Wednesday, 19 November 2014

HOW SOCIETY WORKS:Social order by self-organization

by Dirk Helbing[1]


The invention of laws and regulations is celebrated as great success principle of societies and they are, of course, important. However, a major part of social order is based on self-organization, which builds on simple social mechanism. These mechanisms have evolved over historical times and are the basis of the success or failure of civilizations. Currently, many people oppose globalization, because traditional social mechanisms fail to create cooperation and social order under globalized conditions that are increasingly characterized by homogeneous or random interactions. However, I will show that there are other social mechanisms such as reputation systems, which will work in a globalized world, too.


Since the origin of human civilization, there has been a continuous struggle between chaos and order. While chaos may stimulate creativity and innovation, order is needed to coordinate human action, to create synergy effects and greater efficiency. Our success also critically depends on the ability to produce "collective goods", such as our transportation infrastructures, universities, schools, and theaters, as well as language and culture.

According to Thomas Hobbes (1588-1679), civilization started with everyone fighting against everybody else ("homo hominis lupus"), and it required a strong state to create social order. Even today, the merits of civilization are highly vulnerable, as the outbreak of civil wars illustrate, or the breakdown of social order after a natural disaster. But we are not only threatened by the outbreak of conflict. We are also suffering from serious cooperation challenges, which often arise out of "social dilemma situations."


The challenge of cooperation


To understand the nature of "social dilemmas," let us discuss the problem of producing "collective goods." For illustration, assume a situation in which you and others engage in creating and using something together (the "collective good"). Then, everyone invests a certain contribution that, in some way or another, goes into a "common pot". If the overall investment reaches a sufficient size, it creates a synergy effect and produces benefits. To reflect this, the total investment, i.e. the amount in the pot, is multiplied with a factor greater than 1. Finally, the resulting overall amount is assumed to be equally split among all contributors.

In such situations, you cannot benefit if you don't invest. In contrast, if everyone invests a sufficiently high amount, everyone gains, as the investment is assumed to create a synergy effect. But if you invest much and others little, you are likely to walk home with a loss. Conversely, a low investment may produce an individual advantage, as long as the others invest enough.

Therefore, cooperation (contributing much) is risky, and free-riding (contributing little) is tempting, which destabilizes cooperative behavior. If the situation occurs many times, cooperation erodes, and a so-called "tragedy of the commons" results. In the worst case, nobody invests anymore in the end and nobody will get a benefit. This is pretty much the situation one can find in countries characterized by high levels of corruption and tax evasion.

Therefore, cooperative behavior, even though beneficial for everyone, can break down for similar reasons as free traffic flow breaks down on a crowded circular road: the desirable state of the system is unstable. See the Video. But given that it's possible to stabilize free traffic flows by means of traffic assistance systems, are there any biological or socio-economic mechanisms that can stabilize social cooperation?


An outcome that is bad for everyone


"Tragedies of the commons" are known from many areas of life. Often cited examples are the degradation of our environment, overfishing, the exploitation of social benefit systems, or dangerous climate change. In fact, even though probably nobody wants to destroy our planet, we still exploit its resources in a non-sustainable way and pollute our Earth. For example, a good solution how to safely keep nuclear waste has still not been found. Moreover, nobody wants to be responsible for the extinction of a species of fish, but we are facing a serious overfishing problem in many areas of the world. We also know that public schools, public hospitals, and many other useful investments made by the state require our taxes. Nevertheless, the problem of tax evasion is widespread.

Now, the reader will argue: of course, we can make contracts and establish institutions such as courts to ensure that they will be kept. This is correct, but what institutions shall we establish? How efficient are they? And what are their undesired side effects? In an attempt to address these questions, the following paragraphs will give a short and certainly incomplete overview of mechanisms that support cooperation – and have played a major role in human history.


Family relations


An early mechanism invented to promote cooperation in social dilemma situations, first discussed by George R. Price (1922-1975), is called "genetic favoritism". This means that, the closer you are genetically related with someone else, the more advantages you would give to him or her as compared to strangers. This principle has led to tribal structures and dynasties, which have been around for a long time – and still are in many countries. Today's inheritance law favors relatives, too. However, genetic favoritism has a number of undesirable side effects, such as a lack of fair opportunities for non-relatives, and also ethnic conflicts or blood revenge. The drama of Romeo and Juliet articulates the impermeability of family structures in the past very well. Another well known example is the Indian caste system.


Scared by future "revenge"


What other options do we have? Repeated interactions, as they already occurred in early human settlements, are an example. In case you interact with the same person time and again, you may play tit for tat or some other strategy that teaches your interaction partner that non-cooperative behavior won't pay off. The strategy is thousands of years old and is also known as "an eye for an eye, a tooth for a tooth."

The effectiveness of such revenge strategies was studied in 1981 by Robert Axelrod (*1943) in his famous computer tournaments. It turned out that, if you would interact with me just frequently enough, mutual cooperation would beat exploitation strategies. This effect is also known as "shadow of the future". As a consequence, a situation of "direct reciprocity" is expected to result, in which "I help you and you help me." But what if our friendship becomes too close, such that the result is good for us but bad for others? This could lead to corruption and impermeable, inefficient markets.

Moreover, what should we do if we interact with someone only once, for example, a renovation worker? Will we then have to live with a bad experience that we are likely to make? And what should we do, if a social dilemma situation involves many players? Then, a tit-for-tat strategy is too simple, because we don't know who cheated us and who didn't.


Costly punishment


For such reasons, further social mechanisms have emerged, which include "altruistic punishment". As Ernst Fehr and Simon Gächter have shown in 2002, if players can punish others, this will promote cooperation, even when punishment is costly. This effect is so important that people, if faced with the option to choose between a world without punishment and a world offering a sanctioning option, often decide for the second one. The experimental discovery of this was made as late as 2006 by Özgür Gürek, Bernd Irlenbusch, and Bettina Rockenbach. Punishment efforts are mainly needed in the beginning, to teach people "to behave", i.e. to be cooperative. In the end, the punishment option may be rarely used. But for cooperation to persist, the punishment option still needs to be there (or another mechanims which can stabilize cooperation).

Note that such mutual "peer punishment" is a widespread mechanism. In particular, it is used to stabilize social norms, i.e. behavioral rules. Every single one of us probably exercises peer punishment many times a day – sometimes in a mild way (e.g. by raising eyebrows), and sometimes more furiously (e.g. when shouting at others).


The birth of moral behavior


But why do we punish others at all, if it reduces our own payoff, while others benefit from the resulting cooperation? This puzzle is called the "second-order free-rider dilemma", where "first-order free-riders" mean non-cooperators and "second-order free-riders" mean non-punishers.

To answer the above question, in a study with Attila Szolnoki, Matjaz Perc György Szabo in 2010, I have analyzed a collective goods problem considering four possible behaviors: (1) cooperators who don't punish non-cooperators, (2) cooperators punishing non-cooperators, called "moralists" (green), (3) non-cooperators punishing other non-cooperators, called "immoralists" because of their hypocritical behavior (yellow), and (4) non-cooperators who don't punish (red). We furthermore assumed that individuals imitate the best performing behavior played by their interaction partners.

If the interaction partners were randomly chosen from the entire population, moralists couldn't compete with cooperators – due to the additional punishment costs. Therefore, cooperators ended up in a social dilemma situation with non-cooperators, which they lost, and a "tragedy of the commons" occurred. However, if individuals interacted with a small number of neighbors, we found the emergence of clusters of people, in which the same behavior prevailed (see picture below). Surprisingly, the fact that "birds of a feather flock together" makes a big difference for the outcome of the interactions: it allows moralists to thrive. Assuming all other things to be equal, in the spatial (rather than random) interaction scenario, cooperators lost the battle with non-cooperators in their neighborhood, as expected, but moralists could cope with them (as their punishment of non-cooperators reduced the success of non-cooperative behavior). Therefore, when individuals interacted with neighbors in geographical space or social networks rather than with random interaction partners, moral behavior could win its way and second-order free-riders were crowded out. As a consequence, "moral behavior", i.e. cooperation and the punishment of non-cooperative behavior thrived.

Nevertheless, punishing each other is often annoying or inefficient. This is one of the reasons why we may prefer so-called "pool punishment" over the above "peer punishment". In such a case, we invest into a common pot paying for a police or court or other institution sanctioning improper behavior. But a problem of this approach is that representatives of the punitive institution may be corrupt. Besides this corruption problem, it is often far from clear, who deserves to be punished and who not.

We must certainly be careful not to sanction the wrong people. This requires high inspection and discovery efforts, which might not always be justified by the success rates. Then, lack of success may lead to less inspection and eventually to more crime – a strange interdependency pointed out by Heiko Rauhut, which might explain the crime cycles that have often been observed in the past. To our great surprise, the computer simulations of the spreading and fighting of criminal behavior that I performed in 2013 together with Karsten Donnay and Matjaz Perc suggest that more surveillance and greater punishment, i.e. more deterrence, are not able to eliminate crime. In fact, while there is an almost 10 times higher rate of prisoners in the USA than in Europe, crime rates do not seem to be lower. In fact, I expect a crime prevention strategy based on the consideration of socio-economic factors to be much more effective than one mainly based on deterrence.


Group selection and success-driven migration


Things become even more tricky if we have several groups with different preferences, for example, due to different cultural backgrounds or education. The subject has become popular under the label of "group selection", which was promoted by Vero Coppner Wynne-Edwards (1906-1997) and others.

In fact, it seems that group competition can promote cooperation. Compare two groups: one with a high level and one with a low level of cooperation. Then, the more cooperative group is expected to get higher payoffs, such that it should grow more quickly compared to the non-cooperative group. Consequently, cooperation should spread and free-riding disappear. However, what would happen if there existed an exchange of people between the two groups? Then, free-riders could exploit the cooperative group and quickly undermine the cooperation in it.


The surprising role of migration


Can cooperation only thrive in a world without migration and exchange? Surprisingly, the contrary is true if the conditions are right. In computer simulations I performed in 2008/09 together with Wenjian Yu, we have studied success-driven migration. Our simulations made the following assumptions: (1) Individuals move to the most favorable location within a certain radius around their current location ("success-driven migration"). (2) They tend to imitate the behavior of the most successful interaction partner (neighbor). (3) With a certain probability, the individual migrates to a free location, or it flips its behavior (from cooperative to non-cooperative or vice versa). Rule (1) does not change the number of cooperators, while all the other rules undermine significant levels of cooperation. Nevertheless, when all three rules are applied together, a surprisingly high level of cooperation emerges after a sufficiently long time. This is even true when the computer simulation is started with no cooperators at all, as Thomas Hobbes assumed to be the initial state of society. How is this possible, in spite of the fact that we don't assume a "Leviathan", here, i.e. a strong state, which imposes cooperation in a top-down way?

It turns out that migration disperses individuals in space. However, the rare flipping of individual behaviors creates a few cooperators. After a sufficiently long time, some cooperators are randomly located next to each other by sheer coincidence. In such a cooperative cluster, cooperation is rewarding, and neighboring individuals imitate this successful behavior. Afterwards, cooperation spreads quickly all over the system. Cooperative individuals move away from non-cooperative ones and join other cooperators to form clusters. Eventually, the system is dominated by large cooperative clusters, with a few non-cooperative individuals at their boundaries. Therefore, we find that individual behavior and social neighborhood co-evolve, and individual behavior is determined by the behavior in the neighborhood (the "the milieu").

In conclusion, when people can freely move and live in the places they prefer, this can largely promote cooperation, given that they are susceptible to the local culture and that they quickly adapt to successful local behaviors. Thus, migration isn't a problem, while lack of integration may be. But it takes two to be friends. Integration requires efforts on both sides, the migrants and the receiving society.

Migration is not always welcomed, even though it has always been part of human history. Many countries struggle with migration and integration. However, the United States of America, known as the "melting pot", are a good example for the positive potential of migration. This success is based on the principle that, in the USA, it is relatively easy to interact with strangers.

Another positive example is the Italian village Riace, in which the service sector was gradually disappearing, as young people were moving away to other places. But one day, a boat with migrants stranded. The mayor interpreted this as a divine sign, and decided to use this as an opportunity for his village. And, in fact, a miracle occurred. Thanks to the migrants, the village was revived. Since they were welcomed, they were grateful and gave a lot of good things back to the old inhabitants of the village. As the migrants were not treated as foreigners, but as part of the community, a trustful relationship could grow.


Common pool resource management


Above, I have described many simple social mechanisms that can be and have been tested for their effectiveness and efficiency in laboratory settings. They are also known to play a role in reality. But what about more complex socio-economic systems? Would self-organization be able to create desirable and efficient solutions? This is what Elinor Ostrom (1933-2012) studied – and what she got the Nobel prize for. It is often claimed that public ("common pool") resources cannot be efficiently managed, and that's why they should be privatized.

Elinor Ostrom discovered that this argument is actually wrong. She studied the way in which Common Pool Resources (CPR) were managed in Switzerland and elsewhere, and found that self-governance works well, if only the interaction rules are suitably chosen (read more). One suitable set of rules that provide good conditions for successful self-governance is specified below:
  1. There are clearly defined boundaries between in- and out-group parties (effectively excluding external, un-entitled parties).
  2. Rules regarding the appropriation and provision of common resources exist that are adapted to the local conditions.
  3. The collective choice arrangements allow most resource appropriators to participate in the decision-making process.
  4. There is an effective monitoring by people who are part of or accountable to the appropriators.
  5. A list of graduated sanctions is applied to resource appropriators who violate community rules.
  6. Mechanisms of conflict resolution exist that are cheap and easy to access.
  7. The self-governance of the community is recognized by higher-level authorities.
  8. In the case of larger common-pool resources, the system is organized in the form of multiple layers of nested enterprises, with small local CPRs at the base level.
As it turns out, public goods can even be created under less restrictive conditions. This amazing fact can be observed in communities of volunteers –  from Linux over Wikipedia, Open Streetmap, and StackOverflow to Zooniverse and many other forums.


The problem of globalization


One take home message of this chapter is that self-organization based on local interactions is at the heart of all societies in the world. Since the very beginning of ancient societies until today, a great deal of social order emerges in a bottom-up way, based on suitable interaction mechanisms and institutional settings. This approach is flexible, adaptive, resilient, effective and efficient.

The mechanisms enabling bottom-up self-organization discussed before promote the interaction of agents with mutually fitting behaviors: "birds of a feather flock together". The crucial question is: are these mechanisms also effective in a globalized world?

One may say that the process of globalization creates increasingly "well-mixed" interactions: ever more people or companies interact with each other, often in more or less anonymous or random ways. Such conditions, unfortunately, promote an erosion of cooperation and social order.

This undesirable effect is illustrated by a video. It illustrates a ring of "agents" (e.g. individuals or companies), each engaged with their neighbors in the creation of collective goods. The local interactions initially promote a high level of cooperation. But then we add more and more interaction links with other, randomly chosen agents in the system. While the additional links help to increase the level of cooperation in the beginning, cooperation soon starts to drop, as the connectivity increases, and it finally goes to zero. In other words, when too many agents interact with each other, a "tragedy of the commons" results, where everyone is suffering from a lack of cooperation. Data from Andrew Haldane from the Bank of England suggest that, for example, the financial meltdown in 2008 might have resulted from a hyper-connected banking network.


Age of coercion or age of reputation?


In fact, citizens and politicians all over the world have noticed that our economy and societies, and the foundations they are built on, are destabilizing. In an attempt to stabilize our socio-economic system, governments all over the world have undertaken attempts to establish social order in a top-down way, based on surveillance and powerful institutions such as armed police. However, this approach is destined to fail due to the high level of systemic complexity, as I have pointed out in a previous chapter. In fact, we are seeing a lot of evidence for this failure: signs of economic, social and political instabilty are almost everywhere. Will our globalized society just collapse and break into pieces, thereby re-establishing a decentralized organization? Or is there any chance to live in a globalized world, in which cooperation and social order are stable? Could we perhaps build something like an assistance system for cooperation?

In fact, it is known that reputation systems can promote cooperation through "indirect reciprocity". Here, the principle is that someone helps you, while you help somebody else. Reputation mechanisms support people and companies with compatible preferences and behaviors in finding each other. In a sense, such a reputation system can also serve as a kind of "social immune system", protecting us from harmful interactions.  


Pluralistic, community-driven reputation systems


These days, reputation and recommender systems are quickly spreading all over the Web, which stresses their value. People can rate products, news, and comments, and they do! There must be a good reason that people undertake this effort. In fact, they get useful recommendations in exchange, as we know it from amazon, ebay, tripadvisor and many other platforms. As Wojtek Przepiorka and others have found, such recommendations are beneficial not only for users, who tend to get a better service, but also for companies. A higher reputation allows them to sell products or services at a higher price.

But how should reputation systems be designed? It is certainly not good enough to leave it to a company to decide, what recommendations we get and how we see the world. This would promote manipulation and undermine the "wisdom of the crowd", leading to bad outcomes. It is, therefore, important that recommender systems do not reduce socio-diversity. In other words, we should be able to look at the world from our own perspective, based on our own values and quality criteria. Otherwise, according to Eli Pariser, we will end up in a "filter bubble", i.e. in a small subset of the information society that fits our taste. But we may lose our ability to communicate with others who have different points of views. In fact, some analysts think that the difficulty in finding political compromises between republicans and democrats in the USA is related to the fact that they use increasingly different concepts and words to talk about them. In a sense, they are living in different, largely separated worlds.

Therefore, reputation systems would have to become much more pluralistic. When users post ratings or comments on products, companies, news, pieces of information, and information sources, including people, it should be possible to assess not just the overall quality (as it is often done on a five-star scale or even just by a thumb up or down). The reputation system should support different quality dimensions such as the physical, chemical, biological, environmental, economic, technological, and social qualities. Such dimensions may include popularity, durability, sustainability, social factors, or how controversial something is.

Moreover, users should be able to choose from diverse information filters (such as the ones generating personalized recommendations), and to generate, share, and modify them. I want to call this approach “social filtering”. A simple system of this case has been implemented in Virtual Journal. Then, we could have filters recommending us the latest news, the most controversial stories, the news that our friends are interested in, or a surprise filter. So, we could choose among a set of filters that we find useful. To consider credibility and relevance, the filters should also put a stronger weight on information sources we trust (e.g. the opinions of friends or family members), and neglect information sources we do not want to rely on (e.g. anonymous ratings). For this purpose, users should be able to rate information sources as well, i.e. other raters. Then, spammers would quickly lose their reputation and, with this, their influence on the recommendations made.

Altogether the system of personal information filters would establish an "information ecosystem", in which increasingly good filters would evolve by modification and selection, thereby steadily enhancing our ability to find meaningful information. Then, the pluralistic reputation values of companies and their products (e.g. insurance contracts or loan schemes) would give a pretty differentiated picture, which could also help the companies to develop better customized and more successful products. Hence, reputation systems can be good for both, customers and producers. Customers will get better offers, and producers can take a higher price for better quality. This is serving both sides.

In summary, the challenge of creating a universal, pluralistic reputation system might be imagined as transferring the principles, on which social order in a village is based, to the "global village", i.e. to the conditions of a globalized world. The underlying success principle is the matching of people or companies sharing compatible interests. A crucial question is, how to design reputation systems in a way that would make them resistant to manipulation and provide enough freedom for privacy and innovation? Information Box 1 offers some related ideas.


Social Information Technologies


Reputation systems are just one possibility to promote social order and favorable outcomes of social interactions. As we have seen before, many problems result when people or companies don't care about the impact of their decisions on others. This may deteriorate everyone's situation, as in "tragedies of the commons", or cause mutually damaging conflicts. How do we overcome such problems? How tdo we promote more responsible behaviors and sustainable systems? The classical approach to this is to invent, implement and enforce new legal regulations. But people don't like to be ruled, they can't handle many laws, and they often find ways around them. As a consequence, laws are often ineffective. Nevertheless, with Social Information Technologies it is possible to create a better world, based on local interactions and self-organization. It's actually easier than one might think.

Today, smartphones are increasingly becoming assistants to manage our lives. They guide us to find fitting products, nice restaurants, the right travel connection, and even a new partner. In the future, such personal assistants will be less and less focused on self-centered services. They will pay attention to the interactions between people and companies, and they will produce benefits for all involved parties and the environment, too. Note that, when two people or companies interact, there are just four possible outcomes, among them coordination failures and conflicts of interest.

In the first case, the interactions would be negative for both sides, i.e. it would be a lose-lose situation. It is pretty clear what to do in such situations: one should avoid the interaction in the first place. For this to happen, we need information technologies that make us aware of the negative side effects of the interaction. Similarly, if we knew the social and environmental implications of our interactions on the environment, we could take better decisions. Measuring the externalities of our actions is, therefore, an important precondition for avoiding damage. In fact, if we had to pay for the externalities caused by our decisions and actions, then individual and collective interests would become more aligned. As a consequence, we wouldn’t run so easily into traps where individual decisions cause overall damage.

The second case is that of a bad win-lose situation, i.e. one side would have an advantage of the interaction, while the other side would suffer a disadvantage, and altogether the interaction would be damaging. In this situation, one side is interested in the interaction, but the other side would like to avoid it. Again, increasing awareness may help, but we would also need social mechanisms that would protect the potential loser from exploitation.

The third case concerns good win-lose situations. While the interaction would again be favorable for one side and unfavorable for the other, overall there would be a systemic benefit of that interaction. Consequently, one side would be interested in the interaction, but the other side would want to abstain from it. It is possible though to turn the win-lose situation into a win-win situation, namely by a value transfer. In this way, the interaction becomes profitable for both sides, which would hence engage in it.

Finally, in the fourth situation, the interaction would create a win-win situation. There are nevertheless two things one might do: balance the gains in a fair way, and create awareness of opportunities one would otherwise miss. In fact, every day, we are walking by hundreds people, who might share some interests with us, but we don't even know about it. These may be a hundreds of missed opportunities. If we had social information technologies helping us to interact with each other more successfully, this would unleash unimaginable social and economic potentials. If we had suitable tools to assist us, the large diversity of people with different cultural backgrounds and interests wouldn't be a problem anymore. Rather than producing conflict, diversity would increasingly turn into an opportunity. 

In summary, Social Information Technologies will help us to avoid bad interactions, to discover opportunities for good interactions, to engage in them successfully,  and to turn bad interactions into good ones. In this way, coordination failures and conflicts can be considerably reduced. I, therefore, believe that Social Information Technologies could produce enormous value –  be it material or immaterial. Just remember that Facebook is worth more than 50 billion dollars, even though it is based on a very simple principle: social networking. How much more valuable would Social Information Technologies be? But I don't want to argue for big business here. In fact, if we created these technologies in a crowd-sourced way for the public good or just for fun (as Linus Torvalds, the initiator of the Linux operating system, said) – even better!


Towards distributed cybersecurity, based on self-organization


Since the Arab Spring, governments all over the world have become worried about "twitter revolutions". Are social media destabilizing political systems? Do governments therefore have to censor free speech or influence at least the way tweets or facebook posts are distributed to followers? I don't think so. Biasing free speech will rather affect a society's ability to detect problems and address them early on.

But wouldn't a system based on the principle of distributed bottom-up self-organization be insecure? Not necessarily so! Let me give an example. One of the most astonishing complex systems in the world is our body's immune system. Even though we are bombarded every day be thousands of viruses, bacteria, and other harmful agents, our immune system is pretty good in protecting us for usually 5 to 10 decades.

Our immune system is probably more effective than any other protection system we know. And what is even more surprising: in contrast to our central nervous system, the immune system is organized in a decentralized way. This is not by chance. It is well known that decentralized systems tend to be more resilient to disruptive events. While targeted attacks or point failures can make a centralized system fail, a decentralized system will usually survive the impact of attacks and recover. In fact, this is the reason for the robustness of the Internet. So, why don't we build information systems in ways that protect them by "digital immune systems"? This should also include a reputation system, which could be called a "social immune system".


Managing the chat room



Information exchange and communication on the Web have quickly changed. In the beginning, there was almost no regulation in place. These were the days of the “Wild, Wild, Web”, and people often did not respect human dignity and the rights of companies when posting comments. However, one can see a gradual evolution of self-governance structures over time.

Early on, public comments in news forums were published without previous screening, and this spread a lot of low-quality content. Later, comments were increasingly assessed for their lawfulness (e.g. for respecting human dignity) before they went online. Then, it became possible to comment on comments. Now, comments are rated by readers, and good ones get pushed to the top. The next logical step is to rate commenters and to rate raters. Thus, we can see the evolution of a self-governing system that channels the free expression of speech into increasingly constructive paths. I, therefore, believe it is possible to reach a responsible use of the Internet mainly on the basis of self-organization.

In the end, the great majority of malicious behaviors will be handled by crowd-based mechanisms such as the reporting of inappropriate content and a reputation-based display of user-generated Web content. A small fraction will have to be taken care of by a "chat room master" or moderator, and there will be a hierarchy of complaint instances to handle the remaining, complicated cases. I expect that, only a few cases will have to be taken care of by courts or other institutions, while most activities will be self-governed by social feedback loops in terms of sanctions and rewards by peers. In the following chapters, I will elaborate in more detail, how information technologies allow top-down and bottom-up principles, but also people and companies, to come together in entirely new ways.


INFORMATION BOX 1: Creating a trend for the better


For reputation systems to work well, there are a number of further things to consider: (1) the reputation system must be resistant to manipulation attempts; (2) people should not be terrorized by it, or by rumors; (3) to allow for more individual exploration and innovation than in a village, one would like to have the advantages of the greater freedoms of city life – but this requires sufficient options for anonymity or pseudonymity (to an extent that cannot challenge systemic stability).

First, to respect the right of informational self-determination, a person should be able to decide what kind of personal information (social, economic, health, intimate, or other kind of information) it makes accessible for what purpose, for what period of time, and to what circle (such as everyone, non-profit organizations, commercial companies, friends, family members, or just particular individuals). These settings would, then, allow selected others to access and decrypt selected personal information. Of course, one might also decide not to reveal any personal information at all. However, I expect that having a reputation for something will be better for most people than having none, if only to find fitting people who have similar preferences and tastes.

Second, people should be able to post the ratings or comments entered in the reputation system either in an anonymous, pseudonymous, or personally identifiable way. But pseudonymous posts would have, for example, a 10 times higher weight than anonymous ones, and personal ones a 10 times higher weight than pseudonymous ones. Moreover, everyone who posts something would have to declare the category of information: is it a fact (potentially falsifiable and linked to evidence allowing to check it), an advertisement (if there is a personal benefit for posting it), or an opinion (any other information)? Ratings would always have the category "opinion" or "advertisement". If people use the wrong category or post false information, as identified and reported by, say, 10 others, the weight of their ratings (their "influence") would be reduced by a factor of 10 (of course, these values may be adjusted). All other ratings of the same person or pseudonym would be reduced by a factor of 2. This mechanism ensures that manipulation or cheating does not pay off.





[1] Dear Reader,

thank you for your interest in this chapter, which is thought to stimulate debate.
What you are seeing here is work in progress, a chapter of a book on the emerging Digital Society
I am currently writing. My plan was to elaborate and polish this further, before I share this with anybody else. However, I often feel that it is more important to share my thoughts with the public now than trying to perfect the book first while keeping my analysis and insights for myself in times requiring new ideas.
So, please apologize if this does not look 100% ready. Updates will follow. Your critical thoughts and constructive feedback are very welcome. You can reach me via dhelbing (AT) ethz.ch or @dirkhelbing at twitter.
I hope these materials can serve as a stepping stone towards mastering the challenges ahead of us and towards developing an open and participatory information infrastructure for the Digital Society of the 21st century that would enable everyone to take better informed decisions and more effective actions.
I believe that our society is heading towards a tipping point, and that this creates the opportunity for a better future.
But it will take many of us to work it out. Let’s do this together!
Thank you very much, I wish you an enjoyable reading,
Dirk Helbing  
PS: Special thanks go to the FuturICT community.

Friday, 31 October 2014

QUALIFIED MONEY: A better financial system for the future


by Dirk Helbing [1]


For example, “Qualified Money” would introduce reputation-dependent conversion rates to reward investments into quality. In this post, I will discuss the roots of the EURO crisis and some ideas how to overcome it.



Short history of money and future of Bitcoin






To discuss the future of money, let us first look back a little bit. By inventing a universally interchangeable good, the historical invention of money made the exchange of goods much easier. But while money was based on valuable materials such as gold in the beginning, it was later increasingly replaced by symbolic values, such as paper bills, or even entries in a digital account. Now, money is created in great amounts not only by central banks, but normal banks do it as well. If, one day, we don't trust anymore that we will get valuable goods in exchange, it is obvious that the value of money will be gone. This process is known as "hyper-inflation." In human history, this has happened many times. 

Bitcoin is an attempt to avoid that such a horrible scenario can happen again.It's a peer-to-peer payment system, which does not require banks anymore.But it has other problems. BitCoins are designed such that the overall amount of this digital currency is slowly growing, and it will eventually saturate, thereby establishing something like a new "gold standard." However, we have seen in history that a gold standard is not flexible enough to enable a resilient financial system. If the volume of money grows more quickly than economic output, we will eventually have inflation. Then, the value of money, i.e. its purchasing power, will drop. Conversely, if the volume of money does not grow as quickly as productivity, we run into another problem, called "deflation." Then, money becomes more valuable over time, and people would hoard rather than spend it, as they can buy goods more cheaply in the future. Under such conditions, business cannot thrive. So, the volume of money should grow proportional to productivity, at least on average. Let me add though that the above dependencies regarding inflation and deflation are expected to hold only on the long run. Central banks and other stakeholders can manipulate financial markets, which may create delayed adjustments, biases, and abnormal market behaviors. As a consequence, it becomes increasingly difficult to interpret market signals correctly, and to respond to them in a proper way. On the long run, I think, loss of control is almost inevitable.



A new kind of money - How the idea was born


For the above reasons, Bitcoin will not be the final solution, even though the technology behind it will be crucial for future currencies and other services requiring secure transactions. Altogether, however, we need to fundamentally re-invent money, as it is not adaptive enough for our complex world. We have to ask ourselves, why the financial systems keeps crashing since thousands of years, and what is fundamentally wrong with the way we have set it up.

Thinking about it for a couple of years, I came to the conclusion that, even though money is a great invention, it's outdated. Therefore, it's time to create a better one. The argument goes as follows. Currently, money is a scalar, i.e. the simplest mathematical quantity one can think of. It is neither multi-dimensional nor does it have a memory. But mathematics offers a much richer spectrum of concepts to define exchange processes, such as vectors, i.e. multi-dimensional quantities, and network graphs. In fact, money comes from somewhere and goes somewhere else. Who transfers money to whom defines a network of money flows. Therefore, money should be represented by network quantities. And money should be multi-dimensional to allow other things to happen apart from the eternal ups and downs.

This made me think about "Qualified Money" – multi-dimensional money with a memory. Since Roman times, people have said: "Money doesn't stink!" In other words: it does not matter where it comes from and how it is earned. However, what if we could give it a scent, like a perfume? And what if this would co-determine the value of money? In a discussion during a visit in Zurich, my colleagues Tobias Preis, Dave Rand, and Ole Peters were fascinated by this idea. Later on, I combined it with a reputation system and called the new concept "Qualified Money." Such money could earn reputation and, with this, additional value! This approach has commonalities with local currencies, but is more general and relates to the way, modern stock markets work. However, Qualified Money opens entirely new possibilities.

What's wrong with our financial architecture?


One of the problems of today's financial system is the possibility of cascade effects. What started as a local problem in the Californian real estate market became a world financial and economic crisis, eventually causing social and political unrests. But how could it come that far? For this, see the figure below. The world financial system lacks engineered breaking points to stop cascades. At home, everyone of us has electrical fuses to make sure that a local electrical overload would not cause a larger problem, e.g. the house to burn down. In the financial system, in contrast, the strategy is just the opposite: to ease the load on troubled banks, some of their problems have been taken on by the states, which are now in trouble as well, and so on. Rather than isolating infected "patients" and curing them by a Marshall-plan-like intensive care program, we infected many other countries that were healthy before. In this way, the overall damage became much larger than it could have been, and it is not clear how we will ever recover from the debt levels. If we don't get the problems solved any time soon, cultural values such as tolerance and solidarity, or even peace might be in danger. It's now the very fabric of our society, which is at stake. In societies with mass unemployment, it can take two generations or more until the good relationship between citizens and their state and a healthy social structure recovers.







What worries even more is the fact that we don't currently have a backup financial system. For most other systems that may fail, we have contingency plans – a "plan B" or "plan C." In fact, one might argue that one reason why our current financial system performs badly is the absence of competing financial systems. Given that we believe in competition, why don't we take this seriously and build alternative systems, which could also serve as backups, as plans B or C? It's not enough to complain about having to bail out banks and about lacking alternatives. I also doubt that tougher regulations won't fix the problems. As large banks can handle the additional regulations best, while small and medium-sized banks struggle with them, these regulations may cause big banks to grow even bigger. Therefore, we should rather promote alternatives. In fact, with Bitcoin and peer-to-peer lending systems, some alternatives are eventually emerging, but we need more and better ones.

At the moment, I would say, we cannot take it for granted that the current financial system will still work in 10 or 20 years from now. Most industrial states have debts of the order of 100 to 200 percent of the gross domestic product (GDP), sometimes even a multiple of this. Controlled inflation has been considered to be a recipe to reduce these debts. The trick can work, if applied by a single country or just a few ones. However, if the USA, Europe, Japan, and further countries are all trying to reduce their debts in such a way at the same time, this may trigger an inflationary spiral that can get out of hand. 

Besides, the attempts of central banks to control the level of inflation haven't worked well so far. To save banks, to encourage investments into the real economy, and to increase the level of inflation, central banks have pumped massive, almost unimaginable amounts of money into the financial markets. However, as it turns out, more than five years after the financial crisis started, companies still have difficulties to lend money for investments, many banks are still in a shaky position, and there are even worries about deflationary tendencies. 

Why is this? Banks don't trust that companies would pay their loans back and, besides, they need more capitalization themselves. Most money created by the central banks does not reach the companies in need. Instead, given the low interest rates, money is mainly invested into financial markets. This drives up stock prices even though real economic growth is negligible. Rising stock prices create further incentives for virtual investments at the stock markets rather than real investments into companies. Consequently, central banks have created a gigantic bubble in the financial markets. In some sense, a virtual inflation has happened over there – the stocks have become more expensive even though most companies haven't grown. When this stock market bubble bursts, a large fraction of the money will flee into real values. This will suddenly drive enormous price inflation, as there are not enough material values that these huge amounts of money can buy. Therefore, inflation might easily get out of control. So far, this did not happen as misleading incentives have caused a temporary allocation of money in the stock markets. In the meantime, low interest rates are undermining the perspectives of life insurances and pension funds.




An unfeasible control problem


I have argued above that the central banks haven't been able to reach the effects they wanted. But this is not because they wouldn’t be competent. It's because the control problem they are expected to solve is ill-defined – it's literally unsolvable. The reason is that they don't have enough instruments or, to put it differently, not enough control variables. The central banks can increase the volume of money and they can change the interest rate. That's basically it. They may also buy and sell bonds, but many think they shouldn't. The classical instruments of central banks are apparently not sufficient to do the job. In other words, the weapons of central banks are blunt. New possibilities are urgently needed, and this basically means additional ways of adaptation.

Why is this so? Let's take an example from the world of taxes. Apart from raising money for public engagements and investments, taxes are often used to incentivize or discourage certain kinds of behaviors. For example, many countries have taxes on cigarettes, alcohol, and fuel, to reduce their consumption. They may also offer tax reductions for investments into environmental-friendly heating, better home insulation, or buying solar panels, to promote the production of renewable energy. It is clear that each of these goals can be achieved by suitable taxation-based incentives. But what happens if one tries to reach many goals by one single "control variable," the overall amount of money to be paid for taxes? One may end up investing into solar energy production, while smoking more cigarettes, which altogether would not change the individual tax level. So, on average people may not be very responsive to a multitude of rewards and sanctions. In other words, we are unlikely to reach many goals with a single control variable such as one-dimensional money.

More control variables needed


This problem is actually well-known from control theory. For example, complex chemical production processes cannot be steered by a single control variable such as the temperature or concentration of a certain chemical ingredient. In a complicated production process, one must be able to control many different variables, such as the pressure and the concentrations of all ingredients. It is also instructive to compare this with ecosystems. The plant and animal life in a place will not just be determined by a single control variable such as the amount of water, but also by the temperature, humidity, and various kinds of nutrients such as oxygen, nitrogen, phosphor, etc. Our bodies, too, require many kinds of vitamins and nutrients to be healthy. So, why should our economic system be different? Why shouldn't a healthy financial system need several kinds of money?

If we had different kinds of money, we could probably influence how much of the money handed out by central banks is finally used by companies for real investments. This would require at least one additional kind of money. So, let us assume that, besides cash and goods, we would have two kinds of electronic money: "real" electronic money ("REMO") and "virtual" electronic money ("VEM"). For example, besides real electronic EUROs, we could introduce "AEROs" as virtual electronic money. By law, cash and real electronic money could be invested into goods and real investments, but not into financial products. Virtual electronic money, in contrast, could be invested into financial products, but not into goods. The important point is now that the central bank could hand out REMO and VEM at different interest rates. If REMO were handed out at a lower interest rate than VEM, this would incentivize real investments.

Two kinds of electronic money


Of course, cash, REMO and VEM could be converted into each other. However, by means of conversion fees, one could also create incentives for one kind of money as compared to the other(s). This would create new "degrees of freedom," as a physicists would say, which would enable a better adaptation of the financial system to the actual needs. For example, if REMO earns some interest rate but cash not, or if cash loses value due to inflation, this speaks against saving large amounts of cash. It would be better to spend cash on consumption, or to turn it into REMO or VEM. If lending REMO is cheaper than lending VEM, it will incentivize real investments over virtual investments into financial products. If VEM can be converted into REMO for free, but converting REMO into VEM is costly, this again incentivizes real investments.

So, this little extension of our financial system will allow the central banks to more effectively stimulate real investments into companies' production capacities. Central banks would not have to produce anymore a bubble of cheap money, which will sooner or later overheat the financial and real estate markets. As we have seen in the past, this can cause dangerously large bubbles, which will sooner or later produce large-scale global damage, when they burst.

Europe's "little" mistake


But we should dare to think one step further. While the economy in the USA and the United Kingdom seem to be recovering from the 2008 financial crash and the subsequent economic crisis, most of Europe is still not doing well after several years of struggle – in fact, some indicators are worse than after the great depression in the 1930ies. In January 2014, Nobel prize winner Joe Stiglitz (*1943) summarized the situation in Basel, Switzerland, as follows: before the crisis, Europe was doing very well. It had some of the strongest economies in the world, it had some of the best public infrastructures, best education systems, best health systems, and social systems. However, Europe did a "little" mistake: without creating a sufficiently sophisticated institutional framework, it introduced a new currency, the EURO, which replaced more than a dozen other currencies. Altogether, this created more problems than benefits, he judged.

We are not talking here about the widespread complaints of citizens that the EURO made life more expensive – be they justified or not. Instead, we must talk about the fact that, if we compare all countries on a one-dimensional scale such as the gross domestic product (GDP) per capita, there will be always winners and losers. In this case, Germany happens to be a winner and Greece a loser, but it could have been different as well. We must recognize that, given the different productivity of the countries, it was just a matter of time until economic forces were unleashed, which required adjustments. In the past, this adjustment happened naturally by adapting the currency exchange rates. Now, in more than 15 European countries, this is not possible anymore. As above, this problem can again be solved by adding new "degrees of freedom," i.e. new control variables. But how to introduce these variables without giving up the EURO, which many consider an important peace-building project in Europe?

Vitamins for the financial system


In the following, I will suggest to introduce "Qualified Money". Qualified Money has a number of different qualifiers, which turn money into a multi-dimensional means of exchange. The value of Qualified Money is given not only by its amount, but also a conversion factor that depends on various qualities. For example, if one decided that geographic origin should be a qualifier, one would enable country-specific EUROs, allowing adjustments of the value of money to the respective economic strength. The same approach can be used to define regional or local currencies, if desired. So, one could save the prestige project of the "EURO," by making the currency more flexible. The regional variants of EUROs would be converted into each other similarly as we are currently doing it for different kinds of currencies at the stock markets, such as EUROs, DOLLARs, or YENs.  

However, Qualified Money would not have to be connected to local origin. The concept has potential for extension. For example, the unemployment rate, the Millennium development goals, or any socio-economic-environmental factors considered relevant for human well-being could be used to define qualifiers. In our lives, it's not just money that matters. People care about many things, and this opens up entirely new possibilities!

We could all be doing well


It is important to recognize that both, the self-organization and management of complex dynamical systems require sufficiently many control variables, not just one. Establishing different kinds of money would serve this purpose. Compared to the currency system we have today, these different kinds of money would not be easily convertible. There would be an adjustable conversion tax or fee, to discourage conversion and to encourage earning different kinds of money, instead. This would naturally extend the approach we have discussed above in connection with VEM and REMO, and it would create a multi-dimensional incentive system, rewarding us for different kinds of efforts, including social and environmental ones.

Of course, such a conversion tax or fee would create something like “friction” in the multi-dimensional money system. However, we know from physics that friction can enable important functionality. How would it be to have such a multi-dimensional money and exchange system? Depending on how many dimensions we allow for, everyone could be doing well, each one on the dimensions fitting his or her personal strengths, skills, or expertise.

Today we have many ranking systems to compensate for the lack of such a multi-dimensional money system. Besides the Fortune 500 list of richest persons, we rank tennis players and soccer players. Others collect medals, or scores in computer games. Scientists count citations earned by their publications, etc. Even though some of these ranking scales don't imply any material value, they can motivate people to make an effort. Hence, we can turn this into a mechanism to create a multi-dimensional reward system, as we need it to enable self-organizing socio-economic systems.

One might even consider the possibility to allow everyone to establish a certain number of own currencies. In a sense, this would be the logical next step after allowing banks and BitCoin to create money (rather than just central banks). The value of these personalized currencies would then depend on how much others trust in them and are willing to engage in a related value exchange. I assume that, after some time, there would be just a reasonably small number of successful currencies that are widely used. However, they might have some interesting new properties compared to the currencies we have today. Therefore, opening up money creation to innovations might be really worthwhile.

Money with a memory


Let us now assume that electronic money would be traceable. In this case, we could give electronic money a "memory," and we could make its value dependent on its transaction history. To put it in simple terms, money that went through the hands of Albert Einstein or John F. Kennedy could have more value than money that was earned with "blood diamonds." So, possible qualifiers could be, how the money was earned, its origin or destination location, the reputation of the products bought, or the reputation of the producer or seller. Hence, we can further differentiate electronic money by means of additional qualifiers. This might be imagined as treating money units like stocks or like individual currencies. In other words, a (reputation-dependent) conversion factor would apply, when financial transactions are made.

Benefits of money with reputation


I recognize that some people might feel uneasy about money becoming dependent on reputation. However, in some sense, this is already happening when we go shopping in the Internet. Depending on the country we live in, the type of computer we are using, and perhaps further personal qualifiers such as income, we might get different product offers than others, at different prices. This is part of the logic of personalized recommender systems. One might find it upsetting to pay a higher price than others, but it could also be a lower one. When we book an airplane ticket or a hotel room, we receive different offers, too, depending on when we book and where we book, and whether we are regular customers or not.

In any case, there are quite some benefits of reputation-based Qualified Money. For example, it becomes easier for producers and stores to sell high-quality products at a higher price. Furthermore, to get an idea how future shopping might look like, assume that there is a database, in which information about products is stored, such as the amount of money to be paid, ingredients, durability, level of environmental-friendly production, level of socially friendly production, and much more. Moreover, assume our smartphones knows our preferences, for example, that we give the price a weight of 50%, environmental-friendliness a weight of 30%, and fair production a weight of 20%, and that we want to avoid products with particular ingredients we are allergic against. Then, by scanning product codes and retrieving the related product information, our smartphone will recommend us the best fitting products. Furthermore, if customers were willing to share their preference settings, producers and sellers could better tailor their assortment of products to the customer wishes. Therefore, customers would benefit as well. They would get more products they would really like to have.

Balancing transparency and anonymity


If properly set up, Qualified Money can create a good balance between transparency and anonymity, such that we can have the benefits of both. Transparency can promote more responsible and desirable behavior. It allows ethical values and higher quality to survive in a framework of free economic competition. In fact, a considerable fraction of people cares about ethics and fair products. Even financial investors are getting interested in ethical investments, as they tend to be more sustainable. At the moment we often find ourselves in a situation, where the competition between companies is so harsh that they have to reduce production costs. This will sooner or later decrease salary levels, production standards, product quality and/or sustainability. In the end, we have lower salary levels or lower-quality products. Both will eventually impact producers as well. In contrast, reputation mechanisms could stop the undesirable downward spiral, by rewarding higher quality products and fairer production.

The question is, whether the transparency needed for such reputation systems will ever be reached? In fact, there is currently a trend towards more transparency of money flows. We have recently seen the Swiss banking secrecy melt away. Several times, whistleblowers have sold confidential information about private accounts to public authorities. "Off-Shore Leaks" has made international money flows more transparent as well. Furthermore, there seems to be a "follow the money" program that tracks individual money transactions. And presently, many countries set up agreements for an automatic information exchange allowing public authorities to monitor money flows and to check tax declarations.

Anonymous money exchange is under attack for similar reasons as anonymous information exchange: in many cases, it has promoted crime and misery. Nevertheless, anonymity has still important roles to play. Most of us don't want anybody to know, which toys someone buys in a sex shop. For such and many better reasons, we still need sufficient amounts of cash besides traceable electronic money, even though it should lose its value quickly enough to make traceable transactions sufficiently attractive.

It should be also remembered that anonymity is one of the most important elements of democracies. The principle of anonymous vote is needed for independent decision-making, which is a precondition for the "wisdom of crowds" to work. Academic peer review as well is based on anonymity, to support open criticism without fear of revenge. Organized crime or corruption would also be difficult to fight without protecting the anonymity of witnesses. So, neither full transparency nor full anonymity can work. We need a system that makes it possible to combine and balance both principles. Introducing Qualified Money besides cash is the solution!





[1] Dear Reader,

thank you for your interest in this discussion paper, which is thought to stimulate debate.
What you are seeing here is work in progress. My plan was to elaborate and polish this material further, before I share it with anybody else. However, I often feel that it is more important to share my thoughts with the public now than trying to perfect the content first while keeping my analysis and insights for myself in times requiring new ideas.
So, please apologize if this does not look 100% ready. Updates will follow. Your critical thoughts and constructive feedback are very welcome. You can reach me via dhelbing (AT) ethz.ch or @dirkhelbing at twitter.
I hope these materials can serve as a stepping stone towards mastering the challenges ahead of us.
I believe that our society is heading towards a tipping point, and that this creates the opportunity for a better future.
But it will take many of us to work it out. Let’s do this together!
Thank you very much, I wish you an enjoyable reading,

Dirk Helbing