Tuesday, February 06, 2007

Learning From Chaco: P2P Risk Networks?

Here's a longer potential title, for what may one day become a white paper: "Peer-to-Peer Risk Allocation Strategies: Learning From the Failure of Chaco Canyon to Improve the Resiliency of the Modern Financial System"

For those not familiar with what I mean by "risk-allocation," see my previous post "Financial Wizardry & Collapse."

In Brief:

- Chaco Canyon developed as a centralized marketplace to facilitate the distribution of the risk of crop failure among a system of producers

- The Chacoan System (as best understood) failed for two reasons: 1) The maintenance costs of the central structure created excessive inefficiency in the risk-allocation mechanism, and 2) The centralized and hierarchical topology of the system reduced its resiliency to recover from systemic shock

- The risk allocation mechanisms within our modern financial system (hedge funds, credit default swaps, volatility indeces, etc.) exhibit the same vunerabilities to systemic collapse as the Chacoan System.

- High-maintenance, hierarchical systems for centrally pooling and re-allocating risk developed because centralization allows for the accumulation of surpluses by an elite, who then protected that system against alternatives.

- Today, the legal and communications framework exists to facilitate the development of an alternative risk allocation system based on a peer-to-peer topology.

- Peer-to-peer risk allocation can enhance systemic resiliency by elminating the potential for central place failure.

- Peer-to-peer risk allocation can further enhance systemic resiliency by reducing the rent on transactions extracted to maintained a centralized structure.

- This reduction in transaction costs further enhances systemic resiliency by 1) Decreasing barriers to entry, thereby expanding hte breadth of the risk allocation network to directly include "retail" consumers, and 2) Expanding the depth of the network to new kinds of risk allocation such as exposure to fluctuating energy costs, healthcare needs, life insurance, littigation risk, etc.

So, there's the theory: our financial system primarily performs two functions: matching supply with demand and allocation of risk. The matching of supply and demand is already fully engaged on peer-to-peer topology, but will not move away from centralized production and consumerism until energy costs and diminishing marginal returns force such a move. Risk allocation, however, has yet to make this conceptual move--this represents a huge opportunity, both to lay a rhizomatic grid over the top of our current hierarchical system, and, potentially, to make a fortune as the next "Napster" or equivalent. It really is the kind of situation where a young internet whiz coupled with a few well schooled financial theorists could take down the Wall Street and City (of London) elite, much as Shawn Fanning sent shock waves through the recording labels.

Why would you want to do this? Doesn't it just prop up our current system? Maybe not. What I think it would do is create a "diagonal" market structure. A market structure that could survive dramatic, systemic shocks that will knock down the remnants of hierarchy. It will survive market shocks like the transition away from high-energy consumption to localized consumption. A sufficienty dynamic risk-allocation network (e.g. one not burdened by an invested elite) could survive the decline of the internet and telecommunications if it comes to that--the peer-to-peer topology is quite well suited to festival-based information processing.

Just a thought. It also raises another issue: while it may call into question what definition of "complexity" that one uses, does investment in flat-topology complexity (rhizome) lead to diminishing marginal returns in the same way that investment in peak-topology complexity (hierarchy) does?


Anonymous said...

You say that “Peer-to-peer risk allocation can enhance systemic resiliency by eliminating the potential for central place failure”. While that is true, by decentralizing the structure, you are also decentralizing the risk, which means that the risk will be larger for each decentralized structure (greater chance of statistical anomalies causing a failure). This will lead to more failures, not just in terms of the number of structures (because there are more structures) but it will also lead in an increase in the number of users affected by failure because the risk will not be as widely spread. Because of the increased risk (due to increased chance of failure), the premium charged will be larger. So under this system, fewer users would be able to afford to reduce their risk.

Your example of Napster is not a valid one. There is no way that a system such as Napster could be used for spreading risk in a similar way to music sharing. Digital music is a non-rival commodity (commodities that can be reproduced without cost), whereas money and risk are rival commodities (if you take my $5, I have $5 less). With Napster, there was little cost to others when people took more there fair share of pirated music, whereas with insurance, when people take from the system, others are proportionately effected. For this reason, a centralized administrative system is required to investigate and monitor that the correct amount of money is being paid. In addition, actuaries are required to predict the payouts that will be required, and the payments that will be needed to cover the costs of payouts and administrative fees. How can it be ensured that the administrative system will be as efficient as possible? There must be a group of people (stockholders) that will bare the success or failure of the system. By creating other centralized systems to compete against each other, it can be ensured that competition will keep the administrative costs to a minimum.

The centralized structure serves a purpose (spreading risk and providing efficient administration) that cannot simply be abandoned.

Jeff Vail said...

Why would the risk be larger in a decentralized structure? You don't provide any reasoning to back up that huge leap. I think it would actually be smaller because of the lower associated transaction costs.

Also, I think that risk allocation is a perfectly non-rival commodity. Take the very traditional commodity market for energy: producers want to lock in a price that they can sell at in order to gain project financing, and users want to lock in price that they can buy at for similar reasons. What happens in risk-allocation does not cost anything to affect (the definition of non-rival), but rather is merely an agreement between two parties. Enforcement may have costs, but as I mentioned that is why our extant legal system is a necessary prerequisite to such a P2P system.

It does take a bit of imagination to realize the potential to move beyond our centralized, hierarchical system, but that doesn't mean that it won't work...

nulinegvgv said...

"a centralized administrative system is required to investigate and monitor that the correct amount of money is being paid." This statement seems to assume a level of authoritative responsibility that doesn't exist in large quantities in the current financial system. It also brings up a weakness of such systems over time. When it is small, a centralized system of hierarchy is relatively easy to manage and maintain. As it grows though, more effort is needed to manage and maintain the increasing complexity of the system. Fewer individual participants are willing (and later as it grows able) to leverage the powers of the system to ensure participants are not disproportionately affected. In other words, at a certain point none of the people participating in the car-share program are willing or able to pay for the repairs need to keep the aging automobile operational. At a certain point everyone recognizes that the vehicle will be driven until it breaks down. Those involved try to get as much as possible out of the car before that happens.

Jeff mentioned that, “centralization allows for the accumulation of surpluses by an elite, who then protected that system against alternatives.” I think the elite are also likely to investigate and monitor (protect) only when it is in the best interest of their surpluses. They are likely to be aware of the costs of needed repairs to the system and to try and protect their surpluses by neglecting to pay these costs or by socializing the costs when they must be borne in order to keep the system in place. Take for example some of the airplane industry’s pension plans. First they were under funded and later abandoned- turned over to a government agency that will use taxpayer dollars to partially fund them. The overall result was that the system failed to protect those at the bottom.

I think the idea of shared, peer-to-peer risk as a potentially viable alternative to the possible systematic failure of hierarchical systems is no where more relevant than in a discussion about industrial agriculture. The idea of CSA (Community Supported Agriculture) programs is a system whereby participants share in the upfront cost (and risk) of growing local food. If enough of these programs persisted throughout a neighborhood, then the residents would be less affected by disruptions to the food supply (grocery stores carry only about 3 days worth of essentials). Ecoli contaminations, terrorist attacks, pandemic quarantines, or other catastrophes would have less severe implications to individual households if they not only subscribed to such CSA programs, but also took to growing a bit of their own food as well as supporting neighbor’s efforts and also larger, more traditional farmers near by. The assaults on our current agricultural model serve as great examples of the failure of administration to provide positive (read safe, healthy, and afordable) results to everyone involved. As these cracks become caverns we would do well to foster decentralized, rhizomatic systems of food production and distribution to replace hierarchical food supply systems.

Or, I trust my neighbors way more than I trust Monsanto.

HouseofPolitics said...


political forum

Big Gav said...

Chris Cook (the guy who designed the semi-mythical Iranian oil bourse) has also pondered the value of a more distributed market system, which he calls "Market 3.0":