Triad Guaranty (TGIC) CORRESP Correspondence with SEC

Triad Guaranty Inc.
101 South Stratford Road
Winston-Salem, NC 27104

May 4, 2007

Mr. Jim B. Rosenberg
Senior Assistant Chief Accountant
United States Securities and Exchange Commission
Division of Corporation Finance
Mail Stop 03-09
450 Fifth Street, N.W.
Washington, DC 20549

Re: Triad Guaranty Inc.
Form 10-K for the Year Ended December 31, 2005
File No. 000-22342

Dear Mr. Rosenberg:

We respectfully submit the following in response to our telephone conversation of Friday, April 20, 2007 with Ms. Kei Ino in which she sought further clarification on information contained in our responses dated January 31, 2007 and March 23, 2007. Specifically Ms. Ino requested that:

In disclosure format please provide the following information:

• The actual frequency and severity factor used for the calculation of the estimate of the reserve for losses

• How the factors were determined

• What changed that caused the factors not to be predictive

In addition in our response we were asked to indicate why the cause of the change could not have been anticipated at the time the estimate was made.

Response:
The Company agrees that additional information concerning the calculation of the estimate of the reserve for losses and the following information will be included in future filings as appropriate. We have included information for the year ended December 31, 2006 in our response. The non-emphasized text was included in our Form 10-K for the year ended December 31, 2006 and the items shown in bold are new disclosures not included in previous filings.

1

Reserve for Losses and Loss Adjustment Expenses

We calculate our best estimate of the reserve for losses to provide for the estimated costs of settling claims on loans reported in default, as of the date of our financial statements. Additionally, we provide a reserve for loans in default that are in the process of being reported to us (incurred by not reported) using an estimate based on the percentage of actual reported defaults. Our reserving process incorporates various components in a model that gives effect to current economic conditions and segments defaults by a variety of criteria. The criteria include, among others, policy year, lender, geography and the number of months and number of times the policy has been in default, as well as whether the defaults were underwritten as flow business or as part of a structured bulk transaction.

Frequency and severity are the two most significant assumptions in the establishment of our loss reserves. Frequency is used to estimate the ultimate number of paid claims associated with the current defaulted loans. The frequency estimate assumes that long-term historical experience, taking into consideration criteria such as those described in the preceding paragraph, and adjusted for current economic conditions that we believe will significantly impact the long-term loss development, provides a reasonable basis for forecasting the number of claims that will be paid. Severity is the estimate of the dollar amount per claim that will be paid. The severity factors are estimates of the percent of the risk in force that will be paid. The severity factors used are based on an analysis of the severity rates of recently paid claims, applied to the risk in force of the loans currently in default. The frequency and severity factors are updated quarterly. Economic conditions and other data upon which these factors are based may change more frequently than once a quarter. Therefore, significant changes in reserve requirements may become evident three or more months following the underlying events that would necessitate the change.

The following table provides the average frequency and severity calculations produced by the application of those factors in the calculation of the estimate of the reserve for losses for loans reported in default as of December 31, 2006, 2005 and 2004:

	2006			2005			2004
Frequency — the percent of defaults that are estimated to result in paid claims		36.5	%		31.5	%		32.2	%
Severity — the estimated dollar amount to be paid per claim	$	32,118		$	25,191		$	20,055

The increase in the severity assumption for 2006 was the combination of an increase in the average risk per defaulted loan due to larger loans in the default inventory and the impact of the declining housing market during 2006, with the declining housing market being the more significant factor. The increase in the frequency assumption for 2006 was primarily the result of a change in the default segmentation that was made during 2006 so that the calculation of the reserve factors was better aligned with the characteristics of the inventory of defaults.

The decline in the frequency factors at December 31, 2005 compared to December 31, 2004 was the result of the 2005 default inventory being comprised, on average, of loans that were fewer months in defaults. In our reserve methodology, all other things being equal, the fewer months a loan is in default, the less likely it is to result in a claim and therefore we assign a lower frequency factor to the newer defaults. The fall 2005 hurricane season, primarily Hurricane Katrina, resulted in a significant increase in the number of defaults at December 31, 2005. Since most of these defaults were only 3 or 4 months in default at December 31, 2005 there was a lower frequency factor applied to these loans, which has the impact of reducing the overall frequency factor. The increase in the

2

severity factor from December 31, 2004 was reflective of the higher average paid claim severity during 2005.

The estimation of loss reserves requires assumptions as to future events, and there are inherent risks and uncertainties involved in making these assumptions. Economic conditions that have affected the development of loss reserves in the past may not necessarily affect development patterns in the future in either a similar manner or degree. To provide a measure of the sensitivity on pretax income and loss reserves carried on the balance sheet, we have provided the following table that quantifies the impact of percentage increases and decreases in both the average frequency and severity as of December 31, 2006:

	Sensitivity Analysis
	Effect on Annual Pretax Income
	from Changes in Assumptions
	Decrease in Factors		Increase in Factors
	Resulting in an		Resulting in a
	Increase in		(Decrease) in
	Pretax Income		Pretax Income
	(In thousands)
10% Change in the Frequency Factors Utilized in the Loss Reserve Model	$	7,045	$	(7,145	)
10% Change in the Severity Factors Utilized in the Loss Reserve Model	$	8,325	$	(8,325	)
10% Change in Both the Frequency and Severity Factors Utilized in the Loss Reserve Model	$	14,665	$	(16,185	)

The impact on loss reserves on the balance sheet would be to decrease reserves under favorable development and to increase reserves for unfavorable development. There would be no impact on liquidity resulting from the change in reserves. However, there would be a change in cash or invested assets equal to the increase or decrease in the ultimate paid claims related to the change in reserves. We believe that a 10% increase or decrease in frequency or severity is reasonably likely based on potential changes in future economic conditions and past experience. Economic conditions that could give rise to an increase in the frequency rate could be a sudden increase, or a prolonged period of increase, in the unemployment rate while, conversely, a sudden drop in interest rates or a sustained period of economic and job growth could decrease the frequency rate. A sudden nationwide or widespread regional decline in home prices, or a prolonged period of flat or slowly depreciating home prices could increase the severity while a significant increase, or steady sustained growth in house price appreciation could decrease the severity.

Because the estimate for loss reserves is sensitive to the estimates of claims frequency and severity, we perform analyses to test the reasonableness of the best estimate generated by our loss reserve process. Our loss reserve estimation process and our analyses support the reasonableness of the best estimate of loss reserves recorded in our financial statements.

We monitor the adequacy of our reserves by comparing the actual development of the default inventory to the reserve for losses at the respective financial statement date. Reserves are deemed to be redundant if the actual paid claims plus any remaining reserves are less than the original reserve. If the actual paid claims plus any remaining reserves are greater than the original reserve, the reserves are deemed to be deficient. The information gathered from the monitoring of the reserve development is used to refine our estimates of frequency and severity.

3

For the years ended December 31, 2006, 2005 and 2004 the reserves established as of the end of the prior year developed deficient as follows:

	2006		2005		2004
	Dollars in Thousands
Reserve deficiency based on the reserve for loss and loss adjustment expense as of January 1	$	24,086	$	23,640	$	9,922

The deficiency recognized in 2006 on the December 31, 2005 reserve was the result of higher paid claims (due to both the frequency and severity of the claims) with the severity deficiency resulting primarily from the decline in the housing markets that became apparent during 2006. During the fourth quarter of 2006, we experienced a significant increase in our paid loss severity due, we believe, to changes in the housing market, which resulted in our actual paid loss severity being in excess of the severity assumptions used to calculate the reserve for losses. As a result of the increase in the actual paid severity, we increased the severity assumption underlying the calculation of the reserve for losses. Also during 2006 we observed that defaults were developing into paid claims at a greater rate than was indicated by the frequency factor incorporated in our reserves. After investigation we refined our reserve segmentation to improve alignment of reserve factors with the profile of the changing inventory of defaults, including creating a separate segment for pending claims. Prior to this change our reserve calculation did not include a separate segment for pending claims and our frequency factors incorporated an adjustment to the overall frequency factor for each segment based on the mix of pending claims. During 2006, we became aware that, when a change in the mix towards more pending claims occurs, our historical reserve factors had the potential to understate reserves as the probability of the actual payment of a claim increased. Therefore, in calculating the reserve for losses, consideration was given to the level of pending claims, even though they were not included as a separate segment. However, actual experience indicated that the estimation process could be improved by changing the segments to include a separate segment for pending claims.

Deficiencies recognized in 2005 and 2004 on the respective December 31, 2004 and 2003 reserve for losses were primarily due to greater claims than predicted by the frequency factors. This was driven by changes in the composition of the default inventory, including more pending claims than assumed originally in the reserving model. Actual experience has shown that the frequency factors, calculated from our historical data through December 31, 2004, understated the volume of claims that ultimately would be filed and paid in 2005 on defaults existing at December 31, 2004. During the first half of 2005, filed claims increased significantly above historical levels. This increase in filed claims produced a higher rate of claim payments than implied by the frequency factors used in calculating the reserves for losses at December 31, 2004. The primary cause for the increase in the filed and paid claims during the first half of 2005 was the action of a major customer to accelerate its claim filing process, as well as appeal decisions with respect to certain claims previously denied or coverages rescinded. Concurrent with the increase in filed claims in the first half of 2005, the severity of paid claims increased significantly for a period of time. In the normal course of the claims payment cycle, our ultimate loss is frequently reduced by loss mitigation. The most common mitigation opportunity is for the property to be sold by the borrower prior to foreclosure, or the lender after foreclosure but before we are contractually required to pay a claim under the terms of the policy. In many cases, the sales price of the property is sufficient to reduce, and sometimes eliminate, our loss. Because of the volume of filed claims in the first half of 2005, and customer expectations related to faster processing of filed claims, our ability to reduce our losses through mitigation was diminished. This caused an increase in the severity of paid claims

4

beginning in the second quarter of 2005 over our historical severity rates, also increasing the deficiency from the December 31, 2004 reserve for losses. The increase in the paid claims trend and severity was new information we incorporated in our reserve calculations. The change in estimate for claim frequency increased the reserve for all defaults existing at December 31, 2004 that were still classified as defaults at December 31, 2005, and therefore increased the deficiency of the December 31, 2004 reserves for losses.

Our claims settlement process includes a quality assurance review of the underwriting and loan origination process on most filed claims. As our business grew, the quality assurance review began to slow down the claims settlement process, resulting in an increase in pending claims throughout 2003 and 2004. The increase in pending claims was exacerbated by the increase in the filed claims in the first half of 2005. As previously discussed, prior to 2006, our reserve calculation did not include a separate segment for pending claims and the assumptions used in the calculation of the estimate of the reserve for losses underestimated the impact that an increase in pending claims would have on the ultimate paid claims during 2004 and 2005.

As indicated in our response the reserve deficiencies were the results of 1) changes in economic conditions that occurred after the reserves were established but before the claims were settled, such as the rapid change in the residential real estate market during 2006; 2) changes in customer processes and expectations that impacted our ability to process defaults and claims in a manner consistent with our historical practices on which our reserve factors were partially based; 3) changes in the composition of the default inventory in manner that reduced the predictive ability of the historical experience on which the factors were based. In the case of items 1) and 2) the causes of the deficiency were events occurring, or facts becoming known, after the balance sheet date that could not have been anticipated when the estimate for the reserve for losses was established. In regard to item 3) we have changed our estimation process to improve the segmentation of the defaults and to place more emphasis on recent experience in determine the factors underlying the estimate for the reserve for losses.

We confirm and acknowledge that:

• we are responsible for the adequacy and accuracy of the disclosure in our filing;

• staff comments or changes to disclosure in response to staff comments do not foreclose the Commission from taking any action with respect to the filing; and

• we may not assert 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 our response, please do not hesitate to contact me at (336) 723-1282, extension 1105 or contact Earl Wall, our general counsel, at extension 1108.

Sincerely,

Kenneth W. Jones
Senior Vice President and Chief Financial Officer

5