EDUCATIONAL NOTE DYNAMIC CAPITAL ADEQUACY TESTING PROPERTY AND CASUALTY COMMITTEE ON SOLVENCY STANDARDS FOR FINANCIAL INSTITUTIONS

Transcription

1 EDUCATIONAL NOTE Educational notes are not binding. They are provided to help actuaries perform actuarial work and may include eamples, eplanations and/or options. DYNAMIC CAPITAL ADEQUACY TESTING PROPERTY AND CASUALTY COMMITTEE ON SOLVENCY STANDARDS FOR FINANCIAL INSTITUTIONS AUGUST Canadian Institute of Actuaries Ce document est disponible en français Canadian Institute of Actuaries Institut Canadien des Actuaires

2 TABLE OF CONTENTS Page SECTION 1 INTRODUCTION 1.1 Purpose of Educational Note When DCAT is to be Undertaken Intensity of Investigation Involvement of Management Definitions. 1 SECTION 2 MODEL SPECIFICATIONS 2.1 General Input and Assumptions Suggested Output SECTION 3 SCENARIOS 3.1 Base Scenario Investigation of Adverse Scenarios..6 SECTION 4 ADVERSE SCENARIOS TO BE CONSIDERED 4.1 General Adverse scenarios Single Catastrophic Loss Multiple Catastrophic Losses Multiple Large Losses Loss Ratio (Frequency and Severity) Adverse Development Ecessive Inflation Shift in the Yield Curve Drop in Premium Volume Reinsurer Insolvency Unepected Decrease in Asset Value.. 10 SECTION 5 REPORTING 5.1 Actuary s Opinion Required Reporting Intensity of Investigation APPENDIX 1 SAMPLE MODEL; Ehibits I to IV 13

3 SECTION 1 INTRODUCTION 1.1 Purpose of Educational Note The purpose of the educational note is to provide guidance to APPOINTED ACTUARIES of property and casualty INSURERS (i.e., Canadian companies or Canadian branches of foreign companies) regarding the application of the Standard of Practice on Dynamic Capital Adequacy Testing (DCAT). 1.2 When DCAT is to be Undertaken Generally, it is anticipated that DCAT would be undertaken in conjunction with the INSURER s planning eercise. Normally, the DCAT should be completed as soon as possible after the plan is finalized, but in no case should the DCAT be delayed more than 12 months following the end of the fiscal year. 1.3 Intensity of Investigation The intensity of the investigation and the level of detail required for reporting purposes may vary according to the FINANCIAL POSITION of the company, as discussed in Section 5 of this document. 1.4 Involvement of Management Early involvement of management is important for a number of reasons, including: 1.5 Definitions to promote the development of a business plan which will be a suitable starting point for the determination of a base scenario; to encourage the use of DCAT as a management tool; and to minimize the possibility of management suffering last minute surprises. In this educational note, each term defined in small capitals has the meaning given in Section 1 of the consolidated standards of practice. 1

4 SECTION 2 MODEL SPECIFICATIONS 2.1 General Valuation of Policy Liabilities The starting point for any DCAT eercise should be the balance sheet and the valuation of the policy liabilities of the INSURER as at the immediately preceding year-end. Eperience Period The DCAT standard states that the number of years to be reviewed would reflect the cycle and volatility of the INSURER s operation and that the appropriate period is normally three years. With respect to financial statement information, such as epenses, three years would normally be an appropriate eperience period. In most cases, however, the eperience period for the loss data would be the maimum number of years available in the database underlying the valuation of the policy liabilities. Forecast Period The DCAT standard states that the forecast period for a SCENARIO would be long enough to capture the effect of its adversity, and that for a property and casualty INSURER, the appropriate forecast period would be at least two years. In most cases, it is anticipated that DCAT will be undertaken in the second half of the fiscal year, and that the forecast period will consist of the stub or remaining portion of the current fiscal year, and the following complete fiscal year. Input and Assumptions Generally, the DCAT input and assumptions relating to the base scenario should be consistent with the most recent valuation of the policy liabilities, ecept to the etent that subsequent information would indicate that modifications are warranted. Section 2.2 provides a checklist of additional information required. Output The minimum required output is in the form of financial statement ehibits as discussed in Section 2.3 of this document. Additional output should be tailored to the needs of the INSURER, the regulator and other users of the REPORT. Monitoring and Validity Checking An important element of the model is the facility to compare the results projected by the model with the actual results as presented in subsequent annual statements of the INSURER. Significant differences between projected and actual amounts should be reviewed and used to refine the model or assumptions as appropriate. 2.2 Input and Assumptions Although the final output described in Section 2.3 would be on a company-wide basis, much of the input data and many underlying assumptions would be at a more refined level of detail. In selecting the line of business groupings for the input and assumptions, consideration should be given to the line of business groupings underlying the actuary s valuation of the policy liabilities and to the level of detail reflected in the INSURER s business plan. 2

5 Following is a checklist of input and assumptions which generally should be considered for the model, in addition to information required for the valuation of the policy liabilities: Eternal Conditions - Claims trend - Interest rate (yield curve) - Stock market performance - Market value changes for other significant investment classes Premiums - By line of business - Projected written, earned and unearned premiums - Projected collected and uncollected premiums - Change in underlying eposures - Change in rate level - Gross and ceded or gross and net Epected Loss Ratios - By line of business - By accident or underwriting year - Gross and ceded or gross and net Development on Prior Years Reserves - Development corresponding to the amount of discount, if any - Emergence to date in current calendar year Paid Loss Development Patterns - By line of business - By accident or underwriting year - Gross and ceded or gross and net Reinsurance - Details of reinsurance treaties (e.g., retention level, limits, rates, swing rates, commission, participation clause, reinstatement clause) - Financial information on reinsurer, including solvency rating (from company management or outside sources) Epenses - Assumed to be fied over a reasonable range of premiums/eposures - Variable by line of business - By calendar year Deferred Policy Acquisition Epense - Percentage of deferrable epense by line of business Other Underwriting Income 3

6 - Service charges - Other Other Revenues and Epenses - Ancillary operations - Subsidiaries and affiliates - Fluctuations in foreign echange rates - Other Invested Assets - Investment policy - Bonds: distribution with respect to maturity date, par value, market value, book value, coupon rate, book yield - Other fied income: (e.g., mortgages, preferred stock) same information as bonds - Common stocks: volatility as compared to TSE 300 or other published indices - Investments in subsidiaries or affiliates: details including level of epected income and factors which might affect the level of income Receivables and Payables - Projected levels - Factors that may cause significant changes in the levels - Relationship between receivables and payables Other Assets or Liabilities - Projected levels - Factors that may cause significant changes in the levels Ta - Epected ta rates - Elements of income with permanent or timing differences for income ta purposes Capital and Surplus or Head Office Account - Contributed capital or changes planned by INSURER Dividends - Epected policyholder or shareholder dividends 4

7 2.3 Suggested Output The output from the model should be sufficient to produce the following information for the INSURER in total, as at each year-end in the forecast period: Balance Sheet Statement of Income Solvency Test The output should generally correspond to the following pages of the annual statement: Required Output Balance Sheet Statement of Income Solvency Test Canadian Companies (Reference in P&C-1) Assets (Page 20.10) Liabilities, Capital, Surplus and Reserves (Page 20.20) Statement of Income (Page Federal Minimum Asset Test (Pages 30.10, 30.11) Branches of Foreign Co. (Reference in P&C-2) Assets (Page 20.10) Liabilities, Head Office Account and Reserves (page 20.20) Statement of Income (Page 20.30) Federal Test of Adequacy of Deposit in Canada (Pages 30.10, 30.11) Items not directly relating to or supporting the policy liabilities may be combined. Consideration might also be given to the solvency tests required by provincial regulators in the jurisdiction in which the INSURER is licensed. As a complement to the solvency tests referred to above, consideration might be given to suitable early warning tests, including the results of procedures carried out by the Office of the Superintendent of Financial Institutions. 5

8 SECTION 3 SCENARIOS 3.1 Base Scenario The base scenario should reflect the actuary s best estimate of the assumptions required by the model. Normally, the base scenario is consistent with the INSURER s business plan. In some situations, the business plan assumptions may not represent an appropriate base scenario, for eample: A comprehensive business plan has not been developed by the INSURER. The projected net loss ratio for all lines combined differs materially from the actuary s best estimate. The premium growth rate for all lines combined differs from the actuary s best estimate to the etent it would cause a material change in earnings or solvency test results. The business plan anticipates significant changes which, in the actuary s opinion, cannot be achieved. The actuary would report any inconsistencies between the base scenario and the business plan. 3.2 Investigation of Adverse Scenarios The DCAT standard of practice requires that the actuary investigate the risk posed by the scenarios in Section 4. Normally, no additional scenarios would be necessary, although the actuary should consider whether additional scenarios might be required to address unique aspects of an INSURER. Where no scenario represents a significant risk to the surplus of the INSURER, no further analysis is required. Otherwise, the standards require that the actuary perform a detailed analysis of the three scenarios posing the greatest risk. This results in a two-stage process as described below. Stage 1 Not all of the scenarios listed in Section 4 are material to all INSURERS. The actuary should assess each scenario and determine if the INSURER would maintain a margin of at least 10% on the appropriate solvency test after the occurrence of that adverse scenario. If the actuary can demonstrate that such a margin would be maintained after the occurrence of each of the scenarios listed (considered individually), then no further investigation should be required. Stage 2 Based on the Stage 1 review of each scenario, the actuary should select the three scenarios which have the most adverse effect on the INSURER s solvency test. For each of these scenarios, the actuary should consider related events, eamples of which are provided in Section 4. In addition to the three most adverse scenarios, the actuary should also eamine any other scenarios which result in a solvency margin of less than 5%. The scope of the investigation may vary according to the scenario itself, the nature of the INSURER, and its FINANCIAL POSITION. 6

9 SECTION 4 ADVERSE SCENARIOS TO BE CONSIDERED 4.1 General The standard of practice requires that adverse scenarios be plausible. We believe that any event or set of similar events with a probability of occurrence of one percent (1%) or more is plausible. Therefore, for each scenario, the actuary should determine and test the most adverse event with a probability of 1% or more. In many cases, this will require the actuary to use his/her best judgment. Where an adverse scenario depends on changes in economic factors or capital markets, we have estimated an event with a 1% probability and provided a detailed description for consistency. The actual impact on a company, however, will depend on the circumstances of the company. Although most of the following scenarios are characterized by a single adverse event, many will be accompanied by related events. For eample, major earthquakes are usually followed by increased costs and higher reinsurance premiums in the following year or two. When evaluating the impact of each scenario, it is important to consider the impact of any related event, however, it is not necessary to consider the impact of any unrelated events. In the scenarios that follow, we have identified the primary related events which should be considered. For each scenario, the actuary should assume that the adverse event takes place in the fiscal year following the stub period. Where related events continue beyond the end of the fiscal year, the actuary should use the base case assumptions for all other factors. For most of these scenarios, however, the adverse scenarios and related events can be included in a single fiscal year (i.e., the fiscal year following the stub period). The following is a list of adverse scenarios that should be considered by the actuary. Additional events may be required to address unique aspects of an INSURER. 4.2 Adverse Scenarios Single Catastrophic Loss The actuary should consider earthquakes, windstorms, floods, hail or any other single event which could have a material impact on the INSURER. The analysis should be based on the largest single event. The estimate of the amount should be based on simulation models; however, a rule of thumb may be used if a model is not available to the actuary. Earthquake losses should also include fire following. For convenience, a list of the largest catastrophe losses in Canada since 1984 as compiled by the IBC is included in Appendi 2. 7

10 Related Events Reinsurer insolvency Reinstatement or swing rated premium Post-event inflation Liquidity of assets Multiple Catastrophic Losses The actuary should consider two or more losses whose joint probability is approimately 1%. The actuary should generally consider the combination of losses which has the largest impact on the net results of the INSURER. Related Events Reinstatement premium Multiple Large Losses The actuary should define the size of large loss which will depend on the size of the INSURER and will generally be smaller than the INSURER s net retention. Using historical losses trended to current levels and adjusted for the INSURER s current eposure, the actuary should estimate the frequency and severity distribution of these losses. The cumulative distribution may be estimated using assumed distributions or simulation techniques. The cumulative distribution should be constructed for net and gross losses. The adverse scenario will generally be based on the difference between the losses at a 1% probability and the epected large losses (which we assume are already included in the base scenario). Related Events Adjustable reinsurance premium and commission arrangements Reinstatement premium Reinsurer insolvency Loss Ratio (Frequency and Severity) The actuary should model the accident year loss ratio or frequency and severity of losses. Since catastrophes, large losses and adverse development are considered in other scenarios, the actuary could remove unusual losses from the data prior to his/her analysis. It is generally prudent to eamine the variability of the normal accident year/underwriting year loss ratio or the combined frequency/severity distribution. The actuary should consider the highest loss ratio possible to within a 1% probability. It would generally be appropriate to use the epected loss ratio plus 2.33 standard deviations. Related Events Adjustable reinsurance premium and commission arrangements 8

11 4.2.5 Adverse Development The actuary could model the volatility of the report to report factors by line, or look at the actual development, all lines combined, for ten or more years. When it is not possible to calculate a value for the adverse development, a default value of 20% of net unpaid losses may be used. Related Events Projected accident year/underwriting year loss ratios should be increased consistent with the development on the older accident years/underwriting years Ecessive Inflation In this scenario, the actuary should eplicitly consider a loss trend which is 3.0% above the base case and continues indefinitely. The actuary should consider the impact on accident year/underwriting year losses, adverse development and interest rates using the following or more conservative assumptions: Increase accident year/underwriting year incurred losses in the projection year by 3%. A 150-basis point increase in the yield curve for all fied income assets and the related change in market value. 50% of claim payments on the unpaid claims will be increased by the higher trend. The remaining costs will not change. These higher epected claims payments will be reflected in the claims liabilities at the end of the year Shift in the Yield Curve The actuary should calculate the impact on market values of an upward parallel shift of 300 basis points in the yield curve for all fied income assets. When an INSURER has a significant amount of preferred stocks, the actuary should assess their sensitivity to a 300-basis point increase in interest rates. The actuary should also consider the impact of a 300-basis point decline in interest rates Drop In Premium Volume Since the base scenario could anticipate a drop in volume from the previous year, the actuary should consider the impact of a 30% reduction from the lower of the plan or prior year. For larger companies (over $200 million in net premiums), a 20% reduction may be used. No reduction in the fied costs should be assumed. Related Events Reduced premiums will most likely result from a soft market; therefore, the projected loss ratio should usually be increased. 9

12 Cost of reinsurance might not decrease in dollar terms. Capital loss might result because of the need to liquidate assets to meet current obligations Reinsurer Insolvency The actuary should calculate the eposure to the largest two reinsurers in terms of unpaid claims including IBNR less payable to, and security held from, the same reinsurers. The actuary should evaluate the impact of a 50% default by the largest reinsurer, a 25% default by the second largest reinsurer, and an additional 25% (or more) default by any reinsurer eperiencing financial problems (e.g., if the largest reinsurer is eperiencing financial problems, a 75% default should be used). This scenario should not be combined with other scenarios Unepected Decrease in Asset Value The actuary should consider the impact of an unepected decrease in the value of common shares, real estate, subsidiaries, or other assets. A separate scenario should be evaluated for each asset type. The following should be used as a guide in selecting the market value change for assets: a 25% decline in the TSE 300 inde a 50% decline in the value of all real estate a 75% decline in the value of the largest subsidiary The actual decline in the INSURER s common shares will depend on their volatility (portfolio beta) relative to the TSE. Where it is not possible to calculate a portfolio beta value, the actuary should use his/her best judgment. Similar considerations should be used when calculating the market value change for the remaining asset classes. Related Events Economic conditions related to the decline in asset values may also impact losses. 10

13 SECTION 5 REPORTING 5.1 Actuary s Opinion The actuary would offer an unqualified opinion if, under the base scenario, the model resulted in a satisfactory solvency test, and if the company remained solvent under each adverse scenario considered. As described in Section 2.3, the solvency test would normally be the federal test found in pages and of the INSURER s annual statement (P&C-1 or P&C- 2). A satisfactory test is one where the assets available for test purposes eceed the assets required for test purposes by no less than 5% (i.e.,!or = 5% Margin). Individual INSURERS are free to select a more rigorous standard, although that standard should normally be consistent from year to year. One or more of the adverse scenarios may give rise to results which are unfavourable, although they do not endanger the solvency of the INSURER, as defined herein. Any such results need not be addressed in the actuary s opinion, but would warrant discussion in the actuary s report. 5.2 Required Reporting As indicated in the DCAT standard, the actuary would REPORT to the board of directors or their audit committee, if they so delegate, in the case of a Canadian INSURER, or to the Chief Agent for Canada in the case of a foreign INSURER. For this purpose, an interpretative REPORT is more useful than a statistical REPORT. The format and contents of the REPORT would be determined by the actuary, taking into consideration the needs of the users of the REPORT. In preparing the REPORT, it should be remembered that the purpose of dynamic capital adequacy testing is to identify: Plausible imminent threats to satisfactory FINANCIAL CONDITION Actions which lessen the likelihood of those threats Actions which would negate any threat that materializes Generally, the actuary would include a discussion of the base scenario and each of the selected adverse scenarios. 11

14 5.3 Intensity of Investigation In most cases, it is anticipated that the actuary would undertake a comprehensive investigation of the selected adverse scenarios and provide a complete REPORT thereon. Under certain circumstances, however, it would be appropriate for the actuary to limit the scope of the investigation and REPORT: If each of the selected adverse scenarios could be demonstrated to result in a margin on the solvency test of no less than 10%. Under these circumstances, the actuary would offer an unqualified opinion as to the FINANCIAL CONDITION of the INSURER and would provide a limited REPORT thereon. The actuary would be required to maintain appropriate documentation. 12

15 APPENDIX 1 SAMPLE MODEL (BASE SCENARIO) Background The purpose of this sample model is to illustrate the concepts of dynamic capital adequacy testing as applied to a relatively simple situation. The company s business planning cycle begins in August each year, with DCAT undertaken shortly thereafter. The input to the model (base scenario) comes from the following sources: Valuation at 31 December 1996 Financial statement information for fiscal years 1994, 1995 and 1996, and for the first si months of 1997 (Ehibit I) Investment information from the company and its investment advisors Company s business plan for 1997 and 1998 (preliminary) The output consists of the balance sheet and statement of operations for fiscal years 1996 through 1998 (Ehibit III) and the minimum asset test for fiscal years 1997 and 1998 (Ehibit IV). An important intermediate step is the preparation of a cash flow statement (Ehibit II, Sheet 7). The company is a federally-registered Canadian company, writing personal lines business (homeowners and automobile) in the provinces of Québec and Ontario. The database underlying the valuation has the data grouped according to these two lines of business, but on a countrywide basis. Premium growth has been projected on an all-lines basis, as have all epense assumptions. The company cedes only a small portion of its eposure to reinsurers. The reinsurance program consists of some ecess of loss treaties, and some facultative cessions on larger property risks. No changes in the reinsurance program are anticipated by management. For this eercise, all of the premium and loss projections are done on a net of reinsurance basis. Ceded amounts were computed as a percentage of gross, based on historical relationships. Premiums Ehibit II, Sheet 1, shows the derivation of projected net written, collected, uncollected, earned and unearned premiums. The company plans to increase its premium revenue by 5% per annum until the year Looking at the premiums written in the first si months of 1997 confirms that, to date, the company is achieving its growth plans. The homeowners line of business represents 25% of the company s total written premium, and automobile represents the remaining 75%. Net written premiums are projected by calendar quarter, from the third quarter of 1997 through to the fourth quarter of Historically, yearly written premiums are distributed by quarter as follows: 24%, 26%, 22% and 28%. The written premium projections flow through to the statement of operations. Ceded written premiums are estimated to represent about 5% of gross for homeowners and about 10% of gross for automobile. 13

16 Being a personal lines writer, a large portion of the company s premiums are received in installments. Based on historical eperience, that portion is epected to be 72% of written premiums for 1997 and The installment plan calls for a 10% deposit, with the remaining balance collected over the policy period, which is assumed to be one year. Premiums not paid through installments, representing 28% of the total, are usually collected in the same month, ecept about 15% which is collected in the following month. The overall collecting pattern is therefore as follows: INSTALLMENT PLAN OTHER THAN QUARTER Deposit Remainder INSTALLMENT TOTAL 1 7.2% 8.1% 26.6% 41.9% % 1.4% 17.6% % 16.2% % 16.2% 5 8.1% 8.1% Total 7.2% 64.8% 28.0% 100.0% In conjunction with the written premium projections, the collecting patterns are used to derive the uncollected premium (i.e., amounts receivable from agents and brokers, policyholders and installments) in the balance sheet. Collected premiums are carried forward to the cash flow statement, along with the service charge arising from the installment plan. It is assumed that premiums are earned uniformly over the policy period. Using quarterly data, and assuming 12-month policy duration, the earning pattern is, therefore, 12.5%, 25%, 25%, 25% and 12.5%. The resulting earned premium estimates are used as the basis for projecting ultimate incurred losses. In conjunction with the written premium projections, they are also used to derive the unearned premium amounts in the balance sheet. 14

17 Losses Ehibit II, Sheets 2 and 3, shows the derivation of net ultimate, paid and unpaid losses for homeowners and automobile, respectively. The losses are assumed to include all loss adjustment epenses. The following selected net ultimate loss ratios represent a three-year average of historical ultimate loss ratios, by line of business: LINE OF BUSINESS ULTIMATE LOSS RATIO Homeowners Automobile 55.0% 80.0% These loss ratios are assumed to be appropriate for both accident years 1997 and It is thereby assumed that the effect of changes in loss costs are approimately offset by changes in the average rate level. Historically, the ceded loss ratio has been approimately 50%. Payment patterns used to derive the paid and unpaid losses are based on the valuation at December 31, 1996, although the actual losses for the first si months of 1997 can be used to test the reasonableness of the payment patterns or the epected loss ratios. The resulting projected unpaid claims flow through to the balance sheet. Ceded unpaid claims (i.e., amounts recoverable from reinsurers) are derived as 6.1% of gross, based on historical eperience. Similarly, anticipated salvage and subrogation is estimated to be 2.7% of gross unpaid claims. The projected calendar period paid losses and change in unpaid amounts can be combined to produce calendar period net incurred losses for the statement of operations. Ceded incurred losses are derived as 50% of ceded earned premiums, assuming no development on prior years reserves. Calendar year paid losses are carried forward to the cash flow statement. 15

18 Epenses and Miscellaneous Balance Sheet Items Commissions are assumed to be 20% and 12% of net written premium for homeowners and automobile, respectively, while premium ta is 3.0% of net written premium for all lines. Deferrable acquisition epenses are computed as a percentage of unearned premiums, and the related liability for unearned commissions is assumed to be $0. Analysis of the epected premium liabilities indicate that no reduction to deferrable acquisition epenses is required and no premium deficiency reserve is required. General epenses, including salaries and rent, were derived as 6% of net written premiums plus a fied component growing at 3% per annum, based on the inflation rate. General epenses are epected to be paid evenly throughout the year. Calendar period paid acquisition and general epenses are carried forward to the cash flow statement. On the asset page of the balance sheet, receivables in respect of Facility Association and the P.R.R. and subsidiaries and affiliates are assumed to be level over the projection period, as are other assets. Similarly on the liability page, payables in respect of subsidiaries and affiliates, epenses due and accrued, and other liabilities are also assumed to be level. Investments and Cash Flow Ehibit II, Sheets 4 and 5, shows the projected investment performance for calendar years 1997 and 1998 based on information provided by the company s investment advisors. An important element of the investment projections is the cash flow statement in Ehibit II, Sheet 7, with input from the premium, loss and epenses projections. In accordance with the company s investment policy, most investments are held in bonds and terms deposits, with relatively small amounts in common and preferred shares. Using the investments at December 31, 1996 as a starting point, it is assumed that all investments will be retained by the company, and no capital gains will arise in the projection period. Some of net cash flow indicated in the cash flow statement are reinvested, resulting in a relatively stable level of cash. The book value of the resulting investment portfolio is reflected in the balance sheet. Investment income arising from bonds and term deposits is derived on the basis of coupon rates, while investment income on shares is derived on the basis of dividend rates. Investment income is carried forward to the statement of operations, while the amounts accrued are shown in the balance sheet. 16

19 Income Ta Ehibit II, Sheet 6, shows the income ta calculations. The statutory income used as the starting point in this ehibit is taken from the statement of operations (income or loss before income taes and etraordinary items). Regular taable income is equal to the statutory income from the statement of operations, less permanent differences and timing differences. Permanent differences consist of the dividend income derived in Ehibit II, Sheets 4 and 5. Timing differences are the sum of 20% of the change in unearned premiums, and the change in net unpaid claims, less the sum of the change in deferred policy acquisition costs, and 95% of the change in discounted net unpaid claims. The ta rate is an average rate based on the company s recent eperience. Deferred income ta is based on the difference between regular taable income and GAAP taable income, the latter amount being equal to the statutory income less the permanent differences described above. Current and (changes in) deferred taes are carried forward to the statement of operations, allowing the computation of the net income (loss) for the year. (Changes in) deferred taes are also used to compute the balance for deferred taes as at 31 December 1997 and In this case, the balance appears on the liability page, using the balance at 31 December 1996 as a starting point and applying the changes reflected in the statement of operations. Capital, Surplus and Reserves In light of the favourable income projections, capital stock is assumed to remain level over the projection period. Earned surplus is projected by adding net income to the earned surplus at 31 December 1996, and, in effect, provides the amount necessary to balance the total liabilities, capital, surplus and reserves, to the total assets. 17