American Physicians Service Group, Inc.

January 7, 2008

VIA EDGAR

Mr. Jim B. Rosenberg
Senior Assistant Chief Accountant
Division of Corporation Finance
U.S. Securities and Exchange Commission
Washington, D.C.  20549

Re:

American Physicians Service Group, Inc.

Form 10-Q for the Period Ended June 30, 2007

Filed August 17, 2007

File No. 001-31434

Dear Mr. Rosenberg:

We are in receipt of your letter of December 7, 2007, concerning the above-referenced Form 10-Q. We acknowledge and appreciate your continued comments in regards to this matter. We also appreciate the Staff’s willingness to allow us additional time to complete this response. Our response is set forth below, following the applicable SEC Staff comment which we have repeated for ease of reference.

FORM 10-Q – June 30, 2007

Financial Information

3.

Acquisitions, Page 8

1)

Refer to your response to comment one. We continue to believe that the guidance included in SFAS 141 requires that you apply fair value techniques to any assets or liabilities acquired. It does not appear that you have done this with respect to these reserves. Please revise your financial statements and disclosures to reflect the impact of recording these reserves at fair value at the time of acquisition in accordance with SFAS 141.

The Company has evaluated the impact of applying fair value techniques on reserves and reinsurance recoverables (collectively, herein referred to as “reserves,” or “loss reserves” or “net

1301 S. CAPITAL OF TEXAS HIGHWAY  u SUITE C 300  u AUSTIN, TX 78746 u PHONE: (512) 328-0888  u FAX: (512) 314-4398

Mr. Jim B. Rosenberg
January 7, 2008
Page 2

loss reserves”) acquired in the acquisition of API on April 1, 2007. In our analysis, the Company applied SFAS No. 141, Business Combinations and Statement of Financial Accounting Concepts No. 7, Using Cash Flow Information and Present Value in Accounting Measurements, evaluated industry practice which also considered precedent transactions and also reviewed the most recent guidance in regards to clarification of applying fair value techniques contained in SFAS No. 141 R, Business Combinations, and SFAS No. 157, Fair Value Measurements. While SFAS 141 R and SFAS 157 were not effective at the time of our acquisition of API and SFAS 157 was not adopted early by the Company, both pronouncements provide additional guidance for applying fair value techniques for business combinations. SFAS 141 R also resolves conflicting guidance in SFAS  141 and SFAS No. 60, Accounting and Reporting for Insurance Enterprises, in regards to applying fair value techniques to insurance company business combinations through an amendment to SFAS 60.

In applying fair value techniques, we consulted with our outside actuaries. We also consulted with our external auditors who, in turn, consulted with their actuaries as to the reasonableness and appropriateness of the fair value techniques employed by management.

Due to the complexity of applying fair value techniques in the context of acquiring a medical malpractice company, we have summarized our approach and rationale in some detail below.

As noted in Paragraph B190 of SFAS 141 R, “The FASB decided that insurance and reinsurance contracts acquired in a business combination should be accounted for on a fresh-start (new contract) basis. Accordingly, all assets and liabilities arising from the rights and obligations of insurance and reinsurance contracts acquired in a business combination are recognized at the acquisition date, measured at their acquisition-date fair values.” SFAS 141 R, Paragraphs B246 through 251, highlights implementing fair-value principles in business combinations through reference to SFAS 157. Prior to the issuance of SFAS 157, there were different definitions of fair value and limited guidance in applying those definitions under GAAP. Additionally, the fair value concept guidance was dispersed among many pronouncements. As noted in Paragraph 22 of SFAS 157, “to increa se consistency and comparability in fair value measurements and related disclosures, this Statement establishes a fair value hierarchy that prioritizes the inputs to valuation techniques used to measure fair value into three broad levels. The fair value hierarchy gives the highest priority to quoted market prices (unadjusted) in active markets for identical assets or liabilities (Level 1) and the lowest priority to unobservable inputs (Level 3)…. Assessing the significance of a particular input to the fair value measurement requires judgment, considering factors specific to the asset or liability.”

Insurance loss reserves are not traded in an open market and no active trading market exists. Therefore, the market value of loss reserves should be evaluated on the basis of fair value

Mr. Jim B. Rosenberg
January 7, 2008
Page 3

principles using actuarial methods and models utilizing observable and/or unobservable inputs.¹ We utilized two white papers issued on this topic that were commissioned by the Casualty Actuarial Society for purposes of analyzing both observable and unobservable inputs into the valuations techniques as prescribed under SFAS No. 157, Paragraph 21.² We have provided copies of these reports to the Staff under separate cover. Of the lines of insurance analyzed in the reports, our line of insurance, medical malpractice, was one of the three featured lines of insurance in the analysis along with workers compensation insurance and personal automobile insurance.

Our approach to estimating fair value of loss reserves begins with discounting the reserves to account for the estimated time value of money. We then added an estimated margin for risk and uncertainty.³ We also note that this method is consistent with SFAS No. 157, Paragraph B2, that highlights, in addition to the estimate of future cash flows for the asset or liability being measured and the time value of money, that a fair value measurement of an asset or liability, using present value, should capture the price for bearing the uncertainty inherent in the cash flows (risk premium), other case-specific factors that should be considered by market participants and, in the case of a liability, the nonperformance risk relating to that liability, including the reporting entity’s (obligor’s) own credit risk.

More specifically, the methodology used by the Company with assistance from our outside consulting actuaries was as follows:

Ø

Step One:  We determined an expected payment pattern based on API historical claims data. This was done utilizing the most recent actuarial study (prior to the acquisition) as of December 31, 2006.

Ø

Step Two:  Once derived, the expected payment pattern was applied to the undiscounted reserves by coverage year to obtain a projection of expected future cash flows.

Ø

Step Three:  We then applied a schedule of risk-free interest rates matching the projected cash flows (i.e. a yield curve approach) for each future period. We used

¹

PriceWaterhouseCoopers (PWC), “An Investigation of Practical Matters Related to Implementing Fair Value Accounting for Property/Casualty Loss Reserves”, page 116.

²

Management utilized these reports as a point of reference to its fair market calculations. These two reports served as a guide to the actuarial methods and mathematical calculations employed by management with assistance from its outside actuaries in our analysis. This is consistent with SFAS 157, Paragraph 21 that discusses the use of observable inputs in applying fair value techniques.  The two white papers were part of a report issued by the Casualty Actuarial Society entitled, Fair Value of P&C Liabilities:  Practical Implications, 2004.

The Towers Perrin Report was entitled “How Might the Presentation of Liabilities at Fair Value Have Affected the Reported Results of U.S. Property and Casualty Insurers” and the PriceWaterhouseCoopers Report was entitled “An Investigation of Practical Matters Related to Implementing Fair Value Accounting For Property/Casualty Loss Reserves.”  

³

PWC op. cit., page 118

Mr. Jim B. Rosenberg
January 7, 2008
Page 4

the treasury yield curve rates as of March 31, 2007, to match the expected future cash flows. The imbedded yield rate was calculated to be 4.64%.

Ø

Step Four:  We utilized the imbedded yield rate and a risk premium (described below) to calculate the fair value of the reserves as follows:

o

Utilizing the implied actuarial best estimate of reserves as of the date of acquisition, we discounted reserves utilizing the imbedded yield rate of 4.64%.

o

We then added a risk premium or market value margin (this is also referred to as a “market risk margin” or an “MVM”) which includes two components:

§

A 20% premium associated with reserving risk, or the uncertainty associated with reserves.

§

A 12% expected profit margin which is consistent with management’s understanding of the returns anticipated by the reinsurance market – the reinsurance market representing a willing participant in an arms-length transaction in the purchase of loss reserves.

SFAS 157, Paragraph B5, provides a good description of a risk premium, “Risk-adverse market participants generally seek compensation for bearing the uncertainty inherent in the cash flows of an asset or liability (risk premium). A fair value measurement should include a risk premium reflecting an amount market participants would demand because of the risk (uncertainty) in the cash flows. Otherwise, the measurement would not faithfully represent fair value. In some cases, determining the appropriate risk premium might be difficult. However, the degree of difficulty alone is not a sufficient basis on which to exclude a risk adjustment.” A risk premium or market value margin can also be described as “the estimated margin for risk and uncertainty in excess of discounted loss reserves that would be required by a willing buyer in an arms-length transaction to assume the liabilities the l oss reserves are held to meet.”¹ Additionally, the economic theory behind an MVM is that “an open market exchange would not occur at a price equal to the present value of the expected cash flows at the risk-free rate, unless the cash flows were also risk-free. To the extent that risk is present in the cash flows, rational buyers would demand a higher price to compensate them for that risk.”²

The Company’s methodology is consistent with other precedent property and casualty insurance transactions and the results from the independent studies by Towers Perrin and PWC.

¹

PWC op. cit., page 188

²

Towers Perrin op. cit., page 41

Mr. Jim B. Rosenberg
January 7, 2008
Page 5

We also note that the results of applying risk premiums to reserves varied as well.¹ These companies were also significantly larger, more geographically and product line diverse and in almost all cases either the acquiror or the acquiree had an established outside financial strength rating. At the time of the API acquisition, both the Company and API were unrated by any outside rating agencies. Thus, nonperformance risk or the risk that the obligation (reserves) will not be fulfilled as defined within FAS 157, Paragraph 15, is potentially greater than these precedent transactions.²

There is a wide divergence of opinion as what risks should be included and how market risk margins should be calculated.³ Claim reserves are subject to considerable uncertainty, with significant potential for estimation error, especially in a long-tail line of insurance like medical professional liability (MPL) insurance. The estimation of MPL reserves is inherently difficult. Injuries might not be discovered until years after an incident, or a claimant may delay pursuing recovery for damages. Medical liability claims are typically resolved over an extended period of time, often five years or more. Thus, this uncertainty of reserves must be considered in the application of fair value principles.

In selecting a reasonable risk premium, our goal was to ensure that we evaluated this uncertainty and also provided for a reasonable return on the reserves acquired in the acquisition. In order to select a reasonable risk premium, we evaluated both observable inputs such as the two reports issued by the Casualty Actuarial Society described above and the financial results of the MPL as a whole, as well as unobservable inputs such as historical adverse development of API prior to the acquisition by the Company.⁴

For example, we noted in our review of the PWC report the market value margins were in excess of 20% for all the companies included in the study with a significant number of companies well in excess of 40%.⁵ It was also noted that the testing results were less satisfactory and less stable for medical professional liability than the other two lines analyzed

¹

In our review of precedent insurance business combinations, we noted that the following reporting entity’s employed a similar methodology: Eastern Insurance Holdings, Inc., Quarterly Form 10-Q for the period ending 6/30/07; Safety Insurance Group Inc., Annual Form 10-K for the year ending 12/31/03; United America Indemnity, LTD., Quarterly Form 10-Q for the period ending 9/30/07; First Acceptance Corp; Annual Form 10-K; ProAssurance Corporation, Quarterly Form 10-Q for the period ending 9/30/06; Republic Companies Group Inc., Form 424B1.

²

At time of acquisition, API lacked a financial strength rating. The most common for an insurance company would be an AM Best Rating. Thus, the measure of API’s ability to pay its future obligations would be questioned by a willing participant in this transaction, thus a higher risk premium is warranted.

³

Towers Perrin op. cit., page 42

⁴

As defined in SFAS 157, Paragraph 21, inputs refer broadly to assumptions that market participants would use in pricing the asset or liability.  Observable inputs are inputs that reflect the assumptions market participants would use in pricing the asset or liability developed based on market data obtained from sources independent of the reporting entity.  Unobservable inputs are inputs that reflect the reporting entity’s own assumptions about the assumptions market participants would use in pricing the asset or liability developed based on the best information available in the circumstances.

⁵

PWC op. cit., page 158

Mr. Jim B. Rosenberg
January 7, 2008
Page 6

(Workers Comp and Personal Automobile) due to the volatility of MPL insurance.¹ Furthermore, this is evident in the financial results of the MPL sector of insurance as a whole. The MPL sector has historically operated at combined ratios well in excess of 125% and in only three years in the last twenty-nine has MPL delivered an underwriting profit, or a combined ratio below 100%.² Reserves can vary materially in any given period due to changes in historical settlement patterns; changes in claims handling procedures; trends in loss severity and/or frequency; economic inflation; and legislative changes.

As previously mentioned, the risk premium selected by the Company includes two components. First, a 20% premium associated with reserving risk or the uncertainty associated with reserves. Quite simply, a 20% risk premium was selected since it is reflective of the percentage variance in the range of necessary reserves in API’s most recent actuarial reserve studies. As part of its evaluation of reserve adequacy on a semi-annual basis, API’s outside consulting actuary develops a range of reserves it believes is reasonable. The upper end of the range generally reflects higher estimates of future loss and loss adjustment expenses while the lower end of the range generally reflects lower estimates of future loss and loss adjustment expenses. In general, the width of a range reflects the level of variability in the underlying actuarial projections. While API’s assessment may differ, API’s carried reserves must remain within the actuarial range of the independent actuary’s selected reserve estimate to be considered adequate for the purposes of GAAP. Furthermore, the Company has taken a consistent position of reserving at the upper end of the reserve range.

We also believe that a willing market participant in an arms-length transaction would require a 20% premium for reserve risk given the historical development of API prior to the acquisition. In evaluating reserve development beginning in 1998 through 2004, or the most recent years that are either fully developed or almost fully developed, for accident years 1998 through 2003, the amount of incurred net losses and loss adjustment expenses initially recorded developed adversely by 31% (with 2002 and 2003 as much as 60%) more than originally estimated in undiscounted reserves and for the 2004 accident year developed favorably from the original estimate by 17%. This wide range of uncertainty associated with API’s historical development patterns is further compounded by its lack of product line and geographical diversity. API is a small mono-line and mono-st ate writer of MPL and as such is more exposed than larger, more diversified insurance companies in this highly cyclical and very volatile line of insurance. Tort reforms have made MPL reserves more volatile as companies try to understand the impact of reform.

The second component of a risk premium is the profit margin or the return we believe a willing participant in an arms-length transaction would require in the purchase of loss reserves. For purposes of this analysis, we selected a 12% expected profit margin which is consistent with

¹

PWC op. cit., page 158

²

Thirteen Annual Medical Malpractice Insurers Conference: “National Market Historical Underwriting Results,” page 9, Graph 16

Mr. Jim B. Rosenberg
January 7, 2008
Page 7

management’s understanding of the returns anticipated by the reinsurance market. The reinsurance market was selected since they are a willing partner in the purchase of loss risk through the pricing of reinsurance. We also discussed this profit margin percentage with our reinsurance intermediary who indicated that a targeted return of 10% to 15% was a reasonable expectation for a reinsurer in setting reinsurance prices. We also note that this profit margin percentage is also consistent with the Company’s long-term projected return on equity of 11-14%.

To assess the overall reasonableness of our combined 32% risk premium, we also performed a measure of reserve variability to determine the various confidence levels or risk loads based on an estimated distribution of reserves as of the acquisition date.¹ The starting point of this analysis is the implied actuarial best estimate of API’s reserves at the date of acquisition. With the assistance of our outside consulting actuaries, we determined that at the implied actuarial best estimate, the confidence level of API’s reserves on a discounted basis without adding a risk premium was only 55%. This is hardly reflective of a confidence level that an arms-length market participant would seek as compensation for bearing the uncertainty inherent in the cash flows of reserves for a volatile line of insurance like MPL. Including the 32% risk premium, the confidence level increases to the 85% to 90% range, which we believe is a much more realistic confidence level that a willing purchaser of API’s loss reserves would require.

As an additional test of the overall appropriateness of a 32% risk premium on reserves, we calculated an implied reinsurer margin or an estimated reinsurance-only risk premium. The implied reinsurance margin assumes a risk premium anticipated by the reinsurance market for reserve uncertainty, the underwriting costs of administering and maintaining the treaty and an anticipated profit margin. In evaluating API’s two most recent reinsurance treaty years prior to the acquisition, the average implied reinsurer margin was 34%. Therefore, a 32% risk premium is consistent with the premiums inherent in API’s reinsurance treaties.

In conclusion, discounting API’s reserves utilizing the methodology described above results in a fair value estimate which is somewhat higher, but not materially different than, the carrying basis of loss reserves at the date of acquisition. We noted only a $646,700, or a 0.7%, difference between the fair value estimate of $86,949,000 and undiscounted reserves recorded of $86,302,300 and therefore we did not record an adjustment to the carrying basis of loss reserves at the date of acquisition. It is important to note that this is also a reasonable conclusion based on the Towers Perrin study, which noted only a nominal difference of 2% to 3% between fair value estimates and historical carrying amounts for the 2002 report year in their study.² We

¹

We utilized as the basis of the analysis a paper by Thomas Mack, titled “Measuring the Variability of Chain Ladder Reserve Estimates.” The Mack approach is consistent with fair value techniques/approaches utilized by both PWC and Towers Perrin in their studies. Also, a confidence level can be defined as the probability of settling net loss reserves below a specified amount.

²

Towers Perrin op. cit., page 58

Mr. Jim B. Rosenberg
January 7, 2008
Page 8

also noted three precedent transactions that also drew the same conclusion that fair value estimates of reserves approximated carrying value at the date of acquisition.¹

We believe it would be more appropriate and less confusing to clarify this by additional disclosures in our 2007 annual report. Accordingly, the Company proposes adding the following language in our 2007 Form 10-K to the footnote disclosure regarding the acquisition:  

The fair values of the reserves for losses and loss adjustment expenses and related reinsurance recoverables (the net loss reserves) acquired in the API transaction were estimated as of the date of acquisition based on present value of the expected underlying net cash flows, and includes a risk premium and a profit margin. In determining the fair value estimate, management discounted API’s historical undiscounted loss reserves, to present value assuming discounting patterns actuarially developed from the historical loss data of API. The discount rate used of 4.64% approximates the risk-free treasury rate on the acquisition date for maturities similar to the estimated duration of the reserves being valued. A risk premium was applied to the discounted net loss reserves to reflect management’s estimate of the cost API would incur to reinsure the full amount of its ne t loss reserves with a third-party reinsurer. This risk premium is based upon management’s assessment of the inherent uncertainty in reserving for net loss reserves and their knowledge of the reinsurance marketplace and was confirmed by a reinsurance intermediary. The calculation resulted in a fair value estimate which was not materially different than the historical loss reserves and therefore did not result in an adjustment to the historical reserve amount.

In accordance with the request contained in your comment letter, American Physicians Service Group, Inc. (“APS”) acknowledges that:

o

APS is responsible for the adequacy and accuracy of the disclosure in the Form 10-Q filed by APS for the period ended June 30, 2007 (the “Filing”);

o

SEC Staff comments or changes to disclosure in response to SEC Staff comments do not foreclose the Securities and Exchange Commission (the “Commission”) from taking action with respect to the Filing; and

¹

Safety Insurance Group, Inc. Annual Form 10-K for the year ending 12/31/03; date filed 3/15/2004. We note that in this transaction the acquiree, Thomas Black Corporation also did not have a financial strength rating at the measurement date.  Additionally, ProAssurance Corporation, a much larger and more geographically diverse company within the MPL insurance marketplace reached the same conclusion with respect to two acquisitions as reported in its September 30, 2006 10-Q.

Mr. Jim B. Rosenberg
January 7, 2008
Page 9

o

APS may not assert SEC Staff comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States.

If you have any questions concerning this letter or our responses to your comments, please do not hesitate to contact me at (512) 314-4309 or our Chief Financial Officer, Marc Zimmermann, at (512) 314-4497.

Very truly yours,

/s/ Timothy L. LaFrey

Timothy L. LaFrey, President

TLL/lbw

C:

Mr. Kenneth S. Shifrin

Chairman and Chief Executive Officer

American Physicians Service Group, Inc.

1301 S. Capital of Texas Highway, Suite C-300

Austin, TX  78746-6550

Ms. Pam Roth

Partner

BDO Seidman LLP

333 Clay Street, Suite 4700

Houston, Texas 77002-4180