Regulators hope to make capital calculations more transparent by limiting internal modelling, but the use of more standardised approaches could actually make the process even harder to decipher, writes ISDA’s Mark GheerbrantThe driving force behind much of the regulatory reform agenda enacted over the past several years has been an attempt to make the banking industry more transparent and intelligible to outside observers.
The move against internal modelling for regulatory capital purposes is part of that. Regulators have been concerned that allowing banks to use their own ‘black boxes’ for the calculation of risk-weighted assets has made it much harder to trust the final capital ratio number these calculations produce, and even harder to compare the soundness of one bank against another.
Recent attempts to address this include the removal of the internal model approaches for operational risk and credit valuation adjustment, and restrictions on the use of models for credit risk-weighted assets. The introduction of the non-risk sensitive leverage ratio and the planned rollout of capital floors establish further limits on internal model outputs.
Some of the criticisms directed at models are understandable, but we believe there are other ways to address concerns about comparability and transparency – for instance, through greater consistency in model inputs or regular testing procedures. Choosing instead to clamp down on internal models, or prevent their use entirely, has a number of consequences.
For a start, these changes may actually make transparency and intelligibility even harder to achieve. In seeking to curb internal modelling, it is important to not reintroduce flaws that made their adoption seem so sensible in the first place.
The unique selling point of internal models is that they produce a capital structure that most accurately reflects a bank’s individual risk profile. Less risk-sensitive standardised models and backstops do not allow for such a bespoke approach, and can end up overcharging for high-quality assets and undercharging for low-quality assets. This can lead to a misallocation of resources, and encourage banks to gravitate to higher risk, higher reward strategies. Decisions like these may not be fully reflected in the numbers produced by standardised approaches or the leverage ratio, meaning that bank capital ratios will misstate risk. This is likely to make it harder for investors to understand the true strength of a bank.
The ditching of internal models in favour of standardised approaches could also very quickly leave banks out of step with market risk. Banks constantly update their internal models to reflect market developments such as new products, risks or changes in calibration, like a boxer dodging a barrage of punches. The standard approaches are rarely updated by the Basel Committee on Banking Supervision, and hence cannot capture new, unforeseen challenges.
The ISDA SIMM for modelling initial margin payments is a useful comparison. Regulators require it to be recalibrated at least annually, but standardised models for capital undergo the same treatment every five or 10 years at best. Miscalibration of the model is therefore likely.
One of the classic warnings against the introduction of a one-size-fits-all capital model for the banking industry is ‘herding’. Banks can have very different risk profiles, and bespoke internal models tailored to the particular requirements of that institution are likely to capture risk more effectively. If all banks use the same standardised approaches, they will be driven towards the same types of businesses, thereby reducing diversity in the industry. This means the range of products available to end users will narrow and all banks will become susceptible to the same stress events.
Limits on internal modelling may also impede prudent hedging, as standardised approaches generally do not adequately reflect the risk-reducing effect of hedges or other risk mitigants like initial margin. This could result in extra costs being passed on to end users, and therefore discourage hedging by corporates and pension funds.
Finally, the introduction of more conservative standardised approaches is likely to increase the amount of required capital in the banking sector. This is despite the fact that the Basel Committee and other regulatory bodies have said that further reforms to the capital framework should not produce an overall increase in capital. These increases could be significant. For example, an industry study on the Basel Committee’s Fundamental Review of the Trading Book, conducted by ISDA and other trade associations earlier this year, revealed that market risk capital for FX trading desks could jump by more than six times if that desk loses internal model approval1.
At a time when the cumulative impact of the various reforms to leverage, liquidity and capital is not sufficiently understood, such a capital increase could have a significantly negative effect on the banking industry’s ability to service the real economy. ■
Mark Gheerbrant is head of risk and capital at ISDA in London.