Goodwill and certain intangible assets that have indefinite lives are evaluated at least annually for impairment. All intangible assets are also reviewed for possible impairment whenever changes in conditions indicate that their carrying value may not be recoverable. The annual evaluation is generally performed during the fourth quarter, using forecast information.

In accordance with generally accepted accounting principles, we test goodwill for impairment at the reporting unit level. A component of an operating segment is deemed to be a reporting unit if it constitutes a business for which discrete financial information is available and segment management regularly reviews the operating results of that component.

The identification and measurement of goodwill impairment involves the estimation of fair value of the reporting unit as compared to its carrying value.

The carrying value of goodwill and other intangible assets was $227.1 million and $164.2 million as of December 31, 2011 and 2010, respectively. In 2010, we recorded a $6.4 million goodwill impairment charge for our U.K. Composites reporting unit.. This write-off represented 17.2% of the reporting unit's goodwill and reduced the carrying value of the reporting unit to its fair value. See Note 9, Goodwill and Other Intangible Assets, Net, in the Notes to Consolidated Financial Statements for discussion of this matter.

During 2011, Management adopted Accounting Standards Update ("ASU") No. 2011-08, "Intangibles - Goodwill and Other (ASC Topic 350) - Testing Goodwill for Impairment." ASU No. 2011-08 intends to simplify goodwill impairment testing by permitting assessment of qualitative factors to determine whether events and circumstances lead to the conclusion that it is necessary to perform the two-step goodwill impairment test currently required under ASC 350, Intangibles - Goodwill and Other. Upon completion of this analysis, management concluded that certain reporting units would still be subject to the two-step goodwill impairment test.

Management estimated the fair value of these reporting units using an income methodology based on management's estimates of forecasted cash flows for each business unit, with those cash flows discounted to present value using rates commensurate with the risks of those cash flows. In addition, management used a market-based valuation method involving analysis of market multiples of revenues and earnings before interest, taxes, depreciation and amortization ("EBITDA") for (i) a group of comparable public companies and (ii) recent transactions, if any, involving comparable companies. Assumptions used by management were similar to those that would be used by market participants performing valuations of our reporting units.

In performing our test we used an assumed terminal growth rate of 3.5% for these reporting units. The discount rate utilized to reflect the risk and uncertainty in the financial markets and specifically in our internally developed earnings projections ranged from 13.0% - 15.0% for these reporting units. Changes in these estimates and assumptions could materially affect the results of our tests for goodwill impairment.

For the reporting units subject to the two-step goodwill test, management performed a sensitivity analysis relative to the discount rate and growth rate selected. For our U.K. Composites reporting unit, a decrease of 1% in our terminal growth rate or an increase of 1% in our discount rate would result in a fair value calculation less than the book value for the reporting unit. Additionally, a 10% decrease in the fair value of this reporting unit also would have resulted in a fair value calculation less than the book value. For all other reporting units subject to the two-step goodwill impairment test these changes would still have resulted in a fair value calculation that exceeds the book value.

We do not currently believe there is a reasonable likelihood that there will be a material change in estimates or assumptions used to test goodwill and other intangible assets for impairment losses. However, if actual results are not consistent with our estimates or assumptions, we may be exposed to an impairment charge that could be material. 

The company maintains a Stock Incentive Plan, which provides for share-based payment awards, including non-statutory stock options, restricted stock, stock appreciation rights, and long-term incentive program (LTIP) awards. We determine the fair value of our non-qualified stock option awards at the date of grant using a Black-Scholes model. We determine the fair value of our restricted share awards at the date of grant using an average of the high and low market price of our stock.

LTIP awards provide certain senior executives an opportunity to receive award payments, generally in cash. For each performance cycle, the company's financial results are compared to the Russell 2000 indices for the same periods based upon the following: (a) average return on total capital, (b) earnings per share growth and (c) total return to shareholders. No awards will be payable unless the company's performance is at least in the 25th percentile of the designated indices. The maximum award is payable if performance reaches the 75th percentile of the designated indices. Awards for performance between the 25th and 75th percentiles are determined by straight-line interpolation. Awards will be paid out at 100% at the 50th percentile.

In order to estimate the liability associated with LTIP awards, management must make assumptions as to how our current performance compares to current Russell 2000 data based upon the Russell 2000's historical results. This analysis is performed on a quarterly basis. When sufficient Russell 2000 data for a year is available, which typically will not be until May or June of the following year, management will adjust the liability to reflect its best estimate of the total award. Actual results could differ significantly from management's estimates. The total estimated liability as of December 31, 2011 was $9.5 million.

Option-pricing models and generally accepted valuation techniques require management to make assumptions and to apply judgment to determine the fair value of our awards. These assumptions and judgments include estimating the future volatility of our stock price, expected dividend yield, future employee turnover rates and future employee stock option exercise behaviors. Changes in these assumptions can materially affect the fair value estimate.

Our long-term incentive plan requires management to make assumptions regarding the likelihood of achieving long-term company goals as well as estimate the impact the Russell 2000 results may have on our accrual.

We do not currently believe there is a reasonable likelihood that there will be a material change in the estimates or assumptions we use to determine stock-based compensation expense. However, if actual results are not consistent with our estimates or assumptions, we may be exposed to changes in stock-based compensation expense that could be material.

If actual results are not consistent with the assumptions used, the stock-based compensation expense reported in our financial statements may not be representative of the actual economic cost of the stock-based compensation. A 10% change in our stock-based compensation expense for the year ended December 31, 2011, would have affected pre-tax earnings by approximately $0.7 million in 2011. Due to the timing of availability of the Russell 2000 data, there is a risk that the amount we have recorded as LTIP expense could be different from the actual payout. A 10.0 percentage point increase in the total performance factor earned for our LTIP would result in a reduction of 2011 pretax earnings of $1.0 million.