1000 North Water Street

Suite 1000

Milwaukee, WI 53202

T +414 287 1300   

April 2, 2010

Ms. Denise Wilcox

Vice President Human Resources

Twin Disc, Inc.

1328 Racine Street

Racine, WI 53403

Dear Denise:

As requested, this letter provides information to help explain the approach used by Towers Watson to fulfill Twin Disc’s request for competitive compensation estimates for Twin Disc’s executive officers, including Twin Disc’s Named Executive Officers.   As you know, Twin Disc and Towers Watson have historically conducted this analysis every other year, following the process outlined below.  This analysis was last conducted by Towers Watson on behalf of Twin Disc in June, 2007.

Note that prior to January, 2010 the individuals that conducted work for Twin Disc were part of Towers Perrin (prior to Towers Perrin’s merger with Watson Wyatt, creating Towers Watson)– therefore, references to Towers Watson prior to 2010 should be interpreted as references to Towers Perrin.

The approach consists of the following process steps:

●   Step 1. Screening the compensation database for data linked to a similar position (i.e. position with similar responsibilities) when compared to Twin Disc’s position duties and responsibilities (i.e., similar position content). Comparability can be at the corporate OR business unit level.

●   Step 2. The data that meet the screening criteria in Step 1 are subjected to regression analysis, a well-established and versatile statistical technique. The outcome of the regression is a “regression equation” that most closely defines the relationship between an independent variable (revenue, in this case) and a dependent variable (compensation).

●   Step 3: Twin Disc’s actual revenues for the position are put into the regression equation to calculate a compensation level consistent with the company’s revenue size.

Referring to Towers Watson’s 2007 work conducted for Twin Disc, the basis for the competitive compensation estimates was data culled from the 2006 editions of Towers Watson’s proprietary general industry surveys for each similar position (the most current database available at that time).

The general industry surveys are open year-round to participants who submit data; there are no size or industry restrictions. This means that client analyses are based on the companies that have submitted complete data at the time the analysis is done.  In working with Towers Watson, Twin Disc plays no role in selecting the companies for which data are pulled. Participating companies submit pay information for incumbents in positions that correspond to Towers Watson’s predefined position descriptions, and also include organization information such as revenues for each position. For example, a participant submits corporate revenues for corporate-level positions and sector/business unit revenues for sector/business unit positions. Survey information is consolidated into the database on an as-received basis and, to protect the confidentiality of individual company data, the data is stripped of company-specific identifiers and reported by position rather than by company; in other words, once the information is submitted into the database it cannot be linked back to any particular company and appears merely as data points within the database.

For each Twin Disc position, Towers Watson screened the compensation database for companies that had similar position content. Because the reporting companies varied widely in size, Towers Watson adjusted for size using regression analysis. Essentially, this technique identifies a relationship between one variable (in this case, compensation) and another variable that is closely related to it. In general industry, the firmest linear relationship is to the revenues of the company (or of the sector, in the case of business units). In other words, a larger company would be likely to pay a higher amount of compensation for the same position than a smaller company. Using this relationship, one can predict the level of compensation that any company would pay for a given job based on its revenues.

Denise, please let me know if you have any questions regarding this process and our surveys.  I trust this provides useful information about the use of regression analysis and its application to the historic analysis conducted by Towers Watson on behalf of Twin Disc.

Sincerely,

Towers Watson Pennsylvania Inc., a Towers Watson company