recommended system. On the other hand, if the screening criterion were set at five or more exceptional values out of 26 tests, the current tests and bench marks would identify approximately the same percentage of solvent companies as the recommended system, but the percentage of insolvents correctly identified three years before insolvency would drop from 87 to 52 percent. In Exhibit XXII, the solid lines are the same as the lines for the current and recommended systems in Exhibit XXI. The dotted line represents the effectiveness of the five tests retained from the current system, with exceptional values being those in the extreme 10 percent of solvent company results. It is interesting to note that about half of the improvement in effectiveness from the recommended system as measured by the distance from the line representing the current system to the dotted line in Exhibit XXII is due simply to dropping the ineffective tests. The rest of the improve ment the distance from the dotted line to the line for the recommended system is due to adding and modifying tests and adjusting the cutoffs for determining exceptional values. 14 If page 3, line 30 is zero or negative for either the current or the prior year, test result is -99. A - 2 TEST 2: NET GAIN TO TOTAL INCOME All years Current year: page 5, col. 1, line 33 A × 100 = Result (round to nearest integer) If page 5, col. 1, line 7 is zero or negative, no test result is calculated. A 3 TEST 3: COMMISSIONS AND EXPENSES TO PREMIUM All years Current year: page 5, col. 1, line 21 + Current year: page 5, col. 1, line 23 = A A Current year: page 5, col. 1, line 1 = B B = Result (round to nearest integer) If page 5, col. 1, line 1 is zero or negative, no test result is calculated. A - 4 |