Monday, August 29, 2011

Reasoning about ourselves and our organizations

In this blog post: The tendency for individuals to reason about themselves has arguably increased. Is the same true for organizations ? Top managers are told to ask themselves 'what is this firm about', to become 'paranoid' about what their firm is and could be. Has such anxiety increased? Was there a time when firms had a fixed organizational identity that has now come to pass? 

The modern question: Who am I?
Anthony Giddens has suggested that 'the late modern age' we live in is distinct from the prior times in terms of how we think of ourselves. While previously our identities (who we consider ourselves to be) were defined by our family and our role in the society. Now, we constantly 'try on' different identities, burdened by the knowledge that whatever we are is just one choice out of many possible ones. Identity is a key topic we reason about: We consider whether our observations and 'facts' justify the self-conception we have, and we attend to reasonable implications 'justified' by the self-conception we have chosen. We use our identity to reason about what to do. 
A person may take refuge in a traditional or pre-established style of life as a means of cutting back on anxieties that might otherwise beset her. But, for reasons already given, the security such a strategy offers is likely to be limited, because the individual connot but be conscious that any such option is only one among plural possibilities. (Giddens, 1991: 182).
In late modernity, there is no choice but to reason about who we are, a source of burned and anxiety (and freedom, one may say). Does this apply to organizations? 

Reasoning about Organizational Identity
Just as our conception of ourselves is our identity, the conception an organization has of itself is its organizational identity. Organizational identity is pretty close to what many would understand to be strategy, but I'll use the former concept in order to keep with the argument by Giddens discussed above.Even though Giddens has no interest in organizaitons, we could make analogous argument: in early modernity, organizations were not too concerned to reason about themselves, in late modernity organizations are anxious to consider and reconsider what they are about.

In terms of reasoning, this thesis would mean that the range of beliefs subject to reasoning is broadening. In earlier 'simple times', managerial reasoning incorporated simple premises: how can we do whatever we are about in a way that creates growth or profits? Changes in identity emerged from 'diversification', which added new elements but did not raise thorny questions about the legacy. In late 'complex times', the complexity of reasoning is increased exponentially because the very premises of these prior questions are also subject to reasoning: Should we be this or that? Is there an identity we could assume that we are not aware of? Such complexity can make anyone nervous. Only companies with large irreversible investments are safe from questions concerning the optimality of their current business. 

In conclusion
This blog is just a thought experiment, leading to a rather dull-sounding proposition that 'things have gotten more complex'. But this line of thought may also suggest that managers are becoming increasingly anxious and paranoid. More generally, this implies that the range of topics managers reason about are defined by the broader societal context.  

My colleague Saku told a story behind the book. Giddens got remarried, and his new wife read self-help books. When spending time on the toilet seat, Anthony started browsing these books. They got him thinking about the constant quest to conceive and reconceive the understanding of self, which he associated with late modernity.

Giddens, Anthony. 1991. Modernity & Self-Identity: Self and Society in the Late Modern Age. Polity Press.

Sunday, August 21, 2011

Homes for Africans: How reasonability helps understand social institutions

I'm going to tell a story about David, a smart university graduate who volunteered to help build homes in Africa (and now a CEO of his own firm). This story helps elaborate my own stuff on the role of 'discursive institutional work' in creating, maintaining, and disrupting the prevailing social order.

Warning: This post marks the first instance of explicit promotion of my own work on this blog! 

Selling Houses in Africa
While working at Imperial College London, I met an entrepreneur running an IT company, whom I had to give a guest lecture. While we were chatting, he told me had gone to Africa immediately after he had graduated (from Cambridge) to work as a volunteer. He got involved in a big charity building proper houses for the poor in developing countries, using money and volunteer workers from the West. In really poor countries, 40% of energy is consumed by households, a lot of it goes into heating during winter nights because insulation is terrible. The guy was pretty smart, and he soon figured out that the scale of their activities was non-existing. If he could delegate the house-building to entrepreneurial Africans who would do it for profit, they could leverage the donated money and accomplish a substantive improvement. Moreover, selling subsidized houses would allocate the resources effectively for those families most willing to invest (with safeguards in place to make sure the new tenants were not wealthy). Because continued flow of the Western money was not guaranteed, a for-profit-philanthropy hybrid would lead to a more sustained impact. Great idea, so lets do it?

Not so fast. In the 90s most philantrophists were not overly excited about turning volunteer activities into a franchised business. They hated the idea. The reasoning was pretty straight-forward: entrepreneurs making money out of the poor Africans was wrong. Moreover, because nobody else was doing it, it had to be a bad idea (this is a common element in practical reasoning and probably a very good rule of thumb usually). Today, of course the salient reasoning would be quite different: creating entrepreneurship in an African country is a great way to boost its economy and the well-being of the people. The idea of subsidizing entrepreneurship as a form of philanthropy had not yet been institutionalized, the business plan was "rationally speaking" as good in the 90s as it is today, but the plan did not correspond to an existing social institution. Anyhow, during his 3-year stay in Malawi, the charity became the country's largest home builder.

After Africa and some other charity work, David was dead-set on becoming an entrepreneur. He went to do the MBA at Imperial and got involved into a venture while there. He is now the CEO of his IT firm, a portfolio company of the Imperial Innovations.

Social Institutions
Organization theory is hugely occupied with social institutions -- basically, a term usable for any durable, widespread, observable regularities in behavior. Institutions can be anything: the ubiquitous quality management systems, the tenure system in universities, sales commissions, whatever. Why such preoccupation? Because a lot of behaviors and other stuff taking place in companies (and more broadly in industries and countries) does not seem to result solely from rational decision making or technical concerns. If we can explain rational/purposeful decisions in organizations and the social institutions around those decisions then we have a pretty good understanding of what is going on.

Reasonability and Institutions
The notion of 'institutional work' captures the work done by actors, such as David, to create, maintain, or disrupt social institutions. This theory approach is founded on the notion that practices, entities, or social arrangements do not 'sell themselves' to customers, potential employees, government regulators, or the media. For example, when someone seeks to shape the attitudes or regulations concerning immigration or social welfare they are engaging in 'institutional work'. Institutional work is distinct from 'selling' because acceptability ('legitimacy') gained through institutional work tends to help everyone equally. There tends to be a freeriding problem.

This is a topic of a work (see below) I did with Saku Mantere and Eero Vaara. Our commentary suggests that the theory on institutional work -- by focusing what is done or said rather than the cultural context in which things are said -- has tended to ignore the discursively articulated reasoning around the social institutions in question. Thus:
We argue that reasonability plays at least three crucial roles in institutional work: It provides the main contextual constraint of institutional work, its major outcome as well as the key trigger for actors to engage in it.
These are very basic observations, but they are things we should at least control for when explaining why some actors manage to promote novel social order while others fail. These observations also help explain why discourse matters a lot when firms try to shape their industries or when managers try to change the ways things are done inside their own firm. Institutional work involves the creation and circulation of credible justifications to social institutions. On a meta level, the broader background assumptions (propagated by the Media) define how things can be justified in the first place -- for example, what arguments can be used to defend or attack immigration (an unfortunately hot topic in our increasingly xenophobic Finland).
1. Reasoning as a constraint: Any claims put forth by actors need to be justifiable with acceptable reasons, must have some implications that recipients comprehend, and need to avoid clearly acceptable reasons for their refutation.
For David, the salient beliefs donors and volunteers used to reason about philanthropy mattered a lot. The evaluation criteria and beliefs African regulators used to reason of foreign organizations would also define the conclusions these regulators would draw should the organization experiment with alternative approaches. What matters is how premises (the organization is now doing X) leads to conclusions (the organization is "not philanthropic") in reasoning, not just whether the actual or expected measurable outcomes from X are desirable.
2. Reasoning as an outcome: When actors engage in discursive work, they define and refine the generally accepted conditions for beliefs—the reasons why actors ought to believe one thing or another.

If David would engage in successful institutional work and shape how philanthropists perceive the franchising organization, it would likely lead to a broader change in how they reason -- what type of conclusions they draw from any philanthropic 'business model' that involves local entrepreneurship. Institutional work would make employment and economic growth more salient topics  to reason about.

3. Reasoning as a trigger: When new events and outcomes contradict the existing, accepted linkages between premises and conclusions (the established way of reasoning), institutions can only remain reasonable and legitimate if 'maintenance work' again makes them reasonable. We gave the example of the financial crisis in 2007, which questioned the very reasonability of the existing financial order and the premises on which it had been considered legitimate. As always, various experts were ready to provide reasoning to support the system.

The current story is the Euro crisis. Although the common currency has had many reasonable justifications going for it, the near-default of Greece and the continued crisis with Ireland, Portugal, Spain, and now even Italy and France, is suggesting that the reasoning behind Euro may have been flawed. To retain its legitimacy, the proponents of Euro must engage in what we call 'discursive maintenance work' that diminishes criticism and fears raised by recent events and provides solid line of reasoning to support Euro, set of reasons that must be immune to the recent evidence against the viability of Euro. 
We also pose some research questions in the rest of our brief paper that relate to the role of various actors (such as professions) in creating and maintaining reasonability of practices and arrangements. 

To say some some behavior has reasonability means two different things, which I here call weak and strong reasonability (terms I conveniently invented tonight). Weak reasonability means we are able to provide reasons for that behavior; we can comprehend and evaluate the behavior because there are some reasons for it. A merger lacks reasonability if we cannot understand what the purpose of it is, or how it will change things (to the better or the worse). Strong reasonability means that the justifications are broadly acceptable and aligned with the broader beliefs and interests of the actors involved. A merger may be weakly reasonabile if those who decided on it had clear goals, but still lack strong reasonability (or be 'unreasonable') if reasons exist to discredit the justifications. For example, if the costs of the merger exceed the benefits provided by its goals or if there reasons exist to assume that the goals of the merger cannot be accomplished.

In our paper we say that reasonability is "the existence of acceptable justifying reasons for beliefs and practices". It is not clear that we (or at least I) had in mind the first, weak, sense of reasonability. Social institutions may be sustained to the extent they are weakly reasonable, they can be justified in some comprehensive manner by someone. That does not mean that they are strongly reasonable, that there wouldn't be a plausible way to discredit the social institution. After all, the U.S. has a law about debt ceiling that is justified by some pretty good arguments. However, in a broader sense the law is really bad and even the justifications do not really stand to closer scrutiny. The debt ceiling is a weakly reasonable arrangement (social institution, if you will), but it does not really seem to be strongly reasonable.

The story of our paper
The piece we got published on reasonability and institutional work is not really an article. It is an invited commentary in a special issue. Journal of Management Inquiry was publishing a special issue on Institutional Work -- a new research program that seeks to understand how actors create, maintain, and disrupt social institutions (mainly, how they shape the acceptability of things such as the franchising model for philanthropic construction company). The special issue had a big 'agenda setting' paper by Lawrence, Suddaby & Leca, and some invited pieces by various academics. Because the comments were mainly from Americans, the editor thought we could do a more European type of thing with Saku & Eero.

I knew immediately I wanted to write on reasoning. Luckily Saku was keen on this as well, bringing in the notion of linguistic division of labor (Saku is more knowledgeable on Hilary Putnam, whom I've just read a bit recently). While Eero had his own brilliant ideas as well, we needed to retain some focus so he concentrated on just making our initial ideas much better.We didn't have too much time to fine-tune the commentary but I am pretty pleased about how it turned out. Now I just need to finish writing the proper article on reasoning & social institutions.

The most imporant thing: Reference
Schildt, H.A., Mantere, S., Vaara, E. 2011. Reasonability and the Linguistic Division of Labor in Institutional Work. Journal of Management Inquiry. March 2011, Vol. 20: 82-86.

Monday, August 15, 2011

What is management knowledge and how do B-schools create that?

This blog post discusses how reasoning can shed light to what 'management knowledge' is in order to clarify what makes management research and management education relevant to managers. My interest in philosophy and psychology of reasoning inevitably got me to consider  how managers think about the issues they face. Despite plenty of reflection by management scholars on their practical relevance, the lack of any attention to the ability of managers to think is striking. When a manager is faced with a lot of information (e.g. market share is dropping, a competitor has introduced new products, and key sales people have resigned), managers will try to conceive reasons for these observations, and to come up with with different actions justified by seemingly good reasons.

Reasoning about an investment
Let's think of a manager who must decide whether to invest into an expensive IT project to improve customer satisfaction. Now, one may approach this decision using a simple financial model: if the NPV of increased profits is greater than the NPV of project costs, that is a pretty good reason to make the investment. One of course will then need a way to reason about the costs and benefits. Alternatively,  a manager who has taken a strategic management course might think in terms of capabilities. Can my competitors imitate or substitute the resource I am seeking to develop for a lower cost than I am investing? The IT system may not be a necessary resource, the first mover advantage may be insignificant, and the cost for competitors to copy the system may actually be lower than the initial cost of developing it through trial and error. If these three considerations are all true then there are  pretty good reasons not to invest.

Finance and strategy provide competing ways to reason about the decision. Both of them include some rather explicit inferences (e.g. formula to calculate NPV; the rule "if 2nd mover advantage is greater than 1st mover advantage then do not be the first mover"). These "ways of reasoning"  also require a lot of tacit knowledge (e.g. how to come up with an estimate of costs associated with an IT project; how to figure out the imitability of a potential new capability).

My hypothesis is that wise and stupid (is there a reason why the concept 'stupid' is absent from all theory?) managers are mostly distinguished by substantive (practical) quality of the reasons they conceive for their actions and choices rather than the logical validity of their thinking. Although cognitive psychologists have found clear and stable individual-level differences in the ability for critical reasoning, the lack of attention to imitability or NPV calculations is more likely to distinguish good and bad managers than the number of logical mistakes they make. Philosophy of mind would have us believe that reasoning has in general relatively  little to do with formal logic and a lot to do with the meanings we hold (I believe this view is advocated by Davidson and also John Searle in Rationality in Action - a book that is all about reasoning and very little about typical rationality).

Management education
The above example suggests that management education should enable students to reason about situations they will encounter at work. Managers ought to have the ability to utilize explicit knowledge in their reasoning (e.g. finance formulas, the VRIN criteria of the RBV -- if you believe in it). However, they should also have tacit skills to reason well -- to know what kind of things make resources imitable or substitutable. Is it not a bit strange that probably more than 95% of basic business strategy courses on this planet teach students that  inimitability is a necessary reason to assume a resource to be a sustainable source of value, yet less than 5% of the courses probably address at all how the imitability of a resource might be inferred!

It may seem quite uncontroversial to say that management education should teach people how to think about business problems and business decisions. Yet, there has been a substantial movement towards evidence-based management. This perspective suggests that management is no different from medicine and the key goal of management education should be to provide the right answers (I should mention that evidence-based medicine has been criticized quite a bit too, so medicine may not be about the correct answers either). The idea is that good management research figures out the right answers to problems faced my managers and that management education communicates these answers to people. While this approach may work in many areas, it seems to me that a lot of problems faced by managers are idiosyncratic and require non-trivial amount of reasoning to figure out. Once some management scholar has evidence on whether firms in industry X should  implement the  IT system Y, the window of opportunity has surely already closed.

If reasoning about problems and solutions is what managers do, then  management education needs to provide managers with theoretical concepts (such as imitability, first mover advantage, net present value) and teach the managers how to competently utilize those concepts.

Relevance of management research
I am working on a paper concerning the practical relevance of management research with my colleagues at Hanken: Saku Mantere and Eero Vaara. The paper is based on the conception of management knowledge as explicit and tacit ability to reason using a certain set of concepts. I am slightly concerned about giving out our argument in this blog because there are no clear rules -- at least in principle the journals want to only publish arguments that haven't been published before. In practice, it would probably make little difference if I published the conclusions of our main arguments in this obscure blog.

One conclusion we'll reach that I am particularly fond of concerns the practical relevance of individual articles. Since evidence-based management has limited applicability (basically answering only questions common to all companies and being very limited in its ability to help in thinking about idiosyncratic problems), we should not consider articles with clear prescriptions as more "practically relevant" than studies without prescriptions. In our mind, an article has practical relevance if it provides explicit or tacit examples of how to use theoretical concepts to think about real-life problems in a productive manner. For example, research on organizational identity has provided us a very useful concept that helps understand and solve problems in organizations even though the research provides almost no  prescriptions (it would be silly to expect evidence-based rules such as "companies in industry X should develop an organizational identity with characteristics Y&Z to improve profits").

Monday, August 8, 2011

Can a hedge fund eliminate human biases in reasoning?

One would imagine that organizations that make money through superior reasoning would develop practices and processes that reduce suspectibility to biases. And it turns to be so! The example is provided in recent New Yorker article on Bridgewater, the richest and strangest hedge fund on the planet. The owner of the fund, Ray Dalio, takes biases very seriously. The attempts of the organization to keep human tendencies at bay make the organization seem like a cult. Reporting by John Cassidy, quoted from New Yorker (within fair use):

Dalio asked for another opinion. From the back of the room, a young man dressed in a black sweatshirt started saying that a Chinese slowdown could have a big effect on global supply and demand. Dalio cut him off: “Are you going to answer me knowledgeably or are you going to give me a guess?” The young man, whom I will call Jack, said he would hazard an educated guess. “Don’t do that,” Dalio said. He went on, “You have a tendency to do this. . . . We’ve talked about this before.” After an awkward silence, Jack tried to defend himself, saying that he thought he had been asked to give his views. Dalio didn’t let up. Eventually, the young employee said that he would go away and do some careful calculations.

One day, I drove to Westport and sat in on a management-committee meeting, which had been set up for the purpose of “getting in synch” with a recent recruit, whom I’ll call Peter and who had come from a big financial firm. All nine members of Bridgewater’s management committee were sitting at a long wooden conference table. Peter, a lean man with fair hair, sat stiffly near the front: he looked like somebody anticipating a root canal. Jensen and McCormick were nominally in charge, but Dalio took over, telling Peter that, during a previous management meeting, he had answered emotionally in response to questioning from Jensen. “This is a common thing when somebody’s getting probed,” Dalio said. “Because the amygdala gets stimulated and you have that emotional reaction.” Peter agreed that he had become upset, especially when he sensed he was being accused of misleading his colleagues. “I felt in some sense my integrity was being attacked,” he said. “That’s when things spiralled out of control.” 

Dalio walked to the front of the room, where he wrote on a whiteboard, “FELT,” “INTEGRITY,” and “MISLED.” “?‘Felt’ is the key word here . . . and it’s a challenge for people,” he said. After a bit more discussion, he went on, “What we’re trying to have is a place where there are no ego barriers, no emotional reactions to mistakes. . . . If we could eliminate all those reactions, we’d learn so much faster.” (my emphasis)

Do social processes set up in organizations trump psychology? Or is Dalio's attempt to make humans non-human a futile undertaking? I believe we have good evidence that social contexts do matter, and the social conventions and patterns of conversation to some extent override psychological tendencies when people reason discursively. But there are limits to everything and Bridgewater seems a bit freaky (you can check their 'Principiles' out; they might be handy if you teach org. culture). But they are by some acocunts the most successful organization in their field, and one cannot rule out the possibility that it is because they have engineered a solution that might be accounted for through sociology of reasoning.

Many organizational processes might help actors overcome individual-level 'deficits' in reasoning, such as those identified by behavioral economists. For example, by simply requiring managers to provide explicit justifications (verbal or numeric) for choices they make can reduce the use of biased heuristics associated with non-reflective reasoning. On the other hand, interaction situations tend to lead to all kinds of other unproductive patterns in reasoning (e.g. social desirability bias). There are nice lists of individual-level biases (e.g. by Stancovich, ref at the end), but I am not aware of a good summarizing review on organizational or group-level effects on reasoning.

The passage quoted above is a pretty dystopic vision of future workplace, is it not?

Stanovich, K. E. 2009. 'Distinguishing the reflective, algorithmic, and autonomous minds: Is it time for a tri-process theory?' in In Two Minds: Dual Processes and Beyond. (eds), 55-88. Oxford, UK: Oxford University Press

PS. Thanks to a visually giften friend of mine, I got some tips on making the blog more readable.
PPS. I accidentally called Bridgewater 'Blackwater'. Bridgewater is the hedge fund run by Dalio. Blackwater is the American mercenary group (now known as XE) where employee recreation allegedly involves murdering civilians.

Thursday, August 4, 2011

Self-deception & reasonability

Work on reasoning has traditionally been very rationalistic. Researchers who focus on ways through which actors link beliefs or observations as justifications for further beliefs and actions would like to assume that the consistency that tends to characterize reasoning is universal for human beings. Yet, the real reasoning processes and outcomes are somewhat 'irrational' from the "objective" perspective of the logician.

Davidson on Self-Deception
Irrationality is particularly interesting for analytical philosophers who begin to explain mind and language from assumed logical coherence. I've been recently reading Davidson's essay collection (actually, in four volumes). Donald Davidson is absolutely brilliant and worth reading just for the elegant style of his argumentation (I suppose dummies are mostly not elected presidents of the American philosophical association). His late essay "Who is Fooled?" addresses the question of self-deceit -- how is it that people hold beliefs they know to be contradictory?

The essay begins with the story of Ronald Reagan and George Bush claiming they did not know that the U.S. was offering Iran arms in exchange for hostages. The former Secretary of State Schultz claimed that while G.B. consciously lied to the public, R.R. did not because he had "lied to himself", deceived himself to believe that he did not actually know of the deal (although he had been present in the meetings). What does it mean for R.R. to deceive himself to the effect he was "not aware" of the deal even though he knew the facts? Davidson also brings the example of Columbus claiming he was the first to spot land when reaching America, although he clearly was not (it was the lookout) and most likely knew it. This behavior "has been variously interpreted by some as naked, mean greed, by others as honorable self-deception, born of the arrogance and lust for flame" (D.D. cites F. Fernandez in The London Review of Books). Whatever the initial reason, we know intuitively that some people genuinely begin to believe in propositions they should reasonably reject (ps. D.D. recognizes cognitive dissonance literature as well, something I will not go into here right now).

How to define what is rational and what is irrational?
Davidson takes a humble approach and draws a conclusion on our ability to theorize reasoning (p. 218 in "Problems of Rationality"):
At this point someone is sure to ask who is to be the judge of rationality and consistency. The annoying answer is that this is a bad question, a question without an answer. There is no eternal, absolute standard. At the same time, we are not thrown back on your standards or mine; relativism is not the only alternative to standards independent of all though and judgment. It is clear that in evading the question when a set of attitudes can be recognized as inconsistent, we are quickly driven back to basic logic; there comes a point in which intelligibility is so diminshed by perceived inconsistency that an accusation of inconsistency loses application for lack of identifiable contents about which to be inconsistent.
This insight echos Davidson's truly ingenious resolution of Descarte's problem of scepticism (his essay "Coherence theory of truth"): We are pretty inconsistent in terms of absolute rationality, but we must be quite rational in order for anyone to be able to even point out a logical inconsistency. The fact that we can say someone to be deceiving themselves is only possible if most people most of the time do not deceive themselves: if we stick to conclusions that are warranted in some intersubjectively valid manner by beliefs we endorse as true. WHOAH! Such a logically consistent and clear point, but one that inevitably only gets made by a proper philosopher with a remarkable mind.

The Self-Deception of Madame Bowary
The essay ends in a show of remarkable intellectual breadth, where Davidson elaborates on Flaubert's Madame Bowary.  The protagonist of the story engages in ever-deepening self-deception, going through great deal of effort in order to retain a distinction between the true conditions of her life and the romantic dream image of hers she maintains to motivate herself. Davidson suggests that the key to self-deception is the effortful avoidance of considering the contradictory beliefs at the same time.

I may offer a trivial reading based on modern cognitive psychology. Some research (oh the refs escape me) suggest that all reflective reasoning involves the retrieval of beliefs and memories into working memory (the bit of the mind we are conscious of). Working memory has very limited capacity (you know the popular claim we can only hold "seven things" in our minds at the time), and unless we retrieve our dream-beliefs and the facts that contradict them into our working memory at the same time, we may just be able to deceive ourselves. On the other hand, real self deception might require us to do away with the most basic norms of rationality we have been socialized into. But wouldn't that appear as outright insanity?

And What about Managers?
As you can see, I've been clearly inspired by Davidson. Since this blog is about reasoning in organizations, I might pose the following question: Do managers regularly engage in self-deceit? Do organizations encourage them to do so, or are there perhaps mechanisms in organizations that help prevent or rectify self-deceit?

PS. Fun fact: before turning full-time philosophy superstar, Davidson went to Harvard Business School. His account: "I've always been glad I went to business school because it gave me an insight into how a lot of people think that I would have never known otherwise. And I liked the feelign that I could have done it. But I wouldn't have liked the people. After the war they said come back for a month and you could get the degree. I didn't go back."