Academic journal article International Advances in Economic Research

Assessing Interbank Contagion Risk Using Consolidated Data

Academic journal article International Advances in Economic Research

Assessing Interbank Contagion Risk Using Consolidated Data

Article excerpt

Abstract This study uses the maximum entropy method to estimate bilateral interbank exposure in order to simulate the contagion effect in the UK interbank market using consolidated data. Almost all existing studies use unconsolidated data, which could significantly distort the real contagion effect as the banking sectors of most countries are highly concentrated with most large banks owning a significant number of subsidiaries. The results show that exposure is much more severe using consolidated data, implying that some money center banks or systematically important banks were underestimated by the contagion model before the 2008 financial crisis.

Keywords Entropy maximization * Systemic contagion * Consolidated information

JEL G20

Introduction

The benefit of the interbank market is that it provides an optimal allocation of resources. Funds are distributed efficiently from banks that have a comparative advantage in deposit collecting, but are less skilled in investing capital, to banks that are experts in growing assets, but suffer from lack of funding. However, while such credit linkage between banks provides a risk-sharing mechanism, it leads to an important source of contagion, as manifested by the 2008 financial crisis. Since the late 1990s, the contagion risk in the interbank market has been assessed across many central banks. Many of them resort to a conditional simulation which starts by assuming that a bank is not able to repay its borrowings. Contagious defaults occur if the losses on the exposure to the defaulting bank exceed the capital of a creditor. The losses at the creditor banks are then computed. As each default weakens the surviving banks, it may cause a chain reaction of defaults, resembling a fall of domino pieces. Pre-crisis empirical studies indicate that the scope of contagion is quite limited in some countries. (1)

However, as the simulation is conditional owing to a number of assumptions regarding various issues, including research focus and data restrictions, the results are subject to important caveats and bias. Among the various sources of bias, the issue of estimating consolidated exposure using unconsolidated data mostly has been ignored by researchers.

Underestimation could arise by using consolidated data. Consolidated financial statements present financial information about the undertakings of both a parent company and its subsidiary as a single economic unit, while unconsolidated statements present them as separate units. As the banking industry becomes increasingly integrated and concentrated, large banks usually own a significant number of subsidiaries, which are most likely to stand or fall together, therefore, losses from exposure to one subsidiary are not independent from losses from exposure to another. American International Group, the world's biggest insurance group and an insurer of leveraged debt, is the prime example of this in the late 2008 crisis. The entire company was on the brink of failure, even though a number of its insurance subsidiaries were solvent. Using unconsolidated data in the contagion simulation could seriously affect the robustness of the results of pre-crisis studies and policy implications.

This study analyses this issue by simulating the contagion effect in the UK interbank market using consolidated data. To facilitate the simulation, it is crucial to have the data of bilateral interbank exposure, that is, the amount of interbank lending to and borrowing from a specific bank. However, this type of data is difficult to obtain for many central banks, even at the supervisory level. (2) Banks release only aggregate data revealing total interbank exposure. Following existing studies, this study uses the entropy maximization method (ME) to estimate bilateral interbank exposure. The method assumes that banks seek to maximize the dispersion of their interbank activity, which means each bank symmetrically holds claims on all other banks in the system. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.