle charts display benchmark-relative results over historical periods that may include up and down markets; and quality control charts test active returns against the null hypothesis that the manager has no investment skill.

But investment results cannot be well and truly evaluated without explicitly taking ex post risk into account. As Carl Bacon remarks, “If a manager achieved a reasonable but unexciting return, but took very high risk to achieve this return, then the client should be very disappointed.” Ex post risk analysis can also provide risk managers with valuable feedback on the reasonability of ex ante projections. Bacon says, “Risk controllers in particular need to be grounded in reality; it is essential that they compare realized risk with their previous forecasts of risk.”[1]

Bacon’s new book, Practical Risk-Adjusted Performance Measurement (Wiley, 2012), is an exhaustive compendium of ex post risk measures written for risk and performance practitioners from a buy-side asset management perspective. One of the author’s stated purposes was to document, in a structured format, as many discrete measures as possible, and he succeeded admirably. For instance, in addition to excellent, illustrated explanations of commonplace measures such as skewness and kurtosis, an early chapter on descriptive statistics introduces less familiar indicators, including the Bera-Jarque statistic, which tests for normality; the Hurst index, which tests for mean reversion; and Abdulali’s trade-marked bias ratio, which signals possible return smoothing. Despite its brevity (some 200 pages), Practical Risk-Adjusted Performance Measurement is a hefty book, and it will undoubtedly stand as the reference manual for many years to come. For that reason alone it merits a place in every practitioner’s and system developer’s professional library.

However, this work is not ‘merely’ an encyclopedic treatment of the subject. Like its predecessor, Practical Portfolio Performance Measurement and Attribution (2nd ed., Wiley, 2008), it presents the mathematics with no more complexity than strictly necessary, explains the concepts and calculations in plain language, and provides many worked examples. It is, accordingly, an incomparable instructional resource, one made all the more readable by Bacon’s personal observations. For example, he writes, “Some commentators would suggest that negative Sharpe and information ratios have no value, pointing out that if performance is negative the respective ratios actually reward higher variability and higher tracking error. I disagree—it seems self-evident to me at least, that if you are going to underperform it is better to underperform inconsistently rather than underperform consistently.”[2] Practical Risk-Adjusted Performance Measurement is a technical work written in the author’s own voice, and that’s a rare achievement.

In short, I strongly recommend this book to performance and risk professionals seeking to extend their mastery and improve their practice of both disciplines.

[1] Carl Bacon, Practical Risk-Adjusted Performance Measurement (Wiley, 2012), 163.

[2] Ibid., 57-58.
October 5, 2012 No comments

Assessing the Cost of Basel III

The International Monetary Fund (IMF) recently released a working paper on the prospective credit impact of key Basel III regulations.[1] The authors, Douglas Elliott, Suzanne Salloy, and André Oliveira Santos, reckoned that, after a transitional period, the ongoing cost of bank lending would rise by 13 basis points in Europe, 10 basis points in Japan, and 20 basis points in the U.S. Unsurprisingly, the authors concluded their findings “strongly suggest” that the presumptive benefits “would indeed outweigh the costs of regulatory reforms in the long run.”[2] Because the economic value of financial regulation is contested, and the results of cost-benefit analyses are highly sensitive to the initial assumptions, the IMF-sponsored study merits careful reading.

The scope of the study under consideration is at once broad and limited. It considers the cost of Basel III requirements related to capital and liquidity, derivatives, and taxes and fees, in three distinct markets, but it focuses principally on bank loans, to the relative neglect of other affected segments. It uses a simple formula that reflects the conventional accounting-based approach and reasonably assumes the interest rate on a loan has to cover the cost of funding, expected credit losses, and administrative expenses.

The authors made some crucial and instructive choices in framing their research. For the record, I generally agree with their assumptions. Parti pris. It must nonetheless be observed that their decisions predominately tend to attenuate the economic effects of regulatory reform. Moreover, the resulting reduction in the estimated credit impact of Basel III requirements is substantial. For comparison, in a less forgiving study released a year ago, the Institute of International Finance estimated markedly higher increases in the post-transition costs of lending: 3.28% in Europe, 1.81% in Japan, and 2.43% in the U.S.[3] (See Table 1.) Let’s consider salient positions adopted in the IMF working paper.

First, unlike the IIF report, the IMF-sponsored study heroically leaps the chasm of transitional costs. The authors don’t deny that achieving compliance with Basel III entails costly adjustments. Nor are they cavalier about excluding those costs from the scope of their research. Indeed, they acknowledge, “Good policymaking requires true cost-benefit analyses and it would certainly be a mistake to ignore transitional costs. However, resource constraints required a narrower focus on the central question of the long-term effects on credit.”[4] This is a regrettable limitation. Merely to keep the doors open, banks—many of which are exceedingly complex organizations—are compelled not only to upgrade and integrate their governance, risk management, and compliance (GRC) systems but also to divert business and technological talent from other projects which might improve customer service or operational efficiency. Although ancillary benefits may surface over the long term,[5] implementing system and process changes to cope with regulatory reform on the scale of Basel III is not a trivial undertaking.

Second, the authors select year-end 2010 as the baseline in order to exclude the effects of the higher safety margins attributable to market forces after the subprime crisis. This decision is sound. Choosing a pre-crisis point in time for the baseline, as the IIF researchers appear to have done, arguably overstates the impact of new regulations. It is unfortunate the IMF-sponsored study could not similarly extract the credit impact of the Eurozone crisis.

Third, the authors assume that banks will respond to the new regulations by taking cost-cutting and other measures,[6] and they estimate the resulting economies will offset the unfavorable credit impact of Basel III by 13 basis points in Europe, 10 basis points in Japan, and 20 basis points in the U.S. (These estimated offsets are included in the figures cited above.) In particular, the authors observe, “A substantial portion of the cost of credit provision comes from administrative and marketing expenses, where there is considerable room to cut expenses if necessary.”[7] As a practical matter, then, reducing costs generally means lowering compensation and staffing levels. The authors do not address broader social and economic effects,[8] but it is apparent that a goodly portion of the cost of regulatory reform is borne by financial services employees and their families.[9]  And those who don’t lose their jobs must work all the harder.

Fourth, the authors assume that investors will lower their required rate of return (ROE) on bank equity due to the greater safety presumably resulting from regulatory reform. In applying the credit pricing formula mentioned above, they use required ROEs of 12% in Europe and the U.S. and 7% in Japan. As a reality check, Table 2 displays ballpark estimates of actual and required ROEs for five U.S. banks that have been designated systemically important:[10] Bank of America (ticker symbol BAC), BNY Mellon (BK), Citi (C), JPMorgan Chase (JPM), and Wells Fargo (WFC). In this small sample, the rough-and-ready approximations of required ROE range from 10% to 13%, a fairly tight grouping, and the arithmetic mean is 11.5%. (These values will, of course, change as the banks announce third quarter earnings in the next few weeks.) Without leaning too heavily on the data in this exhibit, the paper’s estimate of 12% seems reasonable for at least a handful of U.S.-headquartered banks.

These notes do not remotely convey the value and interest of the IMF work paper. It is worthwhile to remark, however, that the sharp differences between the IIF and IMF-sponsored studies resurrect long-standing philosophical questions about the credibility of empirical research in economics and the other social sciences.[11] In view of the strongly partisan disagreement over the proper extent of regulation that is currently playing out in the U.S., the philosophical questions of scientific independence and objectivity have real-world consequences. We are all stakeholders. As participants in the capital markets, the most we can reasonably expect of economists is a forthright statement of their assumptions, and of ourselves, critical thinking with openness to the possibility we are mistaken. Regulators charged with presenting cost-benefit analyses to legislative bodies have the far more difficult task of persuading the opposition to accept their conclusions.

[1] Douglas Elliott, Suzanne Salloy, and André Oliveira Santos, “Assessing the Cost of Financial Regulation,” IMF Working Paper WP/12/233 (International Monetary Fund, September 2012).  Accessed October 5, 2012. http://www.imf.org/external/pubs/ft/wp/2012/wp12233.pdf.

[2] Ibid., 6, 68.

[3] Institute of International Finance, The Cumulative Impact on the Global Economy of Changes in the Financial Regulatory Framework (September 2011). Accessed October 5, 2012.  http://www.iif.com/emr/resources+1359.php

[4] Op. cit., 8. The authors avow, “It would be worthwhile to extend the quantitative analysis from this paper to take into account transitional as well as long-term effects.” 67.

[5] For example, automated data collection to monitor liquidity indicators might help the treasury department centrally manage intraday positions. See “Basel’s Proposal for Monitoring Intraday Liquidity,” in Philip Lawton, Middle Office: Managing Financial Institutions in Turbulent Times, 94-96.

[6] In addition to cost-cutting, the mitigants available to management include, among others, improving internal risk models, changing the risk-weighted asset mix, restructuring the portfolio of businesses, lowering expected credit losses by strengthening covenants and rationing credit, and reducing the asset/liability mismatch by shortening assets and lengthening liabilities. Obviously, some of these actions affect banks’ ability to serve the vital economic rôle of transforming maturities and providing liquidity.

[7] Op. cit., 59.

[8] The IIF study extends the analysis to the estimated impact of regulatory reform on real GDP growth and end-point employment levels. The findings are summarized in Table I.1: IIF Cumulative Impact Results – Comparison of Scenarios, IIF, op. cit., 10.

[9] See “The Employment Impact of Securities Regulation,” in Philip Lawton, op. cit., 73-75.

[10]  The complete list of financial institutions that the Financial Stability Oversight Council has designated as systemically important is available at www.financialstabilityboard.org/publications/r_111104bb.pdf

[11] Gunnar Myrdal’s classic study, Objectivity in Social Research (Pantheon, 1969), is indispensable. A recent publication is José Castor Caldas and Vitor Neves, eds., Facts, Values and Objectivity in Economics (Routledge, 2012).
September 23, 2012 No comments

Speed Bumps

There is a long block in a 25 miles-per-hour stretch of Robin Hood Road in Norfolk, Virginia, that has three speed bumps at intervals of about 150 feet. (They’re called “speed humps” in these parts.) Although speeding is dangerous in any residential neighborhood, there is nothing distinctive about this one.

A friend told me the speed bumps in question were constructed in the early 1970s, when a developmentally impaired youngster lived on that block. The child has long since grown up and relocated; he may have children of his own in some other city. But the speed bumps remain and will probably slow traffic for decades or generations to come. “And when folks my age move away or die,” said my friend, “nobody will have any idea why those blasted speed humps are there.”

So think about your operations. Is your group mechanically doing something whose purpose is inapparent? Does anybody remember why? If not, it may be an unnecessary step, and the process can be simplified or streamlined by omitting it. But, to be sure, don’t neglect to inform downstream operational teams before you change the routine!
September 17, 2012 No comments

Real Risk

The probability-based risk measures of conventional financial theory might be termed “idealized,” suggest David Tuckett and Richard J. Taffler. “They may even be viewed on one level as pseudo-defences against uncertainty, or real risk.” Moreover, the definitions of risk employed for analytical purposes do not reflect the actual experience of fund managers who are on “the emotional front line of the asset management industry.” Real risk, say Tuckett and Taffler, is the risk of losing clients, bonuses, and ultimately one’s job due to underperformance in a conflict-ridden role and an inherently unpredictable environment.[1]  

 Soldiers, sailors, combat pilots, firefighters, police officers, spies, drug dealers, racecar drivers, coal miners, and roughnecks on offshore oil rigs might very well offer another point of view on the true meaning of risk. Nonetheless, within the domain of financial services, Tuckett and Taffler insightfully present risk as a lived experience with significant implications for asset management firms and the investment industry as a whole. Their description and analysis of fund managers’ anxieties and coping mechanisms appreciably deepen our understanding of the investment profession.

 Taffler and Tuckett are leading theoreticians and researchers in the domain of emotional finance, an emerging discipline that differs in provenance[2] and content from the more established field of behavioral finance. Rather than focusing on cognitive biases, emotional finance explores the affective consequences of making decisions whose outcomes cannot be known in advance and take time to materialize. The emotions triggered by decision-making under conditions of uncertainty have a dynamic impact on the individual’s thought processes. Because others in the same situation have similar experiences, certain ways of feeling and thinking become normal across the industry. “Emotional finance seeks to incorporate such understanding within a formal theoretical framework that has direct practical relevance to all financial market participants and is close to their personal experience.”[3]

 On the basis of in-depth qualitative interviews,[4] Tuckett and Taffler discern recurrent features of fund managers’ professional experience. Given the relentless pressure to outperform the market and their peers, asset managers must be exceptional. (We will return to the social value of their stardom.) They have to make decisions on the basis of incomplete and ambiguous information. They know that asset valuations rest upon their own and their colleagues’ perceptions and beliefs, and they know, too, that the assumptions behind their claims about portfolio holdings are doubtful. Finally, the investment relationships they have with their assets are highly emotional; in fact, Tuckett and Taffler characterize holdings as phantastic[5] objects—“powerful psychological attractors acting beneath consciousness”[6]—with which fund managers more or less stand in a love-hate relationship. The authors state that, in combination, these themes “have the potential to lead to problematic states of mind and dysfunctional outcomes for investors and society generally.”[7]

 Tuckett and Taffler identify specific areas of concern arising from the unpredictability of investment results. These areas notably include information asymmetry, business risk, and career risk.[8] Fund managers are uneasy about the quality of the information on which they base decisions, and, sometimes, about the trustworthiness of the issuing companies’ executives. They are also apprehensive that clients, including those who say they are long-term investors, will withdraw funds and the firm will penalize or terminate them if their short-term results are disappointing.

 The fund managers’ coping strategies include selective interpretation (e.g., “A benchmark is only a reference point”) and denial through the divided-mind mechanism of knowing-but-not-knowing. In addition, storytelling is one of the most important ways fund managers deal with anxiety. Firms and fund managers construct meta-narratives to describe their fundamental investment philosophy, strategy, or process,[9] and managers tell stories to explain their successes and failures. Tuckett and Taffler do not say that fund managers invariably confabulate, but, quoting Yiannis Gabriel, they indicate that the plausibility and coherence of these narratives may be valued more than their accuracy.[10] Stories about successful investments generally fall into the epic genre, sometimes with a romantic subtext: through perseverance and insight, the fund manager prevailed in a noble quest to find an undervalued security, championed buying it, and held it through many tribulations until the market acknowledged its inherent value. Stories about failed investments tend to evoke the tragic genre, recounting the protagonist’s undeserved misfortune at the hands of villains or malevolent fate. After providing examples, Tuckett and Taffler write,

It is significant and interesting that because of the way the interviewees explained their failures through plausible stories, the failures do not appear to threaten the interviewees’ meta-narratives or underlying investment credos….This conclusion has an important implication: The market as a whole, fund managers, their investment houses, and their clients may have problems learning from experience. Storytelling, in the sense we have described, is a wonderfully flexible way of explaining misfortune and managing anxiety without threatening underlying beliefs.[11]

 Tuckett and Taffler state that the investment industry transforms superior fund managers themselves into phantastic objects. “They are the phantastic objects that their clients, employers, consultants, financial advisers, and the media unconsciously need to be superior to alleviate the anxiety they experience because of the future being unknowable.” And, they argue, “An industry that expects its foot soldiers to be phantastic objects clearly rests on problematic foundations….”[12]

 I strongly recommend Tuckett and Taffler’s fascinating monograph to investment, risk, and human resources professionals. Note, however, that recognizing the incidental psychological utility of quantitative risk measures most emphatically does not mean they are superfluous. Anyone who thinks an investment firm can be effectively managed without rigorous risk measurement techniques and robust risk management policies should have his or her head examined.

[1] David Tuckett and Richard J. Taffler, Fund Management: An Emotional Finance Perspective (Research Foundation of CFA Institute, 2012), 34, 69, 81-82.

[2] Tuckett and Taffler trace key concepts of emotional finance to Sigmund Freud, Melanie Klein, and Wilfred Bion.

[3] Op. cit., 1.

[4] The authors expound their research methodology and defend its validity in Chapter 2. (See especially pages 20-22.) They also note that interviews conducted after “meaning saturation” occurs generally do not produce new insights; rather, they tend to reinforce points that participants have already expressed. Meaning saturation conventionally occurs after as few as 15 to 25 interviews, but the authors interviewed 52 qualified participants for this study.

[5] The spelling matters; unconscious phantasies are distinct from conscious fantasies in the sense of daydreams or wishful thinking. Op. cit., 85, footnote 48.

[6] Ibid., 86.

[7] Ibid., 3-4.

[8] Ibid., 70.

[9] Ibid., 52.

[10] Ibid., 48. See also Yiannis Gabriel, Storytelling in Organizations: Facts, Fictions, and Fantasies (New York: Oxford University Press, 2000), 4.

[11] Op. cit., 67.

[12] Ibid., 91-92.
December 16, 2011 No comments

Pants on Fire

Author’s Note: Thumbing through Douglas R. Hofstadter’s masterpiece, Gödel, Escher, Bach: An Eternal Golden Braid (Vintage, 1979), I came across a two-page typescript headed “The Liar” and signed, “Philip Lawton, Springfield [Massachusetts], 1974.” I’m publishing it now, for the first time, in the hope you’ll find it at least mildly entertaining. This version uses → to represent logical implication because the Windows character map does not appear to include the proper symbol, a sideways U. Here it is:

In the fourth century B.C. a Greek philosopher named Eubulides said, “I always lie.”

Now, we can reasonably assume that Eubulides made this striking confession in all good faith to a close and trusted friend—unless, of course, he made it in bad faith to an enemy. Eubulides’ statement is one historical, and historic, assertion whose truth or falsity is absolutely undecidable, and that, not for any lack of evidence. The crucial question for his interlocutor must surely have been, “Is he telling the truth when he says, ‘I always lie?’” For the statement proves by a reduction to the impossible to be true only if it is false. Which is at the very least to say that it proves nothing at all about Eubulides’ integrity.

The paradox of the liar, now understood as the problem of self-reference, was addressed again in 1931 by the Viennese logician Kurt Gödel. Analyzing the Principia Mathematica, first published some twenty years earlier by Bertrand Russell and Alfred North Whitehead, Gödel established that a so-called meta-mathematical proof of the logical consistency of the whole of pure mathematics is impossible. While we cannot examine in any detail his proof, which consists of 64 cumulative definitions of great complexity, we can appreciate its outline and its significance by reference to the problem of self-reference.

Eubulides’ assertion, “I always lie,” may be translated for the purpose of convenience to read, “This proposition is false.” Expressed in logical notation, it says, “The proposition Gp states that it is the case that the proposition Gp is not the case,” that is, Gp: ~Gp.

Now this is clearly a contradiction; and Gödel, perhaps more critically than anyone before him, appreciated the significance of a formal contradiction: if a self-contradictory proposition is grammatically admissible in a rigorously formalized language, then literally anything follows, and the language becomes “schizophrenic”—it loses all contact with reality. This may be expressed again in logical notation as follows: “Gp and not-Gp implies Q”—whatever Q may be. Or: “If it is the case that both Gp and not-Gp, then it follows that Q,” whatever Q means and whether Q is true or false. That is to say, (Gp·~Gp)→Q.

Partially in response to Gödel’s critique of the Principia, a restrictive effort was made to identify the set of all true propositions in a given formalized language with the set of all propositions provable in that language. Reformulated in these terms, the proposition, “This proposition is false,” or, equivalently, “This proposition is not true,” might be translated again to read, “This proposition is not provable.” As it turned out, however, this restrictive move was not logically justiciable. If the two statements are identified with one another, then the following possibilities must be recognized.

Given:

  • P1: This proposition is not true.
  • P2: This proposition is not provable.
  • P1≡P2

It follows that:

  • If P2 is true, then it is not provable.
  • If P2 is not true, then it is provable.
  • If P2 is provable, then it is not true.
  • If P2 is not provable, then it is true.

What we may conclude from these exhaustive possibilities is that if we have a system rich enough to express Gp then it will contain a sentence (Gp itself) which is true if and only if it is not provable. Thus, the cozy relationship between truth and provability which contemporary logicians attempted to achieve in formal systems—namely, that the set of true sentences and the set of provable sentences be identical—has broken down: both parties to the marriage have filed for divorce on the grounds of systematic infidelity. The liar has left for parts unknown, but his lie can’t be forgotten….

Now, what Gödel proved is that any system rich enough to express primitive recursive arithmetic contains sentences analogous to Gp. Any system, then, which contains primitive recursive arithmetic either proves sentences which are false under some interpretation, or leaves unproved, and unprovable, sentences which are true under some interpretation. This, in very rough outline, is the reasoning and the statement of Gödel’s first incompleteness theorem.

Ernest Nagel and James R. Newman wrote in their classic study, Gödel’s Proof (New York University Press, 1958), that “Gödel’s conclusions bear on the question whether a calculating machine [proceeding in a step-by-step manner] can be constructed that would match the human brain in mathematical intelligence.” Observing that “there are innumerable problems in elementary number theory that fall outside the scope of a fixed axiomatic method,” they conclude, “The human brain may, to be sure, have built-in limitations of its own and there may be mathematical problems it is incapable of solving. But, even so, the brain appears to embody a structure of rules of operation which is far more powerful than the structure of currently conceived artificial machines. There is no immediate prospect of replacing the human mind by robots.” (Pp. 100-101.)

Finally, turning unexpectedly from symbolic logic to religious intolerance, St. Paul, in a perversely noteworthy passage, expressed his opinion that the inhabitants of the island of Crete were cretins. In his First Epistle to Titus, 12-13, he wrote, “One of themselves, a prophet of their own, said, ‘Cretans are always liars, wily beasts, lazy gluttons.’” And “this testimony,” he avers, “is true.” From which we may conclude—well, what?
September 16, 2011 No comments

Begging Your Pardon

It is my habit to read late into the night so as to tire myself enough to sleep soundly rather than drifting in and out of semiconsciousness until daybreak. Selecting the reading material is tricky business; I generally take care not to open a book or article that is likely to start the wheels turning again. Last night, however, I inadvertently read this sentence:

“In the large cap sample, the same frictions are present, albeit in considerably smaller measure, which begs the second and more interesting question: Why do institutional investors not at least overweight the low-volatility quintile?”

What woke me up is not “the second and more interesting question”—good heavens—but the author’s facile misuse of the phrase “begs the question.” Begging the question does not mean urgently seeking an answer.  

Because we hear this useful technical term abused so often, not only in print but in business meetings and on network newscasts, let’s recall the original meaning. In logic, an argument whose premises assume the conclusion that is supposed to be proved is an argument that begs the question.

Here’s an example: Regulators do not understand math because they are innumerate. The premise (regulators are innumerate) implicitly contains the conclusion (therefore they do not understand math).

I hastily apologize to my readers in the regulatory community, especially those with legal training, for repeating this patently unfair stereotype; it was just the first example that came to mind.

So, for the record, begging the question is a logical fallacy also known as circular reasoning or, classically, as petitio principii, a phrase that MS Word’s auto-correction feature evidently finds highly objectionable.

Now I can sleep.

Follow

Get every new post delivered to your Inbox.

Join 201 other followers