CORPORATE GOVERNANCE, ETHICS, AND ORGANIZATIONAL ARCHITECTURE

by James A. Brickley, Clifford W. Smith, Jr., and Jerold L. Zimmerman, University of Rochester*

he recent highly publicized incidents of corporate wrongdoing have opened the floodgates of criticism and recrimination. Blame has been laid at the feet of everything from corporate greed to lax accounting oversight to Wall Streets insistence on ever-rising earnings to a general deterioration in ethics and standards. Corporate governance in general, and the pursuit of shareholder value in particular, have increasingly come under fire for what is perceived to be an excessively narrow focus on the stock price and its surrogate, accounting earnings, as the ultimate measure of corporate performance.1 But the practical implementation of any concept of corporate governance, whether aimed at value maximization or some other corporate objective, rests on the organizational underpinnings of the firm. In our view, the recent corporate scandals stem not so much from a general failure of corporate governance as from flaws in an important facet of corporate governance the organizational design of the firm. We use the term organizational architecture to refer to three key elements of organizational design: The assignment of decision-making authority who gets to make what decisions? Performance evaluationhow is the performance of business units and employees measured? Compensation structurehow are employees rewarded (or penalized) for meeting (or failing to meet) performance goals?

Successful companies assign decision-making authority in ways that effectively link that authority with the knowledge and experience needed to make good decisions. This linkage is bolstered by performance measurement systems that accurately and consistently gauge shareholder value creation, and is then reinforced by compensation systems that provide decision makers with the appropriate incentives to make value-increasing decisions. The three organizational elements are fundamentally interdependent. Indeed, they are like the three legs of a stoolthey must be designed jointly to ensure that the stool is balanced and functional. A weakness in any one leg can undermine a companys organizational architecture and cause the corporate governance system to fail. And because inadequate corporate governance can torpedo even the most brilliant business strategy, managers must structure internal systems that enable a company to achieve its value potential. Enron is a case in point. This sleepy, regulated natural gas company of the 1980s transformed itself in the 1990s by embracing a New Economy corporate culture with a flatter management structure, a dramatically reduced reliance on hard assets, and an entrepreneurial, risk-taking environment open to creative and unconventional products and practices. In August 2000, Enrons market cap reached a peak of nearly $70 billion, and the company was everybodys favorite success story. But then the
1. While corporate governance is not precisely defined here, this term generally refers to the roles of shareholders, managers, directors, and others as laid out by both the underlying legal system (through, for example, state incorporation laws and SEC regulations) and firm-specific mechanisms such as corporate charters, corporate bylaws, and internal operating policies.

*This article is drawn from our book Designing Organizations to Create Value: From Strategy to Structure (New York: McGraw-Hill, 2002). We acknowledge a great debt to Michael Jensen and the late William Meckling, whose pioneering research in organization theory kindled much of our own research. We thank Don Chew and Janice Willett for editorial assistance.

34
JOURNAL OF APPLIED CORPORATE FINANCE

wheels came off, ending in a bankruptcy filing, a near-total evaporation of market cap, and the draining of $1 billion of assets from employee retirement accounts. What went wrong? According to BusinessWeek, Enron didnt just fail because of improper accounting or alleged corruption at the top. . .The unrelenting emphasis on earnings growth and individual initiative, coupled with a shocking absence of the usual corporate checks and balances, tipped the culture from one that rewarded aggressive strategy to one that increasingly relied on unethical cornercutting. In the end, too much leeway was given to young, inexperienced managers without the necessary controls to minimize failures. This was a company that simply placed a lot of bad bets on businesses that werent so promising to begin with.2 In other words, Enrons problems were rooted in a fundamentally flawed organizational design. And each of the three architectural elements was at fault. First, in the course of flattening its management structure, Enron ended up delegating too much decision-making authority deeper into the company without retaining the appropriate degree of control at higher levels. Second, performance was evaluated largely on the basis of near-term earnings growth, which can distort managerial decision-making. Third, the company offered enormous compensation to its top performers, also on the basis of near-term earnings growth, which encouraged excessive risktaking as well as business decisions geared toward propping up earnings.3 Remarkably enough, Enrons own internal risk manual pointed to the core of the problem when it acknowledged that the rules and principles of accounting...do not always create measures consistent with underlying economics. Still, as the manual went on to say, Enrons risk management strategies are therefore directed at accounting rather than economic performance because corporate managements performance is generally measured by accounting income, not underlying economics.4 Although the internal risk management group was charged with reviewing business deals, the perfor2. At Enron, The Environment Was Ripe for Abuse, BusinessWeek, February 25, 2002. 3. Any system in which managers participate in annual profits but not in losses can encourage excessive risk-taking. This perverse incentive is most pronounced when a small bet fails and the employee tries to make it up by doubling the bet.

mance appraisals of the employees in that group were based in part on the recommendations of the very people generating the deals. And the legal staff was decentralized throughout the organization, where they were more vulnerable to pressures to meet their individual business units performance targets. As the Enron story suggests, there is clear potential for troubleand even breaches in ethical behavior when performance measurement and compensation systems are designed around accounting earnings, and a flatter, more decentralized management structure permits freewheeling decision-making. The critical question, however, is whether managers can reasonably be expected to identify these and other potential problems before they materialize, and to structure a productive, well-balanced, value-oriented organization. We believe that the answer to this question is a resounding yes. As described in this article, our approach to organizational architecture provides an integrated framework that can be applied on a consistent basis. Of course, no two companies are likely to adopt precisely the same structureand thats not surprising since the optimal organizational architecture will generally differ from company to company. Depending on its specific circumstances, top management should decentralize certain decisions and centralize others, tailoring the performance measurement system accordingly to detect whether employees are making value-adding decisions, and structuring its compensation system to reward them appropriately. And when a company adopts an incentive compensation plan, it is important that employees have the decision authority to respond to the new incentives and that their performance be measured accurately. As the celebrated architect Louis H. Sullivan, designer of the first skyscraper and founder of the American school of architecture, once observed, Form ever follows function. Applying this same principle to organizational architecture, we see that significant changes in the business environment and hence in a companys strategy will typically call for changes in decision authority, performance measurement, and compensation. Well-designed companies focus onand ensure consistency among all three design elements.
If this second bet also fails, the employee can have a strong incentive to double up again and go for broke. 4. These statements by Enron were presented in testimony by Professor Frank Partnoy, University of San Diego School of Law, in hearings before the U.S. Senate Committee on Government Affairs, January 24, 2002.

35
VOLUME 15 NUMBER 3 WINTER 2003

THE CORPORATE GOAL: SHAREHOLDER VALUE Since the restructurings of the 1980s, maximizing shareholder value has become the premier business mantra. Managers constantly profess their fundamental allegiance to shareholder value, typically measured by the stock price. When a company is rumored to be a possible takeover target, managements first response is to make clear to employees, communities, the media, and stock analysts that their top priority is to increase shareholder returns through value creation. And there is good reason for this development. With increased competition from both foreign and domestic producers, as well as changes in technology that have altered production processes and product demand, companies have had to refocus their attention on shareholder value to survive.5 A senior manager of a German beer company, reflecting on the current business environment in Germany, noted that it used to be about beer, beer, beernow its about shareholders.6 Even the Keidanren, an association of over 1,000 major Japanese companies, recently concluded that it is necessary for Japanese companies to place more importance on shareholder value and recommended the increased use of stock options to compensate managers as well as the addition of outside directors to Japanese boards.7 But what does it mean to maximize shareholder value? Shareholders invest in a companys stock because they expect to earn returns comparable to those available on other similar investments. Ultimately, shareholder returns are determined by the payouts they receive, whether in the form of dividends, stock repurchases, or realized capital gains. But the underlying source of such payouts, of course, is the cash generated by the businessthe operating cash flow that is left after all other bills have been paid, including taxes and debt interest. Therefore, as finance theory says, managers increase shareholder value when they increase the present value of the companys net cash flows.8 A firms stock price is a useful barometer of performance because it reflects not just the value the company has created
5. As we will discuss later, allocating corporate resources among multiple constituencies, such as employees and local communities, does not conflict with maximizing shareholder value. 6. The Wall Street Journal, June 21, 2001, p. A1. 7. See http://www.keidanren.or.jp/.

(or destroyed) historically but also the value of its expected future cash flow. Of course, future cash flows are not known for certain and must be forecast by investors, who look to earnings statements for information that is useful in developing these forecasts. There is obviously a strong correspondence in many companies between reported earnings and cash flows, and in fact managers who work to maximize reported earnings over time will also generally be working to maximize cash flows and share value. Nonetheless, accounting earnings are relevant only to the extent they provide information about a companys ongoing ability to generate cash flow. Window-dressing will at best be disregarded by astute investors and may even be seen as a signal that management is trying to conceal poor operating performance. Whats more, manipulating the timing or reporting of sales or expenses to artificially boost accounting earnings will actually decrease share value if it reduces the cash that can ultimately be paid out to shareholders. Boards of directors have a legal responsibility to make decisions on behalf of shareholders and to adopt performance measures and compensation plans that provide managers with cost-effective incentives to maximize current and future cash flows, or the companys economic bottom line. And as we discuss next, managers should concentrate on developing and executing a solid long-term business strategy, rather than slavishly focusing on accounting earnings. STRATEGY AND ARCHITECTURE Managers must continually seek new ways to create and capture value if their companies are to remain successful. Competitors who work to develop new and better products or production processes will eventually overwhelm any company that fails to innovate. In rapidly changing industries, new ways to create and capture value are likely to be critical. But even a long-time producer of a standard commodity will lose out to the competition if it fails to minimize production and distribution costs and to enhance customer benefits.
8. To be precise, the current value of a firms shares is determined by the expected cash flows that ultimately will be paid out to shareholders, discounted at a rate that reflects the returns that investors could earn on alternative investments with similar risk, liquidity, and tax consequences.

36
JOURNAL OF APPLIED CORPORATE FINANCE

As the Enron story suggests, there is clear potential for troubleeven breaches in ethical behaviorwhen performance measurement and compensation systems are designed around accounting earnings, and a flatter, more decentralized management structure permits freewheeling decision-making.

The central message of the core competency literature is that managers need to identify what their firms are good at and devise ways to leverage those competencies to create and capture value. For example, Sonys skills in electronics might be leveraged in any number of ways, from office equipment to toys. At the same time, managers should avoid investing in businesses with little potential for creating or capturing valuewitness the general lack of success of the unrelated diversifications of the 1970s. Similarly, Enrons demise is attributable in large part to the mistaken belief that its business model and its considerable skills in the energy business would translate into areas like water, wind, or bandwidth projects. It is also critical to consider competitors responses when making major strategic decisions. Sound strategy formulation often requires putting yourself behind your rivals desk. Kodak decided to invest heavily in producing writable CD-ROMs, which had been quite a profitable business for the company, but management failed to anticipate the threat of potential competition and ultimately lost money in that business. According to then-CEO George Fisher, I think we screwed up. We should have known that prices would fall as manufacturers worldwide ramped up production.9 But perhaps most important in any evaluation of strategy is the relationship between strategy and organizational architecture. If a key aspect of an industrys operating environment changes, managers in most companies in that industry will react by reappraising and modifying their strategiestoo often without considering whether the existing organizational architecture will remain effective. Before the rise in foreign competition, large American companies such as ITT, IBM, and General Motors enjoyed substantial market power and faced little external pressure to focus on rapid product development, high-quality production, or competitive pricing. Their organizations were extremely bureaucratic, with centralized decision-making and limited incentive compensation. Competition has forced these companies to rethink their basic strategies and to increase their emphasis on quality, customer service, cost control, and competitive pricing. In the process, decision authority has been pushed lower within the organizationto employees with more

detailed knowledge about customer preferences. This change has been accompanied by the introduction of performance measures tied to quality and customer service as well as an increased reliance on incentive compensation. Not only changes in the competitive environment but changes in technology can precipitate changes in strategy and architecture, too. Purchasing decisions at many retailers used to be relatively centralized, with buyers in New York City selecting a companys clothing lines for the year. But the introduction of satellite communications and closedcircuit television allowed central buyers to display goods to regional store managers, who could then stock their stores based on their knowledge of local fashions and tastes. With the new strategic emphasis on local competition, purchasing decisions became more decentralized. More generally, changes in technology, including email and intranets, have facilitated communication between senior management and lower-level employees, thus reducing the need for mid-level managers and allowing many companies to flatten their management structures. And strategy can in turn be influenced by organizational architecture. A company might decide to enter a new market in part because its decision and control systems are especially well suited for the new undertaking. Before the 1980s, Atlanta was widely acknowledged to be the banking center of the South. Yet at the beginning of the 21st century, Charlotte, North Carolina claims that title. Unlike Georgia, North Carolina had always permitted statewide branching, so when restrictions on interstate banking eased in the 1980s, the North Carolina banksespecially NCNB (now Bank of America) and Wachoviaexploited their experience in establishing and managing statewide systems to forge regional and then national banks. Their organizational architectures were better suited to the new regulatory environment. THE FIRST LEG: DECISION-MAKING AUTHORITY The principal challenge in organizational design is to ensure that decision makers have both the relevant information to make good decisions and the incentives to use their information productively.

9. The Democrat and Chronicle (Rochester, NY), Nov. 2, 1997.

37
VOLUME 15 NUMBER 3 WINTER 2003

The decision-making process generally comprises four steps:10 Initiation: Formulating and then choosing among potential decisions Ratification: Approval of the selected decision with possible modifications Implementation: Executing the selected decision Monitoring: Evaluating the outcome and rewarding the decision makers appropriately Initiation and implementation can be grouped into a category called decision management, while ratification and monitoring constitute decision control. A companys annual budgeting process encompasses all four steps and illustrates how decision authority is assigned within the company. Once managers are given a budget for their operations, they have decision authority over these resources (decision management). The performance of these managers is then judged relative to budget (decision control). As a general principle of corporate governance, decision management and decision control should be separated unless decision makers have a significant ownership stake in corporate cash flows. The most prominent example of this separation is the presence at the top of the corporation of a board of directors with fiduciary responsibility for ratifying and monitoring important decisions initiated by the CEO. 11 If the board of directors does a poor job, it can be replaced through a proxy fight or a corporate takeover. In the U.S. corporate governance system, it is takeover specialists, financial analysts, and large blockholders (such as public pension funds) who perform the role of monitoring the monitor. The principle of separating decision management and control also explains the use of hierarchies within organizations, permitting decisions by certain employees to be monitored and ratified by other employees who are above them in the hierarchy. The same employee might have both decision control and decision management functions, but not for the same decision. For example, division managers might have approval authority over certain initiatives of lower-level employees while at the same time requiring authorization for the divisions capital expenditure plan. In some
10. See E. Fama and M. Jensen, Separation of Ownership and Control, Journal of Law and Economics, Vol. 26 (1983), pp. 301-326.

cases, managers might pre-authorize decisions within a particular rangewhat is known as boundary settingwhile retaining monitoring authority. In smaller organizations, decision management and decision control are often combined. In these cases, the decision maker also tends to be an owner, which internalizes most incentive problemsalthough recent events at Martha Stewart Living Omnimedia, Inc. suggest that value destruction can occur even when principal decision makers have significant ownership stakes. A key issue in designing a companys organizational architecture is how far down within the organization to delegate decision authority. When operating managers have specialized knowledge about markets and processes, decentralization strengthens the link between decision-making and relevant knowledge within the company, and encourages the conversion of employee expertise into shareholder value. Decentralization will tend to add even more value as a company enters more diverse markets because senior managers are less likely to have all the information necessary to make good decisions across the various businesses. A principal drawback of decentralization, however, is that incentive conflicts increase as decision authority is pushed down into the organization to lower-level employees who may not see maximizing shareholder value as their primary goal. For example, managers may continue to operate unprofitable divisions rather than laying off colleagues and friends; and managers who are nearing retirement are less inclined to worry about cash flows that extend beyond their tenure. Fortunately, as we discuss next, the other two legs of the organizational architecture stool can help to control these incentive conflicts. THE SECOND LEG: PERFORMANCE EVALUATION There are at least two reasons to evaluate employee or business unit performance. The first is to provide feedback on whether the company is making the best use of its resources and to guide new resource allocation. For example, additional employee training in particular areas might be
11. Of course, the fact that the CEO has a significant ownership stake does not preclude the existence of a board of directors; yet when the CEO is essentially the sole owner, the firm will tend to operate without a board.

38
JOURNAL OF APPLIED CORPORATE FINANCE

Boards of directors have a legal responsibility to make decisions on behalf of shareholders and to adopt performance measures and compensation plans that provide managers with cost-effective incentives to maximize current and future cash flows, or the companys economic bottom line.

indicated. Employee reviews are also useful to managers who want to assign people to jobs that will make the best use of their skills. Business unit performance should be monitored for ways to increase productivity. Second, performance evaluation provides an indication of an employees or business units contribution to shareholder value, which in turn is useful in determining rewards and penalties for compensation purposes. But its important to keep in mind that these two purposes create different incentives. If performance were evaluated strictly for feedback purposes, employees would have little reason to distort their evaluations to make themselves look better. But when compensation is based on measured performance, employees and managers are more likely to try to find ways to inflate their own or their business units evaluations. Senior managements challenge is to recognize these dysfunctional incentives and to craft reasonably accurate performance measures that are consistent with the strategic objectives of the firmmeasures that help to motivate value-adding effort and decision-making. Evaluating Individual Employees We argued earlier that shareholder value is measured by the present value of all cash flows expected to accrue to the shareholders and that the firms stock price is a reliable barometer of such value. Ideally, then, we would assess an employees performance by his or her individual contribution to the change in stock price after each days work. But this of course is impossible. Shareholder value is determined by the collective actions of employees throughout the company. Even in relatively small companies, the impact of an individual employees efforts simply cannot be detected with any precision. And shareholder value is also subject to random factors from outside the company, such as conditions in the general economytax rates, the Federal Reserve rate, the level of unemployment, oil price levels, and so onthat are well beyond any individual employees control. Even when it appears that an employees contribution to value would be neatly reflected in the stock price, as when a scientist in the research lab discovers a new wonder drug and the stock price goes up $10, that employee also contributes to shareholder value in other ways that are not so readily measured, such as by guiding the work of fellow employees. 39

We therefore look to other measures as proxies for an employees actual contribution to value. These measures typically try to capture the employees productivity or output, as determined by the employees skill level and training and by the amount of effort he or she exerts. Different measures will correspond to varying degrees with the employees actual productivity. If output and quality are easily measured, piecework corresponds quite closely to performance, assuming that output is not affected by factors such as random machine failures and the quality and timely availability of parts and supplies. Other measures such as the number of hours worked will correspond somewhat less closely; for example, employees may be at their desks, but whether productive energy is being expended is another matter. Of course, evaluations based only on the quantifiable aspects of a job can cause the employee to emphasize those aspects to the detriment of other, less quantifiable aspects such as training fellow employees. Subjective performance appraisals can help to capture other dimensions of the employees output that are of value to the company. While no performance measure is perfect, they all provide some basis for assessing an employees contribution to value. In choosing among performance measures, managers must keep several things in mind. First, the costs incurred in evaluating an employees performance must not exceed the gains from doing so. The gains generally stem from the incentive effect of tying compensation to performancebut there are certainly administrative costs associated with using more precise measures, as well as costs related to employees trying to game the system. Second, because an employees productivity is subject to random factors beyond the employees control (weather, delivery schedules, raw material quality, and so on), greater reliance on incentive pay will introduce greater variability in the employees compensation. In these cases, it will make sense to design a (perhaps costlier) measurement system that will gauge employee performance more precisely so as to reduce the compensation impact of random factors. Thus, the degree of incentive pay and the accuracy with which performance is measured are interdependent. Because these two legs of the stool are complements, greater reliance on incentive compensation should be accompanied by an increase in the precision with which performance is measured.
WINTER 2003

VOLUME 15 NUMBER 3

Evaluating Operating Units: ROA versus Residual Income Business unit performance evaluation serves the same two basic purposes: feedback on how the company is doingwhich units are creating value and which are destroying valueand input to the reward system. What ultimately matters is a business units contribution to shareholder value, as determined principally by the present value of its net operating cash flows. But there are different measures of value creation. The most commonly used measure of performance is return on assets, or ROA. ROA is the accounting net income generated by the business unit divided by the total assets employed in that unit. It has intuitive appeal because it can be compared to competitors rates of return to provide a benchmark for a business units performance.12 However, ROA does not accurately measure the business units economic rate of return. For one thing, ROA is calculated on the basis of reported accounting numbers, with all their attendant problems (see, for instance, the article by Bennett Stewart in this issue). Accounting income (the numerator) is not always an accurate measure of economic profit, which represents the change in value over the period, and accounting assets (the denominator) do not necessarily reflect the current value of the divisions investment base. Perhaps the biggest problem with ROA, however, is that a division manager will be inclined to reject value-creating projects whose ROAs would lower the divisions overall ROA. For example, suppose the division has a current ROA of 20%, a 15% cost of capital, and the opportunity to take on a new investment project of similar risk with a 17% ROA. Accepting the project lowers the divisions overall ROA, even though it increases shareholder value and if the division manager is evaluated on the basis of maintaining or increasing division ROA, the manager will reject the project. To overcome this problem, some companies use residual income (or a variant such as economic profit or EVA) to evaluate performance. Residual income is calculated by subtracting the dollar cost of capital employed in the business unit from the
12. For companies with debt financing, it is important to add back interest expense net of taxes to accounting income before comparing ROA with external rates of return.

operating profits of the business unit. In our previous example, the new project will increase residual income (and shareholder value) because the project return is greater than the divisions cost of capital, even though overall division ROA will fall slightly. Nonetheless, both ROA and residual income have several common drawbacks, and any weaknesses must be addressed before these or any other performance metrics can be used as inputs to other systems, particularly the compensation system. First, management must estimate each divisions cost of capitalor even a project-specific cost of capital to control for risk differences (so that managers dont have incentives to plunge their divisions into risky projects to boost measured performance). These risk adjustments may cause division managers to lobby central management to reevaluate their risk and lower their required capital costs, thereby improving their reported performance. Whats more, both ROA and residual income measure performance over a single year only. As a result, they can be increased by cutting maintenance or R&D, but at the expense of future cash flows and hence shareholder value. Or a division manager might take on projects that increase ROA or residual income immediately, even if they are uneconomic in the long run.13 And neither ROA nor residual income reflects the interdependencies among divisionsone division might raise the quality of its products and thereby enhance the perceived quality of the companys brand name, but its divisional ROA or residual income will not capture the additional value created elsewhere in the company as a result. More fundamentally, both metrics rely on accounting measures of operating profits and assets. The internal accounting system plays an important role in organizational architecture, as we discuss in more detail in the box insert. But all accounting (and other) performance measures are prone to manipulation. And because accounting numbers measure performance over only a single period, they reinforce the horizon problem wherein managers emphasize short-term performance at the expense of future cash flows. Therefore, any accounting-based performance measurement system requires careful oversight by senior managers to control behavior by operating managers that is not consistent with maxi13. Of course, many companies have long-run performance plans such as Stern Stewarts bonus banks.

40
JOURNAL OF APPLIED CORPORATE FINANCE

If a key aspect of an industrys operating environment changes, managers in most companies in that industry will react by reappraising and modifying their strategiestoo often without considering whether the existing organizational architecture will remain effective.

The Accounting System and Performance Measurement Many people think of the accounting system in terms of the companys external financial reportsto shareholders, taxing authorities, regulators, and lenders. But these external financial reports (both quarterly and annual) aggregate an enormous amount of internal accounting data. Managers use internal accounting data on expenses, product costs, inventories, customer account balances, and so on for both decision management and decision control. Decision management (initiation and implementation) typically requires estimates of future costs and revenues, and accounting numbers provide a starting point in developing these estimates. Similarly, managers develop forecasts of costs and revenues for the next year in preparing their operating budgets. This process encourages managers to be forward-looking, to coordinate their operations with other managers who are directly affected by their decisions, and to share specialized knowledge of their markets and production technologies. Accounting-based budgets thus provide the framework for knowledge sharing and coordination. Accounting budgets also serve to hard-wire the process of converting employee expertise into shareholder value. For example, suppose a production employee discovers a faster way to set up the machines in the production process. This discovery is translated into a plant production policya new set-up algorithmand becomes part of the companys formula for value creation. Next years budget will be adjusted to incorporate fewer labor hours for machine set-ups. If the set14. S. McKinnon and W. Bruns, The Information Mosaic (Boston: Harvard Business School, 1992).

up efficiencies do not materialize as anticipated, the accounting system will report this in the form of unfavorable variances. Decision control (ratification and monitoring) relies to an even greater extent on accounting systems. In fact, accounting systems evolved primarily for this purpose. They protect against fraud, embezzlement, and theft of company assets. They also provide a scorecard for a business units past performance by measuring its costs, profits, or residual income. Monitoring is by definition a historical function and is well served by the accounting system. But because accounting systems are primarily designed for decision controlto prevent malfeasance and to measure past performancethey are based on historical costs and revenues and in this sense are backward-looking. As a result, they are often found wanting when it comes to providing managers with the information they need for decision management. In response to these deficiencies, operating managers develop their own, often nonfinancial, information systems to provide more of the data that they need for decision management. At the same time, of course, they rely on the output of the accounting system to monitor the managers who report to them. One survey confirmed that managers rely on nonfinancial data (labor counts, units of output, units in inventory, units scrapped) to run their day-to-day operations. But when asked about their most useful report in general, managers cited the monthly income or expense statements that are typically used to judge their own performance.14

mizing shareholder value. (The governance failure at Enron stemmed in part from senior managements failures with regard to oversight.) Most companies employ a single accounting system for multiple purposes, making adjustments as necessary to bring the numbers more into line with economic reality. Shareholder reports, taxes, internal decision management and control, regulatory compliance, debt agreements, and management compensation plans all use accounting-based numbers. Moreover, compa41

nies tend to use the same accounting procedures for all of these different purposes, which helps to control incentives to distort the numbers for any single purpose. And although managers have considerable discretion, particularly for internal purposes, the underlying accounting procedures are regulatedmanagers must choose among methods permitted by generally accepted accounting principles (GAAP). External, independent auditors can then attest to the accuracy and consistency of these accounting reports.
WINTER 2003

VOLUME 15 NUMBER 3

But no accounting system works perfectly; no system eliminates all opportunities for managers to increase their well-being at the expense of the shareholders. The key question is whether the system outperforms the next best alternative after all the costs and benefits are factored in. It is thus important to avoid the Nirvana fallacy, whereby a system is discarded or revamped if it allows any managerial opportunism, no matter how minor, in favor of an unachievable perfect system. By virtue of the simple fact that they have survived, most accounting systems would appear to be useful. Yet all accounting systems are vulnerable, partly because of play in the accounting rulesalthough it is prohibitively expensive to eliminate all managerial discretionand sometimes because of lapses in ethics, which we discuss later. And particular attention must be paid to the use of accounting numbers for compensation purposes, a topic to which we now turn. THE THIRD LEG: COMPENSATION AND INCENTIVE PAY The term incentive pay conjures up images of piece rates, commissions, and cash bonus plans, with the employee paid on the basis of some quantifiable measure of output. More recently, the emphasis has been on stock and stock option awards, although stock ownership is unlikely to provide very powerful incentives to rank-and-file employees in large organizations. Whatever its form, the fundamental purpose of incentive pay is to increase shareholder value by motivating valueadding effort. Rewards do not have to be monetary, but can consist of anything that the employees valuecorner offices, parking places, dining room privileges, and the like. The important question is how to create incentives in a cost-effective wayto design a compensation plan whose benefits outweigh any potential disadvantages. If an employees contribution to shareholder value could be measured precisely, it would be relatively easy to design a compensation program that would motivate the appropriate level of effort. Some of the inefficiencies that result from incentive conflicts can thus be reduced by improvements in performance measurement. This is where adjustments to accounting earnings can be important in motivating value-adding behavior. And although the standard indicator of an employees effort is 42

typically the employees direct output, there are many ways to determine whether the employee has worked hard. In determining the year-end bonus for a salesperson on commission, for example, the sales manager should look at overall sales in addition to individual results. If a particular salespersons performance in a given year was poor, but average sales in the company also declined substantially, it is likely that the salespersons results were simply affected by general market conditions. If other salespeople had great results, however, the salesperson may have slacked off. Similarly, corporate earnings targets can be set with respect to the growth of other companies in the same industry. Appropriate use of relative performance information increases the precision with which an employees or business units contribution to value can be measured and, when included in the compensation contract, increases the effectiveness of the incentive mechanism. But just as employees can be penalized for factors beyond their control, they can also be rewarded for results in which they played no direct part. Strong market conditions can amplify the results of mediocre effort and make normal effort appear spectacular. For example, many executives profited enormously from stock and stock options that increased in value simply as a result of the general bull market of the 1990s. In response, some companies have experimented with industry stock indexes as benchmarks for their stocks performance, with any increase in the firms own stock netted against the change in value of the industry index. Unfortunately, these plans can be administratively unmanageable and difficult to communicate to plan participants. Current accounting ruleswhich require companies to expense indexed options but not standard optionsalso reduce the perceived desirability of indexed options for most companies. Perhaps most important, appropriately designed incentive compensation strengthens the link between the companys organizational architecture and its business strategy. Once the firm has identified certain strategic objectives, those objectives should be factored into the design of the incentive plan just as they should play a role in the assignment of decision authority and the development of performance measures. If quality is an overriding objective, then incentive compensation might be based on measures such as customer satisfaction (based on surveys), the percentage of product returns, the

JOURNAL OF APPLIED CORPORATE FINANCE

Appropriately designed incentive compensation strengthens the link between the companys organizational architecture and its business strategy. In fact, if a company has settled on a strategic courseand has assigned decision authority and established performance measurement systems that will be consistent with its strategic objectivesincentive compensation should fall readily into place.

percentage of warranty claims, and the like. If cost control is a priority, the incentive plan might incorporate measures of cost reduction. In fact, if a company has settled on a strategic courseand has assigned decision authority and established performance measurement systems that will be consistent with its strategic objectivesincentive compensation should fall readily into place. Employees will already know what is expected of them, and the incentive plan will serve to reinforce their understanding and motivate them to work toward achieving the companys goals. Yet there is the danger, as in Enrons case, that incentive compensation works too wellthat it fosters unethical behavior. In the next section, we discuss how corporate ethics can be addressed through organizational architecture as well. ORGANIZATIONAL ARCHITECTURE AND ETHICS People generally have a pretty good idea of what is meant by ethical behavior. Most of us feel an emotional allegiance to the Golden Rule that urges us to treat others as we would have them treat us, and we value such qualities as honesty, integrity, fairness, and commitment to the task at hand. But some behaviors or activities are not so clear-cut witness the debates over affirmative action, animal testing, genetically engineered crops, stem cell research, and sweatshops, to name but a few. In these cases, there is simply no universally accepted code of ethics by which one can readily assess right and wrong. Whats more, behavior that might have been acceptable ten or twenty years ago may not be acceptable today because of changes brought about by movements as disparate as civil rights and womens rights, on the one hand, and corporate restructuring and stakeholder theory, on the other. As a result, ethics and corporate responsibility have increasingly come under scrutiny. U.S. corporations have responded by issuing formal codes of conduct, appointing ethics officers, and instituting training programs in ethics. A corporate code of ethics helps to eliminate uncertainty about ethical standards and how to live up to them; it informs employees that certain activities can damage the reputation of the company and will be severely penalized. At a minimum, a code of ethics will typically proscribe actions such as giving or taking extravagant gifts, bribing government offi43

cials, misrepresenting data, and discriminatory hiring practices. And although many of these acts are also illegal, legality alone is not always sufficient to frame policy. Using child labor in a textile mill may be legal in Pakistan, for example, but American or European customers might object because the practice is illegal in the United States and Europe. Business norms help to codify ethics, and yet these norms can vary in different countries. (A useful approach might be to draft a press release explaining a particular business practice and then consider the potential reaction not only internally but when it appears in The Wall Street Journal.) Of course, a more cynical view is that a corporate code of ethics serves merely to help the company defend itself against charges of illegal behavior. When an individual is found guilty of wrongdoing, his or her employer is also subject to federal penalties, but these penalties can be reduced as much as 50% simply by demonstrating that the employer has a compliance program in place that meets the U.S. Sentencing Commissions standards. (A compliance program minimally consists of a code of ethics and a training program.) In any case, adopting a code of ethics is not sufficient to guarantee ethical behavior. Intangible aspects of corporate culture such as codes of ethics and internal communications must be reinforced by more tangible structures if a company is to become both a value-based and a values-based organization. That is, the formal organizational systems that assign decision authority and measure and reward performance must all be internally consistent and designed to encourage value-adding, ethical behavior. In fact, the critical question is whether the incentives established by the current organizational architecture of the company truly encourage ethical behavior. Strategy, business ethics, and organizational architecture are linked, and it is important to structure the company so that employees work to implement the corporate vision with regard to both strategy and ethics. In other words, all three legs of the stool decision authority, performance measurement systems, and compensation systemsmust be designed to promote the corporate mission. One source of confusion is the concept of corporate social responsibility, which is sometimes used interchangeably with corporate ethics. In 1969, Ralph Nader and several other lawyers launched their Project on Corporate Responsibility with the following statement:
WINTER 2003

VOLUME 15 NUMBER 3

Today we announce an effort to develop a new kind of citizenship around an old kind of private governmentthe large corporation. It is an effort which rises from the shared concern of many citizens over the role of the corporation in American society and the uses of its complex powers. It is an effort which is dedicated toward developing a new constituency for the corporation that will harness these powers for the fulfillment of a broader spectrum of democratic values.15 As this statement suggests, the goal of some advocates of corporate social responsibility is nothing less than to change the purpose of the corporation. In Naders view, the corporation is to be transformed from a means of maximizing investor wealth to a vehicle for using private wealth to redress social ills. The corporate social responsibility movement seeks to make business managers responsible for upholding a broader spectrum of democratic values. Corporate support for such values could take the form of philanthropic activities, the provision of subsidized goods and services to certain segments of the community, or the expenditure of corporate resources on public projects such as education, environmental improvement, and neighborhood reclamation projects. Stakeholder theory, popular today, embodies many of these principles. In contrast, most economists and managers are inclined to endorse Milton Friedmans prescription that the social mission of the corporation is to make as much money for its owners as possible while conforming to the basic rules of society.16 In this viewembodied in value-based managementit is more efficient for the corporation to focus on creating wealth and to let shareholders, employees, and customers undertake their own charitable efforts. By maximizing shareholder value, corporations effectively enlarge the pool of individual (noncorporate) resources available for all stakeholders. But the contrast between the two views is not as pronounced as it might appear. Corporations that wish to maximize shareholder value generally find it in their interest to devote corporate resources to constituencies such as employees, customers, suppliers, and local communities. For example, a company with a large plant in an inner city might decide that investing corporate resources and personnel to
15. T. Donaldson and P. Werhane, Eds., Ethical Issues in Business: A Philosophical Approach, 6th ed. (Englewood Cliffs, NJ: Prentice-Hall, 1979), p. 90.

improve area schools would lead to better-trained job applicants, more productive employees, and thus lower-cost products. Giving money to the local university might benefit the company by improving its research and development, increasing its access to top graduates, or enhancing cultural and educational opportunities for its employees. Improving the environment might make it easier to attract and retain employees as well as lowering the companys legal exposure to environmental damage claims. Creating shareholder value involves allocating corporate resources to all constituencies that affect the process of shareholder value creation, but only to the point at which the benefits from such expenditures do not exceed their additional costs. If the corporation maximizes the size of the pie, each constituency including shareholders, bondholders, managers, employees, customers, suppliers, charities, and local communitiesreceives a larger slice. In short, ethical behavior can help to create shareholder value. And to the extent that the corporate ethics problem involves the issue of incentives, it can be resolved within the context of our organizational architecture framework. In many of the recent corporate scandals, ethical breaches arose from a lack of balance among the three legs of the stool. People may be honest in general, but the promptings of conscience and the desire to maintain a good reputation are neither universal nor constant, and they can certainly be influenced by the firms decision authority, performance measurement, and reward systems. For example, consider the transfer pricing problem faced by a corporation with multiple divisions that buy from and sell to one another. The efforts of division managers to increase their respective divisions profits may come at the expense of the other divisions profits and possibly also at the expense of firmwide profits. Top management might hope that instituting a code of ethics will encourage division managers to adopt a less provincial attitude and look beyond their own self-interest. But as long as the division managers are paid on the basis of the profits of their own divisions, they are unlikely to drastically alter their behavior. The firms organizational architecture will have to be redesigned to change the division managers incentives, perhaps by tying compensation to the overall value of the company as well as to divisional performance.
16. New York Times Magazine, Sept. 13, 1970.

44
JOURNAL OF APPLIED CORPORATE FINANCE

Intangible aspects of corporate culture such as codes of ethics and internal communications must be reinforced by more tangible structures if a company is to become both a value-based and a values-based organization.

If the compensation plan sometimes rewards unethical behavior, then unethical behavior is what the company will get, as we saw with Enron. Corporations develop ethics programs in an effort to persuade employees to put the interests of the organization or its customers ahead of their own. Executive leadership can be equally important in this effort. But internal consistency in the organizational architecture of the firm is crucial. CONCLUSION Effective leadership involves a great deal more than just developing an appropriate strategic or ethical vision for the companyit is also critical to motivate people to implement that vision. And while leadership is clearly important in the internal marketing of a strategic or ethical vision, it plays an even bigger role in recognizing which organizational architecture will best help fulfill that vision and then making the necessary alterations to the current design. Of course, part of good leadership is knowing when and how to work within the existing structure. Not all managers are empowered to make changes in the allocation of decision authority or in the performance measurement or compensation systems. And because too-frequent changes can discourage employees from making long-range decisions and developing effective relationships with colleagues, it is sometimes preferable to exercise leadership within the existing organizational architecture. In fact, before undertaking any architectural changes, managers should understand how their company arrived at its existing structure and, more generally, develop a broader perspective on why specific structures work well in particular settings. Outside consultants may argue that a companys long-standing organizational practices are inefficient
JAMES BRICKLEY is the Gleason Professor of Business Administration at the University of Rochesters William E. Simon Graduate School of Business. CLIFFORD SMITH is the Louise and Henry Epstein Professor of Business Administration at the University of Rochesters William E. Simon Graduate School of Business.

and propose replacing them with popular management techniques such as reengineering, total quality management (TQM), worker empowerment, the Balanced Scorecard, and so on. But most management fads address only one or at most two elements of organizational architecture. For example, reengineering focuses almost exclusively on decision authority. Total quality management ignores the reward system. Activity-based costing (ABC) changes only the performance evaluation systems. Management should resist jettisoning its current architecture without careful analysisparticularly if the business environment has been relatively stable. Uncritical experimentation with the organizational innovation du jour can leave the companys architecture out of balance. The focus should be on ensuring that the three elements of organizational design are synchronized with the companys strategy. The framework outlined in this article accepts peoples self-interest as given and rests on the principle that incentives work when the performance evaluation and reward systems are properly designed. Achieving the right balance among decision authority, performance evaluation, and compensation will ultimately drive shareholder value as well as ethical behavior. Whatever an organizations objectives, our approach to organization provides a solid foundation for creating value. Companies that are successful, particularly over the long run, seem to excel at creating and maintaining effective networks of employees and maximizing their potential. A key factor that these companies have in common is an effective organizational architecturea customized design that promotes and tracks the business strategy and rewards employees for making value-maximizing decisions. And when one digs for the reasons behind spectacular corporate failures such as Enrons, a seriously flawed organizational architecture is underneath the debris.
JEROLD ZIMMERMAN is the Ronald L. Bittner Professor of Business Administration at the University of Rochesters William E. Simon Graduate School of Business.

45
VOLUME 15 NUMBER 3 WINTER 2003