During the crisis, the regulators aimed for model accuracy as their guideline toward normalization of the marketplace. To jump start the process, regulators applied stress tests to aid in price discovery and to rebuild capital. They also set about developing Basel III to permanently close off regulatory capital arbitrage loopholes identified in the crisis.
Their thesis was that first, if everyone could agree on the fundamental value of an asset then, where necessary, troubled assets could be cleared through the system. Secondly, if more sensible capital charges could be assessed, then the public’s confidence could be restored in critical financial institutions.
While not the source of contagion in this latest cycle, commercial real estate (CRE) assets were treated with especial vigor. The stress tests forced relatively aggressive revaluations of legacy CRE assets, and risk-based capital reforms are encouraging capital levels permanently higher along a number of dimensions.
And yet, the Financial Stability Oversight Council (FSOC) and the Office of Financial Research (OFR) have never outed CRE as a source of systemic risk. Each of these bodies publishes an annual report, which is really more of a review of accomplishments during the year combined with a risk assessment of the global markets. None of these reports, at least not as of their most recent, the OFR’s December 2014 annual, suggests that the CRE sector poses a threat to financial stability.
Considering that our European counterparts take a very dim view of CRE, it is interesting to note that for the U.S. regulators, one of CRE’s minuses may be protecting it from a systemic risk label. Unlike other traded credit classes, CRE products resist product standardization, and therefore are limited in the degree to which they can be integrated into faster-paced trading strategies. Other product sets that are more actively traded have attracted a considerable amount of attention from the regulators.
From a systemic risk perspective, CRE debt’s idiosyncratic nature can be viewed as a strength. If all of our financial products were standardized enough to be actively bought and sold, then the system as a whole would be relatively more exposed to herding behavior. By this measure, CRE represents a stabilizing influence, because our investors can be counted upon to be somewhat less reactive in bad times.
But, enough is often not enough. In their April 6, 2015 Liberty Street Blog post, members of the New York Fed’s Research and Statistics Group made the case that over time, the stress tests are yielding some convergence between regulatory and bank models. This would be expected, as banks would naturally try to anticipate their supervisors’ requirements in order to pass the tests. However, according to this recent analysis, the two sides remain farther apart still on net charge-off estimates, and especially those related to CRE loans. That’s not good news for commercial real estate lending.
On the surface of things, variance between the regulatory and the industry models can suggest poorly quantified risk-taking on the part of the banks. Potential model inaccuracies gained attention well before this spring with earlier analyses by the Basel Committee on Banking Supervision (BCBS). As part of their efforts to further refine Basel III, the BCBS has run a series of analyses of bank capital estimations over time to learn more about bank practices. At least on the surface, the BCBS has found that despite the rollout of Basel III, many jurisdictions continue to view capital requirements differently. Looking below the surface, however, these differences do not necessarily mean anything other than the portfolios in question might not be as similar as we think.
To determine whether model variances are a sign of accuracy or of aggressive risk-taking, further work needs to be done. A model used to estimate capital requirements for a CMBS portfolio in the U.S. may act differently than a model used to estimate capital for CMBS in Italy. Both models could be equally robust, yet the underlying risks are very different, which could explain even large variances in capital requirements.
Nuances in product type, market functioning, legal system, and other factors could easily account for differentials in model output across banks and especially across jurisdictions. Concluding that these variances are a sign of problems with the models may be ignoring valid differences in portfolio fundamentals, and the regulators may be promoting conservatism over accuracy. Yet, the regulators typically do not provide the depth of analysis necessary to explain model variances, though international working groups have consistently requested that they do so.
Yogi Berra summed it up pretty well, when he said, “If you don’t know where you are going, you might end up someplace else.” For some, the idea of accuracy may have been more palatable before the propensity for dispersion was made plain.
To get back to that place they had imagined the regulators have a couple of tools at their disposal. Firstly, they can seek convergence through the stress tests by encouraging the banks to agree to more conservative charge-off assumptions. Since their introduction in 2009, the stress tests effectively guided capital estimates higher by forcing banks to use compromise assumptions that blend peak and trough conditions. Secondly, for more consistency in the capital regime ongoing and across a greater number of institutions, the regulators are also preparing to interject a set of floors into the internal models approach for Basel III risk-based capital. Going forward, the Basel floors would effectively set a bottom limit on the amount of capital that could be assessed, no matter what an internal model might suggest.
While the regulators clearly believe that further conservatism at the banks is necessary, there is a growing concern that the authorities are breaching an historic and foundational divide between themselves and industry. Even without further adjustments to the capital regime, regulators have become integrally involved in balance sheet allocation decisions and are therefore influencing strategic decision-making.
Extending this idea out, regulators across jurisdictions are encouraging decision-making at the banks that will lead to a new set of challenges for the financial system:
- At the asset class level, the regulators will likely seek more conservative charge-off assumptions for CRE in the 2016 stress test, since they went out of their way to point out these particular variances. Earnings are a source of stability in and of themselves and there is little proof that the regulatory methodologies are more accurate than the industry’s.
- At the banking system level, model convergence can encourage risk concentrations of assets, maturities and counterparties, all of which can serve as triggers for herding behaviors.
- At the systemic level, the transfer of CRE lending from the banking sector to the nonbanks should gain additional momentum. At the same time, regulators have highlighted this shift in recent speeches and papers, including those delivered at the spring World Bank – IMF meetings, emphasizing that disintermediation requires careful monitoring.
Offsetting variances in internal models will come at a price of additional regulatory burden and changes in the nature of risk. The decision to make convergence the norm risks compromising the progression toward model accuracy and the diversity of behaviors in CRE lending. Without better justification, especially while so many other pieces of regulation are still washing through the system, this is the time to step back and to observe market reactions before fixing reforms that have yet to be fully implemented.