Basel II Pillar 2 Internal Models: Integrating credit and market risk in private equity transactions Erwin Charlier GRM/ERM/Credit Portfolio Modelling 20 November 2006 1
Outline w Private equity market: description w Risk measures: EC and RC w Model features and intuition w Definition of loss w Calibration of the parameters w Results 2
Private equity market 3
Private equity market Main categories: w Leveraged Buy-Outs (LBOs) - Starting point: listed company existing for some time Management buys (part of) the firm and mainly uses debt funding Results in volatile payouts to equity holders (leverage) w Venture Capital (VC) - Small private companies that need money for start-up or growth Difficult to obtain financing through regular channels (banks: no collateral, public offerings too expensive 4
Private equity model focus and purpose w Main purpose: Determine risk for internal decision making - Achieved by quantifying Economic Capital (EC) w Additional spin-off: application for RC - In progress. Requires several modifications (e. g. loss definition, horizon, confidence level) resulting in RC<EC 5
EC at group level: portfolio losses 99. 95% Confidence level ‘Expected’ Loss to be covered by net margin income Risk-based Economic Capital i. e. potential ‘unexpected losses’ against which capital must be held Potential ‘unexpected losses’ against which it would be uneconomical to hold capital Losses 6
Features w Model that recognizes diversification with the rest of the balance sheet w Model that recognizes leverage and combines market and credit risk w Uses Fair Market Values (FMVs) and expected exit proceeds as inputs and is consistent with those w Includes marketability discount w Founded on solid economic and financial theory (structural form model) 7
Model Intuition w (private) equity transactions have option like pay-offs so why not model them as options? w Equity is a call option on assets of the firm w Exact formula dependent on deal structure and leverage w Use the asset correlations that are already there to get the required interdependence in a natural way 8
EC computation procedure: equity and debt values 9
EC computation procedure: definition of loss for EC Based on risk/return framework: w Losses defined as deviation from expected future value (compare: expected return) of our private equity stakes w RAROC and Economic Profit (EP) computations: should include the expected return of the previous item 10
EC computation procedure 1. Estimate parameters 2. Simulate correlated asset evolutions with asset correlations that are already present in current model (based on each investment’s industry and region) 3. In each scenario compute equity values, losses and contributions to portfolio loss 4. Allocate EC based on contribution to extreme portfolio losses 11
EC computation procedure: calibration of parameters If reliable PDs are available: w FMV contains information on current asset value and its volatility w UCR implied PD contains information on current asset value, its volatility and the average return on the assets w Exit proceeds and exit horizon contains information on current asset value, its volatility and the average return on the assets 12
EC computation procedure: calibration of parameters If no reliable PDs are available: w Volatility of assets from region and industry average asset volatilities w Current asset value: from FMV combined with previous point w Average return on the assets: based on exit proceeds and exit horizon and the results from the 2 previous points 13
Results w Risk weight (compared to FMV): 45 - 61% (credit risk: 20% for corporates, small for financials) w Basel 2 regulatory capital: 24 - 32% but different horizon, confidence level and loss definition w Aligning Basel 2 would result in risk weights in line with the 45 – 61% w To compute RAROC and Economic Profit, expected gains should be included 14
Model risks w Model incompleteness: deals have features not fully captured by the model. Issue of materiality. w Parameter uncertainty: historical data are used in the general methodology. What if data not rich enough to generate future scenarios? Will all need to be addressed. BUT: Stress testing will not give you the answers to this! 15
Conclusions Why stress testing: w To clarify the risks that are faced This can also be achieved using other methods than scenario/sensitivity analysis (e. g. summary statistics, statistical analysis/scenarios, visualization) Additional risk to cover: w Model risk BUT: Stress testing will not give you the answers to this! 16
Скачать презентацию Basel II Pillar 2 Internal Models Integrating credit
90f923c1720307c9f303b6fb3f520712.ppt