The Real Deal? Are Performance Awards Really Paying for Performance? October 24, 2013

The Real Deal? Are Performance Awards Really Paying for Performance? October 24, 2013 Dan Kapinos Associate Director, Global Technical Shared Services Team, Aon Hewitt Laura Smith Global Compensation & Benefits, The Travelers Companies, Inc. Carl J. Stegman Sr. Vice President, Stock Plan Services Product Management, Fidelity Investments

Company Background Company Background: The Travelers Companies, Inc. (NYSE: TRV) is a leading provider of property casualty insurance for auto, home and business. The company s diverse business lines offer its global customers a wide range of coverage sold primarily through independent agents and brokers. A component of the Dow Jones Industrial Average, Travelers has more than 30,000 employees and operations in the United States and selected international markets. The company generated revenues of approximately $26 billion in 2012. Award Philosophy: Our compensation program is designed to reinforce a long term perspective, which is especially vital in the property and casualty insurance industry, where the periodic occurrence of catastrophes, changes in estimates of costs for claims and other economic conditions have historically produced results that vary significantly when measured year to year. The portion of total compensation attributable to stock based programs increases with successively higher levels of responsibility. Thus, the senior most executives who are responsible for the development and execution of our strategic and financial plans have the largest portion of their compensation tied to performance based incentives, including equity based compensation, the ultimate value of which is completely or partly dependent on changes in stock price and return on equity. Annual Revenue $26B Total Employees ~30k Employees Receiving Equity Stock delivered globally in ~6k 5 Countries % US Based 95% % Global (Non-US) 5% Award Types: Stock Option: Restricted Stock/Units Performance Shares Employee Stock Purchase Key Statistics VP & Above VP & Below VP & Above No

Why is Performance Equity Here to Stay?

The Say-on-Pay Web Proxy/CD&A Disclosure 2012 Say on Pay Results Proxy Advisory Firms SEC Rule Making 2013 Say on Pay Votes Institutional Shareholders Dodd Frank Performance (TSR) The Media Clawbacks & CEO Pay Ratios Consultant Independence Committee Conflicts Hedging & Pledging Policies

Evaluating the Impact of ISS ISS still holds the largest market share amongst investors Among companies with negative ISS recommendations, only 12.2% failed But, over half passed with less than 70% shareholder support, triggering additional ISS scrutiny next year The impact of Glass Lewis recommendations in 2011 was similar the 10% of firms earning an F grade from Glass Lewis received average shareholder support levels of 73% Voting Outcomes with Negative ISS Recommendations 12.2% 36.6% 51.2% Passed (Above 70%) Passed (Below 70%) Failed (Below 50%) Sources: Radford analysis of data from Institutional Shareholder Services (ISS) and Aon Hewitt s Hot Topics in Executive Compensation Webinar (July 2013)

What Drives Negative Say-on-Pay Recs? Pay-for-performance disconnects are the largest driver of negative Say-on-Pay recommendations for both Glass Lewis and ISS 100% 80% 85% Glass Lewis Drivers for Negative Say-on-Pay Recommendations 2011 2012 60% 40% 20% 41% 57% 23% 49% 33% 45% 43% 32% 32% 16% 16% 16% 18% 29% 28% 24% 25% 17% 8% 0% Source: Radford analysis of data from Institutional Shareholder Services (ISS)

Aon Hewitt s 2013 Hot Topics Survey What is your Top Priority? Competitiveness of the targeted total compensation 75% Redesign/implementation of compensation/benefit programs 24% Degree of pay for performance in executive compensation package 54% Impact of new policies/regulations (ISS Pay-for-Performance Policies) 23% Retention of current talent 53% Shareholder/say-on-pay concerns 18% Alignment of compensation & benefit levels of company performance 46% Drafting of Proxy/CD&A disclosures 15% Communication of compensation and benefit costs 38% Development of (an) appropriate peer group(s) 15% Control of compensation and benefit costs 32% Risk-levels impacting compensation programs 12% Attraction of new talent 26% Governance 9% Source: Aon Hewitt s Hot Topics in Executive Compensation Webinar (July 2013)

Broad-based Equity Granting Trends: 2005-2012 Proportion of Participants by Award Type Proportion of Grant Value by Award Type Options Restricted Performance Options Restricted Performance 2005 2006 2007 2008 2009 2010 2011 2012 2005 2006 2007 2008 2009 2010 2011 2012 Source: Fidelity Stock Plan Services Administration data, 2005-2012

But Are Performance Awards Really Paying for Performance?

Description of Research Methodology Fidelity performance award data 131 clients Average participant size: 185 (Ranging from 1 2,270) Payouts over 6 years ( 08 13) 1-year to 3-year plans Data analysis compared payouts to a company s performance Benchmarked performance against S&P 500 focused on TSR & EPS Payout classification: Payout Above Target considered high pay; Below Target low pay Number of Employees Market Capitalization % of Employees Receiving Performance Award Vesting Years Performance Period Years <2.5K 2.5K 10K 10K 25K 25K 100K >100K 21 41 31 29 9 <$1B $1B $5B $5B $10B $10B $25B >$25B 18 42 21 26 15 <1% 1% 5% 5% 10% 10% 15% >15% 77 43 6 2 3 1 year 2 years 3 years 4 years 5 years 15.2% 18.2% 56.7% 6.6% 3.2% 1 year 2 years 3 years 4 years 5 years 25.2% 15.2% 55.3% 2.8% 1.4% Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Research Highlights Payouts continue to be in directional alignment with relative performance High performing companies tend to demonstrate stronger alignment between performance and award payouts than low performing companies This is particularly true when comparing payouts for awards based on one year of performance vs. three years Raises questions as to whether companies are setting a floor when setting performance goals Performance award cycles ending in the last two years (2011 and 2012) have paid out better than cycles ending in earlier years Significant economic uncertainty at the time performance goals were set may be contributing to this outcome Across most data sets, TSR has risen to the top as the most prevalent performance measure, although earnings measures continue to be widely used Awards based on an earnings measure resulted in above target payouts more often than TSRbased awards Using absolute or relative measures do not result in any materially different payout, but using both resulted in max payouts a vast majority of the time Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Performance Plans Align Pay with TSR Performance Max TSR Performance Versus S&P 500 (n=155) Low Performance - High Pay (n= 31) High Performance - High Pay (n= 67) Overall, payouts aligned with TSR Performance 66% of the time Performance Share Payout Target Low Performance - Low Pay (n= 36) R² = 0.1248 High Performance - Low Pay (n= 21) Key Conclusions: Among high performing companies, 76% of payouts were aligned with performance In contrast, payouts among low performers were directionally aligned only 54% of the time 0% 25% 50% 75% 100% TSR Percentile Rank Relative to the S&P 500 Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Performance Plans Align Pay with EPS Growth Max EPS Growth Performance Versus S&P 500 (n=130) Low Performance - High Pay (n= 30) High Performance - High Pay (n= 52) Payouts aligned with EPS Performance 68% of the time Performance Share Payout Target Low Performance - Low Pay (n= 36) R² = 0.2954 High Performance - Low Pay (n= 12) Key Conclusions: EPS has a stronger statistical correlation to performance compared to TSR, with fewer significant outliers High Performance Low Pay outliers are almost entirely removed 0% 25% 50% 75% 100% EPS Growth Percentile Rank Relative to the S&P 500 Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Performance Plan Cycles Ending in Recent Years Paid Better than Years Prior The vast majority of performance plans ending in 2012 and 2011 paid at target or above Payouts for performance plan cycles ending in 2010 were more evenly distributed between above and below target Performance Plan Payouts by Years 2012 13% 18% 7% 45% 17% n=55 Maximum Payouts have remained consistent Year over Year Approximately 17% 2011 4% 22% 4% 53% 17% n=45 Above Target payouts significantly up ticked in recent years 2010 22% 30% 30% 18% n=38 Approximately 70% Target or above last two years 0% 25% 50% 75% 100% Zero Below Target Target Between Target and Max Max Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Taking a Deeper Dive into Performance Plan Payouts Viewing the data into by performance quartile: Bottom, Middle, and Top Performers Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Performance Plan Cycles Starting to Trend Toward Above Target Payouts Performance vs. Benchmark Below Target Payouts Above Target Payouts Key Conclusions: Top Performers EPS TSR Top Quartile exceed above target payouts at higher rates than Bottom Quartile paid below target Middle Performers Bottom Performers EPS TSR EPS TSR Middle Quartile payouts mirrored the overall profile and paid above target approximately 66% of the time 100% 50% 0% 50% 100%

Performance Plan Cycles Ending in Recent Years Paid Significantly Better than Years Prior (cont d) Our hypothesis: Considering most plans have three-year performance cycles, performance goals for plans ending in 2011 and 2012 were set during a time of great economic uncertainty (in 2009 or 2010) In light of these economic conditions, company-specific outlooks were highly unpredictable, causing many companies to set goals at realistic levels rather than goals with stretch As the economy recovered, many companies faired better than expected S&P 500 TSR for the 3-year periods ending in 2011 and 2012 was significantly stronger than the 3-year period ending in 2010, returning to pre-recession levels Period when Goals Set S&P500 3 Year TSR Through 12/31 Performance Cycle S&P500 3 Year TSR Over Performance Cycle End of 2007 +8.61% 2008 2010 2.84% End of 2008 8.36% 2009 2011 +14.11% End of 2009 5.62% 2010 2012 +10.87%

Pay for What Performance? EPS continues to be a commonly used performance measure, while TSR has gained significant momentum in recent years TSR has outpaced EPS as the most widely used performance measure 60% 50% Prevalence of performance metrics utilized in plans 40% 30% 20% Fidelity CB100 Equilar 10% 0% TSR/Stk Price EPS/Net Inc/EBIT Returns Rev Cashflow Other Financial Non Financial Source: Fidelity Performance Plans 2012 Grants; Equilar.com/knowledge network/researcharticles/2013/pdfs/2013 equilar sp1500 metrics article.pdf; 2012 ClearBridge 100 Data PwC Equity Compensation Study, January 2011

What Influence Does the Performance Measure Selected Have on Payouts? Plans based on EPS and other earnings measures paid out at or above Target approximately twothirds of the time While TSR-based awards paid out above target only slightly more than half of the time Are companies cherrypicking performance measures? TSR 19% Performance Plan Payouts by Performance Measure 25% 56% n=16 Analysis suggests companies are better at predicting internal financial results, or have greater control over the performance outcome when an internal financial measure is used This gives some credibility to institutional investor/proxy advisory push for relative metrics, i.e., TSR EPS Other Earnings Revenue Returns Other Financial Past performance is no guarantee of future results. Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013 9% 16% 22% 29% 38% 16% 27% 25% 43% 27% 56% 0% 25% 50% 75% 100% Zero Below Target Target Between Target and Max Max 64% 14% 38% 36% 22% 14% 4% n=25 n=8 n=7 n=11 n=9

What Influence Does the Performance Measure Selected Have on Payouts? continued There is little difference in how plans pay out when comparing absolute vs. relative plans Plans that are based on a combination of absolute and relative measures, however, paid between Target and Max 83% of the time Performance Plan Payouts Absolute vs. Relative Plans Relative 23% 15% 62% Absolute 14% 26% 2% 38% 19% Both 17% 83% 0% 25% 50% 75% 100% Source: Fidelity Stock Plan Services and ClearBridge Compensation Group Performance Plan Study, 2013

Performance Measurement Period: Implications of Short vs. Longer Term % of Payouts Aligned with Performance 86% 79% 80% 72% 74% 69% 59% 60% 50% 45% 40% 43% High Performers Low Performers 1 Year R 2 = 0.3226 EPS 3 Year R 2 = 0.3034 1 Year R 2 = 0.0182 TSR 3 Year R 2 = 0.2483 1 Year 3 Year R 2 = 0.1160 R 2 = 0.2007 2012 Payouts Key Conclusions: 3 year measurement periods are generally better at creating Performance and Pay alignment across multiple categories versus 1 year periods 1 year measurement periods: Low performers pay low less than 50% of the time (i.e., the majority of payouts are misaligned) 3 year measurement generally reverses that alignment and Low Pay / Low Performance is ~60% of occurrences High Pay / High Performance is strongly aligned across all measurement periods: Over 70% of occurrences

When Business Conditions and Results Soften: How Do Plans Perform? Max Above Target Target Negative Positive Negative Positive Negative Positive Net Income Growth Do performance plans reward poor business performance? Shareholders expect that when business soften, performance plans will react accordingly Key Conclusions: Positive Net Income Growth continues to be strongly associated with Target thru Max Payouts. No Max payouts for negative Net Income Growth. 2013 reduced Positive Growth below target payments increasing Target Payment Below Target Negative Positive Zero Payouts continues to be mostly aligned mostly with Negative Net Income Growth. Zero Negative Positive Number of Payouts Move to the Middle: Target payouts increase for both performance scenarios reducing Above Target Negative occurrences and Below Target Positive occurrences

Focused Look: Do Relative TSR Plans Really Pay for Performance?

Relative TSR Plans Payouts are based upon the company s Total Shareholder Return performance relative to the TSR performance of a peer group or index, aligning with shareholders and theoretically removing the impact of market forces. Target Share Award Relative TSR Peer X Group Performance = Multiplier From Schedule Actual Shares Earned However, Executives feel they have little control and influence on stock price performance, and that TSR is not indicative of actual company performance due to the influence of market forces. Relative TSR plans do not reward actual sustained long-term performance. It s all timing. Why am I being measured on a metric that easily affected by external market forces? I have no control over stock price.

Radford s Balance Study Radford analyzed over 800 public companies, comprising the following indices, studying historical data over a 13-year period, with the assumption that a new 3-year performance period begins every quarter creating 40 rolling performance periods. S&P 500 S&P Midcap 400 NASDAQ 100 Dow 30 We ranked each company s stock price performance against the respective companies in the index, using the 5 metrics below. Metric Definition 13-Year Index TSR TSR from October 1, 1997 to June 30, 2010 Best Average Rank The best average performance for the 40 periods (or the highest of the TSRs on average) Worst Average Rank The worst average performance for the 40 periods (or the lowest of the TSRs on average) Standard Deviation (SD) The square root of the variance. In this analysis, this represents the amount at which the ranks vary from the mean. The reciprocal relationship between the average ranks over the 40 periods and the rank Correlation of the specific 13-year TSR. Goal was to understand the relationship of the relative performance over the individual 3-year periods (hypothetical Relative TSR plans) and the company s long-term performance. Source: 2011 Radford Review Relative TSR Plans: The Balanced Approach to Equity

Balance Study Highlights Performance within Relative TSR plans (and subsequently payouts) is not necessarily skewed by shortterm market performance No consistently extreme performers skewing results Payouts tend to be normally distributed over a long-term period (13 years) Performance within Relative TSR plans (and subsequently payouts) is highly correlated with long-term company performance Essentially, a high performing company over the 13 year period tends to have outperformed the peers and received high payouts over the individual Relative TSR periods Likewise, a low performing company over the 13 year period tends to have underperformed the peers and received low payouts over the individual Relative TSR periods Best Average Rank Worst Average Rank Index 13-Year Index TSR Number of Cos. Standard Deviation Correlation S&P 500 7.88% 418 43.5 371.7 62.31 0.8529 S&P Midcap 400 113.94% 272 36.4 224.4 35.96 0.8499 NASDAQ 100 24.78% 85 17.9 65.7 12.14 0.7854 Dow 30 21.94% 30 7.7 22.1 3.73 0.7366 Source: 2011 Radford Review Relative TSR Plans: The Balanced Approach to Equity

Normally Distributed Payouts S&P 500 Source: 2011 Radford Review Relative TSR Plans: The Balanced Approach to Equity

Snap Shot High Performers in S&P 500 The awards are truly balanced, as long-term performance and Relative TSR performance are correlated, but the individual Relative TSR periods are not skewed. High performing companies over the 13-year period yielded high average ranks, which yielded high average payouts over that time period. Still, the companies ranged from a rank of 1 to 415, indicating that individual extreme performers did not skew the results. 13-Year Period 40 Relative TSR Periods Company TSR Rank Average Rank Best Rank Worst Rank Apple 4573% 3 99.3 1 387 CarMax 624% 26 153.0 1 349 FLIR Systems 1064% 14 81.3 1 415 Jacobs Engineering 244% 99 113.7 10 303 Public Storage 443% 45 113.6 44 262 UnitedHealth 348% 61 145.5 13 394 Urban Outfitters 1554% 10 119.3 1 397 Valero Energy 244% 97 134.3 4 406 Source: 2011 Radford Review Relative TSR Plans: The Balanced Approach to Equity

Snap Shot Low Performers in S&P 500 Balance is maintained with low performing companies. Low performing companies over the 13-year period yielded low average ranks, which yielded low average payouts over that time period. Ranks ranged from 3 to 418, so low performing companies do not skew results either. 13-Year Period 40 Relative TSR Periods Company TSR Rank Average Rank Best Rank Worst Rank Bristol-Myers Squibb -37% 366 297.4 60 404 Dell -3% 325 303.3 44 412 Eastman Kodak -92% 417 371.7 257 414 JDS Uniphase -74% 403 346.2 3 418 Motorola -66% 394 296.0 49 405 New York Times -68% 396 312.9 117 417 Sprint Nextel -86% 415 329.1 63 416 Sara Lee -16% 342 307.7 165 400 Source: 2011 Radford Review Relative TSR Plans: The Balanced Approach to Equity

Important Additional Information FOR PLAN SPONSOR USE ONLY. NOT FOR PLAN PARTICIPANT DISTRIBUTION. The tax information contained herein is general in nature and for informational purposes only. Fidelity does not provide legal or tax advice. Always consult an attorney or tax professional regarding your specific legal or tax situation. Stock plan recordkeeping and administrative services are provided by Fidelity Stock Plan Services, LLC. Aon Hewitt, The Travelers Companies, Inc., ClearBridge Compensation Group and Fidelity Investments are not affiliated. S&P 500 is a registered service mark of Standard & Poor's Financial Services LLC. Fidelity Brokerage Services LLC, Member NYSE, SIPC, 900 Salem Street, Smithfield, RI 02917 621825.2.0

Contacts Dan Kapinos Associate Director, Global Technical Shared Services Team Aon Hewitt dkapinos@radford.com Laura Smith Global Compensation & Benefits The Travelers Companies, Inc. LSMITH@travelers.com Carl J. Stegman Sr. Vice President, Stock Plan Services Product Management Fidelity Investments carl.stegman@fmr.com