September 10, 2009

Mr. Jim Rosenberg
Senior Assistant Chief Accountant
Securities and Exchange Commission
100 R Street N.E.
Washington, D.C. 20549

Re: Erie Indemnity Company

Form 10-K for the Year Ended December 31, 2008
Filed February 26, 2009
File No. 000-24000

Dear Mr. Rosenberg,

On behalf of Erie Indemnity Company (the “Company”) this letter is being submitted in response to comments received from the Staff of the Securities and Exchange Commission (the “Staff”) by letter dated July 27, 2009, with respect to the Company’s Annual Report on Form 10-K for the fiscal year ended December 31, 2008. By letter dated July 30, 2009 to the Staff, we requested and received approval to provide our response by September 11, 2009.

We appreciate your comments with respect to the above listed report and we have carefully considered the Staff’s comments. We are always looking to improve the financial disclosure that we make as part of our public filings and look forward to working with you on the matters outlined in your letter.

Our responses to your comments are provided below. For convenience, each of the Staff’s comments is reprinted below, followed by the Company’s response. We are providing the information that you requested and where you have requested revised disclosure, we have provided the revision we intend to include in future filings.

The adequacy and accuracy of the disclosure in our filings is the responsibility of the Company. We acknowledge that Staff comments or changes to disclosure in response to Staff comments do not foreclose the Securities and Exchange Commission (the “Commission”) taking any action with respect to the filing. We also acknowledge that Staff comments may not be asserted as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States.

1

SEC Comment #1: Management’s Discussion and Analysis of Financial Condition and Results of Operations — Investment Valuation, page 22

Please disclose the significant judgments you made in classifying a particular financial instrument in the fair value hierarchy.

Company Response:

The valuation techniques used to derive fair value of our securities require a degree of judgment associated with the specific characteristics of the asset to be valued, market conditions for that asset, and the type of inputs used in valuation. There are two significant judgments used in classifying the fair value of our investments; the presence of an active or inactive (illiquid) market and the use of observable or unobservable inputs.

The valuation of Level 1 assets is based on exchange traded prices (observable inputs) in active markets and is not subject to adjustments or discounts.

Level 2 assets include securities that rely on observable inputs, but may include quoted prices for similar instruments or prices obtained from third party pricing services that model prices based on observable inputs and can include securities with active or illiquid markets. The judgment involved related to Level 2 securities requires an assessment of the type of security, current market conditions, and available inputs.

The valuation of Level 3 assets is determined based upon unobservable inputs, reflecting our estimates of value based on assumptions used by market participants. Securities are assigned to Level 3 in cases where non-binding broker quotes are significant to the valuation and there is a lack of transparency as to whether these quotes are based on information that is observable in the marketplace. Fair value estimates for securities valued using unobservable inputs require significant judgment due to the illiquid nature of the market for these securities and represent the best estimate of the fair value that would occur in an orderly transaction between willing market participants at the measurement date under current market conditions.

Our valuation techniques maximize the use of observable inputs and minimize the use of unobservable inputs, whenever possible.

As of each reporting period, financial instruments recorded at fair value are classified based on the lowest level of input that is significant to the fair value measurement. The presence of at least one unobservable input would result in classification as a Level 3 instrument. Our assessment of the significance of a particular input to the fair value measurement requires judgment, and considers factors specific to the asset, such as the relative impact on the fair value as a result of including a particular input and market conditions. We did not make any other significant judgments except as described above.

We respectfully propose enhancing our fair value disclosure in future filings as highlighted in bold in Exhibit A.

2

SEC Comment #2: Management’s Discussion and Analysis of Financial Condition and Results of Operations — Investment Valuation, page 22

Please disclose the criteria you used to determine whether the market for a financial instrument is active or inactive (i.e. illiquid).

Company Response:

We used the following definitions as criteria for determining whether a market was active or inactive:

• We define an active market as one in which transactions for the assets being valued occur with sufficient frequency and volume to provide reliable pricing information.

• We define an inactive (illiquid) market as one in which there are few and infrequent transactions, where the prices are not current, price quotations vary substantially, and/or there is little information publicly available for the asset being valued.

We continually assess whether or not an active market exists for all of our investments and as of each reporting date re-evaluate the classification in the fair value hierarchy.

We respectfully propose enhancing our fair value disclosure in future filings as highlighted in bold in Exhibit A.

SEC Comment #3: Management’s Discussion and Analysis of Financial Condition and Results of Operations — Investment Valuation, page 22

Please disclose which financial instruments are affected by the lack of market liquidity (i.e. inactivity), how the lack of liquidity impacted the valuation technique you used, and how you factored illiquidity into your fair value determination of those financial instruments.

Company Response:

Level 1 securities are based on observable inputs in active markets and are not adjusted or discounted for lack of market liquidity.

Both Level 2 and Level 3 securities may contain securities that are considered to be in an inactive market due to limited or no trading activity in the current period.

Fair value for Level 2 securities in an illiquid market are generally determined using comparable securities or non-binding broker quotes received from outside broker dealers or pricing service propriety modeling based on observable inputs. Financial instruments in this category primarily consist of asset backed securities, certain private preferred stock and bond securities, and collateralized mortgage obligations.

3

Fair value estimates for securities in illiquid markets valued using unobservable inputs (Level 3) represent the best estimate of the fair value that would occur in an orderly transaction between willing market participants at the measurement date under current market conditions. Fair value for these securities are generally determined using comparable securities or non-binding broker quotes received from outside broker dealers based on security type and market conditions. Remaining un-priced securities are valued using an estimate of fair value based on indicative market prices that include significant unobservable inputs not based on, nor corroborated by, market information, including the utilization of discounted cash flow analyses which have been risk-adjusted to take into account illiquidity and other market factors. This category primarily consists of certain private preferred stock and bond securities, collateralized debt and loan obligations, and credit linked notes.

We respectfully propose enhancing our fair value disclosure in future filings as highlighted in bold in Exhibit A.

SEC Comment #4a — f: Notes To Consolidated Financial Statements, Note 5. Fair Value, page 63

Based on your disclosure on page 63, it appears that you use a third party pricing service to assist you in determining fair values. Please revise your disclosure to explain the extent to which, and how, the information is obtained and used in developing the fair value measurements in the consolidated financial statements. In your disclosure, please include the following:

a. The nature and amount of assets you valued using broker quotes or prices you obtained from pricing services, along with the classification in the fair value hierarchy;

Company Response:

We used a nationally recognized pricing service to value 96% of our total available-for-sale and trading investment portfolio at December 31, 2008.

Regarding the use of a pricing service, all of our Level 1 and the majority of our Level 2 securities at December 31, 2008 were valued using a pricing service. In Level 2, $547.4 million of the $553.4 million, were valued using a pricing service. In Level 3, $3.4 million of the total Level 3 securities of $26.0 million were valued using a pricing service.

Regarding the use of broker quotes, non-binding broker quotes were used to value $6.1 million of Level 2 securities and $20.0 million Level 3 securities.

In the course of our review, we determined that the table in Note 5 Fair Value, which presents the securities by level of input, included two securities with a total fair value of $2.2 million incorrectly classified as Level 1 securities. These securities should have been classified as Level 2. This reclass had no impact on our net income or shareholders’ equity for the year ended December 31, 2008 and therefore was deemed not to be material. This amount has been properly classified in Exhibit B. We intend to present the reclassified amounts in future filings.

4

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

b. The number of quotes or prices you generally obtained per instrument, and if you obtained multiple quotes or prices, how you determined the ultimate value you used in your financial statements;

Company Response:

For 96% of our total available-for-sale and trading investment portfolio at December 31, 2008, we obtained one price from a nationally recognized pricing service.

In cases in which a price from the pricing service was not available, values were determined by obtaining non-binding broker quotes and/or market comparables. When available, we obtained multiple quotes for the same security. The ultimate value for these securities was determined based on our best estimate of fair value using corroborating market information. Our evaluation included the consideration of benchmark yields, reported trades, issuer spreads, two-sided markets, benchmark securities, bids, offers and reference data.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

c. Whether, and if so, how and why, you adjusted quotes or prices you obtained from brokers and pricing services;

Company Response:

We reviewed the prices obtained from the pricing service continuously. In limited circumstances we adjusted the price received from the pricing service when in our judgment a better reflection of fair value was available based on corroborating information and our knowledge and monitoring of market conditions. At December 31, 2008, we adjusted 3 prices received by the pricing service to reflect a lower fair market value based on observable market data such as a disparity in price of comparable securities and/or non-binding broker quotes. The value of these securities based on prices from the pricing service was $8.4 million. The ultimate value used in our financial statements was $4.2 million.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

d. The extent to which the brokers or pricing services are gathering observable market information as opposed to using unobservable inputs and/or proprietary models in making valuation judgments and determinations;

Company Response:

The pricing service gathers observable market information to the extent possible. Observable market information was used by the pricing service in the fair value measurement of all of our Level 1 securities. Observable inputs from pricing services were used to determine the fair value of $547.4 million of securities that we ultimately classified as Level 2 given the priority of the input used in the valuation. In these instances, the pricing service used multiple

5

verifiable, observable inputs including benchmark yields, reported trades, broker/dealer quotes, issuer spreads, two-sided markets, benchmark securities, bids, offers and reference data. Relative to our securities, unobservable inputs from pricing services were used to value $3.4 million of securities that we classified as Level 3. Pricing service valuations for Level 3 securities are based on proprietary models and used when observable inputs are not available in illiquid markets.

In cases in which a value from the pricing service was not available, values were determined by obtaining non-binding broker quotes and/or market comparables. Securities that were priced using non-binding broker quotes and/or market comparables totaled $6.1 million for Level 2 and $20.0 million for Level 3. There were no non-binding broker quotes and/or market comparables used in the fair value estimates for Level 1 securities.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

e. Whether the broker quotes are binding or non-binding; and

Company Response:

All broker quotes we obtain are non-binding.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

f. The procedures you performed to validate the prices you obtained to ensure the fair value determination is consistent with SFAS 157, Fair Value Measurements, and to ensure that you properly classified your assets and liabilities in the fair value hierarchy.

Company Response:

Prices received from our pricing service are continuously reviewed. Price variances, including large periodic changes, are investigated and corroborated by market data. We have reviewed the pricing methodologies of our pricing service to ensure that investments are properly classified within the fair value hierarchy and believe that their prices adequately consider market activity in determining fair value. Management ultimately determined the pricing for each security that we considered to be the best estimate of fair value.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit B.

SEC Comment #5: Development of Direct Loss Reserves, page 32

Please revise your disclosure to include a reconciliation of the favorable development amounts disclosed in this discussion and disclosed in the discussion on page 24 to the $15.0 million and $20.2 million favorable prior accident year development in fiscal 2008 and 2007, respectively, shown in the table on page 87. Explain in your disclosure the purpose of the reconciliation and each reconciling item.

6

Company Response:

The favorable loss development amounts disclosed on page 24 refer to the entire Property and Casualty Group’s reserves. Our share of these favorable loss development amounts is 5.5% of the Property and Casualty Group’s total reflecting our 5.5% participation in the intercompany reinsurance pooling agreement.

The following table provides the details of the prior year loss reserve development:

(in thousands)	2008			2007			2006
Prior Year Loss Development:
Direct business excluding salvage and subrogation	$	(6,729	)	$	(11,031	)	$	(3,995	)
Assumed reinsurance business		(3,675	)		(3,624	)		(2,722	)
Ceded reinsurance business		167			(902	)		(977	)
Salvage and Subrogation		(104	)		47			(113	)
Total prior year loss development redundancy	$	(10,341	)	$	(15,510	)	$	(7,807	)

Negative amounts represent a redundancy (decrease in reserves) while positive amounts represent a deficiency (increase in reserves).

The favorable loss development amounts disclosed in the discussion on page 24 related to our 2008 retrospective reserve analysis indicated the Property and Casualty Group’s December 31, 2007 direct reserves, excluding salvage and subrogation recoveries, had an estimated redundancy of approximately $122 million, or 3.5% of reserves at December 31, 2007. Our 5.5% share of this development totals $6.7 million.

Our discussion of the development of loss reserves on page 32 was previously limited to our 5.5% share of the loss development of direct reserves only of the Property & Casualty Group. We will expand this disclosure to include the development of loss reserves for our 5.5% share of the Property & Casualty Group’s assumed and ceded business.

In the course of our review, we determined that the amounts previously reported in Note 14 incorrectly reported current year salvage and subrogation as prior year development when the amounts should have been recorded as a reduction in the current year losses. While this did not impact the total incurred losses, it did cause us to overstate our current accident year losses and overstate the redundancy in the prior accident year development. This reclassification had no impact on our net income, shareholders’ equity, financial position or cash flows as of and for the years ended December 31, 2008, 2007 or 2006 and therefore was deemed not to be material. These amounts have been corrected in Exhibit C. We intend to present the reclassified amounts in future filings.

We respectfully propose enhancing our disclosure in future filings as highlighted in bold in Exhibit C.

7

SEC Comment #6: Distribution of Investments — Fixed Maturities, page 37

Please revise your disclosure for the securities in your investment portfolio that are guaranteed by third parties to include the credit rating with and without the guarantee, in tabular format. Also disclose any significant concentration in an individual guarantor, both direct exposure (i.e. investments in guarantor) and indirect exposure (i.e. investments guaranteed by a guarantor).

Company Response:

The Company’s securities that are guaranteed by monoline insurers are limited to securities in our municipal bond portfolio. The fair value of these securities with and without underlying insured ratings are included in tabular format as part of our Credit Risk disclosures in Item 7a. Quantitative and Qualitative Disclosures about Market Risk on page 50 of the Form 10-K for the year ended December 31, 2008.

We have no significant direct investment exposure to the entities providing financial guarantees or other credit support to any security held in our portfolio.

Of the $166.2 million indirect exposure with monoline insurers, 43% relates to MBIA, 39% to FSA, 13% to Ambac and 5% to all other insurers.

We respectfully propose enhancing our disclosure in future filings within Item 7. Management’s Discussion & Analysis under the heading Distribution of Investments and Item 7a. Quantitative and Qualitative Disclosures about Market Risk as highlighted in bold in Exhibit D.

If you have any questions, please do not hesitate to contact me.

Sincerely,

/s/ Marcia A. Dall Erie Indemnity Company

By: Marcia A. Dall
Executive Vice President and Chief Financial Officer

cc: James Tanous, Erie Indemnity Company, Executive Vice President, Secretary and General Counsel
Lisa Vanjoske, Securities Exchange Commission, Assistant Chief Accountant
Tabatha Akins, Securities Exchange Commission, Staff Accountant

8

Exhibit A

CRITICAL ACCOUNTING ESTIMATES

Investment Valuation

We make estimates concerning the valuation of all investments. Valuation techniques are used to derive the fair value of the available-for-sale and trading securities we hold. Fair value is the price that would be received to sell an asset in an orderly transaction between willing market participants at the measurement date.

Fair value measurements are based upon observable and unobservable inputs. Observable inputs reflect market data obtained from independent sources, while unobservable inputs reflect our view of market assumptions in the absence of observable market information. We utilize valuation techniques that maximize the use of observable inputs and minimize the use of unobservable inputs.

For purposes of determining whether the market is active or inactive, the classification of a financial instrument was based on the following definitions.

• An active market is one in which transactions for the assets being valued occur with sufficient frequency and volume to provide reliable pricing information.

• An inactive (illiquid) market is one in which there are few and infrequent transactions, where the prices are not current, price quotations vary substantially, and/or there is little information publicly available for the asset being valued.

We continually assess whether or not an active market exists for all of our investments and as of each reporting date re-evaluate the classification in the fair value hierarchy.

All assets carried at fair value are classified and disclosed in one of the following three categories:

• Level 1 — Quoted prices for identical instruments in active markets not subject to adjustments or discounts.

• Level 2 — Quoted prices for similar instruments in active markets; quoted prices for identical or similar instruments in markets that are not active; and model-derived valuations whose inputs are observable or whose significant value drivers are observable.

• Level 3 — Instruments whose significant value drivers are unobservable and reflect management’s estimate of fair value based on assumptions used by market participants in an orderly transaction as of the valuation date.

9

Level 1 primarily consists of publicly traded common stock, nonredeemable preferred stocks and treasury securities and reflects market data obtained from independent sources, such as prices obtained from an exchange or a nationally recognized pricing service for identical instruments in active markets.

Level 2 includes those financial instruments that are valued using industry-standard models that consider various inputs, such as the interest rate and credit spread for the underlying financial instruments. All significant inputs are observable, or derived from observable information in the marketplace, or are supported by observable levels at which transactions are executed in the marketplace. Financial instruments in this category primarily include municipal securities, asset backed securities, collateralized-mortgage obligations, foreign and domestic corporate bonds and redeemable preferred stocks and certain nonredeemable preferred stocks.

Level 3 securities are valued based upon unobservable inputs, reflecting our estimates of value based on assumptions used by market participants. Securities are assigned to Level 3 in cases where non-binding broker quotes are significant to the valuation and there is a lack of transparency as to whether these quotes are based on information that is observable in the marketplace. Fair value estimates for securities valued using unobservable inputs require significant judgment due to the illiquid nature of the market for these securities and represent the best estimate of the fair value that would occur in an orderly transaction between willing market participants at the measurement date under current market conditions. Fair value for these securities are generally determined using comparable securities or non-binding broker quotes received from outside broker dealers based on security type and market conditions. Remaining un-priced securities are valued using an estimate of fair value based on indicative market prices that include significant unobservable inputs not based on, nor corroborated by, market information, including the utilization of discounted cash flow analyses which have been risk-adjusted to take into account illiquidity and other market factors. This category primarily consists of certain private preferred stock and bond securities, collateralized debt and loan obligations, and credit linked notes.

As of each reporting period, financial instruments recorded at fair value are classified based on the lowest level of input that is significant to the fair value measurement. The presence of at least one unobservable input would result in classification as a Level 3 instrument. Our assessment of the significance of a particular input to the fair value measurement requires judgment, and considers factors specific to the asset, such as the relative impact on the fair value as a result of including a particular input and market conditions. We did not make any other significant judgments except as described above.

Estimates of fair values for our investment portfolio are obtained primarily from a nationally recognized pricing service. Our Level 1 category includes those securities valued using an exchange traded price provided by the pricing service. The methodologies used by the pricing service that support a Level 2 classification of a financial instrument include multiple verifiable, observable inputs including benchmark yields, reported trades, broker/dealer quotes, issuer spreads, two-sided markets, benchmark securities, bids, offers and reference data. Pricing service valuations for Level 3 securities are based on proprietary models and are used when observable inputs are not available in illiquid markets. In limited circumstances we adjust the price received from the pricing service when in our judgment a better reflection of fair value is available

10

based on corroborating information and our knowledge and monitoring of market conditions such as a disparity in price of comparable securities and/or non-binding broker quotes. We perform continuous reviews of the prices obtained from the pricing service. This includes evaluating the methodology and inputs used by the pricing service to ensure we determine the proper level classification of the financial instrument in accordance with the requirements set forth in SFAS 157. Price variances, including large periodic changes, are investigated and corroborated by market data. We have reviewed the pricing methodologies of our pricing service and believe that their prices adequately consider market activity in determining fair value

In cases in which a price from the pricing service is not available, values are determined by obtaining non-binding broker quotes and/or market comparables. When available, we obtain multiple quotes for the same security. The ultimate value for these securities is determined based on our best estimate of fair value using corroborating market information. Our evaluation includes the consideration of benchmark yields, reported trades, issuer spreads, two-sided markets, benchmark securities, bids, offers and reference data.

11

Exhibit B

NOTES TO CONSOLIDATED FINANCIAL STATEMENTS

Note 5. Fair Value

Valuation techniques used to derive the fair value of our available-for-sale and trading securities are based on observable and unobservable inputs. Observable inputs reflect market data obtained from independent sources. Unobservable inputs reflect our own assumptions regarding fair market value for these securities. Although the majority of our prices are obtained from third party sources, we also perform an internal pricing review for securities with low trading volumes in the current market conditions. Financial instruments are categorized based upon the following characteristics or inputs to the valuation techniques:

Level 1	Quoted prices for identical instruments in active markets not subject to adjustments or discounts.
Level 2	Quoted prices for similar instruments in active markets; quoted prices for identical or similar instruments in markets that are not active; and model-derived valuations whose inputs are observable or whose significant value drivers are observable.
Level 3	Instruments whose significant value drivers are unobservable and reflect management’s estimate of fair value based on assumptions used by market participants in an orderly transaction as of the valuation date.

The following table represents the fair value measurements on a recurring basis for our invested assets by major category and level of input as required by SFAS 157:

	December 31, 2008
	Fair value measurements using:
			Quoted prices
			in active		Significant		Significant
			markets for		observable		unobservable
			identical assets		inputs		inputs
(dollars in thousands)	Total		Level 1		Level 2		Level 3
Available-for-sale securities:
Fixed maturities	$	563,429	$	6,272	$	542,940	$	14,217
Preferred stock		55,281		32,969		10,494		11,818
Trading securities:
Common stock		33,338		33,316		0		22
Total	$	652,048	$	72,557	$	553,434	$	26,057

Estimates of fair values for our investment portfolio are obtained primarily from a nationally recognized pricing service. Our Level 1 category includes those securities valued using an exchange traded price provided by the pricing service. The methodologies used by the pricing service that support a Level 2 classification of a financial instrument include multiple verifiable, observable inputs including benchmark yields, reported trades, broker/dealer quotes, issuer spreads, two-sided markets, benchmark securities, bids, offers

12

and reference data. Pricing service valuations for Level 3 securities are based on proprietary models and are used when observable inputs are not available in illiquid markets. In limited circumstances we adjust the price received from the pricing service when in our judgment a better reflection of fair value is available based on corroborating information and our knowledge and monitoring of market conditions such as a disparity in price of comparable securities and/or non-binding broker quotes. We perform continuous reviews of the prices obtained from the pricing service. This includes evaluating the methodology and inputs used by the pricing service to ensure we determine the proper level classification of the financial instrument in accordance with the requirements set forth in SFAS 157. Price variances, including large periodic changes, are investigated and corroborated by market data. We have reviewed the pricing methodologies of our pricing service and believe that their prices adequately consider market activity in determining fair value.

In cases in which a price from the pricing service is not available, values are determined by obtaining non-binding broker quotes and/or market comparables. When available, we obtain multiple quotes for the same security. The ultimate value for these securities is determined based on our best estimate of fair value using corroborating market information. Our evaluation includes the consideration of benchmark yields, reported trades, issuer spreads, two-sided markets, benchmark securities, bids, offers and reference data.

For certain structured securities in an illiquid market, there may be no prices available from a pricing service and no comparable market quotes available. In these situations, we value the security using an internally-developed risk-adjusted discounted cash flow model.

The following table sets forth the fair value of our fixed maturity and preferred stock securities by pricing source as of December 31, 2008:

	December 31, 2008
(in thousands)	Total		Level 1		Level 2		Level 3
Fixed maturity securities:
Priced via pricing services(1)	$	548,034	$	6,272	$	539,097	$	2,665
Priced via non-binding broker quote/market comparables (2)		12,786		0		3,843		8,943
Priced via internal modeling (3)		2,609		0		0		2,609
Total fixed maturity securities	$	563,429	$	6,272	$	542,940	$	14,217
Preferred stock securities:
Priced via pricing services (1)		42,008		32,969		8,256		783
Priced via non-binding broker quote/market comparables (2)		13,273		0		2,238		11,035
Priced via internal modeling (3)		0		0		0		0
Total preferred stock securities	$	55,281	$	32,969	$	10,494	$	11,818
Total available-for-sale securities	$	618,710	$	39,241	$	553,434	$	26,035

(1)	Pricing service valuations for Level 3 securities are based on proprietary models and used when observable inputs are not available in illiquid markets.
(2)	All broker quotes obtained for Level 3 securities were non-binding.
(3)	Internal modeling using a discounted cash flow model was performed on 2 fixed maturities representing less than 1.0% of the total portfolio.

13

Exhibit C

CRITICAL ACCOUNTING ESTIMATES

Property/casualty insurance liabilities

We also perform analyses to evaluate the adequacy of past reserve levels. Using subsequent information, we perform retrospective reserve analyses to test whether previously established estimates for reserves were reasonable. Our 2008 retrospective reserve analysis indicated the Property and Casualty Group’s December 31, 2007 direct reserves, excluding salvage and subrogation recoveries, had an estimated redundancy of approximately $122 million, or 3.5% of reserves at December 31, 2007. Our 5.5% share of the Property and Casualty Group’s favorable development of prior accident year direct losses, after removing the effects of salvage and subrogation recoveries, was $6.7 million in 2008.

Insurance Underwriting Operations, page 32

Development of loss reserves

Our 5.5% share of the Property and Casualty Group’s favorable development of prior accident year direct losses, after removing the effects of salvage and subrogation recoveries, was $6.7 million in 2008 and improved the combined ratio by 3.2 points. Of the $6.7 million, $4.3 million related to the personal auto line of business. The Property and Casualty Group reduced reserves in 2008 on prior accident years as a result of improvements in frequency trends and slight improvements in severity trends on automobile bodily injury and on uninsured/underinsured motorist bodily injury. For private passenger auto comprehensive coverages in 2008, the severity trend outpaced improvements in frequency, causing an increase in the loss cost trend. For all other private passenger auto coverages, the Property and Casualty Group’s frequency improvements either offset or outpaced increases in severity, causing flattening or improving loss cost trends. Overall, loss costs for private passenger auto continue to remain relatively flat. The Property and Casualty Group also experienced favorable development of prior accident year loss reserves on its assumed book of business totaling $66 million. In 2008, our 5.5% share of this favorable development totaled $3.7 million. This favorable loss reserve development was a result of reductions in IBNR reserves on involuntary reinsurance and to a lesser extent, reduced reserve levels related to our exposure to World Traded Center (WTC) losses.

In 2007, our share of the Property and Casualty Group’s favorable development of prior accident year direct losses, after removing the effects of salvage and subrogation recoveries, was $11.0 million and improved the combined ratio by 5.3 points. Of the $11.0 million, $8.1 million related to the personal auto line of business. The Property and Casualty Group reduced reserves in 2007 on prior accident years as a result of sustained improved severity trends on automobile bodily injury and on uninsured/underinsured motorist bodily injury. The Property and Casualty Group also experienced favorable loss development on its assumed business in 2007, caused by reduced IBNR reserves on involuntary business and a decrease in the IBNR reserves related to WTC losses. Our share of the favorable development on the assumed book of business totaled $3.6 million in 2007.

In 2006, our share of the Property and Casualty Group’s net positive development totaled 1.9 points, or $4.0 million, on direct losses and was the result of improved frequency trends for automobile bodily injury and uninsured/underinsured motorist bodily injury, from the 2004 and 2005 accident years. The total gross favorable development of 3.9 points, or $8.2 million was offset by 2.0 points, or $4.2 million, for reserve strengthening of the pre-1986 automobile

14

catastrophic injury liability reserve. Our positive reserve development on our assumed reinsurance business of $2.7 million was primarily driven by favorable development in our involuntary book.

The following table provides the details of the prior year loss reserve development for our wholly-owned property/casualty insurance subsidiaries for the years ended December 31:

(in thousands)	2008			2007			2006
Prior Year Loss Development:
Direct business excluding salvage and subrogation	$	(6,729	)	$	(11,031	)	$	(3,995	)
Assumed reinsurance business		(3,675	)		(3,624	)		(2,722	)
Ceded reinsurance business		167			(902	)		(977	)
Salvage and Subrogation		(104	)		47			(113	)
Total prior year loss development redundancy	$	(10,341	)	$	(15,510	)	$	(7,807	)

Negative amounts represent a redundancy (decrease in reserves) while positive amounts represent a deficiency (increase in reserves).

Note 14. Unpaid losses and loss adjustment expense, page 87

Note 14. Unpaid losses and loss adjustment expenses

The following table provides a reconciliation of beginning and ending loss and loss adjustment expense liability balances for our wholly-owned property/casualty insurance subsidiaries for the years ended December 31:

(in thousands)	2008			2007			2006
Total unpaid losses and loss adjustment expenses at January 1, gross	$	1,026,531		$	1,073,570		$	1,019,459
Less reinsurance recoverables		834,453			872,954			827,917
Net balance at January 1		192,078			200,616			191,542
Incurred related to:
Current accident year		147,508			141,413			147,437
Prior accident years		(10,341	)		(15,510	)		(7,807	)
Total incurred		137,167			125,903			139,630
Paid related to:
Current accident year		85,029			78,276			78,509
Prior accident years		57,463			56,165			52,047
Total paid		142,492			134,441			130,556
Net balance at December 31		186,753			192,078			200,616
Plus reinsurance recoverables		778,328			834,453			872,954
Total unpaid losses and loss adjustment expenses at December 31, gross	$	965,081		$	1,026,531		$	1,073,570

Driving the favorable development in 2008, 2007 and 2006 on prior year accident reserves were improved frequency and severity trends for automobile bodily injury and uninsured/underinsured motorist bodily injury claims.

15

Exhibit D

Management’s Discussion & Analysis

FINANCIAL CONDITION

Distribution of investments

Our municipal bond portfolio accounts for $211.4 million, or 37.5%, of the total fixed maturity portfolio. Of this $211.4 million, $166.2 million, or 78.6%, of the total municipal bond portfolio is insured. This insurance guarantees the payment of principal and interest on a bond if the issuer defaults. Our municipal bond portfolio is highly rated and includes all investment grade holdings (BBB or higher). The overall credit quality rating of our municipal bond portfolio is AA. Using the underlying rating of the bonds without consideration of insurance, the overall credit quality rating of our municipal bond portfolio would be AA- . The following table presents an analysis of our municipal bond ratings at December 31, 2008.

Municipal Bond Portfolio at December 31, 2008
(in thousands)

Ratings with insurance						Underlying ratings without insurance
Rating	Fair value		Fair value %			Rating	Fair value		Fair value %
AAA	$	32,179		15.2	%	AAA	$	23,340		11.0	%
AA		119,419		56.5		AA		89,360		42.3
A		58,483		27.7		A		93,939		44.4
BBB		1,339		0.6		BBB		2,745		1.3
Not rated		0		0.0		Not rated		2,036		1.0
Total	$	211,420		100.0	%	Total	$	211,420		100.0	%

We have no significant direct investment exposure to the entities providing financial guarantees or other credit support to any security held in our portfolio.

Of the $166.2 million indirect exposure with monoline insurers, 43% relates to MBIA, 39% to FSA, and 13% to Ambac.

16