DOCUMENT> <TEXT>[Letterhead of Bryan Cave LLP] Heather R. Badami Direct: (202) 508-6082 hrbadami@bryancave.com August 3, 2005 Mr. Oscar Young Ms. Keira Ino Securities and Exchange Commission 100 F Street, N.E. Washington, DC 20549 RE: JAMES RIVER GROUP, INC. FORM S-1 REGISTRATION STATEMENT FILE NO. 333-124605 Dear Mr. Young and Ms. Ino: This letter sets forth the draft responses of James River Group, Inc. (the "Company") to comments provided during our telephone conversation with you on August 2, 2005 with respect to the above-referenced Registration Statement on Form S-1 (the "Form S-1"). Draft responses to each comment are set forth below. Page references are to Pre-effective Amendment No. 3 to the Form S-1. 1. Reserve for losses and loss adjustment expenses - sensitivity, page 32 The Company proposes the revisions to pages 32 and 33 attached to this letter. 2. Reserve for losses and loss adjustment expenses - best estimate, page 31 The Company proposes the revisions to pages 31 and 32 attached to this letter. 3. Application of FASB Statement 123(R), page 60 As outlined in the Company's response dated August 1, 2005, the Company became a public company for purposes of Statement 123 and 123(R) on May 3, 2005, and prior to that date, the Company was a nonpublic entity that used the minimum value method. Paragraph 83 of Statement 123(R) states that nonpublic entities that used the minimum value method of measuring equity share options and similar instruments for either recognition or pro forma disclosure purposes under Statement 123 shall apply Statement 123(R) prospectively. Accordingly, upon adoption of Statement 123(R), the Company will not recognize any expense in the income statement related to awards issued prior to May 3, 2005. Instead, those awards will continue to be accounted for using the provisions of APB Opinion No. 25 and its related interpretive guidance. Footnote 34 of Statement 123(R) states that a nonpublic entity shall apply the method (in this footnote, the "method" is referring to the modified retrospective method) to all prior years for which Statement 123's fair value based method was adopted for recognition, or pro forma disclosure if that date is later than when Statement 123 was effective. The Company believes that Footnote 34 also applies to the modified prospective method by analogy because the intent of Statement 123(R) is that stock option compensation expense subsequent to the date of adoption of Statement 123(R) should be the same under either the modified retrospective method or the modified prospective method. The Company adopted the fair value method for pro forma disclosures under Statement 123 on May 3, 2005, which is later than the date that Statement 123 was effective for the Company. Further, paragraph 74 of Statement 123(R) states that, in applying the modified prospective method, "compensation cost for the portion of awards for which the requisite service has not been rendered that are outstanding as of the required effective date shall be recognized as the requisite service is rendered on or after the required effective date". Based on the interpretation section 74-6 of Deloitte & Touche LLP's "Roadmap to Applying Statement 123(R)", the Company understands that the FASB staff has confirmed that the modified prospective application method is only applied to awards issued while a company was a public company. In addition the Company consulted with their independent registered accounting firm, Ernst & Young LLP, and they concur with this interpretation. Accordingly, the Company will apply the modified prospective application to awards modified, repurchased or canceled on or after May 3, 2005. In further response to this comment, the Company has inserted the word "remaining" prior to "vesting period" on the fifth line of the fourth paragraph of the Statement 123(R) discussion on pages F-11, F-37 and 60. 4. Fair value of stock options, pages F-23 and F-24 This response reflects changes to the footnotes following the table in reponse to your comment. The table below summarizes the Company's history of option grants by date and by quarter from inception. These amounts do not give effect to the ten-for-one stock split the Company intends to effect prior to completion of the offering. The Company's estimation of the fair value of its common stock from inception has been driven principally by (1) equity offerings in which independent parties have Mr. Oscar Young Ms. Keira Ino August 3, 2005 Page 2 participated and (2) the Company's operating results. The Company consistently estimated the fair value of its common stock at $100 per share for the period from inception through April 19, 2004 based on private equity offerings during the same period, as described on pages F-20 and F-21. These offerings involved the participation of the persons described under "Certain Relationships and Related Transactions", as well as independent third parties. The Company increased the estimated fair value of its common stock to $125 per share in November 2004. The $125 per share price was estimated based on the application of valuation metrics for publicly traded commercial lines insurers published in the October 2004 edition of Monthly Deductible by Keefe, Bruyette & Woods, Inc. in which the $125 per share estimated fair value was approximately equal to the median of the implied fair values per share. This increase was further supported by discussions with third party investors regarding a potential equity transaction. From those discussions, the Company believes the transaction could have been completed at a $125 per share price on terms similar to prior equity offerings by the Company. Though the transaction was not completed because circumstances changed the Company's need for the capital, the $125 per share estimated fair value was deemed the appropriate fair value for option grants during that quarter. Based on the above, the Company believes that the estimated fair values used at each option grant date represent the best estimate of the Company's value as of each respective grant date. The estimated fair value of the common stock used for all option grants other than the fourth quarter 2004 option grants corresponds to an actual offering price of shares issued near the grant date. Though the Company believes that the estimated fair value used for the fourth quarter 2004 option grants is appropriate, the application of the initial public offering price of its common stock (expected to be between $160 and $180 per share on a pre-split basis) to the fourth quarter 2004 option grants would be immaterial to earnings and stockholders' equity. The Company attributes the increase in the value of its common stock during 2005 to the increase in profitable operations during 2005, specifically, the Company's reported $9.2 million of net income for the six months ended June 30, 2005 which is a significant improvement over the $8.8 million of net income for the year ended December 31, 2004 and the $5.2 million loss for 2003. Mr. Oscar Young Ms. Keira Ino August 3, 2005 Page 3 ----------- -------------- ------------- -------------- ------------ ----------- Per Share Option Net Income/ Price of Quarter Option Exercise Options (Loss) (in Equity Ended Grant Price Granted 000's) Raised ----------- -------------- ------------- -------------- ------------ ----------- 12/31/02 None None None $(278) None ----------- -------------- ------------- -------------- ------------ ----------- 3/31/03 None None None $(1,158) $100 (1) ----------- -------------- ------------- -------------- ------------ ----------- 6/30/03 5/15/03 $100 110,443 $(1,246) $100 (2) ----------- -------------- ------------- -------------- ------------ ----------- 9/30/03 9/4/03 $100 17,302 $(1,541) None ----------- -------------- ------------- -------------- ------------ ----------- 12/31/03 None None None $(1,243) $100 (3) ----------- -------------- ------------- -------------- ------------ ----------- 3/31/04 3/25/04 $100 29,685 $581 None ----------- -------------- ------------- -------------- ------------ ----------- 6/30/04 4/19/04 $100 8,850 $1,947 $100 (4) ----------- -------------- ------------- -------------- ------------ ----------- 9/30/04 None None None $2,968 None ----------- -------------- ------------- -------------- ------------ ----------- 12/31/04 11/15/04 $125 3,936 $3,259 None 12/17/04 $125 711 ----------- -------------- ------------- -------------- ------------ ----------- 3/31/05 None None None $4,606 None ----------- -------------- ------------- -------------- ------------ ----------- 6/30/05 None None None $4,587 None ----------- -------------- ------------- -------------- ------------ ----------- (1) The $8.5 million January 2003 offering of Series A Convertible Preferred Stock was distributed as follows: $4.6 million, or 54%, to 28 independent third parties (including $1.7 million, or 20%, to three independent third parties who were collectively granted a Board position concurrently with the investment along with the right to maintain their collective proportionate ownership percentage in future private equity offerings (collectively, the "Funds")); $3.6 million, or 43%, to eight Board members and their affiliates; $100,000, or 1%, to one executive officer; and $203,000, or 2%, to seven other employees. (2) The $50.0 million June 2003 offering of Series B Convertible Preferred Stock was distributed primarily to existing investors as follows: $22.2 million, or 44%, to independent third parties and the Funds (including $16.1 million, or 32%, to the Funds); $26.6 million, or 53% to 13 other Board members and their affiliates; $370,000, or 1%, to one executive officer; and $780,000, or 2%, to five other employees. (3) The $20.0 million October 2003 offering Series B Convertible Preferred Stock was distributed primarily to existing investors as follows: $9.0 million, or 45% to independent third parties and the Funds (including $6.0 million, or 30%, to the Funds ); $10.4 million, or 52%, to nine Board members and their affiliates; $270,000, or 1%, to two executive officers; and $350,000, or 2%, to nine other employees. Mr. Oscar Young Ms. Keira Ino August 3, 2005 Page 4 (4) The $1.350 million May 2004 offering of Series B Convertible Preferred Stock was distributed to new investors as follows: $500,000, or 37%, to one independent third party; and $850,000, or 63%, to one executive officer. * * * * * If you have any questions or comments concerning the above, or if you require additional information, please do not hesitate to contact me at (202) 508-6082. Sincerely, /s/ Heather R. Badami Heather R. Badami cc: Mary K. Fraser, Esq. James Atkinson J. Adam Abram Michael T. Oakes Kenneth L. Henderson, Esq. John M. Schwolsky, Esq. We issued 200,000 shares of Series B convertible preferred stock during the fourth quarter of 2003, with proceeds, net of notes receivable from employees and directors, of $19.7 million. This transaction was the primary reason for the increase in stockholders' equity from September 30, 2003 to December 31, 2003. CRITICAL ACCOUNTING POLICIES We have identified the accounting policies below as critical to the understanding of our results of operations and financial position. We use significant judgments and estimates concerning future results and developments in applying these accounting policies and in preparing our financial statements. Actual results may differ from these estimates. We evaluate our estimates on a continual basis using information that we believe to be relevant. For a detailed discussion of our significant accounting policies, see Note 1 to the 2004 audited consolidated financial statements included in this prospectus. Reserve for Losses and Loss Adjustment Expenses The reserve for losses and loss adjustment expenses represents our estimated ultimate cost of all reported and unreported losses and loss adjustment expenses incurred and unpaid at the balance sheet date. We do not discount the reserve for losses and loss adjustment expenses. We estimate the reserve for losses and loss adjustment expenses using individual case-basis valuations of reported claims. We also use statistical analyses to estimate the cost of losses that have been incurred but not reported to us. Those estimates are based on historical information and on estimates of future trends that may affect the frequency of claims and changes in the average cost of claims that may arise in the future. As a relatively new company, our historical loss experience is limited. We utilize two primary actuarial methods to arrive at our loss reserve estimates for each line of business. One of the primary methods, the incurred Bornhuetter-Ferguson method, utilizes our initial expected loss ratio (the ratio of losses and loss adjustment expenses incurred to net earned premiums), expected reporting patterns for losses based on industry data and our actual reported losses and loss adjustment expenses to estimate the reserve. The other primary method, the loss ratio method, estimates the reserve by applying an expected loss ratio to net earned premiums by line of business. For both methods, the expected loss ratio is established using judgment, and is based primarily on industry data and the experience of our management team in writing similar insurance coverages. The expected loss ratio used in the loss ratio method is judgmentally adjusted up or down when the level of reported losses for a specific accident year for a line of business leads us to conclude that the ultimate loss ratio will differ from our initial expected loss ratio. The expected loss ratio used in the incurred Bornhuetter-Ferguson method is typically not adjusted after it is initially set, because the mechanics of that method already incorporate departures from expected reported losses into the reserve calculations. We did not change our actuarial methods during the three months ended March 31, 2005. We have generally selected the method that yields the highest reserve for each line of business. At March 31, 2005, our aggregate reserve is the sum of the selected estimates for each line of business. This aggregate reserve reflects our best estimate. Since we have selected the higher of the two methods at each financial statement date, the aggregate reserve includes components calculated using the incurred Bornhuetter-Ferguson method and components calculated using the loss ratio method. This combination of methods varies by line of business and period. The primary factor contributing to the difference between the methods is that the incurred Bornhuetter-Ferguson method incorporates assumptions regarding industry reporting patterns for losses and our reported loss and loss adjustment expenses which are not reflected in the loss ratio method. This means that the incurred Bornhuetter-Ferguson method is more sensitive to our level of reported losses than the loss ratio method. The estimated net reserve at March 31, 2005 using the incurred Bornhuetter-Ferguson method was $48.2 million, and the estimated net reserve at March 31, 2005 using the loss ratio method was $55.6 million. Each of these reserve estimates are individual point estimates and do not constitute an actuarial range. Given that our insurance companies have limited historical experience and that our two primary acturial methods utilize industry data, that losses on our casualty business often take a number of years to develop and that we write difficult classes of business which typically do not meet the risk criteria of standard carriers, we believe that we do not have sufficient evidence at this time to conclude that our 31 ultimate loss ratios will be significantly better than our expected loss ratios. In most instances we have recorded the reserve estimate derived using the loss ratio method as our best reserve estimate because the expected loss ratio used in this method is derived from our management's experience in the lines of business that we write and our current underwriting criteria. We record the reserve estimate derived from the incurred Bornhuetter-Ferguson method as our best estimate when the level of actual reported losses provides sufficient evidence that the ultimate loss ratio will exceed our expected loss ratio. When the level of actual reported losses provides sufficient evidence that the ultimate loss ratio will be less than our expected loss ratio, we generally reduce the expected loss ratio used in the loss ratio method. We believe that generally selecting the higher of our two primary methods by line of business as the basis for our best reserve estimates at this early stage in our history is appropriate because more evidence is needed before we can reasonably conclude that our loss ratios will be significantly better than our expectations. In the future as more information becomes available to us and we develop more experience in our casualty business, we may adjust or change our actuarial methods. Such future methods may not necessarily result in selecting the highest estimate of our actuarial methods. Reserve estimates derived using the incurred Bornhuetter-Ferguson method are driven by our assumptions related to the expected loss ratio and the expected reporting pattern for losses, while reserve estimates derived using the loss ratio method are driven by our assumptions related to the expected loss ratio and are not sensitive to reporting patterns. Reporting pattern means the estimated percentage of the ultimate losses for a particular line of business that have been reported at the valuation date. The table below quantifies the impact that reasonably likely changes in these two variables, the expected loss ratio and the expected reporting pattern for losses, would have on the recorded reserve for losses and loss adjustment expenses at March 31, 2005. The reporting pattern scenarios below only apply to the incurred Bornhuetter-Ferguson analysis. In the table below, the expected loss ratio and the expected reporting pattern refer to assumptions related to those two variables that we made in calculating the net reserves that we recorded at March 31, 2005. <TABLE> IMPACT ON NET CHANGE FROM LOSS REPORTING RECORDED RECORDED NET RATIO PATTERN RESERVES RESERVES -------------------------------------------- ---------------------------- --------------- ------------- (IN THOUSANDS) Increase expected loss ratio by 5% ......... Slower reporting $ 5,586 9.8% Increase expected loss ratio by 5% ......... Expected reporting pattern 5,144 9.0% Increase expected loss ratio by 5% ......... Faster reporting 5,076 8.9% Expected loss ratio ........................ Slower reporting 623 1.1% Expected loss ratio ........................ Expected reporting pattern 0 0.0% Expected loss ratio ........................ Faster reporting -396 -0.7% Reduce expected loss ratio by 5% ........... Slower reporting -4,238 -7.4% Reduce expected loss ratio by 5% ........... Expected reporting pattern -4,880 -8.6% Reduce expected loss ratio by 5% ........... Faster reporting -5,498 -9.7% </TABLE> The sensitivity analysis above was based on the weighted-average assumptions presented in the table below. For the expected loss ratio assumptions, the expected loss ratio represents a weighted average of the expected loss ratios used for each of our lines of business, with the expected loss ratio for each line weighted based on net earned premiums. For the expected reporting pattern assumptions, the expected reporting pattern represents a weighted average of the percentage of ultimate losses assumed to be reported to us at March 31, 2005 for each of our lines of business weighted based on net earned premiums. 32 <TABLE> LOSS RATIO METHOD INCURRED BORNHUETTER-FERGUSON METHOD -------------------------------------------------- ---------------------------------------------------------- WEIGHTED-AVERAGE WEIGHTED-AVERAGE SCENARIO ASSUMPTION SCENARIO ASSUMPTION ------------------------------ ----------------- ------------------------------------ ------------------- C> Expected loss ratio ......... 58.2% 35.8% of losses Increase expected loss ratio Expected reporting pattern ......... reported by 5% .................... 63.2% 32.8% of losses Reduce expected loss ratio Slower reporting ................... reported by 5% .................... 53.2% 39.2% of losses Faster reporting ................... reported Expected loss ratio ................ 60.6% Increase expected loss ratio by 5% ........................... 65.6% Reduce expected loss ratio by 5% ........................... 55.6% </TABLE> We believe that loss ratios 5% above or below our expected loss ratio constitute a reasonable range of expectations for each major line of business. In addition, we believe that the adjustments that we made to speed up or slow down the reporting patterns in the sensitivity analysis above are reasonably likely outcomes. Under the incurred Bornhuetter-Ferguson method, slowing the reporting pattern results in an increase in the estimate of reserves, because reported losses are seen as a smaller percentage of the ultimate reported losses, resulting in an increase in estimated unreported losses. Similarly, using a faster reporting pattern results in a decrease in the estimate of reserves. In determining the amounts by which to slow or accelerate the reporting patterns for the sensitivity analysis presented above, we applied judgment to reflect the fact that the variability of reasonably likely reporting pattern outcomes varies by major line of business. There is very little variability around our expected reporting patterns for our excess and surplus insurance property lines, because property claims are typically reported relatively quickly. Conversely, there is considerably more variability around our expected reporting patterns for our excess and surplus insurance casualty lines, because losses on casualty business often take a number of years to develop. Reporting patterns for workers' compensation insurance are generally slower than reporting patterns for our excess and surplus insurance property lines and generally faster than reporting patterns for our excess and surplus insurance casualty lines. Accordingly, the variability around our expected reporting patterns for workers' compensation insurance is greater than it is for our excess and surplus insurance property lines and less than it is for our excess and surplus insurance casualty lines. The impact of recording the net reserve for losses at the highest value from the sensitivity analysis above would be to reduce net income by $3.6 million and to reduce stockholders' equity at March 31, 2005 by 4.4%. The impact of recording the net reserve for losses at the lowest value from the sensitivity analysis above would be to increase net income by $3.6 million and increase stockholders' equity at March 31, 2005 by 4.3%. Such changes in the net reserve for losses and loss adjustment expenses would not have an immediate impact on our liquidity, but would affect cash flow in future periods as the incremental or reduced amount of losses are paid. Our reserves are driven by a number of important assumptions including litigation and regulatory trends, legislative activity, social and economic patterns and claims inflation assumptions. Our reserve estimates reflect current inflation in legal claims settlements and assume we will not be subject to losses from significant new legal liability theories. Our reserve estimates also assume that we will not experience significant losses from mass torts and that we will not incur losses from future mass torts not known to us today. While it is not possible to predict the impact of changes in this environment, if new mass torts or expanded legal theories of liability emerge, our incurred but not reported, commonly referred to as IBNR, claims may differ substantially from our IBNR reserves. Our reserve estimates assume that there will not be significant changes in the regulatory and legislative environment. The impact of potential changes in the regulatory or legislative environment is difficult to quantify in the absence of specific, significant new regulation or legislation. In the event of significant new regulation or legislation, we will attempt to quantify its impact on our business. Our reserve estimates assume that the inflation assumption implicitly built into our expected loss ratio will continue into the future. Unexpected changes in loss cost inflation can occur through changes in general inflationary trends, changes in medical technology and procedures and changes in legal theories of liability. The estimated impact of an additional one percent increase in loss cost inflation as of March 33
