Future of Asset Liability Management Abstract This paper identies, describes, and evaluates the analytical tools that CEOs and corporate boards have used to understand capital and earnings exposure to interest rate risk. Furthermore, the paper examines lessons learned as a result of the market turbulence that began in late 2007 and explores how the nature of nancial risk management is evolving. Most banks actively monitor structural asset and liability risk; however, unlike market or credit risk, there are no standard metrics to assess asset and liability risk. In fact, practitioners do not even agree on whether interest rate risk metrics should be based on earnings sensitivity, market value sensitivity, or the traditional cash ow gap model. Given this lack of consensus and standardization, it is not surprising that there is a wide range of Asset Liability Management (ALM) practices and sophistication. Moreover, while no ALM risk management tool is ideal, each has its strengths and weaknesses. To effectively manage their balance sheet, nancial institutions must select the risk management tools that are appropriate for the institutions size, complexity and risk management objectives in the context of a dynamic regulatory environment. The global regulatory community has indicated that ALM was not adequate and is forcing banks to improve their risk management practices. MODELING METHODOLOGY Author Robert J. Wyle, CFA Senior Director - Product Management Rob joined Moodys Analytics from KPMG where he was the Director of the National ALM Practice. Prior to joining KPMG, Rob was Director of Market Risk Management at E*TRADE Financial where he was responsible for the daily quantication of interest rate risk. Prior to that, Rob was the ALM Product Manager for SunGard Trading and Risk Systems. Other professional risk management roles include the Dime Savings Bank and the Federal Home Loan Bank of New York. Contact Us Americas +1.212.553.1653 clientservices@moodys.com Europe +44.20.7772.5454 clientservices.emea@moodys.com Asia (Excluding Japan) +85.2.2916.1121 clientservices.asia@moodys.com Japan +81.3.5408.4100 clientservices.japan@moodys.com 2 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS CONTENTS 1. INTRODUCTION ...............................................................................................................................................3 2. TRADITIONAL INTEREST RATE RISK MANAGEMENT TOOLS ....................................................................5 2.1 The Periodic Gap Model .................................................................................................................................. 5 2.2 Net Interest Income (NII) Simulation and Earnings at Risk (EaR) ........................................................ 7 2.3 Market Value of Portfolio Equity ................................................................................................................. 11 2.4 Value-at-Risk (VaR) ......................................................................................................................................... 16 VAR CASE STUDY ...............................................................................................................................................18 3. ADVANCES IN ALM RISK QUANTIFICATION ..............................................................................................19 4. CONCLUSION ................................................................................................................................................ 22 FURTHER READING ........................................................................................................................................... 23 ABOUT MOODYS ANALYTICS ALM SOLUTION .......................................................................................... 23 APPENDIX A VAR RISK FACTOR DECOMPOSITION ................................................................................... 24 APPENDIX B VAR INTEREST RATE RISK DECOMPOSITION ...................................................................... 25 REFERENCES ........................................................................................................................................................ 26 3 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS 1. Introduction ALM is dened in several ways. Traditionally, ALM has been associated with the management of structural balance sheet interest rate risk (IRR). The traditional denition of IRR is the exposure of a banks nancial condition to adverse movements in interest rates 1 . This denition, however, considers only a single risk factor. Consistent with this denition, the tools for measuring and monitoring IRR have historically been the Repricing Gap Model, Net Interest Income (NII) Simulation, and the Sensitivity of Market Value of Portfolio Equity. At some banks, due to the concentration of skills and cash ow models, the ALM function is also responsible for performing a variety of other balance sheet management analyses including: Liquidity Risk Funds Transfer Pricing (FTP) Capital Management Risk Policy Setting Therefore, depending on the institution, the full scope of ALM can be much broader than just IRR. And given the lack of standardization in the industry, it is not surprising that there is a wide range of sophistication in ALM. Approaches to ALM can be broadly categorized as simple or sophisticated: Simpler Approaches to ALM Periodic Gap Model Calculating the impact of parallel and instantaneous interest rate shocks on static earnings or market value using discounted cash ow analysis. Sophisticated Approaches to ALM Dynamic simulation of the balance sheet under multiple interest rate scenarios Option Adjusted Valuation (OAV) Volatility-based risk metrics, that include Value at Risk (VaR), stochastic Earnings at Risk (EaR), Risk Adjusted Return on Capital (RAROC), or Economic Value Added (EVA) While regulators and practitioners might agree on what ALM risk management tools are available, they do not necessarily agree on which ALM tools should be used to quantify risk (that is, an earnings-based sensitivity or market value sensitivity.) Regulators are demanding stronger risk management practices and so banks are increasingly looking at more sophisticated approaches to ALM. This lack of consensus on how to measure ALM risk was a major reason why interest rate risk outside the trading book was not subjected to an explicit (Pillar 1) capital charge under Basel II but is covered under Pillar 2 instead (BCBS 2005, 762). Lessons Learned The industry and regulatory response to the market dislocations that began in late 2007 have sparked renewed interest in ALM. Traditional approaches to ALM did not foresee or prevent the credit crisis. Some of the key lessons learned include 2 : The need for effective rm-wide risk identication and analysis: Firms that were better able to share quantitative and qualitative information across the enterprise were better able to identify the sources for inherent risks sooner. These rms identied risks earlier which gave them more time to develop rm-wide solutions rather than wait and hope that the lines would make decisions that benet the rms exposures collectively. Firms that performed less well did not effectively share information, and the lines were left to make decisions in isolation. 1 Principles for the Management and Supervision of Interest Rate Risk; Bank of International Settlements; July 2004 2 Senior Supervisors Group; Observations on Risk Management Practices during the Recent Market Turbulence; March 6, 2008 4 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS The consistent application of independent and rigorous valuation practices across the rm: Firms that performed better throughout the credit crisis generally had a more disciplined enterprise valuation process. These rms developed in-house infrastructure and governance capability to quantify and share the intrinsic value of complex potentially illiquid securities. These rms identied future risks sooner and had more time to frame enterprise risk management strategies. The effective management of funding, liquidity, capital, and the balance sheet: Firms that implemented and enforced enterprise risk management control systems for the management of capital and liquidity tended to perform better during the credit crisis. For example, rms that aligned their Treasury functions more closely with risk management processes provided internal incentives for individual business lines to control activities that might lead to unexpected losses primarily by charging the business lines for contingent liquidity risk. Informative and responsive risk measurement and management reporting and practices: Better performing rms tended to look at risk using metrics that were based on different underlying assumptions and were better able to update their modeling assumptions to better reect current market conditions. These rms were better able to evaluate forward-looking scenarios under changing market circumstances and pursue opportunities as they emerged. In contrast, rms that experienced more difculty were dependent on specic risk measures which might have been based on outdated information. Emerging Trends in ALM The acknowledgment of industry weaknesses and an atmosphere of strong regulatory reform has signaled the incentive for change, post 2008 banking crisis. Governance: Enterprise risk management governance practices are gaining ground driven both by lessons learned by nancial institutions during the credit crisis and as a result of regulatory pressure. Liquidity Risk: The measurement, management, and supervision of liquidity risk is now considered to be as important as capital management. 3 Funds Transfer Pricing: Funds transfer pricing is enjoying a renaissance particularly with regards to incorporating liquidity costs in product pricing, performance measurement, and new product approval. Prior to the credit crisis, many banks treated liquidity like a free good for transfer pricing purposes. This behavior was reportedly one of the causes for the poor liquidity outcomes during the credit crisis. 4 Data Infrastructure: The inevitability of Basel III compliance is forcing banks to invest in data infrastructure. The Basel III framework hinges on integrated asset, capital, and funding management. Basel III liquidity data requirements span multiple functional and organizational silos necessitating the implementation of the enterprise risk management Datamart. ALM: Moving Out of the Back Ofce: ALM is evolving from a back-ofce risk management cost center to an integrated front ofce balance sheet management function. The convergence of market and credit risk has accelerated post crisis partly due to regulatory pressure. Firms that avoided signicant losses appear to have a better ability to integrate exposures across businesses for both market and counterparty risk management. 5 To remain competitive, banks must keep up with the latest developments in risk measurement and management. Ultimately, rms that tie risk exposures to capital more effectively will be better able to integrate risk-taking decisions into their strategic and tactical decision making. 6 3 The Turner Review: A regulatory response to the global banking crisis; March 2009 4 Liquidity Transfer Pricing: a guide to better practice; Bank for International Settlements; December 2011 5 Observations on Risk Management Practices During the Recent Market Turbulence; Senior Supervisors Group; March 2008 6 Governor Susan Schmidt Bies; Board of Governors of the Federal Reserve; March 29th, 2006 5 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS 2. Traditional Interest Rate Risk Management Tools Changes in interest rates can have adverse effects both on a banks earnings and its economic value which has given rise to three separate, but complimentary, perspectives for assessing a banks structural balance sheet interest rate risk exposure: the Periodic Gap Model, Market Value of Portfolio Equity (MVPE), and Net Interest Income Simulation. 7 2.1 The Periodic Gap Model Gap analysis is a useful asset/liability management tool to manage interest rate risk, make funding decisions, and allocate capital along the yield curve. Interest rate gap models were rst produced in the 1950s when main frame computers became more affordable. These models were further rened and used widely through the late 1970s and are still used to this day across Asia, Europe, and at small institutions globally. Gap is simply the post-hedge difference between rate sensitive assets and rate sensitive liabilities, bucketed into the sooner of reprice, maturity, or expected call date. Floating rate instruments are usually bucketed according to their next reprice date only one time per reporting date. Such treatment is consistent with the cash ow treatment of maturing assets or liabilities and because gap models implicitly assume that each interest rate reset date are treated separately and in sequence. However, the gap treatment of oating rate instruments can be broken down into rate-sensitive and rate-insensitive components in some regions of the world. When periodic gap is zero, net interest income in that period is hedged against changes in interest rates. However, when there is a positive gap, earnings decline as interest rates decrease and rise as interest rates increase. Figure 1 overleaf, shows a typical periodic gap report. 8 All of the cash ows converge to generate two risk metrics: periodic gap and cumulative gap. These gap patterns are interpreted in order to understand interest rate risk exposures. The periodic gap is the difference between rate-sensitive assets and rate-sensitive liabilities net of hedges. It provides an indication of the direction of earnings relative to changes in interest rates. When viewed over time, periodic gap shows whether or not there are asset/liability mismatches. Lumpy periodic gap patterns indicate that earnings will be volatile. The literature advocating the periodic gap model recommends that net interest income be hedged by setting each periodic gap equal to zero. Therefore, in the course of routine balance sheet management, if funds need to be invested, sold, or borrowed, then gap patterns provide a guide as to what maturities should be purchased, sold, or borrowed in order to smooth earnings or take bets on the direction of interest rates. Cumulative gap is less useful from a tactical interest rate risk management perspective because any cumulative gap arises as the result of many preceding periodic gaps. Despite the convenience of a single index number for interest rate risk, cumulative gap is at best ambiguous. However, from a strategic perspective, cumulative gap provides an indication for how the banks capital is invested across the yield curve. Thus, given the short-comings of this tool, both tactical and strategic risk management decisions can be made. 7 Principles for the Management and Supervision of Interest Rates; Basel Committee on Banking Supervision; Page 6; July 2004 8 Interest Rate Risk, Comptrollers Handbook; Comptroller of the Currency (OCC); June 1997 6 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Figure 1: A Typical Gap Report < 1 Mo. 1-3 Mos. 3-6 Mos. 6-12 Mos. 1-2 Yrs. 2-3 Yrs. > 3 Yrs. Total Loans 100 10 20 45 5 20 30 230 Investments 5 5 10 20 20 50 110 Other Assets 5 15 20 Total Assets 105 15 25 55 25 40 95 360 Nonmaturity Deposits -65 -30 -50 -145 CDs and Other Liabilities -35 -35 -45 -30 -10 -10 -20 -185 Total Liabilities -100 -35 -45 -30 -40 -10 -70 -330 Equity -30 Net Periodic Gap 5 -20 -20 25 -15 30 25 0 Cumulative Gap 5 -15 -35 -10 -25 5 30 0 Strengths of the Periodic Gap Model The development costs for implementing periodic gap models are small. In many instances, gap models are produced in spreadsheets. However, for detailed gap reporting, complex simulation software is needed. Periodic gap models are easy and intuitive to understand. (Senior managers who are less familiar with advanced nancial risk management concepts often request Gap models, which are often included as tables in the annual report or 10-K.) Risk managers can easily set risk-limit policies based on the relative size and sequence of gap patterns. Weaknesses of the Periodic Gap Model Sizeable cash ow mismatches can be hidden in gap buckets as short as one month. For example, periodic gap might be positive indicating that earnings should rise when rates go up. However, if the majority of the liability cash ows occur at the beginning of the month while the asset cash ows occur towards the end of the month, then income might actually fall when rates go up. The obvious solution would be to produce gap reports with daily time buckets using powerful cash ow simulation software. However, even with chunky gap buckets (i.e., daily, monthly, quarterly, semi- annual, and annual) gap reports are difcult to interpret. Therefore, a gap report with numerous daily time segments might be difcult if not impossible to interpret. Gap does not provide a single index number for risk exposure. Therefore, different risk professionals can have different interpretations of the same gap report. There is a xation on plugging Gaps to zero which restricts ALM risk management choices because many different non-zero Gap combinations immunize net interest income as well. A risk manager that plugs all gap buckets to zero will unnecessarily incur hedging costs that are too high. Gap models cannot adequately measure risk associated with options sensitive to interest rate risk. For example, difculties arise when non-maturing deposit redemptions or prepayments are sensitive to the spread between contractual and market rates. Gap mistakenly assumes that interest rates change by the same amount for all assets and liabilities. As a result, gap does not sufciently quantify basis risk. 7 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Gap cannot measure or manage more than one risk factor simultaneously. That is, gap models cannot quantify risk exposures for market value capitalization, nancial leverage, or total return on equity simultaneously like other ALM risk metrics. In addition, shareholders care more about a banks stock price. Therefore, their preference would be to position the interest rate risk of their equity based on their outlook of risk and return. Since gap models tend to focus on earnings more than risk exposures of capital, these preferences are not addressed. Interest cash ows are typically not included in gap reports and are therefore ignored. Cash ows of non-interest income and non-interest expense are ignored. Conclusion: The Periodic Gap Model can be Useful Gap can be a useful tool for balance sheet managers, senior managers, and boards of directors depending on the circumstances of use. For example, Gap is more useful for balance sheets that have few embedded options. Second, Gap provides some means of risk control to banks that have meager nancial or information technology resources. Third, Gap models are intuitive and easily understood by senior managers and boards of directors with little nancial expertise. However, Gap models have many disadvantages. Therefore, a risk management program that benets from the positive qualities of gap might be stronger when supplemented with other risk metrics based on alternative underlying assumptions. 2.2 Net Interest Income (NII) Simulation and Earnings at Risk (EaR) From an earnings perspective, the focus of analysis is the impact of changes of interest rates on accrual or reported earnings, usually net interest income. This is the traditional approach to interest rate risk assessment taken by many banks. Variation in earnings is an important focal point for interest rate risk analysis because reduced earnings or outright losses can threaten the nancial stability of an institution by undermining its capital adequacy and by reducing market condence. 9 The simulation model starts with the current balance sheet, including detailed maturity or repricing schedules and the associated rates and yields of those balances, and forecasts the income statement, balance sheet, and cash ow schedules for a series of future time periods (typically 12 - 36 months and even as long seven years). This is accomplished by simulating the repricings, maturities, rollovers, and new business originations for all balance sheet activities of the bank. To generate a plausible set of nancial statements, you must make a number of assumptions about important issues, including target balances, maturity schedules for new business, yield curve behavior, non-yield curve rate assumptions, and pricing assumptions for new business. 9 Principles for the Management and Supervision of Interest Rates; Basel Committee on Banking Supervision; July 2004. 8 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Figure 2: A Typical NII Sensitivity Analysis 3.00% 2.00% 1.00% Base -1.00% -2.00% -3.00% HFS Portfolio 178,560 162,560 146,560 130,560 114,560 98,560 82,560 Investment Securities 53,384 52,758 52,131 51,505 50,879 50,252 49,626 Mortgage Backed Securities 284,649 275,080 264,715 252,192 237,042 219,480 201,044 Residential Loans 625,980 611,801 592,547 563,991 510,149 432,001 376,180 Commercial Mortgages 365,879 342,567 319,258 295,714 270,377 247,222 224,094 Consumer Loans 262,678 251,508 240,328 229,120 217,908 206,678 195,460 Business Loans 104,970 97,971 90,955 83,934 76,888 69,872 62,877 Total Interest Income 1,876,099 1,794,244 1,706,494 1,607,016 1,477,804 1,324,065 1,191,841 Deposit Expense 728,224 679,052 603,002 544,298 468,162 446,370 381,753 Borrowing Expense 595,110 527,611 456,727 383,032 299,256 213,908 152,459 Total Interest Expense 1,323,334 1,206,664 1,059,728 927,331 767,418 660,278 534,212 Net Interest Income 552,765 587,581 646,766 679,685 710,385 663,787 657,628 Provision for Losses - - - - - - - Servicing Hedges Receive - - - - - - - Servicing Hedges Pay 32,000 32,000 32,000 32,000 32,000 32,000 32,000 Non Interest Income - - - - - - - Total Non Interest Income 32,000 32,000 32,000 32,000 32,000 32,000 32,000 Total Non Interest Expense 573,860 573,860 573,860 573,860 573,860 573,860 573,860 Income before taxes 10,905 45,721 104,906 137,825 168,525 121,927 115,768 Income Taxes 4,144 17,374 39,864 52,374 64,040 46,332 43,992 Net Income Before Extraordinary Items 6,761 28,347 65,042 85,452 104,486 75,595 71,776 Extraordinary Items 127,390 140,619 163,110 175,619 187,285 169,578 167,238 Net Income 134,151 168,966 228,151 261,071 291,771 245,172 239,014 % Change Interest Income 17% 12% 6% -8% -18% -26% Interest Expense 43% 30% 14% -17% -29% -42% Net Interest Income -19% -14% -5% 5% -2% -3% 9 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Strengths of NII Net interest Income simulation models produce their results in a dynamic or forward-looking manner whereas market value and gap models produce their results statically. Since simulation models are both forward looking and interactive, they require greater emphasis on assumptions and managerial behavior. Therefore, simulations are unique in that they help managers anticipate future events, perform assumption-sensitivity analysis and provide a means by which managers can test the effect of different external shocks and strategies on income. Unlike the periodic gap model, simulation model results can be unambiguously interpreted. In the example in Figure 2, it is clear that a parallel and instantaneous 100 basis point increase in interest rates will reduce net interest income by 5%. Weaknesses of NII The software, hardware, and personnel requirements necessary to run an ALM model are costly. ALM software models are strategic balance sheet simulation tools for banks. By denition, they require signicant setup and maintenance requiring highly specialized and trained personnel and are capable of performing very large quantities of calculations within one compute. Simulations depend on assumptions and data analyses that place strenuous demands on the operator. For example, for every reporting date, the model operator must parameterize the model with assumptions for new volumes, new volume contractual characteristics, new volume pricing assumptions, prepayment assumptions, and so on. Some of these assumptions are provided by business units or are extrapolated from empirical data. In addition, assumptions become less meaningful as a function of time. Therefore, the simulation results are only as good as the analyst operating the model. Since simulation solves problems by trial and error, a thorough examination of current risks can be clumsy, time consuming, and labor intensive. For example, to produce interest rate risk results that reect a targeted nancial leverage, the operator might reproduce and adjust the results many times before achieving the desired level. Simulation models produce large amounts of output that requires great insight and skill to analyze, interpret, and summarize. ALM models can be black boxes. The internal structure and calculations of models might not reect the bank or institution being modeled. In addition, econometric models and relationships might become invalid with time. NII simulation models can be gamed depending on the scenarios chosen for quantifying risk and the forecast horizon. For example, if the user leverages interest rate ramp scenarios instead of shocks, they can time new volumes of say interest rate swaps in order to reduce income volatility. However, such ALM strategies may reveal no risk where risk is present. For example, lets assume a bank uses an ALM risk management strategy that short funds originations of xed rate assets. Assume that the bank originates a ten year asset and funds it with 50% ve year liabilities and 50% 3 month and 1 month funding. To hedge risk the bank swaps both the asset and the xed rate liability into oating. If the forecast horizon is less than ve years, it is difcult to detect the funding liquidity risk and counter-party credit risk. Stochastic Earnings at Risk (EaR) EaR quanties the maximum earnings decline over a forecast time horizon, within a given level of condence. The idea behind EaR is similar to Value-at-Risk (VaR) in that they are both volatility-based metrics. The major difference is that EaR evaluates earnings sensitivity dynamically over a forecast horizon whereas VaR evaluates risk at a point in time. While the technology is known and readily available, EaR is rare among industry practitioners representing perhaps only 10% of ALM industry users. Most stochastic analysis is limited to VaR based methodologies or Monte Carlo valuation also known as option adjusted valuation (OAV). 10 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS EaR uses a stochastic interest rate term structure model to generate probable interest rate paths to compute an earnings distribution. Stochastic generation of interest rates is considered to be a more robust and exible alternative to the deterministic approach. By simulating many interest rate paths, you can estimate earnings or market value losses over a specied time period and condence level by capturing balance sheet growth or yield curve movements. You can use stochastic EaR for a variety of analyses. For example, beyond generating an earnings distribution based solely on probable interest rate paths, you can construct a rich, multi-factor analysis by quantifying earnings sensitivity that incorporates non-interest rate risk factors like spread risk, index- specic volatility, or even credit risk. Finally, stochastic EaR is also useful in framing hedge strategies that minimize portfolio risk using non-duration based risk metrics. This approach is more robust because one of the limitations of duration is that it assumes parallel movements in the term structure of interest rates. Strengths of Stochastic EaR Considers multiple risk factors (in addition to interest rate). Shows aggregate and component risk factors (that is, in addition to interest rate risk) facilitating risk management and improved hedging decisions. Permits risk managers to express the potential earnings decline within a given probability rather than just the relative earnings decline. Relies on items other than the deterministic base case scenario for NII. Weaknesses of Stochastic EaR Stochastic income modeling is complex and heavily reliant on user-dened assumptions and data, typically resulting in increased operations, compliance, and oversight demands. Stochastic income modeling and the related assumptions require constant updating and back-testing. Stochastic income modeling can be excessively short-term focused and might not capture embedded options risk. The growing complexity of many products and investments heightens the need to consider interest rate risk from a long-term perspective (that is, fair value). For example, the impact of common options such as periodic and lifetime interest rate caps on adjustable-rate mortgages, the prepayment option on xed-rate mortgages, or embedded calls in funding structures might not be discernable if the impact of interest rate changes is evaluated over only the short term. Evaluating interest rate risk solely from exposure to income streams might be insufcient for banks with large positions that reprice or mature in the intermediate-term or long term. That is the horizon for a balance sheet mismatch might be beyond the forecast horizon. Therefore, if a balance sheet mismatch is beyond the forecast horizon, this risk management metric can be gamed. Some securities offer relatively high initial coupon rates to investors at the expense of potentially lower-than-market rates of return beyond the 2-year term point. Failure to consider the value of potential total cash ows under a range of interest rate paths might leave the bank with an instrument that underperforms the market or provides a rate of return below the banks funding costs. Therefore, the fact that NII does not consider the entire life of the balance sheet can lead to a short sided understanding of structural balance sheet interest rate risk. Conclusion: NII Simulation and Stochastic EaR A risk management framework that only considers a short-term risk perspective or does not adequately evaluate embedded options can lead a bank to make nave risk management decisions. Therefore, there are clear benets to using EaR. However, less than 10% of banks use stochastic earnings analysis. The most common reason for focusing on deterministic analysis is that most management teams do not perceive a favorable cost/benet relationship. That is, some individuals do not fully understand and know how to interpret more advanced metrics. 11 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS 2.3 Market Value of Portfolio Equity The sensitivity of the market value of portfolio equity (MVPE) is a useful risk management tool for measuring and managing interest rate risk. MVPE is dened as the market value of assets minus the market value of liabilities net of derivatives. The balance sheet is marked-to-market under multiple interest rate scenarios; typically +/- 300 bps parallel and instantaneous interest rate shocks in 100-bps increments. The table in Figure 3 is an example of an MVPE analysis for a medium-sized balance sheet that has signicant embedded options both on the asset side of the balance sheet and the liability side of the balance sheet. Figure 3: Duration of Equity. A long-term measure of risk in a single index -300 Bps -200 Bps -100 Bps Base Rates +100 Bps +200 Bps +300 Bps Assets Total Invest- ments &MBS 4,210,438 4,086,497 3,962,398 3,828,600 3,686,892 3,542,992 3,400,794 Total Loans 18,930,851 18,700,001 18,449,300 18,143,494 17,761,872 17,307,224 16,830,622 Total Interest Earning Assets 23,141,288 22,786,498 22,411,698 21,972,094 21,448,764 20,850,216 20,231,416 ORE, Other Assets 2,116,185 2,355,803 2,554,619 2,673,208 2,733,383 2,760,875 2,760,091 Off Balance Sheet Positions (273,704) (227,759) (136,195) 12,233 201,353 419,808 644,944 Total Assets 24,983,769 24,914,542 24,830,122 24,657,535 24,383,499 24,030,899 23,636,452 Liabilities and Stockholders Equity Total Deposits 13,649,749 13,398,552 13,189,325 13,037,661 12,921,253 12,815,022 12,710,552 Total Borrowed Funds 9,142,738 9,108,658 9,075,009 9,042,338 9,011,119 8,981,450 8,953,361 Total Interest Bearing Liabili- ties 22,792,487 22,507,210 22,264,334 22,080,000 21,932,372 21,796,472 21,663,913 Other Liabilities 485,158 485,158 485,158 485,158 485,158 485,158 485,158 Off Balance Sheet Positions 39,881 26,313 13,235 1,215 (9,067) (17,899) (25,923) Total Liabilities 23,317,526 23,018,681 22,762,728 22,566,372 22,408,463 22,263,731 22,123,148 Stockholders Equity 1,666,243 1,895,861 2,067,394 2,091,163 1,975,036 1,767,168 1,513,304 Total Liabilities & Stockholders' Equity 24,983,769 24,914,542 24,830,122 24,657,535 24,383,499 24,030,899 23,636,452 MVPE Change from Base (424,920) (195,302) (23,769) (116,127) (323,995) (577,860) % Change -20.3% -9.3% -1.1% -5.6% -15.5% -27.6% Duration of Equity (10.6) (4.7) 2.2 8.2 13.1 12 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Key Facts about Market Value of Portfolio Equity: There is generally an inverse relationship between interest rates and changes in market value. When duration of equity is negative, there is a positive relationship between interest rates and changes in market value. The changes in market value are not linear nor symmetrical. There are various reasons why the market value prole for the price/yield relationship behaves as it does as described in the previous section and illustrated in Figure 4; most notably, the impact of market volatility on embedded options. The market value results can be summarized in a number of ways that provide useful insights into the risk prole of the balance sheet. Some examples include duration, convexity, and VaR.
Figure 4: Market Value of Equity Duration of Equity Duration is a measure for the market value sensitivity of an asset or liability. In this context, duration does not simply mean average life but refers to the change in market value for a parallel change in rates. For example, the concept of duration suggests that an asset with duration of 6 will suffer a 6% price decline for every 100 bps increase in rates. Thus, duration is the rst derivative of the price yield relationship or more formally Price/Yield. Just as the concept of duration can refer to the price sensitivity of an individual nancial instrument or a portfolio of xed income instruments, it can also be applied to the net market value of equity for a banks balance sheet. There are various forms of duration; common ones include: Macauley duration Modied duration Effective duration (most common) Effective duration is more widely used for risk management practices because it incorporates the impact of market volatility on embedded options whereas the other methods do not. Effective duration can then be calculated from the base market value calculation and one or more parallel interest rate scenarios: The interpretation of the percent change in MV to a change in rates using effective duration is D * i. Figure 4 Figure 8 6.66% 6.94% 6.97% 6.39% 5.04% Gap Analysis Beta weighted Gap; Crude Simulation Category-level simulation; Static cash ow market value Probabilistic Evaluation; More detailed simulation; Integration of systmes 4.0% 4.5% 5.0% 5.5% 6.0% 6.5% 7.0% 7.5% 8.0% -300 Bps +200Bps +300Bps +100Bps Base Rates -100Bps -200Bps late 1960s Indicators 1970s Indicators Estimates 1980s Indicators Estimates & Measures 1990's onwards Advance Measures Market Value of Equity 13 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Note: The price yield relationship for an instrument with embedded options is not linear and might not be symmetrical. This is due to the cash ow characteristics of embedded options otherwise known as behavior or correlated risk. Behavior refers to circumstances where contractual terms might be varied by custom or implication due to: The short prepayment option for loans The redemption option of non-maturing deposits Call and put options Caps and Floors These risks are said to be correlated risks because they are dependent upon movements of some underlying variable (that is, interest rates). Therefore, the price/yield relationship of a xed mortgage would have a negatively convex prole shown in Figure 5. Duration of Equity
Figure 5: The First Derivative of the Price Yield Relationship -300 Bps -200 Bps -100 Bps Base Rates +100 Bps +200 Bps +300 Bps Assets Total Investments & MBS 3.0 3.3 3.6 3.9 4.0 Total Loans 1.3 1.5 1.9 2.4 2.7 Total Interest Earning Assets 1.6 1.8 2.2 2.6 2.9 ORE, Other Assets (9.3) (6.2) (3.3) (1.6) (0.5) Off Balance Sheet Positions (0.5) (0.9) (1.3) (1.6) (1.7) Total Assets 0.3 0.5 0.9 1.3 1.6 Liabilities Total Deposits 1.7 1.4 1.0 0.9 0.8 Total Borrowed Funds 0.4 0.4 0.4 0.3 0.3 Total Interest Bearing Liabilities 1.2 1.0 0.8 0.6 0.6 Other Liabilities - - - - - Off Balance Sheet Positions 1.1 1.0 0.9 0.8 0.7 Total Liabilities 1.2 1.0 0.8 0.7 0.6 Duration (10.6) (4.7) 2.2 8.2 13.1 Duration of Equity The net duration of the market value of equity is a single index for risk across the entire life of the balance sheet. The greater the magnitude of DOE (in a positive or a negative direction), the greater the interest rate risk exposure for MVPE. The asset/liability mismatch scaled by the institutions leverage determines the duration of a nancial institutions MVPE. The impact of the change in the market value of assets and liabilities due to changes in interest rates has entirely the opposite impact on MVPE. Therefore, the notation for DOE can be summarized as: 14 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Where, DOE = Duration of equity MV A = Market value of assets MV L = Market value of liabilities D A = Duration of assets D L = Duration of liabilities Table 3 is the corresponding MVPE prole for Table5 above. Duration of equity is 2.2 years: In addition, DOE can be restated to illustrate the impact of market value capitalization on the risk sensitivity of the balance sheet: For example, if market value capitalization is cut in half, then risk doubles. Conversely, if market value capitalization is doubled, then risk declines by 50%. MV Duration MV Duration MV Duration Assets 95 2.0 95 2.0 95 2.0 Other Assets 5 0.0 0 0.0 15 0.0 100 1.9 95 2.0 110 1.7 Liabilities 90 1.5 90 1.5 90 1.5 Equity 10 5.50 5 11.00 20 2.75 Strengths of Duration of Equity Shows magnitude and timing of cash ows Provides a single exposure number to manage against Identies which transactions drive exposures (risks off-set because duration is additive based on market value, also known as the Taylor series expansion) Easily accommodates unusual security types and correlated interest rate risks (that is, effective duration) Quanties more than one target account at a time (that is, economic leverage and price risk) 15 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Weakness of Duration of Equity Assumes all interest rates change in equal amounts Only effective for small rate changes (convexity) Does not capture future business volumes, managerial decisions, and so on. Convexity The convexity of the price/yield relationship for MVPE is not linear. Convexity is clearly visible in the preceding graph and must be evaluated in order to have a prospective understanding of the price yield relationship. There are many interpretations of the concept of convexity. Mathematically, convexity is the second order derivative of the price/yield relationship (Duration/Yield). That is, it is the change in price not explained by duration or, said another way, the rate at which duration changes for a given change in market rates. The table in Figure 6 contains the corresponding convexity measurements for the preceding two tables. Convexity of Equity
Figure 6: The second derivative of the price yield relationship -300 Bps -200 Bps -100 Bps Base Rates +100 Bps +200 Bps +300 Bps Assets Total Investments & MBS (0.00) (0.24) (0.21) (0.06) 0.05 Total Loans (0.11) (0.30) (0.42) (0.41) (0.13) Total Interest Earning Assets (0.09) (0.29) (0.38) (0.35) (0.10) ORE, Other Assets (1.73) (3.14) (2.19) (1.20) (1.02) Off Balance Sheet Positions - - - - - Total Assets (0.06) (0.36) (0.41) (0.32) (0.17) Liabilities Total Deposits 0.31 0.44 0.27 0.08 0.01 Total Borrowed Funds 0.00 0.01 0.02 0.02 0.02 Total Interest Bearing Liabilities 0.19 0.26 0.17 0.05 0.02 Other Liabilities - - - - - Off Balance Sheet Positions - - - - - Total Liabilities 0.19 0.26 0.17 0.06 0.02 Duration (3.06) (7.15) (6.69) (4.65) (2.60) Note: The duration measurements for some line items are positive while others are negative. Negative convexity is characteristic of the price yield behavior of short prepayment options or in the money call options. A low positive convexity is characteristic of optionless bonds or commercial real estate loans with yield maintenance agreements. A higher degree of positive convexity can be associated with the short redemption option for NMDs depending on the corresponding deposit products and relevant behavior factors. 16 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Conclusion: Keeping Market Value Sensitivity Risk Measures in Perspective Measures of market value sensitivity for the net equity of a balance sheet are not universally accepted around the globe. Consequently, some regulators and nancial institutions favor the sensitivity of earnings as their primary risk metric and do not consider economic value. However, it is important to bear in mind that both give different perspectives of risk because they are based on different underlying assumptions and, while fundamentally different, they are interrelated. First, earnings sensitivity only provides perspective into risk over the forecast horizon whereas market value is a long-term risk measure incorporating all cash ows of the balance sheet through maturity. Therefore, having both the long-term perspective and the short-term perspective adds value because it strengthens a banks risk management framework making it harder to game. For example, what often happens is that banks enter into risk management strategies that maximize short-term earnings without realizing that they are cannibalizing their net equity into the current earnings stream. The effect is that market value of equity will decline over time. Second, earnings volatility and market value sensitivity are interrelated. Consistent with xed income mathematics, if you hedge earnings then market value will become volatile. Conversely, if market value sensitivity is perfectly hedged, then earnings will become volatile. Therefore, beyond the individual strengths and weaknesses of both measures of earnings and market value sensitivity, it is important to understand both. 2.4 Value-at-Risk (VaR) Regulatory requirements have necessitated holding capital to support trading activities since the 1990s. The requirement to quantify market risk for the trading portfolios using VaR was specied in the Basel Market Risk Amendment. Value-at-Risk is a statistical estimate of the maximum potential loss of a nancial transaction or inventory of transactions over a specied time horizon for a given condence level, based upon an adverse move in market factors. It provides a common measure of risk across diverse nancial instruments and markets in terms of potential loss per unit of currency. Although non-trading activities like management of the structural interest rate risk of the balance sheet, which arises primarily from banking activities and the management of the investment portfolio, do not require regulatory capital to support market risk or VaR measurement, some of the tools used to adjust the exposure are often commonly used in non-trading areas. 17 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS How Market Risk is Measured and Managed
Figure 7: The Trading and Banking Books are Each Modeled and Managed Separately Sources of Market Risk Interest Rate Maturity Yield-Curve Risk Basis Risk Options Risk Exchange Rate Commodity Equity Models Used for Market Risk Value-at-Risk Variance/Co-Variance Historical Simulation Monte-Carlo Simulation Models Used for ALM NII Simulation Economic-Value-of-Equity GAP Internal Controls Board & Senior Management oversight Policies, Procedures & Limits Risk Measurement, Monitoring & Information Systems Internal Controls & Audit Trading Book Banking Book Common uses of VaR analysis for the banking book include determining the component of economic capital for structural balance sheet interest rate risk. This form of analysis has application to capital allocation and Risk Adjusted Return on Capital. The properties of VaR can also be used to prepare multi- factor analysis that can isolate the risk contribution of individual factors over multiple holding periods. 18 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS VaR Case Study The VaR analysis in the attached appendices (pages 22-23) was performed by calculating the maximum loss resulting from disturbances in key stress factors related to the assets, liabilities, and off balance sheet items of a Bank. After the computation results were summarized, the one standard deviation (STD) loss arising from a one STD factor disturbance is simply: Price Volatility = Factor Duration * Factor Volatility 10 By applying the Pythagorean Theorem to the constituent price movements, the net portfolio loss is calculated as: Price Volatility = where n is the total of risk factors. The results for an example of this type of analysis are summarized in Appendices A and B (pages 22-23). The stress factors that contribute most to the expected loss are interest rate risk, mortgage spread risk, and deposit retention (see Appendix A). The largest factor is interest rate risk contributing 51% to the total maximum loss within a one-month interval. This risk factor grows in signicance over time explaining 80% of the six-month VAR. The second largest contributor is mortgage spread risk which accounts for 29% of the one-month VAR and 6% of the six-month VAR. Deposit retention uncertainty contributes roughly 17% to the one- month VAR. The interest rate risk factor is composed of short rate risk, curve slope (yield curve twist risk), and a mixed affect. The larger of the three is yield curve twist risk accounting for 82% of the one month variance based loss for this factor. This risk factor decreases in importance to 57% of the six month variance based loss. This trend is explained by the relationship between short rates and the curve slope. That is, these two factors are chosen because they are independent in the very short term. However, over the long term, they become more highly correlated. Therefore, curve twist risk tends to decrease and a mixed effect grows as interest rate risk created by broad movements of the term structure grows in prominence. The relatively large curve slope component of this VaR calculation is caused by (A) the term structure model specication, namely the factor selection with a higher volatility of the slope factor and a lower volatility of the short rate; (B) the allocation of the Banks assets to the intermediate sector funded in part by liabilities in the short sector of the term structure which causes them to be more sensitive to the slope factor; and (C) the potential imperfection of this institutions hedging program that was traditionally based on the conventional single-factor rate model. One of the most interesting outcomes of this analysis was that mortgage spread risk (basis risk) contributes such a large portion to the Banks overall risk prole. None of this Banks interest rate risk controls adequately quantied or hedged this risk. From the factor attribution analysis, it is clear that this factor diminishes in signicance over time. Therefore, this item is a short-term problem with greatest implications for the warehouse, the pipeline, and the mortgage servicing rights portfolio. 10 For purposes of this example, we are making the assumption that the various risk factors of the balance sheet are independent. However, in reality, they have correlation. For a more robust analysis, correlation has to be accounted for and factors can be orthogonalized. 19 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Strengths of VaR Provides common denominator upon which to manage risk Attaches probability to magnitude of loss Allows for decomposition of risks Weaknesses of VaR Susceptibility to assumptions Provides likely picture, but not extreme event results because the underlying assumption behind some forms of VaR is a normal distribution Computational and resource requirements can be high VaR Conclusion VaR technology is somewhat mature since these models have been a regulatory requirement used to quantify regulatory capital to support trading activities since the 1990s. The regulatory requirements impose similar standards on all large banks as to the quantication of trading risk using VaR supplemented by stress scenarios. Meeting the regulatory requirement assures a reasonable level of quality. While non-trading activities (that is, management of structural balance sheet interest rate risk or management of the investment portfolio) do not require regulatory capital to support market risk and therefore do not require VaR measurement, this volatility-based risk measure does offer more insight into the nature of structural balance sheet interest rate risk. Therefore, rather than a replacement for other ALM risk management tools, VaR enhances the universe of ALM risk management tools since it views the balance sheet from different underlying assumptions. 3. Advances in ALM Risk Quantication Prior to the credit crisis, ALM was focused almost entirely on a single risk factor: interest rate risk. This framework was inadequate because it did not predict capital and earnings exposure to market volatility. The global response to these events has imposed new regulatory requirements and has quickened the pace of innovation to create new tools and methods to more effectively assess and manage emerging risks. In many ways, both Basel III and Dodd Frank are forcing banks to implement a more holistic data infrastructure and risk management framework. For example, Basel III imposes a strengthening of the capital risk coverage to include: An integrated management of market and counterparty credit risk The assessment of risk due to deterioration in counterpartys credit rating (the Credit Valuation Adjustment (CVA)) The strengthening of capital requirements for counterparty credit exposures arising from banks derivatives, repo, and securities nancing transactions The elevation of counterparty credit risk management standards by including an explicit wrong-way risk Pillar I capital charge In addition, Comprehensive Capital Analysis and Review(CCAR) and Basel III require a more holistic view when performing stress testing. All of these factors as well as many other parallel regulatory initiatives are overwhelming nancial institutions compliance and IT teams. The industry response to these requirements and lessons learned are starting to take shape. Both the Enterprise Risk Management (ERM) Platform and the integration of risk across the risk taxonomy are emerging as regulatory deadlines near. 20 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS 3.1 The Necessity of an ERM Data Platform The rst step towards refactoring the legacy data architecture to meet imminent regulatory and risk management requirements is to build a centralized risk and nance datamart. In the past, risk management was done in silos. Basel III is now breaking down the walls of those silos. For example, there are now common data elements between Risk, Finance, and Compliance 11 : 10 Liquidity ratios require asset/liability cash ows as well as counterparty details. For example: Securities issuers asset classes, ratings information, and standardized risk weight for LCR buffer For NSFR, identify fully secured mortgage positions eligible for 35% risk weight For leverage ratio, netting is allowed for OTC derivatives and REPO transactions but follows the Basel Current Exposure Approach used for credit risk RWA. The leverage ratio requires information from the risk system Accurate collateral and guarantee information optimizes regulatory capital, and new information is required for Basel III (for example, inception ratings when hedging securitizations) Regulators are even starting to implement consistency checks between risk and nance regulatory reports (FRY 9C vs. FRY 14) A holistic view is now required when performing stress testing or when assessing regulatory costs for new trades at origination Consolidating at a nancial group level requires information about concentration risk, including direct and indirect risks coming from collateral issuers, guarantors, or funds underlying the assets. (This is a signicant challenge for many institutions.) New regulatory requirements are challenging for many nancial institutions that are still organized into silos. For example: Liquidity risk was traditionally managed by nance/treasury teams in charge of ALM. Risk and compliance teams must be involved. Holistic stress testing is very challenging for many institutions. In addition, regulators challenge stress test results more routinely now. In sum, the ALM function is evolving from a silo-based back ofce risk management function to more of an integrated IRR, credit, and liquidity front-ofce balance sheet management function. A centralized risk and nance datamart is a required and fundamental element to navigate the regulatory risk that will dominate the attention of the global nancial services industry over the next decade. Furthermore, metrics that are used for risk management decisions will be more consistent with regulatory reporting if the underlying data has the same source. A good rule of thumb is to focus on data and data quality and invest in a centralized ERM system that meets most of the Basel III and Balance Sheet management analytics and reporting requirements. Lastly, the DataMart must interface with capital markets data, pricing sources, the channels of origination, the channels of distribution (i.e., securitization), and perform holistic stress testing. 3.2 Advanced Analytics That Tie Risk Exposure to Capital ALM tools have evolved since the invention of the mainframe computer which made the rst crude Gap reports possible. Then, gradually, as credit markets and nancial instruments became more complex and computing resources became more powerful and less expensive, IRR measurement became more sophisticated. Simple gap reports gave way to more complex simulation models which eventually lead to Multi-path Probabilistic distributions of earnings and value. Today, we are on the verge of a new era in the quantication of IRR. 11 Regulatory Capital and Liquidity Risk Compliance; Moodys Analytics Risk Practitioners Conference; October 2012; Pierre Etienne Chabanel and Robert J Wyle 21 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Timeline of IRR Measurement Evolution Source: New York Institute of Finance, 1997 Market risk and credit risk have been converging for the last decade. There are numerous examples of credit risk factors that are standard features of risk management and pre-trade analytics systems (for example, Intex, Moodys Analytics Structured Analytics and Valuation, the Andrew Davidson, and Company Loan Dynamics module). The credit crisis and regulatory response has accelerated the evolution of integrated analytics across the taxonomy of risk types. In addition, the new technologies that are being created are being applied to traditional problems like funds transfer pricing, linking pay with performance, capital allocation, or dening risk appetite. The future requirement in Asset and Liability Management will be stress testing on an integrated IRR, credit, and liquidity risk platform. Regulatory requirements are forcing banks to evaluate a more holistic stress testing regime based on severe macro economic scenarios. Understanding the joint dynamics between credit and interest rate risk is critical if banking organizations want to stay competitive. The empirical evidence suggests that at the aggregate level, there is signicant negative contemporaneous correlation between changes in short interest rates and default rates, and signicant positive contemporaneous correlations between the changes in the slopes of term structure and default rates. Over time, changes in interest rates and default rates show negative correlations. In addition, the correlations seem stronger around nancial crises. 1211 Therefore, understanding the relationship between credit and interest rate risk is critical to many applications in nance, from valuation of credit and interest rate-sensitive instruments to risk management and stress testing. Given the lessons learned as a result of the credit crisis, there are a number of applications for an ERM analytical platform with integrated market and credit analytics across the various disciplines of Asset and Liability Management risk functions: Analytics based on the same market data (rates, prices, volatilities, FX), transaction-level bank data, behavior models, new volume assumptions, and cash ow engine provide more consistency. That is, silo-based risk management based on different underlying data, analytics, and assumptions are likely to produce results that imply different risk quantication answers. For example, if the IRR prole and the FTP models are different, the incentives implied through transfer pricing can be inconsistent. 12 The Relationship Between Default Risk and Interest Rates: An Empirical Study; Andrew Kaplin et al; Moodys Analytics; October 2009 Figure 4 Figure 8 6.66% 6.94% 6.97% 6.39% 5.04% Gap Analysis Beta weighted Gap; Crude Simulation Category-level simulation; Static cash ow market value Probabilistic Evaluation; More detailed simulation; Integration of systmes 4.0% 4.5% 5.0% 5.5% 6.0% 6.5% 7.0% 7.5% 8.0% -300 Bps +200Bps +300Bps +100Bps Base Rates -100Bps -200Bps late 1960s Indicators 1970s Indicators Estimates 1980s Indicators Estimates & Measures 1990's onwards Advance Measures Market Value of Equity 22 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Principle 4 from the Basel Liquidity Risk Working Group publication Principles for Sound Liquidity Risk Management and Supervision states that a bank should incorporate liquidity costs, benets and risks in the internal pricing, performance measurement and new product approval process for all signicant business activities (both on- and off-balance sheet), thereby aligning the risk-taking incentives of individual business lines with the liquidity risk exposures their activities create for the bank as a whole. A robust method to compute the contingent liquidity risk transfer price involves the evaluation of the joint dynamics of interest rate risk and credit risk. 1312 The facilitation of capital management for concepts like capital allocation provision for loan and lease losses all require computations of economic capital that benet from the evaluation of the joint dynamics of interest rate and credit risk. 4. Conclusion The nature of risk management is evolving rapidly in the wake of the credit crisis and the corresponding regulatory response. Regulatory pressure to integrate across the taxonomy of risk types is forcing banks to improve their ERM practices and invest in centralized data infrastructure and software. In addition, practices that more closely associate Treasury processes (that is, liquidity risk management, capital management, and balance sheet management) with risk practices across the enterprise are being re- evaluated. More effective risk assessment and risk reporting processes are required. As it was in the past, no individual risk metric is ideal. Rather, they all have strengths and weaknesses. Institutions use those tools that quantify risk consistent with the complexity of the balance sheet. However, viewing risk using metrics based on different underlying assumptions can provide insight into evolving market conditions. In particular, rms that were able to quickly adjust forward-looking scenario analysis or integrate measures of market risk and counterparty credit risk into their positions across businesses were better able to assess evolving market conditions. Thus the tying of risk to losses in capital and earnings exposures and explain those losses will be a major determinant for banks that want to remain competitive. 13 See Implementing High Value Funds Transfer Pricing Systems; Wyle and Tzaig; 2011 23 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Further Reading Additional materials are available to download at www.moodysanalytics.com/riskcondence The New Path to Shareholder Value Robert J. Wyle, CFA. World Finance Magazine, May-June 2012. Implementing High Value Funds Transfer Pricing Systems - Robert J. Wyle, CFA, Yaakov Tsaig, Ph.D. September 2011. Risk Integration: New Top-down Approaches and Correlation Calibration Nan Chen, Andrew Kaplin, Amnon Levy and Yashan Wang. January 2010. The Relationship Between Default Risk and Interest Rates: An Empirical Study - Andrew Kaplin, Amnon Levy, Shisheng Qu, Danni Wang, Yashan Wang and Jing Zhang. October 2009. About Moodys Analytics ALM Solution RiskCondence is Moodys Analytics enterprise asset and liability management (ALM) software solution. It goes beyond the traditional ALM feature set by creating a high value front ofce balance sheet management function. The Solution integrates ALM, liquidity risk management, credit, and regulatory reporting functions onto a single platform with a common data source and single engine strategy. For more information, please visit www.moodysanalytics.com/riskcondence 24 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Appendix A VaR Risk Factor Decomposition Monthly Variance Quarterly Variance Semi-Annual Variance Interest Rates Deposits Retention Mortgage Spread COFI Rates Volatility Assumption Prepayment Error 28.5% 0.4% 16.6% 3.1% 0.0% 0.0% 6.1% 0.5% 11.2% 2.5% 79.8% 11.7% 0.5% 15.3% 3.1% 69.4% 51.3% 0.1% 25 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS Appendix B VaR Interest Rate Risk Decomposition Monthly Variance Quarterly Variance Semi-Annual Variance Short rate Mixed Effect Curve slope 13.60% 15.87% 26.74% 57.39% 17.77% 68.62% 81.60% 7.55% 10.85% 26 MARCH 2013 AN EVALUATION OF INTEREST RATE RISK TOOLS AND THE FUTURE OF ASSET LIABILITY MANAGEMENT MOODYS ANALYTICS 2013 Moodys Analytics, Inc. and/or its licensors and afliates (collectively, MOODYS). All rights reserved. ALL INFORMATION CONTAINED HEREIN IS PROTECTED BY LAW, INCLUDING BUT NOT LIMITED TO, COPYRIGHT LAW, AND NONE OF SUCH INFORMATION MAY BE COPIED OR OTHERWISE REPRODUCED, REPACKAGED, FURTHER TRANSMITTED, TRANSFERRED, DISSEMINATED, REDISTRIBUTED OR RESOLD, OR STORED FOR SUBSEQUENT USE FOR ANY SUCH PURPOSE, IN WHOLE OR IN PART, IN ANY FORM OR MANNER OR BY ANY MEANS WHATSOEVER, BY ANY PERSON WITHOUT MOODYS PRIOR WRITTEN CONSENT. All information contained herein is obtained by MOODYS from sources believed by it to be accurate and reliable. Because of the possibility of human or mechanical error as well as other factors, however, all information contained herein is provided AS IS without warranty of any kind. Under no circumstances shall MOODYS have any liability to any person or en- tity for (a) any loss or damage in whole or in part caused by, resulting from, or relating to, any error (negligent or otherwise) or other circumstance or contingency within or outside the control of MOODYS or any of its directors, ofcers, employees or agents in connection with the procurement, collection, compilation, analysis, interpreta- tion, communication, publication or delivery of any such information, or (b) any direct, indirect, special, consequential, compensatory or incidental damages whatsoever (including without limitation, lost prots), even if MOODYS is advised in advance of the possibility of such damages, resulting from the use of or inability to use, any such information. The ratings, nancial reporting analysis, projections, and other observations, if any, constituting part of the information contained herein are, and must be construed solely as, statements of opinion and not statements of fact or recommendations to purchase, sell or hold any securities. NO WARRANTY, EXPRESS OR IMPLIED, AS TO THE ACCURACY, TIMELINESS, COMPLETENESS, MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OF ANY SUCH RATING OR OTHER OPINION OR INFORMATION IS GIVEN OR MADE BY MOODYS IN ANY FORM OR MANNER WHATSOEVER. Each rating or other opinion must be weighed solely as one factor in any investment decision made by or on behalf of any user of the information contained herein, and each such user must accordingly make its own study and evaluation of each security and of each issuer and guarantor of, and each provider of credit support for, each security that it may consider purchasing, holding, or selling. References 1. International Convergence of Capital Measurement and Capital Standards: A Revised Framework; Basel Committee on Banking Supervision; November 2005 2. What We Know, Dont Know, and Cant Know about Bank Risk: A View from the Trenches; Kuritzkes sand Schuermann; Princeton University Press; February 2006 3. Principles for the Management and Supervision of Interest Rates; Basel Committee on Banking Supervision; July 2004 4. The Handbook of Fixed Income, sixth Edition; Frank J. Fabozzi; McGraw-Hill, 2001 5. Implementing High Value Funds Transfer Pricing Systems; Wyle and Tzaig; September 2011 6. Observations on Risk Management Practices During the Recent Market Turbulence; Observations on Risk Management Practices during the Recent Market Turbulence; Senior Supervisors Group; March 2008 7. Governor Susan Schmidt Bies; Board of Governors of the Federal Reserve; March 29th, 2006 8. Interest Rate Risk, Comptrollers Handbook; Comptroller of the Currency (OCC); June 1997 9. The Turner Review: A regulatory response to the global banking crisis; March 2009 10 Liquidity Transfer Pricing: a guide to better practice; Bank for International Settlements; December 2011 11. Regulatory Capital and Liquidity Risk Compliance; Moodys Analytics Risk Practitioners Conference; October 2012; Pierre Etienne Chabanel and Robert J. Wyle 12. The Relationship Between Default Risk and Interest Rates: An Empirical Study; Andrew Kaplan; Moodys Analytics; October 2009 13. Implementing High Value Funds Transfer Pricing Systems; Wyle and Tzaig; September 2011 SP21911