Asset connectedness is the concern that the failure of one financial institution will provoke a chain reaction of failures by other financial institutions with direct credit exposures to the failed institutions. This phenomenon was not observed in the financial crisis. Most important, the losses imposed on firms by Lehman were not large enough to push them into bankruptcy. Liability connectedness refers to the connection between the providers and recipients of short-term funding, whereby if a funding institution fails, the failure of its dependent recipient institutions will also result. Like asset connectedness, liability connectedness defaults were not a problem during the crisis, as Lehman, nor other financial institutions, were an important source of short-term funding for other financial institutions. Most of the connectedness literature is focused on liabilities rather than assets.
Much of the literature on connectedness is theoretical in nature, given the rarity of financial crises and the tendency of governments to intervene when large financial institutions become distressed. Network theory is emerging as a powerful tool to analyze how asset and liability connectedness influence the propagation of shocks throughout the financial system. However, most of the literature that uses network theory to analyze financial stability makes unrealistic assumptions about bank behavior or assumes implausibly large idiosyncratic shocks. Some models do incorporate behavior consistent with the contagious runs witnessed in the 2008–2009 crisis, and may therefore shed light on the relationship between asset connectedness and contagion. Connectedness is tinder that might allow a small spark to ignite contagious runs.
The economic literature on asset connectedness (or connectedness) supports the conclusion that asset connectedness is not likely to be a major source of systemic risk. A simple theoretical structure of asset connectedness would be the following: Bank B has direct exposure to Bank A (e.g., owning debt extended to Bank A). Bank C has direct exposure to Bank B. If Bank A fails, then the subsequent loss to Bank B through its asset exposure to Bank A causes Bank B to fail. Similarly Bank C fails due to its asset exposure to Bank B. These failures can permeate throughout the financial system via asset connectedness. Such an asset connectedness model of systemic failure has been widely studied and universally rejected as a plausible cause of the financial crisis.1 In addition these models generally consider just the fixed credit exposures without taking into account how such exposures are reduced in practice through the use of hedging collateral. For example, if Bank B’s credit quality declines, Bank A may purchase credit default swaps, which will pay off if Bank B does fail, to reduce its exposure to a Bank B failure. The literature concludes that while it is theoretically possible to have chain reactions of default, there would have to be implausibly large shocks for this to occur. This conclusion is supported by the historical record, as no large bank has ever failed as a result of losses incurred in the interbank lending market.2 Furthermore, even the existence of asset connectedness does not imply the presence of substantial risk, since much of the risk from asset connectedness exposure can be reduced through collateral, hedging, and diversification. For a detailed overview of the academic literature on asset connectedness, see the appendix.
Modern financial markets are a highly complex system of financial institutions with a high degree of interdependence and interconnections. Financial institutions are not only connected through exposure on the asset side of the balance sheet, as discussed above, but also on the liability side through funding relationships, referred to herein as “liability connectedness.” As part II demonstrates, Lehman’s failure did not present liability connectedness issues, primarily because Lehman was not a significant funder of the US financial system. Moreover money market funds, which some have suggested are potential sources of liability connectedness as heavy investors in the short-term liabilities of banks, were not, as shown in part II, a significant source of systemic risk. For a more detailed overview of the academic literature on liability connectedness, see the appendix. Overall, the real story of liability connectedness is contagion.