chapter 10

PAY-FOR-PERFORMANCE: SHORT TERM INCENTIVE PLANS (INDIVIDUAL-BASED)

Merit Bonuses aka Lump-Sum Bonuses - Merit bonuses differ from merit pay increases in that employees receive an end-of-year bonus that does not build into base pay. Because employees must earn this increase every year, it is viewed less as an entitlement than as merit pay.17 It also provides employers a way to make pay vary more in line with variations in company performance by reducing fixed salary costs that can grow rapidly through merit pay increases. As Exhibit 10.6 indicates, merit bonuses can be considerably less expensive than merit pay over the long run. - Notice how quickly base pay rises under a merit pay plan. After just five years, base pay is almost $14,000 higher than it is under a merit bonus plan. It should be no surprise that cost-conscious firms report switching to merit bonuses. It also should be no surprise that employees aren't particularly fond of merit bonuses. After all, the intent of merit is to cause shock waves in an entitlement culture. By giving merit for several years, a company is essentially freezing base pay. Gradually this results in a repositioning relative to competitors. The message becomes loud and clear: "Don't expect to receive increases in base pay year after year—new rewards must be earned each year." Consider the bonus system developed by Prometric Thomson Learning call centers, which register candidates for computerized tests. The centers have very clear targets that yield specific employee bonuses, as shown in Exhibit 10.7. At Prometric, each day is a new day when it comes to earning bonuses. Individual Spot Awards - Spot awards are seen by many organizations as being effective.18 Usually these payouts are awarded for exceptional performance, often on special projects or for performance that so exceeds expectations as to be deserving of an add-on bonus. The mechanics are simple: After the fact, someone in the organization alerts top management to the exceptional performance. If the company is large, there may be a formal mechanism for this recognition, and perhaps some guidelines on the size of the spot award (so named because it is supposed to be awarded "on the spot"). Smaller companies may be more casual about recognition and more subjective about deciding the size of the award. The Pharmacy School at the University of California at San Francisco has a pretty typical spot award program. Awards are given for such behaviors as "effectively resolved a complaint situation," or "went beyond the expected by staying late to get a grant out on time."19 Individual Incentive Plans - These plans differ from the merit and lump sum payments because they offer a promise of pay for some objective, preestablished level of performance. For example, offering students financial incentives for getting better grades actually works. These students end up having better first- and second-year grade point averages. The "bribe" worked even better for impatient students. (Yeah, we know, get on with the story.)20 - All incentive plans have one common feature: an established standard against which worker performance is compared to determine the magnitude of the incentive pay. For individual incentive systems, this standard is compared against individual worker performance. Because it's often difficult to find good, objective individual measures, individual incentive plans don't work for every job. How, for example, would you come up with an incentive plan for construction laborers? Maybe this wouldn't be difficult if they did the same thing all day: Your goal is to dig 5 feet of trench, 2 feet wide by 18 inches deep, every hour. But construction laborers aren't limited to shovel jobs. They also help pour concrete, assist carpenters and masons framing buildings, etc. The job is too complex for an individual incentive plan. Even a repetitive job like working on an assembly line isn't well suited to individual incentives. One of us (Newman) used to work on a Ford assembly line building Lincolns. Even if workers wanted to build faster to make more money, the line went by with a new car frame every 55 seconds. There is no room here for individual differences, we would argue. Despite this constraint, a number of different individual incentive plans exist. Their differences can be reduced to variation along two dimensions and can be classified into one of four cells, as illustrated in Exhibit 10.8. The first dimension on which incentive systems vary is in the method of rate determination. Plans set up a rate based either on units of production per time period or on time period per unit of production. On the surface, this distinction may appear trivial, but, in fact, the deviations arise because tasks have different cycles of operation.21 Short-cycle tasks, those that are completed in a relatively short period of time, typically have as a standard a designated number of units to be produced in a given time period. For example, a book distributor we worked with had an incentive plan for packers. Number of books packed is a short-cycle task, with only seconds taken to get a book from a supply stack and place in a shipping box. For long-cycle tasks, this would not be appropriate. It is entirely possible that only one task or some portion of it may be completed in a day. Consequently, for longer-cycle tasks, the standard is typically set in terms of time required to complete one unit of production. Individual incentives are based on whether or not workers complete the task in the designated time period. Auto mechanics work off a blue book that tells how long, for example, a fuel injection system should take to replace. Finish faster than the allotted time and the full pay is awarded. The second dimension on which individual incentive systems vary is the specified relationship between production level and wages. The first alternative is to tie wages to output on a one-to-one basis, so that wages are some constant function of production. In contrast, some plans vary wages as a function of production level. For example, one common alternative is to provide higher dollar rates for production above the standard than for production below the standard. Each of the plans discussed in this section has as a foundation a standard level of performance determined by some form of time study or job analysis completed by an industrial engineer or trained personnel administrator. The variations in these plans occur in either the way the standard is set or the way wages are tied to output. As in Exhibit 10.8, there are four general categories of plans: 1. Cell 1: The most frequently implemented incentive system is a straight piecework system. Rate determination is based on units of production per time period, and wages vary directly as a function of production level. The major advantages of this type of system are that it is easily understood by workers and, perhaps consequently, is more readily accepted than some of the other incentive systems. 2. Cell 2: Two relatively common plans set standards based on time per unit and tie incentives directly to level of output: (1) standard hour plans and (2) Bedeaux plans. A standard hour plan is a generic term for plans setting the incentive rate based on completion of a task in some expected time period. A common example we introduced earlier can be found in any neighborhood gasoline station or automobile repair shop. Let us assume that you need a new transmission. The estimate you receive for labor costs is based on the mechanic's hourly rate of pay, multiplied by a time estimate for job completion derived from a book listing average time estimates for a wide variety of jobs. If the mechanic receives $40 per hour and a transmission is listed as requiring four hours to be removed and replaced, the labor cost would be $160. All this is determined in advance of any actual work. Of course, if the mechanic is highly experienced and fast, the job may be completed in considerably less time than indicated in the book. However, the job is still charged as if it took the quoted time to complete. The "surplus" money is split between the employee and the service station. Standard hour plans are more practical than straight piecework plans for long-cycle operations and jobs that are nonrepetitive and require numerous skills for completion.22 3. A Bedeaux plan provides a variation on straight piecework and standard hour plans. Instead of timing an entire task, a Bedeaux plan requires division of a task into simple actions and determination of the time required by an average skilled worker to complete each action. After the more detailed time analysis of tasks, the Bedeaux system functions similarly to a standard hour plan. 4. Cell 3: The two plans included in cell 3 provide for variable incentives as a function of units of production per time period. Both the Taylor plan and the Merrick plan provide different piece rates, depending on the level of production relative to the standard. The Taylor plan establishes two piecework rates. One rate goes into effect when a worker exceeds the published standard for a given time period. This rate is set higher than the regular wage incentive level. A second rate is established for production below standard, and this rate is lower than the regular wage. The Merrick system operates in the same way, except that three piecework rates are set: (1) high for production exceeding 100 percent of standard; (2) medium for production between 83 and 100 percent of standard; and (3) low for production less than 83 percent of standard. Exhibit 10.9 compares these two plans. 5. Cell 4: The three plans included in cell 4 provide for variable incentives linked to a standard expressed as a time period per unit of production. The three plans include the Halsey 50-50 method, the Rowan plan, and the Gantt plan. The Halsey 50-50 method derives its name from the shared split between worker and employer of any savings in direct cost. An allowed time for a task is determined via time study. The savings from completion of a task in less than the standard time are allocated 50-50 (most frequent division) between the worker and the company. The Rowan plan is similar to the Halsey plan in that an employer and employee both share in savings resulting from work completed in less than standard time. The major distinction in this plan, however, is that a worker's bonus increases as the time required to complete the task decreases. For example, if the standard time to complete a task is 10 hours and it is completed in 7 hours, the worker receives a 30 percent bonus. Completion of the same task in 6 hours would result in a 40 percent bonus above the hourly wage for each of the 6 hours. The Gantt plan differs from both the Halsey and the Rowan plans in that the standard time for a task is purposely set at a level requiring high effort to complete. Any worker who fails to complete the task in the standard time is guaranteed a preestablished wage. However, for any task completed in standard time or less, earnings are pegged at 120 percent of the time saved. Consequently, workers' earnings increase faster than production whenever standard time is met or exceeded. Individual Incentive Plans: Returns (But Also Risks) - Although individual incentive plans receive much attention (probably because they carry risks—see below), it turns out that they are not widely used. One estimate is that fewer than 7 percent of U.S. employees are covered by individual incentive plans and almost half of those are in sales occupations. Thus, outside of sales occupations, fewer than 4 percent of employees work under such plans.23 There is strong evidence that individual incentives, on average, have substantial positive effects on performance.24 - However, besides not fitting many jobs in the new economy, another reason for their limited use is that with such plans, things can go wrong—sometimes spectacularly wrong.25 For example, as we have noted, incentive plans can lead to unexpected, and undesired, behaviors. Certainly Sears, one of our examples in Chapter 9, did not want the public relations nightmare of having mechanics sell unnecessary repairs, but the incentive program encouraged that type of behavior. Please don't think Sears is an isolated example. Workers in a subsidiary of Caterpillar were placed on an incentive system to find problems with rail cars that could be repaired and charged back to their respective owners. Under pressure employees—just like at Sears—manufactured problems (smashing brakes with hammers, ruining wheels with chisels) and then making the "repairs."26 This is a common problem with incentive plans: Employees and managers end up in conflict because the incentive system often focuses only on one small part of what it takes for the company to be successful.27 Employees, being rational, do more of what the incentive system pays for. Sales staff provide a perfect example. Try developing a unit-based sales system that doesn't prioritize products by offering different levels of incentives. In this case the smart salesperson sells the easiest product to unload (e.g., discounted and bargain prices products).28 As a sad example, New York teachers and administrators were put on an incentive system to get student pass rates and graduation numbers up. Both these rates increased, not because of better performance, but because standards were lowered so the incentive would be paid out. In another example, evidence suggests that hospitals using some types of pay for performance are more likely to "upcode" medical conditions into more complex categories, which brings higher reimbursements from Medicare.29 The cause of the Great Financial Crisis is typically attributed in significant part to the fact that incentives were improperly designed such that they drove mortgage loan originators to sell/approve more mortgage loans to make more money for themselves and the company, but without an adequate incentive to be careful not to sell/approve mortgages for people unlikely to be able to afford them. Wells Fargo has paid substantial fines for creating incentives for its customer-facing employees to open more customer accounts to drive more revenue, even opening accounts for customers without their knowledge! Exhibit 10.10 outlines some of the general potential problems, as well as potential advantages, with individual incentive plans. Individual Incentive Plans: Examples - Even though pure individual incentive systems are not as widely used as sometimes thought, there are notable successes. Of course, most sales positions have some part of pay based on commissions, a form of individual incentive. One of today's biggest success stories is the merger of individual incentives with efforts to reduce health care costs. For example, Jet Blue deposits $400 into employee health reimbursement accounts for participating in various activities such as smoking cessation programs or running in Ironman contests. In general, health incentives are on the rise: 57 percent of companies used them in 2009, while over 80 percent are expected to use them this year.30 Perhaps the longest-running success with individual incentives, going back to before World War I, belongs to a company called Lincoln Electric. In Exhibit 10.11, the compensation package for factory jobs at Lincoln Electric is described. Notice how the different pieces fit together. This isn't a case of an incentive plan operating in a vacuum. All the pieces of the compensation and reward package fit together. Both culture and the performance review system support the different pay components. Lincoln Electric's success is so striking that it's the subject of many case analyses.31

PAY-FOR-PERFORMANCE: SHORT-TERM INCENTIVE PLANS (TEAM-BASED)

When we move away from individual incentive systems and start focusing on people working together, we shift to team or group incentive plans. The group might be a work team. It might be a department. Or we might focus on a division or the whole company. Or it might even be a pirate ship! Around 1750 Captain Henry Morgan, infamous pirate, recruited and motivated his men using a simple group incentive plan: an even split of all "booty" captured (of course, after distribution of shares to Captain Morgan and his top men). "No prey, no pay" was an early tagline, and it worked: A good day for a pirate was 1,000 pound sterling, far more than the 13 to 33 pounds sterling earned by merchant seamen in more legitimate vessels.32 The basic concept is still the same, though. A standard is established against which worker performance (in this case, team performance) is compared to determine the magnitude of the incentive pay. With the focus on groups, now we are concerned about group performance in comparison against some standard, or level, of expected performance. The standard might be an expected level of operating income for a division. Or the measure might be more unusual, as at Litton Industries (now a part of Northrop Grumman Corp). One division has a team variable-pay measure that is based on whether customers would be willing to act as a reference when Litton solicits other business. The more customers willing to do this, the larger the team's variable pay.33 In a second study, four food processing plants moved to team-based structures and implemented team incentive plans. The team incentive component resulted in productivity increases of 9-20 percent.34 Some group incentive plans have even higher success goals—like winning a war. Napoleon took over an army that was underfed and demoralized. He developed one of the earliest profit sharing plans by promising soldiers a share of any bounty achieved. When Napoleon defeated Italians in what is now the Piedmont region, he demanded gold and silver from the defeated foe. Morale soared when Napoleon then shared this with his troops. Despite an explosion of interest in teams and team compensation, many of the reports from the front lines are not encouraging.35 Companies report they generally are not satisfied with the way their team compensation systems work. Failures of team incentive schemes can be attributed to at least five causes.36 First, one of the problems with team compensation is that teams come in many varieties. There are full-time teams (work group organized as a team). There are part-time teams that cut across functional departments (experts from different departments pulled together to improve customer relations). There are even full-time teams that are temporary (e.g., cross-functional teams pulled together to help ease the transition into a partnership or joint venture). With so many varieties of teams, it's hard to argue for one consistent type of compensation plan. Unfortunately, we still seem to be at the stage of trying to find the one best way. Maybe the answer is to look at different compensation approaches for different types of teams. Perhaps the best illustration of this differential approach for different teams comes from Xerox. Xerox has a gain-sharing plan that pays off for teams defined at a very broad level, usually at the level of a strategic business unit. For smaller teams, primarily intact work teams (e.g., all people in a department or function), there are group rewards based on supervisory judgments of performance. Units that opt to have their performance judged as teams (it is also possible to declare that a unit wouldn't be fairly judged if team measures were used) have managers who judge the amount to be allocated to each team based on the team's specific performance results. For new teams, the manager might also decide how much of the total will go to each individual on the team. More mature teams do individual allocations on their own. In Xerox's experience, these teams start out allocating equal shares, but as they evolve the teams allocate based on each worker's performance. Out of about 2,000 work teams worldwide at Xerox, perhaps 100 have evolved to this level of sophistication. For problem-solving teams and other temporary teams, Xerox has a reward component called the Xerox Achievement Award. Teams must be nominated for exceptional performance. A committee decides which teams meet a set of predetermined absolute standards. Even contributors outside the core team can share in the award. If nominated by team members, extended members who provide crucial added value are given cash bonuses equal to those of team members. A second problem with rewarding teams is called the "level problem" or "line of sight" problem. (See Expectancy Theory in Chapter 9.) If we define teams at the very broad level—the whole organization being an extreme example—much of the motivational impact of incentives can be lost. As a member of a 1,000-person team, I'm unlikely to be at all convinced that my extra effort will significantly affect our team's overall performance. Why, then, should I try hard? Conversely, if we let teams get too small, other problems arise. TRW found that small work teams competing for a fixed piece of incentive awards tend to gravitate to behaviors that are clearly unhealthy for overall corporate success. Teams hoard star performers, refusing to allow transfers even for the greater good of the company. Teams are reluctant to take on new employees for fear that time lost to training will hurt the team—even when the added employees are essential to long-run success. Finally, bickering arises when awards are given. Because teams have different performance objectives, it is difficult to equalize for difficulty when assigning rewards. Inevitably, complaints arise.37 The last three major problems with team compensation involve the three Cs: complexity, control, and communications. Some plans are simply too complex. Xerox's Houston facility had a gain-sharing plan for teams that required understanding a three-dimensional performance matrix. Employees (and these authors!) threw their hands up in dismay when they tried to understand the "easy-to-follow directions." In contrast, Xerox's San Diego unit has had great success with a simple program called "bet the boss." Employees come to the boss with a performance-saving idea and bet their hard effort against the boss's incentive that they can deliver. Such plans have a simplicity that encourages employee buy-in. With a good line of sight as it's called, employees can see a clear link between their effort and the rewards they receive. The second C is control. Praxair, a worldwide provider of gases (including oxygen) extracted from the atmosphere, works hard to make sure all its team pay comes from performance measures under the control of the team. If mother nature ravages a construction site, causing delays and skyrocketing costs, workers aren't penalized with reduced team payouts. Such uncontrollable elements are factored into the process of setting performance standards. Indeed, experts assert that this ability to foretell sources of problems and adjust for them is a key element in building a team pay plan.38 Key to the control issue is the whole question of fairness. Are the rewards fair given our ability to produce results? Recent research suggests that this perception of fairness is crucial.39 With it, employees feel it is appropriate to monitor all members of the group—slackers beware! Without fairness, employees seem to have less sense of responsibility for the team's outcomes.40 The final C is a familiar factor in compensation successes and failures: communication. Team-based pay plans simply are not well communicated. Employees asked to explain their plans often flounder because more effort has been devoted to designing the plan than to deciding how to explain it. Conversely, the more transparent the plan, the more employees trust management and respond positively to the incentive effects of the plan. Although there is much pessimism about team-based compensation, many companies still seek ways to reward groups of employees for their interdependent work efforts. Companies that do use team incentives typically set team performance standards based on productivity improvements (38 percent of plans), customer satisfaction measures (37 percent), financial performance (34 percent), or quality of goods and services (28 percent).41 For example, Kraft Foods uses a combination of financial measures (e.g., income from operations and cash flow) combined with measures designed to gauge success in developing managers, building diversity, and adding to market share.42 Exhibit 10.12 summarizes some of these measures. As Exhibit 10.12 suggests, the range of performance measures for different types of corporate objectives is indeed impressive.43 For example, if the corporate objective is to reward short term performance, the measures outlined in Exhibit 10.13 could be used. Historically, financial measures have been the most widely used performance indicator for large group incentive plans. Increasingly, though, top executives express concern that these measures do a better job of communicating performance to stock analysts than to managers trying to figure out how to improve operating effectiveness.44 Whatever our thinking is about appropriate performance measures, the central point is still that we are now concerned about group performance. This presents both problems and opportunities. As Exhibit 10.14 illustrates, we need to decide which type of group incentive plan best fits our objectives. Indeed, we should even ask if an incentive plan is appropriate. Recent evidence, for example, suggests that firms high on business risk and those with uncertain outcomes are better off not having incentive plans at all—corporate performance is higher.45 Comparing Group and Individual Incentive Plans - In this era of heightened concern about productivity, we frequently are asked if setting up incentive plans really boosts performance. As we noted in Chapter 9, the answer is yes. And individual—rather than group—incentives win the productivity "medal." We also are asked, though, which is better in a specific situation—group or individual incentive plans. Often this is a misleading question. Individual incentives yield higher productivity gains, but group incentives often are right in situations where team coordination is the issue. One study found that changing from individual incentives to gain sharing resulted in a decrease in grievances and a fairly dramatic increase in product quality (defects per 1,000 products shipped declined from 20.93 to 2.31).46 - As we noted in Exhibit 10.14, things like the type of task, the organizational commitment to teams, and the type of work environment may preclude one or the other type of incentive plan. Exhibit 10.15 provides a guide for when to choose group or individual plans. When forced to choose the type of plan with greater productivity "pep," experts agree that individual incentive plans have better potential for—and probably better track records in—delivering higher productivity. Group plans suffer from what is called the free-rider problem. See if this sounds familiar: You are a team member on a school project and at least one person doesn't carry his or her share of the load. Yet, when it comes time to divide the rewards, they are typically shared equally. Problems like this caused AT&T to phase out many of its team reward packages. Top-performing employees quickly grew disenchanted with having to carry free riders. End result—turnover of the very group that is most costly to lose. - Research on free riders suggests that the problem can be lessened through use of good performance measurement techniques. Specifically, free riders have a harder time loafing when there are clear performance standards. Rather than being given instructions to "do your best," poorer performers who were asked to deliver specific levels of performance at a specific time actually showed the most performance improvement.47 Large Group Incentive Plans - When we get beyond a small work team and try to incentivize large groups, there are generally two types of plans. Gain-sharing plans use operating measures to gauge performance. Profit sharing plans use financial measures. Gain-Sharing Plans - Our discussion of team-based compensation often mentioned gain-sharing plans as a common component. As the name suggests, employees share in the gains in these types of group incentive plans. With profit-sharing plans (surprise) the sharing involves some form of profits. Realistically, though, most employees feel as if little they can do will affect profits; that's something top-management decisions influence more. So gain sharing looks at cost components of the income ledger and identifies savings over which employees have more impact (e.g., reduced scrap, lower labor costs, reduced utility costs). It was just this type of thinking that led the United States Post Office to an annual cost avoidance of $497 million under its gain-sharing plan.48 Other studies of gain sharing report similar positive results. Indeed, the empirical evidence on gain sharing appears to be quite favorable.49 One study of 1,600 employees in an auto parts company showed gain sharing over five years reduced labor, material, tool purchase, scrap, rework, and supply costs. The total savings were $15 million over the five-year period. There were also decreases in absenteeism (by 20 percent) and grievances (by 50 percent).50 - In a particularly good study of a major retailer, stores with gain-sharing incentives had 4.9 percent higher sales, 3.4 percent higher customer satisfaction, and 4.4 percent higher profit than stores without the incentive plan. In our experience these effects are pretty typical of gain-sharing plans—improvements in the 4-5 percent range. Keep in mind, though, gain-sharing plans can lead to the sorting effect we talked about in Chapter 9. Good employees want to be rewarded for their individual effort and performance. Changes to group plans, like gain-sharing, can lead to turnover. Just ask AT&T. They found that very-high- and very-low-performing individuals had much higher turnover rates under gain-sharing than other employees.51 The following issues are key elements in designing a gain-sharing plan:52 1. Strength of reinforcement: What role should base pay assume relative to incentive pay? Incentive pay tends to encourage only those behaviors that are rewarded. For example, try returning an unwanted birthday present to a store that pays its sales force solely for new sales. Tasks carrying no rewards are only reluctantly performed (if at all). 2. Productivity standards: What standard will be used to calculate whether employees will receive an incentive payout? Almost all group incentive plans use a historical standard. A historical standard involves choice of a prior year's performance to use for comparison with current performance. But which baseline year should be used? If too good (or too bad) a comparison year is used, the standard will be too hard (or easy) to achieve, with obvious motivational and cost effects. One possible compromise is to use a moving average of several years (e.g., the average for the past five years, with the five-year block changing by one year on an annual basis). One of the major problems with historical standards is that changing environmental conditions can render a standard ineffective. For example, consider the ice cream company that based profit sharing on a financial measure that had been easily achieved the prior year. What they didn't expect was a dramatic increase in milk costs. It was clear by the third month of the fiscal year that the goal wasn't going to be met, and grumblings among employees were mounting. The targets were adjusted and ice cream makers rejoiced.53 Such problems are particularly insidious during economic swings and for organizations facing volatile economic climates. Care must be taken to ensure that the link between performance and rewards is sustained. This means that environmental influences on performance, which are not controllable by plan participants, should be factored out when identifying incentive levels. 3. Sharing the gains split between management and workers: Part of the plan must address the relative cuts between management and workers of any profit or savings generated. This also includes discussion of whether an emergency reserve (gains withheld from distribution in case of future emergencies) will be established in advance of any sharing of profits. 4. Scope of the formula: Formulas can vary in the scope of inclusions for both the labor inputs in the numerator and the productivity outcomes in the denominator.54 Recent innovations in gain-sharing plans largely address broadening the types of productivity standards considered appropriate. Given that organizations are complex and require more complex measures, performance measures have expanded beyond traditional financial measures. For example, with the push for greater quality management, we could measure retention of customers or some other measure of customer satisfaction. Similarly, other measures include delivery performance, safety, absenteeism, turnaround time, and number of suggestions submitted. Four specific examples are:55 5. Great care must be exercised with such alternative measures, though, to ensure that the behaviors reinforced actually affect the desired bottom-line goal. Getting workers to expend more effort, for example, might not always be the desired behavior. Increased effort may bring unacceptable levels of accidents. It may be preferable to encourage cooperative planning behaviors that result in more efficient work. 6. Perceived fairness of the formula: One way to ensure the plan is perceived as fair is to let employees vote on whether implementation should go forward. This and union participation in program design are two elements in plan success.56 7. Ease of administration: Sophisticated plans with involved calculations of profits or costs can become too complex for existing company information systems. Increased complexities also require more effective communications and higher levels of trust among participants. 8. Production variability: One of the major sources of problems in group incentive plans is failure to set targets properly. At times the problem can be traced to volatility in sales. Large swings in sales and profits, not due to any actions by workers, can cause both elation (in good times) and anger (in bad times). As stated above, a good plan ensures that environmental influences on performance, which are not controllable by plan participants, should be factored out when identifying incentive levels. The second author once worked with an ice cream producer that experienced huge unexpected increases in milk costs. End result? The original profit goal was unattainable. To their credit, the company adjusted the profit target to reflect the uncontrollable cost change. One alternative would be to set standards that are relative to industry performance. To the extent data are available, a company could trigger gain sharing when performance exceeds some industry norm. The obvious advantage of this strategy is that economic and other external factors hit all firms in the industry equally hard. If our company performs better, relatively, it means we are doing something as employees to help achieve success. Exhibit 10.16 illustrates three different formulas that can be used as the basis for gain-sharing plans. The numerator, or input factor, is always some labor cost variable, expressed in either dollars or actual hours worked; the denominator is some output measure such as net sales or value added. Each of the plans determines employees' incentives based on the difference between the current value of the ratio and the ratio in some agreed-upon base year. The more favorable the current ratio relative to the historical standard, the larger the incentive award.57 The three primary types of gain-sharing plans, differentiated by their focus on either cost savings (the numerator of the equation) or some measure of revenue (the denominator of the equation), are noted below. Scanlon Plan - Scanlon plans are designed to lower labor costs without lowering the level of a firm's activity. Incentives are derived as a function of the ratio between labor costs and sales value of production (SVOP).58 The SVOP includes sales revenue and the value of goods in inventory. To understand how these two figures are used to derive incentives under a Scanlon plan, see Exhibit 10.17. In practice, the $50,000 bonus in Exhibit 10.17 is not all distributed to the workforce. Rather, 25 percent is distributed to the company, 75 percent of the remainder is distributed as bonuses, and the other 25 percent is withheld and placed in an emergency fund to reimburse the company for any future months when a "negative bonus" is earned (i.e., when the actual wage bill is greater than the allowable wage bill). The excess remaining in the emergency pool is distributed to workers at the end of the year. To look at the impact of Scanlon plans, consider the retail chain that adopted a Scanlon plan in six of its stores and compared results against six control stores chosen for their similarity.59 Presence of a Scanlon plan led to stores having higher customer satisfaction, higher sales, and lower turnover. Rucker Plan - The Rucker plan involves a somewhat more complex formula than a Scanlon plan for determining worker incentive bonuses. Essentially, a ratio is calculated that expresses the value of production required for each dollar of total wage bill. Consider the following illustration:60 1. Assume accounting records show that the company expended $.60 worth of electricity, materials, supplies, and so on, to produce $1.00 worth of product. The value added is $.40 for each $1.00 of sales value. Assume also that 45 percent of the value added was attributable to labor; a productivity ratio (PR) can be allocated from the formula in item 2. 2. PR (labor) × .40 × .45 = 1.00. Solving yields PR = 5.56. 3. If the wage bill equals $100,000, the expected production value is the wage bill ($100,000) × PR (5.56) = $555,556. 4. If actual production value equals $650,000, then the savings (actual production value minus expected production value) is $94,444. 5. Since the labor contribution to value added is 45 percent, the bonus to the workforce should be .45 × $94,444 = $42,500. 6. The savings are distributed as an incentive bonus according to a formula similar to the Scanlon formula—75 percent of the bonus is distributed to workers immediately and 25 percent is kept as an emergency fund to cover poor months. Any excess in the emergency fund at the end of the year is then distributed to workers. Implementation of the Scanlon/Rucker Plans - Two major components are vital to the implementation and success of a Rucker or Scanlon plan: (1) a productivity norm and (2) effective worker committees. Development of a productivity norm requires both effective measurement of base-year data and acceptance by workers and management of this standard for calculating bonus incentives. Effective measurement requires that an organization keep extensive records of historical cost relationships and make them available to workers or union representatives to verify cost accounting figures. Acceptance of these figures, assuming they are accurate, requires that the organization choose a base year that is neither a "boom" nor a "bust" year. The logic is apparent. A boom year would reduce opportunities for workers to collect bonus incentives. A bust year would lead to excessive bonus costs for the firm. The base year chosen also should be fairly recent, allaying worker fears that changes in technology or other factors would make the base year unrepresentative of a given operational year. - The second ingredient of Scanlon/Rucker plans is a series of worker committees (also known as productivity committees or bonus committees). The primary function of these committees is to evaluate employee and management suggestions for ways to improve productivity and/or cut costs. Operating on a plantwide basis in smaller firms, or a departmental basis in larger firms, these committees have been highly successful in eliciting suggestions from employees. It is not uncommon for the suggestion rate to be above that found in companies with standard suggestion incentive plans.61 - Scanlon/Rucker plans foster this type of climate, and that is perhaps the most vital element of their success. Numerous authorities have pointed out that these plans have the best chance for success in companies with competent supervision, cooperative union-management attitudes, strong top-management interest and participation in the development of the program, and management open to criticism and willing to discuss different operating strategies.62 It is beyond the scope of this discussion to outline specific strategies adopted by companies to achieve this climate, but the key element is a belief that workers should play a vital role in the decision-making process. Similarities and Contrasts Between Scanlon and Rucker Plans - Scanlon and Rucker plans differ from individual incentive plans in their primary focus. Individual incentive plans focus primarily on using wage incentives to motivate higher performance through increased effort. While this is certainly a goal of the Scanlon/Rucker plans, it is not the major focus of attention. Rather, given that increased output is a function of group effort, more attention is focused on organizational behavior variables. The key is to promote faster, more intelligent, and more acceptable decisions through participation. This participation is won by developing a group unity in achieving cost savings—a goal that is not stressed, and is often stymied, in individual incentive plans. Even though Scanlon and Rucker plans share this common attention to groups and committees through participation as a linking pin, there are two important differences between the two plans. First, Rucker plans tie incentives to a wide variety of savings, not just the labor savings focused on in Scanlon plans.63 Second, this greater flexibility may help explain why Rucker plans are more amenable to linkages with individual incentive plans. Improshare - Improshare (Improved Productivity through Sharing) is a gain-sharing plan that has proved easy to administer and to communicate.64 First, a standard is developed that identifies the expected hours required to produce an acceptable level of output. This standard comes either from time-and-motion studies conducted by industrial engineers or from a base-period measurement of the performance factor. Any savings arising from production of the agreed-upon output in fewer than the expected hours is shared by the firm and by the workers.65 For example, if 100 workers can produce 50,000 units over 50 weeks, this translates into 200,000 hours (40 hours × 50 weeks) for 50,000 units, or 4 hours per unit. If we implement an Improshare plan, any gains resulting in less than 4 hours per unit are shared 50-50 between employees and management (wages times number of hours saved).66 - One survey of 104 companies with an Improshare plan found a mean increase in productivity during the first year of 12.5 percent.67 By the third year the productivity gain rose to 22 percent. A significant portion of this productivity gain was traced to reduced defect rates and downtime (e.g., repair time). Profit-Sharing Plans - Profit sharing is also positively related to productivity and productivity growth. One study of 6 million employees across 275 firms found 3.5-5.0 percent higher profits in companies that used profit sharing than in those that didn't.68 Productivity was much higher in plans where payouts were that year than in plans where payment was deferred (as in profit sharing used for a pension program). Also, these plans worked much better in smaller (less than 775 employees) companies. - Even in companies that don't have profit sharing plans, many variable pay plans still require a designated profit target to be met before any payouts occur. Our experience with chief executive officers is that they have a hard time giving employees extra compensation if the company isn't also profiting. Thus, many variable pay plans have some form of profit "trigger" linked to revenue growth or profit margins or some measure of shareholder return such as earnings per share or return on capital. Despite modestly positive results, profit sharing continues to be popular because the focus is on the measure that matters most to the most people: a predetermined index of profitability. When payoffs are linked to such measures, employees spend more time learning about financial measures and the business factors that influence them. - On the downside, most employees don't feel their jobs have a direct impact on profits. A small cog in a big wheel is difficult to motivate very well. For example, before the big crunch in the auto industry Ford Motor and GM gave profit sharing checks of about $7,500, compared to $2,250 at Fiat-Chrysler.69 You can bet the Chrysler employees wondered if their counterparts at Ford and GM were working more than three times as hard. If you guessed that they blamed the difference on bad management decisions, you're right on target. - The trend in recent variable-pay design is to combine the best of gain-sharing and profit-sharing plans.70 The company will specify a funding formula for any variable payout that is linked to some profit measure. As experts say, the plan must be self-funding. Dollars going to workers are generated by additional profits gained from operational efficiency. Along with having the financial incentive, employees feel they have a measure of control. For example, an airline might give an incentive for reductions in lost baggage, with the size of the payout dependent on hitting profit targets. Such a program combines the need for fiscal responsibility with the chance for workers to affect something they can control. Earnings-at-Risk Plans - We probably shouldn't separate earnings-at-risk plans as a distinct category. In fact, any incentive plan could be an at-risk plan. Think of incentive plans as falling into one of two categories: success sharing or risk sharing. In success-sharing plans, employee base wages are constant and variable pay adds on during successful years. If the company does well, you receive a predetermined amount of variable pay. If the company does poorly, you simply forgo any variable pay—there is no reduction in your base pay, though. In a risk-sharing plan, base pay is reduced by some amount relative to the level that would be offered in a success-sharing plan. AmeriSteel's at-risk plan is typical of risk-sharing plans. Base pay was reduced 15 percent across the board in year 1. That 15 percent was replaced with a .5 percent increase in base pay for every 1 percent increase in productivity beyond 70 percent of the prior year's productivity. This figure would leave workers whole (no decline in base pay) if they only matched the prior year's productivity. Each additional percent improvement in productivity yielded a 1.5 percent increase in base wages. Everyone in AmeriSteel, from the CEO on down, is in this type of plan and the result has been an 8 percent improvement in productivity.71 - Clearly, at-risk plans shift part of the risk of doing business from the company to the employee. The company hedges against the devastating effects of a bad year by mortgaging part of the profits that would have accrued during a good year. Not surprisingly, a key element of incentive design is to identify possible risks and incorporate design features that minimize them.72 At-risk plans appear to be met with decreases in satisfaction with both pay in general and the process used to set pay.73 In turn, this can result in higher turnover. Group Incentive Plans: Advantages and Disadvantages - Clearly, group pay-for-performance plans are gaining popularity in today's team-based environment. Other factors play a role, though. One factor with intriguing implications suggests that group-based plans, particularly gain-sharing plans, cause organizations to evolve into learning organizations.74 Apparently the suggestions employees are encouraged to make (how to do things better in the company) gradually evolve from first-order learning experiences of a more routine variety (maintenance of existing ways of doing things) into suggestions that exhibit second-order learning characteristics—suggestions that help the organization break out of existing patterns of behavior and explore different ways of thinking and behaving.75 - Exhibit 10.18 outlines some of the general positive and negative features of group pay-for-performance plans.76 Group Incentive Plans: Examples - All incentive plans, as we noted earlier, can be described by common features:(1) the size of the group that participates in the plan, (2) the standard against which performance is compared, and (3) the payout schedule. Exhibit 10.19 illustrates some of the more interesting components of plans for leading companies.

WHAT IS A PAY-FOR-PERFORMANCE PLAN?

Good question! Many different compensation practices are lumped under the name pay for performance. When your Mom told you last summer she would give you 20 dollars to cut the grass, that's a pay-for-performance plan! When you didn't do a very good job because you were rushing to make a pickup game on the local court, and she paid you anyway, that's still a pay-for-performance plan, just not a very good one. Indeed, it sounds like your Mom learned about pay for performance from many of the managers we've met, whose plan seems to be "Don't distinguish between good and bad performance and pay everyone about the same." Listen long enough and you will hear about incentive, variable pay plans, compensation at risk, earnings at risk, success sharing, risk sharing, and others. (Recall our discussion of these in Chapter 9.) Sometimes these names are used interchangeably. They shouldn't be. The major thing all these plans have in common is a shift in thinking about compensation. We used to think of pay as primarily an entitlement. If you went to work and did well enough to avoid being fired, you were entitled to the same size check as everyone else doing the same job as you. Pay-for-performance plans signal a movement—sometimes a very slow movement—away from entitlement and toward pay that varies with some measure of individual or organizational performance. Of the pay components we discussed in Chapter 9, only base pay and across-the-board increases don't fit the pay-for-performance category. Curiously, though, many of the surveys on pay for performance tend to omit the grandfather of all these plans, merit pay, which, as we will see, is very widely used. Exhibit 10.1 provides another way to classify pay-for-performance plans—in this case, short-term incentive/variable pay plans—and shows the breadth of these types of plans in use. How Widely Used is Pay for Performance (PFP)? - Exhibit 10.1 also provides data on the use of the short-term incentive plans (where the performance period is 12 months or less) in organizations. We see that 99 percent of organizations surveyed use some form of short-term incentive plan for at least some of their employees, and also that most have multiple short-term incentive plans. What in the past was primarily a compensation tool for top management (and is still significantly more likely to be used for them) is now often used for lower-level employees too. The use of variable pay in general has increased. For example, in 1990 the average budget for base salary increases was 5.5 percent. More recently, it was down to 3.0 percent. One reason is a decline in wage and price inflation. However, that does not explain the simultaneous increase in merit bonus/variable pay budgets from 4.2 percent in 1990 to roughly 13 percent today.1 - The greater interest in variable pay probably can be traced to two trends. First, the increasing competition from foreign producers forces American firms to cut costs and/or increase productivity. Well-designed variable pay plans have a proven track record in motivating better performance and helping cut costs. Plus, variable pay is, by definition, a variable cost. No profits, or poor profits, means no extra pay beyond base pay—when times are bad, compensation is lower.2 Second, today's fast-paced business environment means that workers must be willing to adjust what they do and how they do it. There are new technologies, new work processes, and new work relationships. All these require workers to adapt in new ways and with a speed that is unparalleled. Failure to move quickly means market share goes to competitors. If this happens, workers face possible layoffs and terminations. To avoid this scenario, compensation experts are focusing on ways to design reward systems so that workers will be able—and willing—to move quickly into new jobs and new ways of performing old jobs. The ability and incentive to do this come partially from reward systems that more closely link worker interests with the objectives of the company.3 - Other evidence points to the very strong overall reliance on pay for performance (PFP), including variable pay, especially in private sector organizations. World-at-Work surveys its members (compensation professionals) about practices in organizations. Its survey results indicate that pay for performance (PFP) is very widely used. Specifically, 94 percent of organizations have merit pay programs, 99 percent (as we have seen) have short-term incentive plans (payment based on attainment of financial, operational, and individual goals during a period of 12 months or less), and 88 percent have long-term incentive plans (payment based on attainment of goals over a period of longer than 12 months usually related to performance in terms of company stock price/return). However, these percentages underestimate the use of PFP in the private sector, given that 22 percent of responding organizations were in the nonprofit, not-for-profit, or public sectors, where we know that the use of PFP (as well as its intensity when it is used) is considerably lower than in the private sector. Also, the typical long-term incentive plan, which, as noted, is based on company stock performance, is not possible in organizations where there is no stock/ownership. Thus, the typical U.S. private sector company relies especially heavily on PFP, and the percentage of private sector organizations using two or more of the above PFP plans is probably close to 100 percent. - One important caveat is that the roughly 13 percent merit bonus/variable pay percentage applies only to organizations that use such plans and only to the employee groups in such companies covered by such plans. (By contrast, merit pay plans cover almost all employees and, as we will see shortly, nearly all organizations.) Exhibit 10.2 reports short-term incentive/variable payouts and the performance basis used for them as a percentage of base pay, by employee group, in organizations using such plans and in all organizations (on average), adjusted for the fact that not all organizations use such plans and that organizations using such plans do not use them for all employee groups. Again, we see that short-term incentive payouts as a percentage of salary are larger than merit increases. Moreover, for some employee groups (those at higher job/pay levels), short-term incentive/variable payouts are much larger than merit increases. We also see that the most common performance basis for short-term incentive/variable payouts is a combination of corporate, unit, and individual objectives. - The details on performance objectives beyond the split across corporate, business unit, and individual reported in Exhibit 10.2 gets a bit confusing, which reflects the great variety in design specifics of plans in different organizations and for different employee groups. But here goes: In almost all (97 percent of) short-term incentive plans, payouts are based to some degree on financial (i.e., organization-level) performance—most commonly revenue (43 percent), followed closely by various (organization-level) profit measures. In addition, "overall individual performance (e.g., performance evaluation or rating)" was used by 48 percent of organizations. (These plans would be closer to the definition of a merit bonus plan, specifically.) Operational measures of performance were often used also, with the most common being customer satisfaction (28 percent). - Long-term incentive plans (where the performance period is more than 12 months) are also more likely to be used for officers/executives and other higher job levels. Indeed, Exhibit 10.3 shows that, on average, the lowest job levels, where most employees are, receive only a very small percentage of the value of long-term incentives granted by organizations. Long-term incentive plans (especially for top executives) will be covered further in Chapter 14. Here, we note that they include stock grants, stock option grants, performance share grants, performance units, and other programs where the payout depends on shareholder return and/or other corporate financial performance measures. The Important Role of Promotion in Pay for Performance - As noted above, merit pay is widely used by organizations and for all types of employees. As we will see in our discussion below, the average merit pay increase is about 3 percent per year. At that rate, it would take an employee about 23 years to double his/her salary from say $70,000 to $140,000 or from $100,000 to $200,000. Yet, we know that many of you will go on to earn much more than $140,000 or even $200,000 per year. Although it is true that if you are a higher performer, you will get larger merit pay increases and that your salary will increase faster, it will still take a while. So, how will some of you end up making much higher salaries than described here? The answer is that you will get promoted to higher job levels (i.e., salary ranges/grades) and that salary increases due to promotion are much larger than 3 percent. For instance, if you return to Chapter 3 and examine the pay structure for engineers at Lockheed Martin, you will see that a promotion brings a salary increase of roughly 21 to 23 percent. At that rate, one's salary doubles after three or four promotions. Except at higher job levels, promotion increases are probably more in the range of 15 percent (based on looking at typical midpoint progressions—the percentage difference in salary midpoints of adjacent salary ranges).4 Using 15 percent, an employee's salary would double after 5 promotions. Because promotion is based importantly on performance, any discussion of how strongly pay and performance are related must recognize that it is about much more than merit pay increases, which are within-grade increases. For many employees, their performance will be rewarded with higher pay more strongly through their high performance leading to promotions to higher levels, either within their current organization or within an organization that they move to.5

PAY-FOR-PERFORMANCE: MERIT PAY PLANS

A merit pay system links increases in base pay (called merit pay increases) to how highly employees are rated on a performance evaluation. (Chapter 11 covers performance evaluation.) However, most organizations use a merit increase grid, which determines merit pay increases not only on the basis of performance rating, but also on the basis of an employee's position in the salary range/grade (i.e., how close to the minimum versus maximum). Position in range is nicely captured by the compa-ratio, defined as employee salary divided by salary range midpoint. Thus, in a salary range that goes from a minimum of $50,000 to a maximum of $70,000 with a midpoint of $60,000, an employee with a current salary of $53,000 would have a compa-ratio of $53,000/$60,000 = .88, whereas an employee in that same salary range with a salary of $67,000 would have a compa-ratio of $67,000/$60,000 = 1.12. If both employees have the highest performance rating, the employee with the compa-ratio of .88 would get a larger merit increase than the employee with a compa-ratio of 1.12 because the lower-paid employee (with the compa-ratio of .88) is farther away from where his/her salary should be in the salary range if his/her performance were to stay at that high level over time. For example, in the following merit increase grid, the first employee (with the compa-ratio of .88) would receive a 7 percent merit increase (resulting in a new salary of $56,710), whereas the second employee (with the compa-ratio of 1.12) would receive a merit increase of 3 percent (resulting in a new salary of $69,010). One can imagine what would happen without the compa-ratio element of the merit increase grid. If a high-performing employee received a 7 percent salary increase each year, his/her salary would double every 10 years! That would be fine if the value generated for the organization of that same level of high performance also doubled every 10 years. However, that may not be likely in most jobs. Typically, to double one's contribution to the organization, one must get promoted to higher-level jobs that allows more impact on organization performance. Exhibit 10.4 provides an example of a merit increase grid. We can also look at the degree to which companies award different merit increases to employees with different performance ratings. Exhibit 10.5 shows that, on average, in companies using a 5-point rating scale, employees with the highest rating of 5 receive a 4.5 percent merit increase on average, compared to, for example, a 2.6 percent merit increase for an employee in the middle performance category. Exhibit 10.5 also shows that about 7 percent of employees receive the highest rating and about 56 percent of employees receive the middle category performance rating. Of course, this distribution of employees across performance rating categories will have major cost implications. In some companies, as we discuss further in Chapter 11, most employees fall into the top two performance categories, which would translate into more higher merit pay increases and more cost. Finally, Exhibit 10.5 shows that higher performance ratings translate not only into higher merit pay increases, but also into higher short-term incentive payouts. At the end of a performance year, the employee is evaluated, usually by the direct supervisor. A key feature of a merit pay increase is that, unlike variable pay programs (short-term incentives and long-term incentives), the increase is added into base pay. This point is important. In effect, what an employees does this year in terms of performance is rewarded every year for as long as the employee remains with the employer. Once awarded, that merit pay increase is there forever. With compounding, this can amount to tens of thousands of dollars over an employee's work career.6 Year after year, there are concerns about merit pay. One concern is that it increases fixed compensation costs over time. One response has been to use merit bonuses and/or other forms of variable pay plans. Another concern is that merit pay becomes costly if too many high performance ratings are awarded. A response is to control the number of high ratings and/or improve the accuracy and credibility of performance ratings. (See Chapter 11.) Another concern is that merit pay differentials based on performance are too small to motivate performance. That is a challenge. However, larger differentials can be used. Additionally, as discussed above, the strength of merit pay differentials will be greatly underestimated if the role of performance in promotions and resulting salary growth over time is not included. Another potential problem, true of any pay-for-performance program, is that individual performance is a deficient measure in organizations where work is interdependent and requires cooperation to achieve team/organization objectives. A possible solution is to broaden performance criteria to include cooperation and other factors that contribute to team/organization success. Our view is that it has been difficult to study and document the effects of merit pay plans on performance. However, the evidence that does exist is positive.7 Further, when one considers the theory and evidence on pay-for-performance plans broadly, it appears to us that organizations use such plans with good reason: in their absence, an environment is created that does not reward excellence and employees who aspire to excellence will decide to look elsewhere.8 In this vein, it is important to again keep in mind the idea of incentive and sorting effects, first introduced in Chapter 1. Most discussion of merit pay focuses on incentive effects: how does merit pay influence performance of current employees. However, merit pay may have a significant sorting effect in that people who don't want to have their pay tied to performance don't accept jobs at such companies or leave when pay for performance is implemented. By contrast, people who do prefer to be paid for their performance (likely higher performers on average) are more likely to join and stay with companies that more strongly link pay to performance via merit pay and other pay-for-performance programs.9 Meanwhile, at the state and municipal levels, public schools in Minnesota, Ohio, Denver, and Philadelphia led the way to merit pay for teachers.10 In Cincinnati, for example, teachers began to be held accountable for things they control: good professional practices. Teachers argue they should be held to standards similar to those for doctors: not a promise of a long healthy life but a promise that the highest professional standards will be followed. To assess teacher professional practices in Cincinnati, six evaluations were conducted over the school year, four by a trained teacher evaluator (essentially a trained teacher) and two by a building administrator. The size of pay increases was directly linked to performance during these observational reviews. However, in this case the plan did not survive.11 Perhaps because of these pioneering attempts, in 2006 Congress appropriated $99 million per year to school districts, charter schools, and states on a competitive basis to fund development and implementation of performance-related pay programs for principals and teachers.12 In another vein, the public sector is also experimenting with bonuses for better student test scores. Teachers who show improved student scores can receive up to $8,000 in annual bonuses in Chicago and up to $15,000 in Nashville.13 Be careful what you wish for, though! New York State teachers and principals recently were caught cheating on the grading process for regents' exams. The lure of bonuses totaling as much as $3,500 for teachers, and increased funding for principals' schools if pass rates improved, was too tempting.14 Of course, we could always design a system like that used in South Korea, a country whose students consistently outperform ours. Teachers in private schools often are paid on incentive. The dollar amount can be staggering if you're judged to be a good teacher. Kim Hi-Koon is an after-school tutor who makes 4 million per year because of his performance. He teaches three hours of lectures then spends 50 or so hours tutoring students online, developing lesson plans, and writing texts.15 If we want merit pay to live up to its potential, it needs to be managed better.16 This requires a complete overhaul of the way we allocate raises: improving the accuracy of performance ratings, allocating enough merit money to truly reward performance, and making sure the size of the merit increase differentiates across performance levels. To illustrate the latter point, consider the employee who works hard all year, earns a 5 percent increase as our guidelines above indicate, and compares herself with the average performer who coasts to a 3 percent increase. First we take out taxes on that extra 2 percent. Then we spread the raise out over 52 paychecks. It's only a slight exaggeration to suggest that the extra money won't pay for a good cup of coffee. Unless we make the reward difference larger for every increment in performance, many employees are going to say, "Why bother?"

PAY-FOR-PERFORMANCE: LONG-TERM INCENTIVE PLANS

All of the individual and group plans we have discussed thus far focus on short time horizons for performance and payouts. Usually the time horizon is a year or less. Now we shift to variable pay plans where the time horizon is longer than a year. Such programs force executives to think long term, and develop strategic plans that don't sacrifice tomorrow's riches for today's small gains. Exhibit 10.20 shows different types of long-term incentives and their definitions. These plans are also grouped by the level of risk faced by employees having these incentives, as well as the expected rewards that might come from them. Long-term incentives (LTIs) focus on performance beyond the one-year time line used as the cutoff for short-term incentive plans. Recent explosive growth in long-term plans appears to be spurred in part by a desire to motivate longer-term value creation.77 The empirical evidence that stock ownership by management leads to better corporate performance varies by study.78 There is some evidence, though, that stock ownership is likely to increase internal growth, rather than more rapid external diversification.79 All this talk about stock options neglects the biggest change in recent memory. As of June 2005 companies were required to report stock options as an expense.80 Prior to this date, generally accepted accounting rules didn't require options to be reported as an overhead cost. They were (wrongly) viewed as a free good under old accounting rules.81 Think about the executive issued 500,000 shares with a vesting period of five years (the shares can be bought in five years). After five years, the CEO can purchase the stock at the initial-offer price (if the market price is now lower than that, the stock option is said to be "underwater" and is not exercised).82 If the executive bought the shares, they were typically issued from a pool of unissued shares. The money paid by the CEO was treated like money paid by any investor . . . found money? Not really. Options diluted the per-share earnings because they increase the denominator applied to net profits used to figure per-share earnings. (OK, OK we promise, no more accounting terms!) Cases like Enron, which did not expense options, gave an unrealistic picture of profits and helped to elevate stock prices. The publicity from this case increased pressure to change accounting rules and led to the changes that began affecting most companies in 2006.83 Microsoft asserts that it will still continue to offer stock options to rank and file employees, but insiders admit that the modest movement in their stock prices have lessened the appeal of this vehicle for retaining top talent. No more instant multi-millionaires like the roaring 90s.84 As a direct result of the changing rules, some companies, like Coca Cola, Dell Inc., Aetna Inc., Pfizer Inc., McDonald's Corp., Time Warner Inc., ExxonMobil Corp., and Microsoft Corp either stopped granting options or only grant them to executives.85 Employee Stock Ownership Plans (ESOPs) - Some companies believe that employees can be linked to the success or failure of a company in yet another way—through employee stock ownership plans.86 At places like PepsiCo, Lincoln Electric, DuPont, Coca-Cola, and others, the goal is to increase employee involvement in the organization, and hopefully this will influence performance. Toward this end, employees own 28 percent of the stock at Lincoln Electric. At Worthington Industries, an oft-praised performer in the steel industry, employees augment wages by 40-100 percent between profit sharing and stock ownership.87 - Despite these high-profile adoptions, ESOPs don't make sense as an incentive. First, the effects are generally long-term. How I perform today won't have much of an impact on the stock price at the time I exercise my option.88 Nor does my working harder mean more for me. Indeed, we can't predict very well what makes stock prices rise and this is the central ingredient in the reward component of ESOPs. So if the performance measure is too complex to figure out, how can we control our own destiny? Sounds like ESOPs do poorly on two of the three Cs we mentioned earlier as causing incentive plans to fail. Why then do over 6,000 companies have ESOPs covering more than 14 million employees with assets of well over $1.3 trillion in the stock of their companies?89 The answer may well be that ESOPs foster employee willingness to participate in the decision-making process.90 And a company that takes advantage of that willingness can harness a considerable resource—the creative energy of its workforce. - If we just look at the impact of ESOPs on productivity or financial outcomes, leaving aside the positive effect on employee participation, the results are very modest. ESOPs have little impact on productivity or profit.91 Critics of ESOP argue that companies don't use these programs effectively. If more firms would combine ESOPs with high goal setting, improved employee communication with management, and greater participation in decision making by employees, maybe ESOPs would have more positive results.92 Performance Plans (Performance Share and Performance Unit) - Performance plans typically feature corporate performance objectives for a time three years in the future. They are driven by financial earnings or return measures, and they pay out for meeting or exceeding specific goals. Broad-Based Option Plans (BBOPs) - For the past 15 years broad-based option plans represented a growing trend. BBOPs are stock grants: The company gives employees shares of stock over a designated time period. The strength of BBOPs is their versatility. Depending on the way they are distributed to employees, they can either reinforce a strong emphasis on performance (performance culture) or inspire greater commitment and retention (ownership culture) of employees. There is growing evidence that the incentive effect of BBOPs is relatively small, and declines as the number of employees included grows. The smaller size of the effect in larger plans is explained by the free rider effect—"I can coast and get the same size benefit as others, and not get caught!"93 Some of the best-known companies in the country offer stock grants to employees at all levels: Southwest Airlines, Chase Manhattan, DuPont, General Mills, Procter & Gamble, PepsiCo, Merck, Eli Lilly, Kimberly-Clark, Microsoft, and Amazon.com.94 For example, Starbucks has a stock grant program called Beanstock, and all employees who work at least 500 hours per year, up to the level of vice president, are eligible (broad-based participation).95 If company performance goals are reached, all employees receive equal stock grants worth somewhere between 10 and 14 percent of their earnings. The grants vest 20 percent each year, and the option expires 10 years after the grant date. This program exists to send the clear signal that all employees, especially the two-thirds who are part-timers, are business partners. This effort to create a culture of ownership is viewed as the primary reason Starbucks has turnover that is only a fraction of the usually high turnover in the retail industry. - Colliding with this trend, though, is increasing shareholder pushback against equity awards for all but the top 1 percent of employees. The public simply doesn't believe that lower level employees can affect stock price with their performance! Bowing to this pressure, Coca-Cola is eliminating stock options for all but the top 1 percent. They will shift to cash awards for performance instead.96 Combination Plans: Mixing Individual and Group - It's not uncommon for companies to use both individual and group incentives. The goal is to both motivate individual behavior and to insure that employees work together, where needed, to promote team and corporate goals. These combination programs start with standard individual (e.g., performance appraisal, quantity of output) and group measures (e.g., profit, operating income). Variable pay level depends on how well individuals perform and how well the company (or division/strategic business unit) does on its macro (e.g., profit) measures. A typical plan might call for a 75-25 split. Seventy-five percent of the payout is based on how well the individual worker does, the other portion is dependent on corporate performance. An alternative might be a completely self-funding plan, often favored by CEOs who don't like to make payouts when the company loses money. These plans specify that payouts only occur after the company reaches a certain profit target. Then variable payouts for individual, team, and company performance are triggered.

DOES VARIABLE PAY (SHORT-TERM AND LONG-TERM INCENTIVES) IMPROVE PERFORMANCE RESULTS? THE GENERAL EVIDENCE

As the evidence pointed out in Chapter 9, variable pay-for-performance plans (short-term incentives and long-term incentives) seem to have a positive impact on performance if designed well. Notice that we have qualified our statement that variable-pay plans can be effective if they are designed well. Too often though the plans have too small a payout for the work expected, unattainable (or too easy) goals, outdated or inaccurate metrics, or even too many metrics making it hard to determine what is important.97 A quite different problem is that the payouts are quite large for high performance, but the behaviors taken to achieve those particular aspects of performance cause major problems. As such, incentive pay plans are sometimes described as high return, but high risk plans. When they go wrong, they go wrong in a big way. In the preceding sections, we have further addressed issues in design and the impacts they can have.