In March 2016, the largest Danish real estate credit union Nykredit/Totalkredit sent out a letter to all its clients/members informing them that the contribution margins on their loans would be raised, i.e. the fee the members are to pay Nykredit/Totalkredit for administering their loans. Needless to say that the letter spurred discontented reactions among the members. Of interest here is how the reaction was made possible by virtue of the available information structures. The available information channels were incorrectly construed and thus fallaciously utilized by the sender, Nykredit/Totalkredit. This was in turn advantageously put to use by the recipients culminating in a corollary of great distress for the credit union.
1.0. Misunderstanding the Structure: The Pigeon Hole (PH)
The letter of individual notification, which Nykredit/Totalkredit chose to use as a messaging system for informing its clients/members regarding the increase in contribution margins may graphically be represented in an umbrella-like figure: One sender of one piece of information to many, independent members. Metaphorically, one may refer to this message system as the pigeon hole (PH) (Figure 1)
Figure 1. The Pigeon Hole. The information structure where one node sends information to many nodes.
A prerequisite for the PH structure is that multiple receivers cannot interact with each other while receiving the individually addressed letter. Members are separate receivers of information, and dispersed in such a way that it is not likely that communication between them occur immediately. As such, the receivers are an aggregate conglomerate connected to the sender on a one-to-one basis.
The PH information structure may be used with expediency when identical pieces of information need to be communicated to the corresponding aggregate conglomerate, denoted a cluster. A cluster is often homogenous because all its members, qua membership, have the same status and are hence largely qualitatively non-differentiable. Periodic updates such as newsletters or software tweaks are often effectuated by organizations through the PH structure when relatively harmless or easy-digestible information is be communicated or downloaded by clients.
However, this mode of communication does not actually correspond to the type of organization that Nykredit/Totalkredit is. It is not an organization whose clients or members are dispersed to such a degree that they refrain from communicating. It is an association (or at least was until recently) where its members are gathered around a particular area of interest, in this case loans and mortgages. Members of an association are, by the very nature of being associated, connected through a cause, a profession, theme, interest or hobby (e.g. medicine, charity, politics or boat-sailing). Not only do members have an established and apposite interest (in their agglutinated association) to exchange opinions and viewpoints pertaining to their common cause or interest, but they certainly also have knowledge about one thing: That every other member of the association has taken out a loan or mortgage with this real-estate credit union. By logical consequence this is merely the definition of what is means to be a member or client of this very union or association.
2.2 Common knowledge
If members therefore receive a letter in the capacity of their membership, as mortgagor, then they know that every other member correspondingly possess this knowledge as they also received the letter, as well as they know that all other members know that they know (i.e. possess this type of knowledge) and so on. This reciprocal exercise qualifies as common knowledge: The common element is that everybody knows that every other member knows that every other member possesses information relating to the forthcoming increase in his or her contribution margins.
In an interconnected network, common knowledge is valuable as it makes the network informationally transparent with respect to known facts among the agents. Common knowledge assures this transparency of knowledge on all higher levels as opposed to a network in which everyone simply believes that every other member also thinks the same thing; or as opposed to a network in which knowledge is mutual (Figure 2) among the members but without the transparent higher order element that common knowledge (Figure 3) implements.
Figure 2. Mutual knowledge of proposition p for agent a and b but without a knowing that b knows p.
Figure 3. Common knowledge of proposition p for agent a and b with a knowing that b knows p, b knowing that a knows p, a knowing that b knows that a knows p, and so forth for all higher order levels.
Say that knowledge is mutual in a network if every agent knows the same-some proposition p (Figure 2). That every agent knows p does not reveal nor entail anything about what the individual agents know about the knowledge of other agents, their potential knowledge of some proposition p or their knowledge of knowing p and so forth. This shortage of higher order commonality, assurance and conjoined transparency make networks of mutually knowing agents less informationally potent, and powerful, than they would have been had they enjoyed common knowledge. An authoritarian CEO or statesman is rarely ousted even when everyone desires him or her dethroned if there is a lack of common knowledge of this desire in the group. Every agent wants to get rid of the leader but does not know that everybody else feels the same way about her nor knows that everybody shares this sentiment.
The lack of common knowledge makes coordination of a joint effort to get rid of the leader tricky as members may be in doubt as to what others think, and mutual sentiments of repugnance do not suffice for joint action. The same goes even when it comes to distributed knowledge as the aggregate knowledge of the entire population still falls short of a ticket to joint action given the depletion of higher order information about knowledge of knowledge of other agents and the adherent scarcity of transparency and assurance.
The successful implementation of organizational strategies in a firm, the mobilization of masses in politics, religion and businesses as well as the possibility of collective deliberation, decision and action is in no small measure reliant on the imparted or stored knowledge that is available within a population: Mutual, distributed or common knowledge.
Given the similar interests of the members or mortgagors in the Nykredit/Totalkredit case, one may expect a certain aligned joint reaction to the announced increased contribution margins. Each member may voice their individual dissatisfaction; one signal of dissatisfaction per member may be returned to Nykredit/Totalkredit but that does not make for a bull-horn joint voice to be heard by the association; neither by the board of directors, not by the other members and mortgagors nor by the public. The disparage knowledge network needed, so to speak, an outlet or a platform which could collect all the voices, and distribute the common knowledge of dissatisfaction to other members of the network and coordinate a collective response and joint action back to the source of the problem.
Facebook was the answer; a single member created a Facebook group for all mortgagors of Nykredit/Totalkredit to join. Not only could all of the 500.000 clients or members thus communicate – getting further assurance that almost all members of the association too were dissatisfied with the advertised increase, but every member could also invite their Facebook friends, who did not belong to the internal network of the Nykredit/Totalkredit membership quota. In this way, members organically organized a rapid extension of the network through Facebook by conjoining non-associative networks: These networks however still shared the discontent for whatever the reason – sympathy with their friends or fury over the greed of the financial world.
Figure 4. The mobilized discontent assembled on a Facebook-enabled platform via common knowledge about the discontent among members and “friends” of members.
This self-organizing mobilization broke fiercely with Nykredit/Totalkredit’s PH strategy that informed its members on an individual basis. Yet, at the same time, this strategy was the initial source which enabled common knowledge to permeate the system: Clients/members were informed in distinct, but not unconnected information spaces. The connection, to be sure, was simple deductive reasoning and the sufficing channel of a single information-mobilizing valve: The Facebook-group for discontented clients/members. Through this group the rest of the members were mobilized along with the members’ networks as well as media, journalists and other interested parties who suddenly raised the case to a level of public awareness until the politicians also got involved.
What should have been a one-to-one correspondence became in fact one-to-many communication, which mobilized itself from many-to-even-more, and returned a signal of discontent of a magnitude never expected which again was followed up by general political reverberation. (Figure 5). That’s how the pigeon-hole message box landed right on top of the sender.
Figure 5. The aggregated information signal sent back from the Facebook-platform to Nykredit/Totalkredit, press/media, politicians, other stakeholders . . .
Featured image courtesy of Flickr.