This discussion will address some of those differences in an attempt to answer most of the questions.

Transcription

1 Discussion of Earnings at Risk Model for Measuring IRR Examiners have raised questions related to the TCT ALM model. Many of the questions are the result of an attempt to apply concepts from a NEV model to the EAR (Earnings at Risk) model. While these two models include many of the same inputs in their calculations, they each address the measurement of IRR in distinctly different ways. Not all assumptions for the two models will be consistent with each other. This discussion will address some of those differences in an attempt to answer most of the questions. To start this discussion I address the following foundation concept related to valid research. A key axiom related to valid research is the requirement to minimize the number of dependent variables in the experiment. When multiple dependent variables (those being affected by the experiment) are included in the experiment, they are likely to confound the results and threaten the validity of the findings. To minimize this risk we control changes to the independent variables (e.g., loan balances and amortization rates, investment balances and maturities and deposit balances). This control establishes a base line for the analysis and allows the manipulation of interest rates and cash flows to arrive at a valid measure of IRR in the current balance sheet. In regards to questions related to changes in the balances sheet (i.e., decay rate of deposits). Discounted cash flows method, (commonly referred to as NEV) is designed to create a liquidation value of a credit union. In other words, it assigns a purchase price by discounting loans, investments and deposits. One key variable that must be present for discounting to work is a maturity date for all discounted accounts. Since non-maturity shares, by their very name, do not mature, an imaginary maturity is assigned to allow the discounting to work. In an effort to minimize the purchase of shares that may run off, a decay, or runoff rate, is applied that assumes some shares will be lost. In the EAR method, all assets and liabilities on the balance sheet are held constant so as to measure the actual IRR in the current balance sheet. If any of the assets or liabilities were changed, we would not be measuring the IRR in this balance sheet, but rather in one with varying inputs. These changes threaten the validity of the output and clouds the results. For example, a run off of deposits would also reduce assets and artificially increase the resulting equity ratio. Hence the threat to the validity of the output. For example, if the current balance sheet contains $55,000,000 in total deposits with $25,000,000 in regular shares and we change the balances (due to an assumed decay) to $50,000,000 in total deposits with $20,000,000 in regular shares we are no longer measuring the risk in the original balance sheet. We have altered the independent variables and created a new construct inconsistent with the original one. These changes will affect total assets, cost of funds and interest margin making them inconsistent with the current status. This variability of inputs creates the threat to validity in the output. However, the model provides multiple additional simulations where management may vary the inputs and examine the impact of those changes on IRR, equity and income. For example once the base IRR is established, simulations may be generated that includes the following modifications:

2 Increasing loans Increasing investments Increasing deposits (with specific types or generally applied) Decreasing deposits (with specific types or generally applied) Changing mix of deposits Changing mix of loans By using the additional Simulations to examine changes to the balance sheet, we maintain the validity of the original balance sheet IRR measure and may use it to compare the impacts of designated balance sheet adjustments. Questions related to the application of Shocks to Rates The EAR model employs an application of the Price Elasticity of Demand, to determine the percent of shock that must be applied to each deposit type in order to maintain the current balance of deposits. The Price Elasticity measure employs a regression model to predict runoff at distinct interest rate increases. This algorithm informs the model on the magnitude of cost of funds increases that are required to continue to fund loans and investments to maturity, based on the current balance sheet configuration. Questions related to Amortization Schedules A credit union s ability to mitigate the effect of a shock is, in great part, dictated by its ability to reprice its loan portfolio. The amount of loan balances that are available for repricing is the total principle payments submitted by members each month. Because the principle reduction is set in the loan repayment schedule, and loans are generally within a set range in maturity a straight-line calculation with even cash flows in each period appears valid and applicable in this model. The EAR model uses Weighted Average Maturity of loans to create an anticipated amortization schedule for the principle balances of loans. Loans with common amortization patterns, as determined by WAM are input into the model. Using the WAM a straight-line calculation of amortization is derived. The amount of amortization is then applied to each quarter and year as a baseline for repricing. A statically derived algorithm is then applied to the amortization schedule in shocks to allow for anticipated changes in prepayment speed resulting from shocks of increasing magnitudes. This adjustment allows credit union managers to choose different shocks in simulations to test the outer limits of IRR on their balance sheets.

3 Distinct Shock Applications The table below shows a history of the benchmark US interest rate from 1974 through It can be noted that rates have been volatile over most of that 39 year period. From this overview, on can see that regulatory concern regarding IRR appears well founded. Indeed shocks of all types and directions are displayed. Because of the very fact that different shocks are shown it is important to discuss two distinct types and their impact on the measurement of IRR. The period from 1974 to 1976 was characterized by unstable economic conditions principally influenced by the OPEC and the oil embargo. President Nixon experimented with price controls to combat inflation resulting from the crisis. From 1976 through 1980 inflation was a serious problem within the economy. Interest rates were out of control and President Carter return to price controls and austerity in attempt to stabilize the economy. The period from 1980 through 1988 were years of adjustment as President Reagan initiated deregulation in multiple industries, including the financial industry. Leading the more stable and prosperous years from 1988 through 2004 with rates decreasing substantially from 2000 through Then in 2004 rates began a steady climb as Fed Chairman Greenspan tightened monetary policy in an effort to cool down the economy. This led to a sustained multi-year pattern of increasing rates.

4 Immediate Shock An immediate, or sometimes called Overnight Shock, is displayed below. This type occurred in 1980 and The 1981 occurrence was related to the Savings & Loan crisis that led to their collapse. To understand this type of shock we need to a brief review of economic policy and history. In 1933 in attempt to resuscitate the financial industry Congress implemented Regulation Q as part of Section 11 of the Banking Act of Regulation included a number of stipulations directly affecting credit unions. Among those stipulations, Reg Q: Assigned deposit types to institutions (savings only for credit unions) Set Maximum rates on deposits Prohibited interest on checking One of the primary objectives of Reg. Q was to promote house lending by assuring an ongoing source of low cost funds. The ceiling on deposit rates created a situation of minimal IRR in which S&Ls could comfortably extend 20 and then 30 year mortgages. This condition remained stable for almost 50 years. As mentioned before the period from was characterized by high inflation, instability in the economy and interest rate tension in the financials markets. The prime rate rose to over 16% with mortgage rates starting at 18% and consumer loans starting at 21%. In spite of these high rates the interest ceiling imposed by Reg Q stayed in place.

5 In the midst of this tension between low deposit rates and high loan rates, two distinct events occurred both of which would directly affect the resulting interest rate shock. First, Merrill Lynch seeking to capitalize on the Reg. Q ceilings introduces and a new Overnight Investment Accounts. Called Money Market accounts it offered interest rates well above the Req. Q controlled rates and overnight availability of funds. The new account was a direct competitor to regular savings accounts which were primary source of funds for both all financial institutions but especially Credit Unions and Savings & Loans. The difference between the savings rates and Money Market rates rose to 3% over this period as resulted in what was called disintermediation as money moved from the traditional financial institutions into Money Market accounts. The second event was the initiative toward deregulation which began in The goal appeared to create a more market driven interest rate environment. Phased in over time it resulted in the elimination of Req. Q rate ceilings in With the elimination of the ceilings savings rates at banks, Credit Unions and Saving & Loans increased immediately to match Money Market rates, hence the immediate shock. The results proved disastrous for S&Ls. The increased interest rates drove the cost of funds above the yield on loans which created an unsustainable negative margin. The key was the long-term nature of earning assets which allowed for minimal amortization of principal balances and extreme IRR. The outcome of this immediate shock is well reference is the literature and provides a warning to lenders to avoid excessive exposure to long-term maturities. However, what factors, which led to 1981 immediate shock, are present in our current economic condition? Do we have regulated interest rates? No we do not. Are we currently experiencing high inflation? Do we are not. Do we have limited access to multiple types of deposits such as checking and CDs? No, we can offer them all. Since the answer to each questions is no, then what is the likelihood that we will experience this an Immediate Shock the current economic conditions? The answer is much less than probable. If it is unlikely, then why would we use an immediate shock as our primary analysis of IRR? Sustained for Stepped Shock The alternate type of Shock is the Sustained or Stepped Shock. The history of interest rates in the US provides multiple examples of this type of shock. The example in the table below is the most recent example of this type of shock. The economy has been running clean and stable since the early 1990s. Unemployment was low and the economy was expanding at a steady rate. Then in mid-1994 things began to change.

6 From the mid-1990s the government had been intervening heavily in housing market with the goal of expanding home ownership. Credit was easy with loans such as stated income and negative amortization making home purchases cheap and easy for almost all Americans. Cheap and easy credit led to high appreciation of values in the housing market. The boom also accelerated the demand for consumer credit and the rapid appreciation also resulted in expanded consumer deposits. These effects resulted in inflationary pressures and increased cost of funds for credit unions. With inflation increasing the Fed initiated an ongoing response to slow its effect. Under the direction of Chairman Greenspan the Fed began increasing rates. These increases were down in multiple adjustments to ratchet down demand and cool inflation. The stepped approach was intended to control inflation while taking care not to kill economy. This scenario seems much more consistent with the current status of the economy. And is a much more probable scenario for examining IRR in credit unions today. To cover both possibilities, CostPro ALM provides reports addressing both types of shocks. The immediate is shown in a report called Base Case. The Sustained/Stepped shock is show in a report called Simulation 1. Base Case is presented for reference and review, however, since the most likely shock scenario is the Sustained/Stepped shock Simulation 1 is employed for policy applications.