Keywords: systemic risk, behavioral finance, economic complexity, evolutionary computation, computational trust, blockchain, cryptocurrencies, market microstructure.
JEL Classification: G02, F63, B17, C53, C58
Introduction
There have been numerous cases documented of malfunctions in the cryptocurrency markets; some have been related to the internal operation of exchanges, such as the glitch in the cryptocurrency exchange GDAX that crashed the price of Ether from $319 to 10 cents \cite{trade}; others have to do with the dissemination of information, such as the negative impact on bitcoin prices after the de-listing of Korean exchanges in Coinmarketcap, a popular price tracker \cite{warning}.
While the bitcoin network is decentralized by design, the web and other peripheral networks where the services that enable the bitcoin economy operate are not. This extends to any other cryptocurrencies with a sufficiently high degree of decentralization, such as Ether. Roll \cite{ROLL_1984} demonstrated that volatility is affected by market microstructure, and by applying this idea to the off-chain side of a crypto-economy we are able to develop real-life risk metrics for the blockchain financial system.
The themes covered by this study are diverse, nevertheless, together they operate in the realm of complexity economics: i) Dynamics (time series analysis and prediction, temporal correlations; interaction between internal dynamics and external inputs; programmability of self-organization), ii) Methodology (data-driven research; machine learning; soft computing), iii) Human Social & Economic Systems (economies and markets; financial systems; risk management).
Our research question is whether factors related to market structure and design, price formation and price discovery, transaction and timing cost, information and disclosure, and market maker and investor behavior, are quantifiable to the degree that can be used to price risk in digital asset markets. This research is pertinent to the regulatory function of governments, that are actively seeking to understand and develop policies for crypto markets, and for investors, who are in need of expanding their understanding of market behavior beyond explicit price signals and technical analysis.
Literature
Authors \cite{Zenil_2011} have demonstrated the nature of the market as a rule-based system with an 'algorithmic' component that can be studied using the theory of algorithmic probability and Kolmogorov complexity \cite{Solomonoff_1964}. However, due to the computational expense other approaches are possible, such as genetic programming \cite{Kotanchek_2012} that is suitable for real-world data modeling and also is useful to search for solutions in the space of algorithms; besides, soft computing can deal with uncertainty and approximation to achieve practicability in a similar fashion that the human mind would (i.e. using symbolic regression representations).
The evolutionary foundations of economics have been studied by authors \cite{Dopfer} seeking to understand the processes which generate particular forms of economic activities and structures. In this context, a soft computing approach can express metaphors from socio-economic activity; for instance, the population balance modeling \cite{2014a}, a balance on the number of particles of a particular state, can be seen as the balancing of attention units, such as visits or calls to a web service. The flow and allocation of attention resources and transactional activity can be explained using both established frameworks that have been updated with novel evolutionary perspectives, such as the Coase theorem \cite{Mikami_2013}, or with entirely novel approaches based on Econophysics, such as fields finance \cite{Krabec_2015}. The "stock of trust" that those flows build (the economic activity of agents) can be aggregated following established frameworks such as economic complexity \cite{2014} , that can be adapted to the case of crypto economic complexity \cite{Venegas}, to form the basis to understand how "crypto nations" learn.
An analysis of robustness and fragility in crypto markets should necessarily include as well aspects of the behavioral modeling of networks; given that such approaches usually have universal properties, as with the controllability of complex networks \cite{Liu_2011}, or have applications across disciplines from business to information security \cite{Altshuler_2014}, and law \cite{Zander}, where even high-benefit, high-probability outcomes do not outweigh the existence of low probability, infinite cost options \cite{Taleb2014}. This "ruin problem" treatment is pertinent for the study of cryptocurrency markets, especially after the introduction of bitcoin futures in derivatives markets of systemic importance to the world financial system \cite{futures,group}.
Results
Methodology
Data for this section includes digital assets historical daily prices and volumes (Coincheckup.com, Coinmarketcap.com), and off-chain web and social network analytics (EconomyMonitor.com and click-stream data providers). The period of study is August 2016 to January 2018, using daily data points. Data processing was performed in Mathematica \cite{2017}.