Essays on Banking and Capital: An Agent-Based Investigation

How do institutional arrangements in banking affect the occurrence of crises?The first two chapters present an endeavor in using new modeling techniques toanswer this question. Even if the results are not widely accepted, the way inwhich the problem is tackled here is offered for consideration and debate. Iscapital the result of an evolving process that takes advantages of entrepreneurialnetworks? The last chapter put forth a model wherein firms develop economicties with one another. By doing so, a market network unfolds along time as anspontaneous process.In the first chapter, I explore the occurrence of bank runs by developing asensitivity analysis to the model in Diamond and Dybvig (1983). I implement anagent-based economic model to analyze different modifications and extensionsto the original. In 36 experiments based on three different versions of the one-bankmodel the frequency of bank runs dropped from 42% to 17%. This was dueto changes in the payoffs structure and social network effects whereby depositorsgo to the bank if at least three of their proximate neighbors went previously.What is the role of interbank markets and central banks in coping with bankingcrises? In experiments using an agent-based framework with multiple banks andan interbank market. I found that when banks cannot interact, then runs inisolated banks occur with a higher frequency than when banks have equalmarket shares. That is, there are no runs escalating to systemic panics. Incontrast, if one bank has a market share twice as big as the rest, runs spread.The presence of a central bank may unexpectedly increase the occurrence ofbank runs. Institutional complexity helps to reduce the frequency of bank runs.Hence, decentralized institutional structures perform better than centralized ones.The objective in this chapter is to implement a parsimonious agent-basedcomputational model of economic networks whereby agents make strategicdecisions based upon profits and information generated through their immediatesocial network. In this model firms are represented by nodes and the linksbetween each pair of them are the result of a mutually advantageous economicdecision. Therefore, links are two-sided or undirected. The economic decision isbased on two elements, namely: a myopic profit motive and local informationchanneled through collaborating firms. Here I endogenize the formation anddeletion of links. Furthermore the number of firms (nodes) in the network at eachtime by allowing firms (nodes) to enter and exit the market. Centrality measuresare reported together with firms’ profits. The evolution of the network yieldshigher connectivity and profits when the (positive) externality is high and the ruleto exit the market more strict. The higher the network connectivity, the higher theoverall profits of firms.