An agent-based modelling platform for financial crisis simulation

Abstract

The prediction and the consequences of banking crises continue to be a fab in academic and political discussions. Researchers attempt to describe the causes and the effect of these crises on the other branches of the economic environment. Despite the frequent and the gravity of financial crises, authorities have proved to be unable to prevent them or to contain their negative consequences. In this dissertation, we present an agent-based modelling platform that attempts to explain the links between the financial system and the rest of the economic ecosystem. We begin by discussing major financial crises which date as far back as the 18th century. We then present our modelling platform and the literature that supports it and show how it can be applied to simulate the economic system. Then, we execute simulations, employing our agent-based model to examine the effects of banking regulation and of banking crises on various aspects of the economy. We examine the consequences on financial stability and on the real economy and find that the suggestions of Basel III are inadequate to prevent banking crises and, additionally, have negative effects on the banking sector and on the total output. Our findings suggest that some of the criticism against the proposed measures is justified, since neither economic crises nor contagion are diminished under Basel III. At the same time, our findings support that the stability goal is met, at least in part. We also assess the adequacy of various solutions to bank distress with respect to subjective well-being and we find the detrimental effects of instability on happiness. We also show that societal preferences should be taken into account, the different choices of response carry a different opportunity cost in terms of welfare. To this effect, we propose an ideal policy mix, which depends on the preferences and on the macroeconomic goals of authorities. We also test for the effects corporate governance factors on the ability of financial institutions ability to weather the storm during times when the banking system experiences distress. In this part of our research, we test existing risk estimation models and show that governance factors should be taken into account when evaluating the state of banking institutions. Our empirical results put further emphasis on the inclusion of these factors in the regulation regarding banking supervision, particularly in context of the Eurozone banking union. We modify our modelling platform to simulate two current events: the banking crisis in Greece and the oncoming Brexit. In Greece, we examine whether capital controls were enforced at the appropriate time and attempt a prediction on the end of these restrictions. We demonstrate their destructive effects on both the financial system and the real economy and show that the response of Greek authorities was delayed. Regarding Brexit, we test for the short-term and long-term effects on both sides of the Channel, taking into account the relative strength of the UK economy and the banking sector vis-à-vis the EU. Our results in this part of the research confirm predictions regarding the output cost of Brexit, but show that there is big burden to be borne by the EU as well, with the banking system suffering significant losses, particularly over the longer term. We also propose that that policymakers should take into consideration the dynamic effects that may be caused by UK bank assets moving to the EU after Brexit, since our model shows that, as the UK banking system loses its assets, the end state of the UK economy is deteriorated while the end state of EU economy is improved.