Gaming goals can kill businesses

The presiding deity over executive target setting and pay for performance is Loki, the god of mischief. Most performance management and incentive systems assume SMART and equitable target setting an accurate and honest measurement of the results actually achieved. But what if the targets could be achieved without actually making the effort the end objective actually demands? Any worthwhile objective is too complex for its achievement to be measurable by one or two quantitative measures but (for reasons of practicality) we end up doing precisely that. We thus set up a system that is open to gaming and, when linked to high variable compensation and other rewards, it becomes extremely tempting to do so. Let me pick an almost trivial example with which every reader of this column is familiar. Whenever we are asked to lecture about training evaluation, we trot out the Kirkpatrick model¹ or one of its more recent avatars. Yet, the number of training days per person (an eminently gameable statistic, which can be raised without imparting much worthwhile training) is still the most frequently used KPI for training managers in India. Obviously, at the level of the organization, both the temptation and the impact of manipulation are of a much higher order of magnitude. This is where gaming skills are at a premium and their use leads to great divergence between the lofty original purposes of organizations and what they actually land up doing – and for which their people are munificently rewarded.

Games people play - With goals

The seemingly simple path to goal-setting is strewn with many pitfalls – both conceptual and practical. In 'Goals Gone Wild: The Systematic Side Effects of Over-Prescribing Goal Setting', four leading academics conclude with a damning indictment of goal-setting: "For decades, scholars have prescribed goal setting as an all-purpose remedy for employee motivation. Rather than dispensing goal setting as a benign, over-the-counter treatment for students of management, experts need to conceptualize goal setting as a prescription-strength medication that requires careful dosing, consideration of harmful side effects, and close supervision. Given the sway of goal setting on intellectual pursuits in management, we call for a more self-critical and less self-congratulatory approach to the study of goal setting."²

Similar concerns are behind a piece titled 'Stop Paying Executives for Performance^'3 in HBR which emphasizes that:

• Contingent pay only works for routine tasks
• Fixating on performance can weaken it 
• Intrinsic motivation crowds out extrinsic motivation
• Contingent pay leads to cooking the books
• All measurement systems are flawed 

The problem is compounded when the same individual (or organization) has some goals that can be accurately measured and others that can’t. As John Roberts points out: "Providing comparably intense incentives for different activities becomes problematic... when the available measures of the two tasks differ greatly in their precision or timeliness… Giving incentives for some desirable activities can then be a very bad idea because these become negative incentives for other activities that cannot be similarly rewarded."⁴ Thus "goals in areas where quantification is difficult often go unspecified. The organization therefore often is in a position where it hopes for employee effort in the areas of team building, interpersonal relations, creativity, etc., but it formally rewards none of these."⁵

Real champions don't wait for targets to be set before gaming them. The target-setting process itself provides an enormous scope for making actual achievement far easier

Given these infirmities of goal-setting, it is not surprising that executives with high variable compensation linked to their target achievement are tempted to 'game' the system. Their efforts at manipulation are aided by the fact that accurate and timely measures are available for only part of the total desired performance (and that part becomes a decreasing proportion of the whole as roles – and firms – become more complex). In such 'synecdochic' (taking a part to stand for the whole) situations, Goodhart’s Law has free play. As initially framed by Goodhart, the law states: "Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes".⁶ Thus employees whose performance in a company is measured by some known quantitative measure, will attempt to maximize that measure regardless of whether or not this results in overall sub-optimal performance. As Bevan and Hood explain: "The extent of gaming can be expected to depend on a mixture of motive and opportunity. Variations in the motives of producers or service providers can be described in various ways, of which a well-known current one is LeGrand’s dichotomy of 'knights' and 'knaves' .… LeGrand argues that governance by targets can turn 'knights' into 'knaves' by rewarding those who produce the right numbers for target achievement, even if it means avoidance or evasion and neglect of [the real objectives]."⁷

As Harris and Tayler put it in a recent HBR: "Of course, we all know that metrics are inherently imperfect at some level. In business the intent behind metrics is usually to capture some underlying intangible goal – and they almost always fail to do this as well as we would like. Your performance management system is full of metrics that are flawed proxies for what you care about."⁸ Simply put, when the effort and cost of achieving a proxy target is significantly lower than for meeting the real organizational objective which the surrogate represents, rational actors ignore the main objective in favor of meeting or exceeding the indicative target. 

A universal malaise

Gaming surrogate indicators is neither a recent nor a peculiarly Indian phenomenon. It is generally believed (though now contested by some historians) that Roman generals could hope to receive the honor of a 'triumph' only if the number of enemies killed exceeded 5,000. The effect on the conduct of battles and on the reporting of casualties can well be imagined. Centuries later, when England imposed the window tax (in 1696), it "was not intended as a window tax, but as a property tax, as a house was considered a safe criterion of the value of a man’s property, and the windows were only assumed as the index of the value of houses"⁹ The response of citizens to evade the tax led to serious consequences. "… owners of dwellings attempted to reduce their tax bills by boarding up windows or by constructing houses with very few of them. In some dwellings, entire floors were windowless, leading to very serious and adverse health effects."¹⁰ Economists call this phenomenon 'the cobra effect' based on the British colonial government’s attempt to curb the number of cobras in Delhi. The government is said to have offered a bounty for every dead cobra. While this helped for a while, our smart countrymen soon began to breed reptiles for revenue. When the authorities realized this, the reward program was stopped, causing the cobra breeders to set the now worthless snakes free. As a result, the wild cobra population increased even more. An even more bizarre instance took place in Vietnam. When the colonial French administration wished to exterminate rats and offered people a bounty for each rat tail handed in, it led to rat tails being cut and submitted for the reward and the tail-less rats being used to breed more offspring.¹¹ A more recent example from the corporate world is the collapse of Fannie Mae which is significantly attributable to "Fannie Mae’s compensation arrangements [that] richly rewarded its executives for reporting higher earnings without requiring them to return the compensation if the earnings turned out to be misstated, thus providing incentives to inflate earnings."¹² 

Having established the ubiquity of the target manipulation malaise, perhaps we can still claim some pride of place for India by virtue of the early training we give our youngsters in aiming for the highest marks in examinations rather than for mastery of the discipline itself. With friends, teachers and guides telling children to study just for exam excellence, it is no wonder India is a favorite in the World Cup Championship for Corporate Gaming.

Real champions don’t wait for targets to be set before gaming them. The target-setting process itself provides an enormous scope for making actual achievement far easier. For instance, the almost ubiquitous use of measures which set targets relative to the previous year have a reverse-ratcheting effect. A year when performance is low, for whatever reason, makes future target achievements that much easier. The gaming consists in pushing past and future negatives into an already bad year and conserving efforts for the years when they convert to targets being exceeded significantly. Another variant is to take on targets that should have been the minimum requirement for that role and then getting rewarded for reaching that threshold. 

Of course, the bread and butter of target gaming is to find ways of minimizing effort and maximizing ratings once proxy targets have been established for critical objectives. When such primacy is given to production volumes, quality short-cuts, neglect of machine maintenance and excessive usage of consumables are among the many negative side-effects. Making sales volume both the kharif and rabi crop invites pests such as curtailment of advertising investment, sacrifice of margins, ignoring customer complaints and, of course, dumping. Dumping on dealers, distributors and retailers is a particularly hardy and noxious weed which has brought the fortunes of many a company low, yet which has reappeared, year after year, on corporate acreage ever since I started my career more than four decades ago. Even when production and sales are substituted by profit as the prime measure, the problem is merely elevated by one level. Now it is investment in research, new product introduction and upgradations of people and plant capabilities that are sacrificed at the altar of short-term profits. I shall not multiply illustrations but anyone who has ever worked in industry will recall goal gaming variations that were no less creative than the ones named by Beethoven after Diabelli.

Of course, target distortions don’t happen only at the individual level. Unions sometimes play a major role in the progressive 'flaccification' of targets (and making the engagement of permanent workers very unattractive in the process). A huge manufacturing unit that I was closely involved with became thoroughly uncompetitive over the years because the union had acquired consultation (read veto) rights over the establishment or change of the standard times ie the optimal time determined by Industrial Engineering for any given element of a work operation. Completing the task faster than the standard resulted in higher incentive payouts. By refusing to allow tightening of the standards when new equipment or automation was introduced, the union not only made it unattractive to invest in improvements but gradually drove capacity expansions to other units that were not plagued with such handicaps. Gaming has serious consequences.

Before my HR friends get too pious about these instances, some introspection will reveal that HR has at least two gaming innovations for everyone our line colleagues ever thought up. The core organizational objective most HR departments are expected to meet is to provide an adequate number of suitably skilled and committed people and to create a working environment that makes them productive, innovative and happy over a sustained period of time. When this goal is broken down into surrogate sub-targets like number of people recruited, cost per recruit and number of training programs held, all the issues (covered in the previous paragraph) of poor quality, short-termism and getting inordinate rewards while neglecting (or even liquidating) assets, rear their heads again. As for attrition, as a group HR head, whenever I looked at such a statistic for each business, I insisted it be broken down by rating and tenure. I wonder how well such a drill-down would reflect on the HR effectiveness of (for instance) our software majors which frequently claim credit for restraining or reducing employee turnover. But surely, I hear some naïve readers say, composite anonymous scores, such as those yielded by engagement surveys and 360º feedback, can’t be fudged. While researching this article, the most innovative ideas I heard from my senior HR friends were for improving engagement survey scores (without necessarily enhancing engagement). Those ways of gaming engagement scores will have to be held over for a future column. Or, even more lucratively, converted (after suitable euphemization) into a juicy consultancy offering!

Not by bread alone

Defeatist as it may sound, the first requirement for minimizing the manipulation of targets is to lessen the reliance on the variable compensation with which they are linked. We rely on target-linked monetary rewards to carry far too much of the heavy-duty carrying load for motivating people. Once a lucrative variable compensation scheme is in place, we tend to leave it on autopilot, inviting the same kind of malfunction as when faulty sensor measurements couldn’t be overridden and brought down two Boeing 737 Max aircraft. There are at least four other contenders for primacy of place over strong monetary incentives in unleashing motivation. They are:

• Organizational purpose
• Inspirational leadership
• Intrinsically exciting work and
• Multiple options to learn and grow

As we have noted earlier, the problem with powerful monetary incentives for individuals is greatly compounded when jobs consist of multiple tasks (as is the case for most corporate jobs, particularly at middle and senior levels), not all of which can be measured with the same degree of precision and timeliness. One can be more aggressive with introducing strong incentives for uni-goal roles, with minimal team dependence, such as some field sales jobs. Baron and Kreps¹³ provide a useful check-list for identifying where strong individual incentives should be eschewed. "In general, pay for performance is less likely to be effective:

• The more complex the technology
• The more ambiguous the tasks
• The more the culture emphasizes cooperation
• The more the strategy centers on hard-to-measure quality or emphasizes innovation
• The more tenuous the connection between inputs and outputs
• The more one can rely on the intrinsic motivation of the workforce 
• The more workforce diversity and/or technological diversity will encourage perceptions of inequity or illegitimacy in a pay-for-performance regime
• The more the general social culture, and the specific culture of the workforce, militate against 'crass monetary distinctions'." 

The same authors point out the additional dangers when there is a mechanistic linkage between target achievement and variable compensation payouts. "Difficult trade-offs arise when those being evaluated have the ability to manipulate or obscure the data on which performance evaluations are made. If the performance evaluation scheme is formulaic, the firm is inviting a great deal of attention to and (attempted) manipulation of the numbers on which the formula is based."¹³ Even at the cost of introducing an element of subjectivity into the process, such gaming can be minimized by delinking the hard connect between target achievement measurement and performance rating as well as between the rating and the quantum of bonus or merit increase. "When an employee's contribution to firm value is not objectively measurable, it often can be subjectively assessed by managers or supervisors who are well placed to observe the subtleties of the employee's behavior and opportunities. Even if such subjective assessments of an employee's contribution to firm value are imperfect, they may complement or improve on the available objective measures."¹⁴ Supervisor judgment (including an assessment of how far real organizational objectives have been attained and whether any long-term goals or values have been contravened along the way) must intervene at both points in the sequence of events. In judging whether target achievement is adding to organizational capabilities or liquidating them, supervisors should be aided by information from disaggregated engagement surveys, 360º feedback and other qualitative data that HR gathers through its employee interaction programmes. 

The core organizational objective most HR departments are expected to meet is to provide an adequate number of suitably skilled and committed people and to create a working environment that makes them productive, innovative and happy over a sustained period of time

Another way to reduce individual gaming is to link bonuses to collective performance though their traction in pulling individual motivation is doubtful given the lack of line-of-sight between effort and outcome and the 'free rider' type of problems in any but the smallest groups. All the same, company-wide rewards can be highly efficacious for non-economic reasons which build a sense of belongingness from bottom to top rather than just furbishing the fortunes of a few. "The economic arguments against rewards based on large-group performance take no account of the symbolic content of such reward systems, which can powerfully affect the extent to which the individual internalizes the welfare of the entire organization."¹⁵ I believe Telco (now Tata Motors) was the pioneer in Indian industry when, in 1980, it offered convertible bonds to ALL employees, gave them loans on easy terms to buy the bonds and ran a campaign to educate workers who had never owned shares before. While clearly not the only reason, this was certainly contributory to the feeling of belongingness and ownership that characterized the Telco workforce for years thereafter.

Difficult trade-offs arise when those being evaluated have the ability to manipulate or obscure the data on which performance evaluations are made

To revert to the issue of gaming prevention, nothing can substitute for the continuing vigil of the target setters, evaluators, HR business partners and checking agencies like Internal Audit. For the highest levels these roles have to be discharged almost entirely by the independent directors on the NRC. When they are incompetent, susceptible to optical wool coverings or complicit, we get scandals such as Satyam and IL&FS. 

McNamara's fallacy

All things considered, is the use of misconceived surrogate targets, that can be easily exceeded or fudged, such a serious matter? To take a non-corporate example again, many historians believe the magnitude of the US disaster in Vietnam was partly due to pursuing metrics that were not germane to what should have been its primary goal. As a result, the world’s most powerful superpower had to ignominiously retreat from its conflict with one of the puniest and poorest nations in the world. 

"When the last helicopter rose above the American embassy in Saigon on April 29, 1975, the US had been winning the Vietnam War for over a decade. The data said so. The strategy had been driven by a simple hypothesis, proven by history: Wars were won by inflicting damage on an enemy until they surrendered. The Pentagon set up metrics to measure that progress, the primary data point being kills (dead enemies), which was reviewed as an absolute number and expressed as a ratio against our own dead. The bigger ratio, the better the war was going, and Viet Cong casualties were generally 2x or more those of Americans dead.... The approach was led by Robert McNamara... It was also wrong..."¹⁶ "Preoccupied with searching for and destroying enemy formations, the Americans overlooked that much of the Vietcong's power derived from its political organization in rural villages and hamlets of Saigon... Reducing the Vietcong’s fighting power had not diminished their political influence within the local hamlets and villages. Killing the enemy was one thing. Defeating him politically was something altogether different.”¹⁷ 

Charles Goodhart would have smiled. Pursuing a misconceived proxy goal can be very costly indeed. Gaming corporate targets may not lead to as many deaths as the war in Vietnam did but it can be equally fatal to the health and success of any commercial enterprise.

Reference:

1. Serhat Kurt, Kirkpatrick Model: Four Levels of Learning Evaluation, in Educational Technology, 24 October 2016.

2. Lisa D. Ordóñez, Maurice E. Schweitzer, Adam D. Galinsky and Max H. Bazerman, Goals Gone Wild: The Systematic Side Effects of Over-Prescribing Goal Setting, Academy of Management Executive, 23 January 2009.

3. Dan Cable and Freek Vermeulen, Stop Paying Executives for Performance, Harvard Business Review, 23 February 2013

4. John Roberts, The Modern Firm: Organizational Design for Performance and Growth, Oxford University Press, 2004.

5. Steven Kerr, On the Folly of Rewarding A, While Hoping for B, The Academy of Management Executive; February 1995.

6. C. A. E Goodhart, Problems of Monetary Management: The U.K. Experience, Papers  in Monetary Economics, Reserve Bank of Australia, 1975.

7. Gwyn Bevan and Christopher Hood, What's Measured Is What Matters: Targets and Gaming in the English Public Health Care System, Public Administration, August 2006.

8. Michael Harris and Bill Tayler, Don’t Let Metrics Undermine Your Business, Harvard Business Review, September-October 2019.

9. House of Commons Debates, Motion for the Repeal of the Window Tax in Ireland, 5 May 1819, Vol. 40 cc 126–148.

10. Wallace E Oates and Robert M Schwab, The Window Tax: A Transparent Case of Excess Burden, Land Lines, April 2014.

11. Michael G Vann, Of Rats, Rice, and Race: The Great Hanoi Rat Massacre, an Episode in French Colonial History, French Colonial History 4(1):191-203, January 2003.

12. Lucian A Bebchuk and Jesse M Fried. Executive Compensation at Fannie Mae: A Case Study of Perverse Incentives, Nonperformance Pay, and Camouflage, Journal of Corporation Law, 2005.

13. David M Kreps, James N Baron, Strategic Human Resources: Frameworks for General Managers, Wiley Student Edition, 2009.

14. George Baker, Robert Gibbons and Kevin J Murphy, Subjective Performance Measures in Optimal Incentive Contracts, The Quarterly Journal of Economics, Vol. 109, No. 4, November 1994.

15. Artur Victoria, Group Incentives in Human Resources

16. Jonathan Salem Baskin, According To U.S. Big Data, We Won The Vietnam War, Forbes, 25 July, 2014.

17. Gregory A Daddis. No Sure Victory: Measuring U.S. Army Effectiveness and Progress in the Vietnam War, Oxford University Press, 2011.