Organizations and their Effectiveness (2016)

This is a page to collect resources, links, and supplementary information related to the 2016 Summer Institute on Organizations and their Effectiveness held at the Center for Advanced Study in the Behavioral Sciences (CASBS) at Stanford University.

Please create an account (reload the page if it gives you a hard time), be bold, and add/organize/discuss as you see fit.

If you need help with wiki-markup or anything else, feel free to email Aaron.

Workshop Information

• The workshop website has more detailed information about (note that a login is required to access some resources).

Our esteemed leaders

Papers

Rules Versus Statutes and Type I versus Type II Incomplete Law

This is a random selection. More of the important ones coming.

• Baier, Vicki Eaton, James G. March, and Harald Saetren. "Implementation and ambiguity." Scandinavian Journal of Management Studies 2, no. 3 (1986): 197-212.
• Pistor, Katharina, and Chenggang Xu. "Law enforcement under incomplete law: Theory and evidence from financial market regulation." LSE STICERD Research Paper No. TE442 (2002).

Standards and Standard-Setting Organizations

• Bonatti and Rantakari (2016) on Standard-Setting Organizations, The Politics of Compromise (from Bob)
• Timmermans and Epstein's (2010) review article on sociological research on standards: A World of Standards but not a Standard World
• Ahrne and Brunsson's (2008) theory of Meta-Organizations, an organization with organizations as members
• Bromley and Meyer's (2015) book on hyper-organization, and an article (2013) that summarizes the main argument

Cooperation

• Kollock (1998) "Social Dilemmas: The Anatomy of Cooperation"

Definitions

Culture

Dan W

Definition: An enduring system of widespread beliefs and values in a given society that guides individuals’ expectations and behavior in everyday social life. (Inspired by: Geertz (1973), DiMaggio and Powell (1983), Morris (2015))

Important elements of my definition:

- “enduring” - culture can change, but it changes slowly.

- “widespread beliefs and values” - subjective notions about what is appropriate behavior and important to living a good life (in Cicero’s sense!) that are shared by some non-trivial proportion of a bounded population.

- “guides” - important that culture only guides instead of enforcing conformity. People have various degrees of attachment to a given culture.

- “expectations” - how one ought to feel about a certain action based on what is appropriate in a given culture.

- “everyday social life” - culture pervades every social interaction because it delineates the terms by which we come to a shared understanding through communication.

Example:

- Guiso, Pistaferri, and Schivardi (2015) (EIEF Working Paper 1512) — How might one detect culture? Well, here's how. The authors in this paper find that Italian adults who moved from a more entrepreneurial region (i.e., higher % self-employed) to a less entrepreneurial region (i.e., lower % self-employed) as adolescents are more likely to become entrepreneurs than the average resident of the current less entrepreneurial region in which they live.

Bob

As mentioned last week, I would distinguish between an external culture that might seep into an organization versus an organizational culture that might arise or be built within an organization. See the Martinez et al. (2015) short essay in AER P&P from the readings for my second session for more on this distinction. On the former, in addition to the Hofstede (1980) and Bloom, Sadun, and Van Reenen (2012) papers mentioned last week and cited in Martinez et al., there is also Ichino and Maggi (QJE 2000) about absenteeism by Northern and Southern Italians who move between northern and southern branches of a large Italian bank. On the latter, Schein (1985) is a key reference and gives a definition of organizational culture quoted in Martinez et al., and I would offer the unpublished paper mentioned last week by the Martinez team plus additional co-authors as first-differenced evidence: analyzing ICUs in Michigan in 2004 and again in 2006, the bloodstream-infection rate fell the most in ICUs where the nurses' answer improved the most to the statement "I frequently have trouble expressing disagreement with staff physicians in this ICU."

Network

Dan W

Definition: A non-random set of meaningful relationships distributed among individuals (or organizations, or any other purposive entities) that together do not reduce to the aggregate properties of those connected individuals (Inspired by: Scott (1991) -- the one and only textbook on social networks anyone ever needs!).

Important elements of my definition:

- “non-random” - social relationships emerge between entities in a social environment for some articulable, and often observable, reason.

- “meaningful” - a set of relationships that contributes interpretive richness to the entities connected by those relationships beyond their individual characteristics.

- “purposive” - entities connected by network relationships have their own desires which are sometimes liberated and sometimes shackled depending on the pattern, strength, and nature of the networks in which they are embedded.

Example:

- What does a real network look like with real consequences? Padgett and McLean (2006) show that a centuries-long intermarriage network between families representing different guilds after the Ciompi revolt around the time of Renaissance Florence redefined the economic and social actions of those families, giving birth to the partnership system of economic organization

Romain

Definition: a network is a collection of nodes and a collection of ties among these nodes.

Types of networks: Several common distinctions are often made when talking about networks. A network can be directed or undirected (tie from $i$ to $j \neq$ tie from $j$ to $i$), and weighted or unweighted. One can also consider static vs. dynamic networks (ties or nodes are created/disappear over time), or multiplex networks (there are several types of ties).

When talking about networks, one should always define: (1) what type of network are we talking about, (2) what is a node, (3) what is a tie. Examples: friendship network [type = undirected, unweighted network, node = person, tie = $i$ and $j$ are friends], airport traffic [type = weighted, directed network, node = airport, tie = number of daily flights from airport $i$ to airport $j$]

(Theoretical) network models: Two broad distinctions: (1) network formation vs. process on a network, (2) non-strategic [no game theory] vs. strategic [game theory].

Examples:

Non-strategic game of network formation: Barabassi and Albert's model of preferential attachment. Question: Why do networks often exhibit a hub and spoke structure [few nodes with many connections, many nodes with few connections]? Model: start with one node. At each time period, a new node is created and forms 1 tie with some old node selected at random. Old nodes with more connections are selected with higher probability.

Strategic game of network formation: Matt Jackson's coauthor's model. Question: how does a population of scholars pick coauthors? Authors choose whether to form ties with other authors. Two authors with a tie are coauthors. Players' utility increases with the number of coauthors, but busy coauthors (ie coauthors with many connections) give less benefits less than non-busy coauthors.

Non-strategic process on a network: probabilistic diffusion model. Some disease diffuses on a network. Question: which structures lead to a pandemic (everybody is infected), which do not? On a network, an initial node is infected (the seed). She infects her neighbors with some probability. At each time period, newly infected nodes infect their neighbors with some probability.

Strategic process on a network: Chwe's model of social movements. Question: which network structures lead to a revolution (every node protests)? On a network, each node has a threshold. A node protests if the number of her neighbors who protest is above that threshold.

Power

Dan W

Definition: The probability that one can carry out some desire on the basis of resource dependence, legitimacy, or normative commitment in a social setting despite potential resistance to that desire. (Inspired by: Weber (1905), Cook and Emerson (1983))

Important elements of my definition:

- “social setting” - power is only meaningful when one can exercise it over at least one other person.

- “potential resistance” - the other party in a social setting need not actively resist the use of power for power to be manifest.

- “resource dependence” - Source #1 of power: the other party depends on a focal individual for some resource, which is why the focal individual can exercise power despite the other party’s resistance.

- “legitimacy” - Source #2 of power: the other party acknowledges the widespread legitimacy of the focal individual’s power, even if the other party disagrees with the individual’s use of it.

- “normative commitment” - Source #3 of power: the other party disavows the legitimacy of an individual’s power and does not rely on the individual for some resource, but can exercise no real challenge because the individual’s incumbency is woven into the very norms that guide social action.

Example:

- All of Dan Carpenter's readings??

Mara

Power is the #1 obsession in IR, so it's difficult to formulate a one-sentence definition. But roughly, I think the various understanding of power can be distilled along two dimensions.

1) "Materialist/Agentic": in the sense that power is something agents possess and exert, having “power over”

• The most famous example here is Dahl’s (1957) definition: power is “power is the ability of A to get B to do something he or she would otherwise not do”

However, this reflects a rather limited conception since it views power as: agentic; direct (A acts on B); intentional (A means to do something to B); compulsive/coercive; and requiring a change in behavior, which assumes that the effects of power will be observable & measurable.

To address some of these limitations, we can bring in a second dimension of power

2) "Relational/Constitutive": power is something that creates agents and defines them in relation to one another about having “power to”.

• The best summary here is Barnett & Duvall (2005), who develop a 2x2 typology:
  • "Compulsory power": direct control by one actor over another in relations of interaction (Dahlian power)
  • "Institutional power": control actors exercise over others indirectly & diffusely over within formal & informal institutions—i.e. exercise of power via institutional rules that prescribe and proscribe certain conduct; actors control others in indirect ways (ex. agenda-setting power)
  • "Structural power": constitution of subjects’ identities, capacities and interests in direct, dialectic social relation to one another (e.g. master/slave dialectic, civilized/uncivilized) (Hegel)
  • "Productive power": socially-diffuse production of subjectivity in systems of meaning and signification that shapes all actors alike in their environment (not in direct relation to one another)—i.e. constitution of all subjects together through various systems of knowledge and discursive practices, and networks of social forces perpetually shaping one another (e.g. categories like “civilized” or “terrorist”) (Foucault)

Mike

This is not a formal definition, but it is illustrative of how some economists think about power. In a situation in which there are two or more people with conflicting preferences over a decision to be made, a person has "all the power" if their preferred decision is made. Compromises can of course result in decisions that are not the preferred decision of any person involved, so in many situation, nobody has "all the power." (Though according to this definition, in any single-person decision problem, the decision-maker has "all the power.") A person has "more power" if the decision made is something they prefer to the decision that would be made if they had "less power."

Sources of power in economics: (1) patience and risk tolerance in bargaining (these are the standard components of "bargaining power" in models of bargaining), (2) information (for a particularly clear illustration, see Aghion and Tirole (1997) on "real authority" in which a formal decision-maker makes my preferred decision if I am informed, she is not, and she prefers making my recommended decision to inaction or taking a stab in the dark. Models of delegation, going back to at least Simon (1951)--although I'm not sure others view his model as a model of delegation, are also relevant here.), (3) possession of legally granted control rights (see Grossman and Hart (1986) and Hart and Moore (1990) for a clean operationalization), (4) commitment power (this is akin to the "first-mover advantage," and it begs the question of where commitment power comes from), (5) rewards for past performance (see Li, Matouschek, and Powell (2016)).

Utility

Romain

Real-valued function representing preferences over choices. That is, suppose there are many alternatives $x_1, ..., x_n$. Suppose I have well-defined preferences over those choices (that is, for each pair of choices, I can say which one I prefer). A utility function is a function over those choices is function such that if I prefer $x_i$ to $x_j$, then $u(x_i) \geq u(x_j)$.

Mike

Utility is just a representation of an individual's preferences. If the preferences are well-behaved, then utility can be represented as a continuous function (as Romain describes above). We like preferences that can be represented by a utility function, because then the notion of "optimal choice" is well-defined (as long as the set of things the individual can choose from is well-behaved).

On the board, it said, "utility (with no condition)," presumably referring to the use of the word in sentences like, "Choice A gives him more utility than choice B." This is just another way to that the individual prefers A to B.

Note that preferences are defined on the set of all possible choices and not just on the set of choices that are feasible to the decision maker. Also, utility functions do not need to always be increasing--representing preferences with a utility function does not imply that "more is always better."

Rationality

Romain

I am rational when I select the choice that maximizes my utility.

Manuel

Definition: Coherent utility maximization given a certain utility function and situational constraints.

Examples: Expected utility theory (von Neumann & Morgenstern, 1944) but potentially also less obvious candidates like satisficing (Simon, 1956). Often things that look irrational at first sight are not.

Mike

Definition (formal): An individual's preferences are rational if they are complete (meaning that for any two outcomes, one is preferred to the other, or neither is preferred to the other) and transitive (meaning that if cows are preferred to telephones, and telephones are preferred to Wonder Bread, then cows are preferred to Wonder Bread). This is the standard definition that comes up in any graduate-level microeconomics textbook (such as Mas-Collel, Whinston, and Green).

Definition (less formal): An individual choice is rational if we can think of it coming from a cost-benefit analysis made by the decision maker. This cost-benefit analysis does not need to be made consciously, and this definition does not require that the decision maker not make any mistakes. It just means that the decision maker does not systematically make decisions she knows are a bad idea (in the sense that, according to whatever preferences she has, she does not systematically make decisions for which the costs outweigh the benefits). The costs and benefits of the decision do not need to be immediate, financial, or even concrete.

Tastes

Manuel

Definition: Preferences that are inherent to an individual and that are in some way exogenously given, that is their source cannot be further tracked down or their specific shape truly explained. It is therefore useless to discuss about the nature of a set of (coherent) preferences: “de gustibus non est disputandum”. We simply have to accept them as given.

Examples: Gary Becker’s (1957) taste-based discrimination model is probably the most striking (and controversial) example.

Trust

Manuel

Definition: Willingness to rely on (and make oneself vulnerable to) the actions of another person because of a belief that the other person will honor this leap of faith and act in ways that benefit the relationship (and not only according to pure monetary self-interest). Note: this definition of trust differs from "economic trust" in repeated games, in which trust is sustained via the benefits of future cooperation.

Examples: Trust game (Berg, Dickhaut & McCabe, 1995 Games and Econ Behav) or gift-exchange literature (Akerlof, 1982 QJE; Fehr, Kirchsteiger & Riedl, 1993 QJE)

Bob

I agree that it is important to save "trust" to mean something beyond the purely consequentialist logic sometimes called "calculative trust" in repeated games. Last week I called the repeated-game argument "assurance," following Yamagishi and Yamagishi (1994). See the Gibbons-Henderson (Org. Sci. 2012) reading from my second session last week for more discussion and references, as well as quarter-baked queries about whether the consequentialist logic might complement or crowd out real trust.

That said, I think trust and assurance remain hard to separate, at least for me. For example, to my eye, the first part of Manuel's definition applies to both: "Willingness to rely on (and make oneself vulnerable to) the actions of another person because of a belief that the other person will honor this leap of faith and act in ways that benefit the relationship." It is the second part that then distinguishes between the two -- namely, " (and not only according to pure monetary self-interest)" -- although I would delete "monetary" from this phrase because there are of course repeated-game arguments that are entirely consequentialist but do not involve money (such as a repeated PD).

So I guess I have moved from needing one definition to needing another: if you and I are playing a two-move game and I play "trust" in my first move (and it is common knowledge that the game is definitely over after your second move, and that we will never see each other again, and that you will never interact with third parties who knowhow you played in our game), is it the same thing to say (a) that I am trusting you and (b) that I have a view about your "type" that leads me to predict how you will act after I play "trust"?

Rationalization

Manuel

Definition (as a research approach): Explaining observed behavior via a rational utility maximization (“functionalist”) account.

Examples: Again taste-based discrimination (Becker, 1957), but also recent models of self-confidence (e.g., Bénabou & Tirole, 2002 QJE), ego-utility (Koszegi, 2006 JEEA) or reference-dependence (Koszegi & Rabin, 2006 QJE).

Legitimacy

Mara

IR scholars fight volumes of battles over this one too. Rather than try to give an overview, I focus on Thomas Franck’s (1988) piece. Franck is an international lawyer and I think this definition has some interesting implications for our discussions of formal and relations contracts.

Franck defines legitimacy as “the quality of a rule which derives from a perception on the part of those to whom the rule is addressed that it has come into being in accordance with the right process”. He identifies four indicators of legitimacy:

• ”Determinacy”= clarity, letting people/states know exactly what is expected of them. The more determinant a rule, the more difficult to resist compliance and to justify non-compliance & the less room for “flexible” interpretation

• “Symbolic validation/ritual/pedigree”= for a rule to be legitimate, it needs to be able to communicate its authority—“the authority of the rule, the authority of the originator of a validating communication and, at times, the authority bestowed on the recipient of the communication.” It does this through:
  • ”Symbolic validation”: when some kind of signal or act is used as a cue to elicit compliance with a command. The cue is a surrogate for articulating the reasons for obedience. Ex. a salute reinforces a soldier’s deference to his commander
  • “Ritual”: a specific form of symbolic validation marked by ceremonies, often—but not necessarily—mystical, that provide unenunciated reasons or cues for eliciting compliance with the commands of persons or institutions…often presented as drama, to communicate to a community its unity, its values, its uniqueness in both the exclusive and inclusive sense”. Ex. taking community
  • ”Pedigree”: “a subset of cues that seeks to enhance the compliance pull of rules or rule-making institutions by emphasizing their historical origin, their cultural or anthropological deep-rootedness”. Ex. the act of Parliament bringing a bill to the Queen for her approval

• ”Coherence”: “a rule is coherent when like cases are treated alike in application of the rule and when the rule relates in a principled fashion to other rules of the same system….requires that a rule, whatever its content, be applied uniformly in every ‘similar’ or ‘applicable’ instance.” Ex. in the US, we express this through tenets like “justice is blind” and “all men are equal before the law”

• ”Adherence” (to a normative hierarchy): “rules tend to achieve compliance when they, themselves, comply with secondary rules about how and by whom rules are to be made and interpreted” and these secondary rules, in turn, must comply with a “unifying rule of recognition” that “specifies the sources of law and provides criteria for the identification of its rules”, which is the ultimate authority (e.g. US Constitution)

Norms

Mara

Norms are intersubjective standards defining socially-appropriate behavior for a given type of actor in a given situation. They can have regulative, constitutive, permissive, prescriptive & proscriptive effects. Norms don't guarantee that agents will behave in certain ways; they only make certain behaviors more or less likely. Relatedly, norms are "counterfactually valid", meaning that specific incidences of non-compliance doesn't invalidate the norms (i.e. rules can be honored in the breach).

Russ

Social norms are informal rules that govern behavior in social groups. Honesty and reciprocity are common examples. Norms are an important mechanism for maintaining social order and facilitating cooperation. Deviation from norms is often met with sanctions, which may be informal (e.g., scorn) or formal (e.g., exclusion from the group). Robert Merton's classic article on "The Normative Structure of Science" offers a nice illustration of norms. In the article (attached), he argues that science is characterized by four norms:

Universalism---scientific contributions are evaluated independently of the person or persons who made them (e.g., anyone can potentially make a contribution to science)

Communism---scientific findings are the property of the community (i.e., there should be no private ownership of scientific knowledge)

Disinterestedness---science is done to advance the enterprise of science, not for the personal gain of contributors

organized skepticism---scientific claims should be subjected to criticism and scrutiny before being accepted

Bob

Can a repeated-game equilibrium be a norm? If so, are all norms such equilibria? How does this distinction relate to trust and socialization?

Functionalism

Russ

Functionalism (particularly structural functionalism), within sociology, is a perspective that social phenomena in terms of the function or purpose they serve. The perspective offers a very static picture of society. Moreover, it can be used to justify undesirable social phenomena (e.g., inequality) and therefore has been sharply criticized. The article, "Some Principles of Stratification" by Kingsley Davis and Wilbert E. Moore (1945) offers a good example. Davis and Moore argue that stratification (i.e., social classes) is functionally necessary in society to motivate people to do things that are important but would otherwise be undesirable, e.g., medical school: "Modem medicine, for example, is within the mental capacity of most individuals, but a medical education is so burdensome and expensive that virtually none would under- take it if the position of the M.D. did not carry a reward commensurate with the sacrifice."

• Davis, Kingsley, and Wilbert E. Moore. "Some principles of stratification." American sociological review 10, no. 2 (1945): 242-249.

Voice

Russ

Voice, as defined in Albert O. Hirschman's book, Exit, Voice, and Loyalty (attached), is one of two options people (e.g., employees, customers, citizens) have in response to the declining performance of their organizations (e.g., employers, firms, states). In contrast to the other option, exit, in which a person dissolves his or her relationship with the organization, voice is an attempt to repair the relationship through criticism and feedback. Voice is potentially more informative for organizations because it gives reasons for declining performance, whereas exit is just an indicator. The book focuses on (among other things) determining when exit or voice is most likely to be used (e.g., when options for exit are higher, voice is likely lower; when people have loyalty to the organization, voice is likely higher).

As an interesting bit of context, the book was written while Hirschman was a fellow at CASBS: "This is an unpremeditated book. It has its origin in an observation on rail transport in Nigeria which occupied a paragraph in my previous book, reproduced here at the start of Chapter 4. One critic objected to that paragraph because, as he charitably expressed himself, 'there must be a lot of assumptions hidden there somewhere.' After a while I decided to pursue these assumptions into their hiding places and was soon on an absorbing expedition which lasted the full year that I had planned to spend in leisurely meditation at the Center for Advanced Study in the Behavioral Sciences...The Center provided a particularly favorable environment for this sort of project. I made ample use of the 'right to buttonhole' the other Fellows which is, I believe, an essential part of the oral tradition at the Center. My intellectual debts to those who spent the year with me are generally acknowledged in footnote references. Special gratitude is owed to Gabriel Almond who contributed important critical points while being permanently supportive of my enterprise; to a comment by Richard Lowenthal that led me to write Chapter 6; and to Tjalling Koopmans 'who helped sharpen some of the technical arguments, as did Robert Wilson of the Stanford Business School."