Where the Models Are Wanting Part 1: Banking Sector Stocks and Modern Investment Theory

by Dr Constantin Gurdgiev, Adjunct Assistant Professor of Finance with Trinity College, Dublin

You can refer to part two of my post here.

Traditional financial models treat banking stocks similarly to the shares issued by companies trading in other sectors. This is true under the various multi-factor models (the Arbitrage Pricing models, and the Fama-French and broader multi-factor models) as well as in the case of the balance sheet-based analysis. However, key insights from the Global Financial Crisis offer us the grounds for reconsidering such approaches to pricing banks’ shares.

The first insight relates to the importance of the loops of risk contagion between banks and sovereigns. The second, to the networked nature of the banking sector. Both are still missing in the mainstream equity valuations models deployed by the financial analysts today, and both remain under-developed within the CFA curriculum.

In this post, lets consider the first issue, leaving the discussion of the networks effects for the subsequent post.

January 2014 research paper, titled “The negative feedback loop between banks and sovereigns” claimed that years since the beginning of the banking and sovereign debt crises worldwide the banking systems of several Euro area countries “remain exposed to the vagaries of government bond markets” characterised by “the different channels through which sovereign risk affects banking risk (and vice versa)”.¹

In simple terms, the risk position of the sovereign borrowers has both direct and indirect effects on banks exposures to sovereign debt.

The direct effect, or the first order effect, is risk transmission from the value of the sovereign debt (yield-price pair) to the asset base held by the banks. As yields rise and bond prices fall, banks sustain asset value losses that, sooner or later, have to be recognised. This effect is routinely priced directly into equity valuation models as simple sovereign risk factor, proportional to the volume of bonds holdings by the banks.

However, the same source of risk also has indirect, or second order effects on the banks that can be significant in size, and non-linear and multi-dimensional in nature.

Firstly, sovereign bonds represent banks’ collateral available for repo operations with the central banks and for securing interbank lending. As the value of the bonds falls, their collateral qualities change non-linearly. Until a certain point, usually determined by the investment grade rating of the bonds, their repo value via-a-vis central banks operations remains, generally, unimpaired. Once ratings fall below investment grade, there is a cascading effect or a discrete jump in terms of collateral risks. This non-linearity of risk transmission to the banks’ balance sheets makes any valuation of the banking shares highly uncertain and methodologically problematic.

Secondly, bonds quality (or put differently, risks associated with sovereign bonds) can have a potential impact on bank capital. Today, the Basel rules governing regulatory capital definitions use zero risk-weighting on sovereign bonds. In simple terms, banks are not required to set aside any capital to cover their government bonds’ exposures. However, the Basel Committee for Banking Supervision is currently reviewing this rule with a view of adopting some sort of risk-linked weighting measure that would require banks to provide capital cover against some of the riskier Government bonds.²

Thirdly, sovereign risks that materially impact the banking system can also generate contagion back to the sovereign whenever the state becomes either the guarantor or the lender of last resort to the banks. Using state funding can link banks future operations, strategies and performance to the state policy. It also can further undermine the solvency of the state, triggering more losses on bonds held by the banks.³

Once again, the core problem here is that none of these second order risks can be modelled robustly using traditional factor analysis or balance sheet models. The reason for this is simple. The first and the last sets of effects relate to rare events for which data is never sufficiently available to allow for rigorous econometric analysis. In addition, the third set of risks relates, at least in part, to political risks, which are virtually totally absent from valuations models used by the majority of analysts.

The second set of effects is subject to the potential new regulatory changes. While giving some definition to the sovereign bonds-banks risks links, the new Basel rules can leave banks and analysts under the assumption of false safety. If so, we are likely to replay the same failures as we have witnessed before with the AAA-rated securitization products and large lower tier capital buffers. All of these have passed the regulatory tests for safety prior to the crisis, yet have proven to be useless as risk buffers in the crisis.

References

1. Paolo Angelini, Giuseppe Grande and Fabio Panetta, “The negative feedback loop between banks and sovereigns“, Banca d’Italia, Occasional Paper Number 2013.

2. See “Bank supervisor to review sovereign debt rules” by Caroline Binham, January 23, 2015.

3. Perhaps not surprisingly, the third dimension of risks is also linked to the first one: cascading nature of sovereign risks impacts not only banks within the country, but also international banks, and this can result in cross-border cascades across both sovereigns and banks. For example, see “Systemic risk and sovereign debt in the Euro area” by Radev, Deyan.

CFA Training - Free Trial with Learn Signal for WSO Members