Fundamental Review of the Trading Book (FRTB) Shearin Cao 12 th June 2019 Disclaimer: The views expressed here are those of the author and do not represent the views of Standard Chartered Bank 8 th Annual Risk EMEA 2019: FRTB 0
FRTB: the Basel initiative to “fix” Market risk capital • History and the purpose of the “Fundamental Review of the Trading Book” (FRTB) § Quick History on Market Risk Rules at Basel, spanning 20+ years o 1996: First introduced market risk capital requirement under Basel II (BCBS 24) o 2009: Major revisions to the market risk, also known as the Basel 2.5 reforms (BCBS 193) o 2012-2014: 3 consultations on the earlier draft of the FRTB rules (BCBS 219, 265 & 305) o 2016: First go at FRTB under Basel IV for 1-Jan 2019 go-live (BCBS 352) o 2017: FRTB go-live extend to 1-Jan 2022 (BCBS 424) o 2018: Consultation on 2016 FRTB rules (BCBS 436) + 1 st set of FAQs (BCBS437) o 2019: Second go at FRTB to address fundamental issues (BCBS 457) § Since the last financial crisis, FRTB is intended to fix the weaknesses in market risk capital measures for both the standardised approaches and the internal models. § Basel believes some aspects of the FRTB rules, i.e. “the overall calibration” or the risk weights for specific asset classes, may be justifiable to be re-calibrated later. Reference: 1. “Explanatory note on the minimum capital requirements for market risk”, Basel Committee on Banking Supervision, January 2019. 8 th Annual Risk EMEA 2019: FRTB 1
FRTB framework at a glance • FRTB framework at a glance 1. Introduces the risk sensitivity to the fall-back Sensitivities-Based Standardised Approach (“SA”) to incentivise appropriate risk management behaviour for all banks. 2. Restricts incentives to use the Internal Models Approach (“IMA”) by increasing the complexity and higher the bar to obtain and continuously retain the IMA approval. 3. Tightens the Trading/Banking book boundary , i.e. more product-driven for designation. • Dynamics between Standard Rules and IMA has been the focus of 2019 “calibration” § FRTB is intended to deliver a broadly similar level of capital across the industry (weighted average or aggregate bank basis) where only some positions have IMA for big banks. § Standard Rules is intended to be approximately 1.5x as conservative as the IMA to avoid further reduction in banks’ market-making capacity and as a credible fallback to IMA. § Challenge remains on how to strike the right balance between the complexity of the IMA (which significantly increases change-the-bank costs for implementation and run-the-bank costs post go-live) and the sufficient capital incentives to invest in widespread and high-quality IMA modelling. 8 th Annual Risk EMEA 2019: FRTB 2
FRTB framework has been significant “re-calibrated” • Dynamic relationship amongst key components offer options for ongoing calibration over time : § Standard Rules relies on sufficient risk sensitivity and its accompanying risk weights to serve as a credible fallback and a floor to the IMA. o Both the risk weights (GIRR and FX) and the capital formula (i.e. Low Correlation Scenario) were “re-calibrated” in the Jan 2019 ruleset. o Curvature is also key to capture the risk sensitivity of the SA which capitalises the higher order risks in addition to Delta and Vega, thus also re-calibrated in the Jan 2019 ruleset. § Non Modellable Risk Factors (“NMRFs”) is the IMA capital add-on to reflect an appropriate liquidity assessment of modelled risk factors. o NMRF capital as it stands could lead to disproportionately high IMA capital and could be highly variable (i.e. due to seasonality). o Changes and ongoing calibration are required to achieve the goal of consistent identification and capitalisation of material risk factors. § P&L Attribution Test (“PLAT”) is the regulatory model performance measure that determines capital regime at the desk-level. o Alternative metrics (i.e. Spearman Correlation with Kolmogorov–Smirnov test) replaced the originally proposed test metrics in Jan 2019 ruleset, yet future re-calibration remain key. 8 th Annual Risk EMEA 2019: FRTB 3
Standard Rules as a credible fallback to the IMA Like-for-like comparison between Standard Rules (“SA”) and IMA varies materially across risk classes. § To deliver the 1.5X target ratio of SA/IMA capital is non-trivial*, due to the capital formula and aggregation. § Components Appetite for Calibration • Jan 2019 ruleset revised the risk weights down for GIRR and FX in particular (IR down 30%, FX down 50%, high-yield sovereign bonds down 33.3%, covered bonds down 37.5%) relative to the 2016 ruleset. Risk Weights • Jan 2019 ruleset also introduced an index bucket for EQ and Credit Spread risks without mandatory decomposition. • The “re-calibration” better aligns the risk weights with the market shocks observed in historical stress period. • Different jurisdictions could introduce new buckets or further revise risk weights as part of the local calibration . • The scenario (out of the three settings - “High/Medium/Low”) which drives the maximum capital determines the SA outcome. The Low Correlation Scenario is defined as the medium correlation multiplied by 75%. Low Correlation • In Rates, high concentration of basis trades used for hedging and market making could attract disproportionately higher capital charge under Low Correlation Scenario. • Jan 2019 ruleset incorporated a formula change (adding in a flooring element) relative to the original 2016 ruleset. • Jan 2019 ruleset introduced a formula change that now applies consistent shocks to similar risk factors and removes Curvature a potential cliff effect in aggregation. • Jan 2019 ruleset revised the risk weights down for Covered bonds by 37.5%. Covered Bonds • near-final CRR2 draft carves out the treatment for bonds issued by credit institutions in Member States and re-assign the risk weight to 200bps (rather than 400bps). • Jan 2019 ruleset reduces the FX risk weights by 50% and allow for “one step” triangulation for crosses between 20 FX Treatment currencies which are recognised as “liquid” pairs. Note: EBA’s Basel III Monitoring exercise report for Europe for QIS 10 (COB 30-June 2018) suggests the capital requirements increase by 81%, assuming current IMA coverage; and by 170% if all banks are capitalised under Standard Rules only. 8 th Annual Risk EMEA 2019: FRTB 4
NMRF is a critical and “novel” IMA component • Banks must evidence that risk factors comply with strict criteria to be included in internal models, with those deemed to be “Non-Modellable” subject to a significant IMA capital add-on. • A significant population of risk factors could be excluded from Expected Shortfall (ES) models which may see more capital “outside” models than “in” models. • The NMRF framework would be subject to ongoing calibration to adequately reflect an appropriate liquidity assessment of risk factors (where “non-modellable” ones have a tougher but more basic model). Definition for NMRF NMRF Framework (Observability and Liquidity) • Data pooling : External vendors collating and providing aggregated data across risk factors and markets to improve “observability”. Capitalisation of Individual NMRFs Aggregation across NMRFs • Seasonality criteria : Where • Liquidity horizon now more • Recognise zero correlation for observable prices may not be available consistently across 12 aligned with the methodology Equity idiosyncratic risk. months, the “Max Gap Rule” is used for the Expected Shortfall • Constant correlation aggregation now revised to “4 observations in model, with a floor of 20 days. approach for flexibility in providing 90 days” to address seasonality • A common stress period is now a procedural tool to regulate the effects, or at least 100 “real” price allowed for all risk factors relevant impact on remaining NMRFs, with observations over a 12-month to a particular risk class. correlation parameter set at 0.6. window in Jan 2019 ruleset. 8 th Annual Risk EMEA 2019: FRTB 5
PLAT determines IMA coverage PLAT is a new requirement designed to assess how close the profit and loss (P&L) representation is § between the front office and the risk engine (i.e. assess the completeness of risk factors and accuracy of the P&L representation in the risk engine). Its objective is to invalidate internal models when risks are not adequately captured in the risk engine. § It is essential that PLAT is developed as a stable and predictable test to create the right incentives to build § the infrastructure capability across all desks in scope and ensure an appropriate (re-)calibration using consistent real portfolio data is put in place for PLA pass/fail rules. 100% 0.00 Once a desk falls in the Red zone, it can only be 80% 0.09 modelled again if all metrics 70% 0.12 are back in the Green band. Note, amber desks attract more capital add-on to incentivise path-to-green. -100% 1.00 Spearman Correlation Kolmogorov-Smirnov Path to Green – Process Illustration - Correlation between - Similarity of the distributions RTPL and HPL of RTPL and HPL 8 th Annual Risk EMEA 2019: FRTB 6
Recommend
More recommend
