3. EVA in Group-level controlling
This chapter discusses how EVA should be defined and used in Group-level controlling of operations. First, it is examined in detail how EVA should be defined in order to balance easiness, theoretical correctness and right steering. Thereafter the chapter deals with arguments for and characteristics of EVA bonus systems. Finally the chapter discusses things that are vital in implementing phase of EVA controlling.

3.1 A rational definition of EVA in business unit management
The most important reason for making EVA-concept simpler is to facilitate the learning process of operating people. There are plenty of adjustments that make EVA theoretically and/or practically a better measure or a better guideline in assessment of different units. The question is whether it is worth to do these adjustments or not. Every adjustment increases the complexity of the concept although some of them might be technically fairly easy to execute. When the organization has first adopted the basic concept well, it might be good to slightly modify the concept later on.

Avoiding additional costs in drawing the routine reports is also an important reason to simplify the concept. Stewart (1993, p.8) suggests plenty of adjustments to the basic residual income concept in order to avoid some accounting distortions. These adjustments include e.g. changes in depreciation schedule, inflation adjustments, capitalization of R&D and other strategic investments, currency translation etc. As Stewart admits, it is not wise to do all of these adjustments because of the marginal effects with some fields. Many of these adjustments cost something by increasing the workload in reporting. The problem whether to make some individual adjustment to EVA figures or not, can be approached e.g. by answering the following five questions: Will the operating managers understand the change? Will it influence their decisions? How big difference does it make with this company? Can the necessary data be obtained? How much does it cost?

There can also be other reasons to deviate from the theoretically correct way of calculating EVA than to only simplify the concept. For example, it might be for the SBU-managers difficult to realize that equity is costly capital. They might also have an approach that since they have earned the equity in their balance sheet, they also have the right to use it in their own investments. In this kind of situation, it might be useful - at first, in the early years of implementing EVA - to emphasize the cost of equity capital even with the ways that are not theoretically correct. An example of this kind of procedure is given later in the section 3.1.3. "Average cost of capital".

The degree of complexity in EVA can depend on the use of EVA and also on a business unit's accounting systems resources to make the adjustments needed. If a unit has a new and flexible information system which can easily make few adjustments to EVA then there is naturally no need not to do them. However, there are possibly a lot of units where the accounting systems are not very sophisticated nor very flexible and for those units there is a strong need to make EVA as simple as possible to prevent extra workload for internal reporting staff. The cost and benefits of information should be considered unit by unit or more accurately system by system.

We should perhaps start seeking the practical definition of EVA by examining the individual terms that EVA consists of. In Chapter 2.1.2 EVA was defined with formula 2 to be as:

EVA = (Rate of return-Cost of capital) x Capital =

(NOPAT/Capital - Cost of capital) x Capital

Following paragraphs seek to discuss what the individual terms might include in business practice.

3.1.1 Capital, NOPAT and Rate of return
Stewart (1990) defined capital to be total assets subtracted with non-interest bearing liabilities in the beginning of the period (year)9. Rate of return e.g. ROI is however typically calculated as return on average net assets, because it is a better estimate of the capital employed than the beginning capital (Telaranta 1997, p. 26). Although using average capital seems to be estimate of the capital employed, the method has also its weaknesses. Average assets include part of the return generated during the year. Yet, calculating rate of return should not include return in the capital side (in denominator) but only in the return side (numerator)10. That is because the people have used to understand and express return in relation to the initial investment and not in relation to investment's value in the end of the period. For example an investment in stock markets that was a year a ago 100 and is now 110 is said to have earned a return of 10% (10/100) and not a return of 9,1% (10/110). Simply excluding the profit from the ending balance sheet in calculating average assets can prevent this error. In practice the reporting happens at least once a month and so the average assets can be calculated as average of individual months instead of average from the beginning and ending balance sheet.

Net operating profit (NOP) is quite straightforward item. Of course it can and should be also adjusted according to the unique characteristics of the company in question, but normally there is no need to that. NOPAT is derived from NOP simply by subtracting calculated taxes from NOP: NOPAT = NOP x (1-Tax rate). These calculated taxes do not correspond the taxes actually paid because e.g. interest on debt decreases real taxes. The tax shield of debt is however taken into account with the capital costs.

Rate of return is NOPAT divided by capital, so both the definitions of capital and NOPAT affect rate of return. As stated in previous chapter, there are some problems in assessing rate of return with accounting book values. The rate of return might be somewhat distorted because of e.g. inflation and it could also be periodized wrongly. Using current value of assets instead of book values can radically reduce the distortions caused by inflation. This is still not necessarily a sound procedure because it is surely much more costly and difficult than using book values. Furthermore the benefits (the changes of EVA-figures into right direction) could be quite small especially if the company has a big proportion of current assets and the economic life of the fixed assets is relatively short.

The problem of wrong periodizing can be remedied with using different depreciation method in internal accounting. Normal straight-line depreciation will tend to underestimate the true internal rate of return in the early years and overestimate it in the later years. Using an economic depreciation schedule, known as the "sinking fund" method, will eliminate this distortion. Under sinking-fund depreciation, an asset is written off in the same way that a banker amortizes the principal on a mortgage. This means that in the early years most of the cash the asset generates is used to provide for the return on capital, and only a small fraction amortizes the capital balance. In the latter years it is the opposite. This schedule records little depreciation early on and more later on, but a steady rate of return and hence EVA is recorded over the life of the asset. Also all other depreciation methods that weight the depreciation more heavily into the later years will reduce the problem of wrong periodizing. However all these methods increase the subjectivity of EVA because they sort of bring some part of future profitability into present. Furthermore they are in most cases unnecessary because with steady capital spending program the periodizing problem is immaterial. (Stewart 1993, p.15-16)

3.1.2 Taxes in EVA-formula	
Although taxes are without excess depreciation, increase in reserves etc. about one-fifth of the net income and thus even bigger part of EVA, they are often totally ignored in EVA-control and reporting. E.g. Löyttyniemi (1996b) considers this as a sound approach. This approach can be justified because taxes are not a part of operative activities that should be measured and improved. If the pre-tax EVA is improved, then also the wealth of shareholders is improved. So including taxes in reporting does not change the situation in that sense. It only complicates the concept and calculations. Without taxes the reported income statement is simpler.

However, if taxes are totally ignored, then the minimum acceptable target can not be that pre-tax EVA=0. In order to achieve EVA=0, the pre-tax EVA should naturally be somewhat positive. This is a disadvantage, particularly if EVA is used in bonus systems. The bonuses should in that case be calculated based on the above target EVA. Especially if the bonus is paid for all employees, the zero target would be desirable. Major part of the employees can not comprehend EVA precisely, because they do not know the basic concepts of accounting and finance. On the other hand, it is sufficient for them to know that EVA is somehow (but consistently) calculated net result and if it is positive, it means bonus. A measure with target as zero is also psychologically and conceptually better than some other target level even though EVA would not be used in bonus systems at all.

Taxes in EVA formula according to theory

Normally in calculating EVA taxes are subtracted straight form Net operating profit and the tax shield of debt is taken into account in capital costs:

EVA =	net operating profit*(1-tax rate) - WACC*capital

This formula does not take into account that excess depreciation and reserves often decrease the amount of taxes paid in real life. At least taxes can be deferred into a distant future. With continuously growing operations the net reserves increase all the time. The practical tax rate is thus lower than the nominal tax rate.

Adapted way to calculate taxes

It is not very difficult to calculate a good estimation of taxes paid in the period. This can be done by simply subtracting the increase in reserves from Net operating profit before calculating taxes.

EVA =	[ Net operating profit - ((Net operating profit - excess dereciation - other increase in reserves)*(tax rate)) ] -WACC*capital

This way the taxes are calculated in the same way as tax authorities do it.

Other method that produces about the same result is to decrease the tax rate somewhat according to the estimated average impact of increased reserves. The tax rate could e.g. be 20% instead of the nominal 28%, if the estimated average impact of reserves would be something like that. This is naturally a method of simplification, but might produce a sufficiently accurate result. Furthermore it would perhaps decrease fluctuations from year to year in tax component.

When the taxes can be deferred into distant future, then considering inflation and the time value of money, it is not of great importance if they have to be paid ever or not. In this case the reserves can be viewed as equity and it is reasonable to deduct the increase in reserves from net operating profit before calculating taxes. However, in some occasions the taxes deferred with reserves have to be paid quite soon if e.g. tax regulation changes or if the company can not do new investments. Then the reserves can not be viewed as equity.

The connection between deferred taxes in income statement and in balance sheet

The question whether or not to subtract increase in reserves from operating profit when calculating taxes and the treatment of deferred taxes in balance sheet should be linked to each other. If reserves are noticed in calculating taxes, then the change in deferred taxes should be treated consistently as profit and the deferred taxes in the balance sheet should also be viewed as equity. This means that deferred taxes should be attached with capital costs. If the reserves are not noticed in calculating taxes then the deferred taxes in balance sheet should consistently be treated as non-interest bearing tax-debt. This means that they should be subtracted from capital as other non-interest bearing liabilities.

It is difficult or impossible to say in advance which of these two methods reflects better the actual situation. However the results of these two methods are not so far away from each other: the other has bigger NOPAT but also bigger capital costs, so the end result (EVA) does not be that much different. From the practical viewpoint: it is easier to ignore reserves in income statement and view deferred taxes as non-interest bearing debt.

3.1.3 Average cost of capital
As explained in chapter 2, the cost of capital is defined as weighted average cost of both equity and debt. The tax shield of debt is noticed with the cost of debt:

Cost of capital = Cost of Equity x (Solvency ratio) + Cost of debt x (1- Solvency ratio) x (1-tax rate)

There are, both in defining the cost of debt and the cost of equity, few different methods and also some variation in results. They are however mainly estimation problems and are of little interest in this context. In other words they do not have anything to do with simplifying the EVA-concept or making it a better functioning controlling tool. Some of these problems are discussed in chapter 4.

The calculation formula of average cost of capital (WACC) includes also the solvency ratio. The solvency ratio usually changes according to business cycles and other factors. Financial theory suggests (Copeland&Weston 1992, p.443-444) that when solvency changes the costs of the equity and debt shift so much that the WACC itself does not change (or it would not change without different tax treatment to debt and equity). When the solvency- or equity-to-debt -ratio decreases, the risk of equity increases. So when the relative proportion of debt from capital increases, the return on equity becomes more volatile and thus also the true cost of equity capital increases. Also the lenders demand higher premiums on debt when the leverage increases. So when solvency ratio decreases both the costs of equity and debt increase and visa versa. The increase in costs of equity and debt cancel out the decrease in WACC caused by bigger relative proportion of cheaper debt capital. Hence the change in WACC is zero (This is illustrated in figure 2 (Alternative 1) on the page after next page.)

The reason why average capital costs do not change according to leverage becomes more intuitive if we think of expected returns. Cost of capital (WACC) reflects the expected return on capital with similar risky businesses because it is an opportunity cost i.e. expected return on similar risky investments. If change in leverage does not affect the expected return of the company (expected ROI) then WACC can not change. Well it is obvious that expected ROI does change according to changes in solvency since solvency does not affect operating profit. Changing only the liabilities-side of the balance sheet, e.g. replacing equity with debt, does not affect the expected return on assets. The expected ROE in turn changes according to changes in leverage. Decreased solvency raises expected ROE because increased financial leverage raises return on equity capital (as well as risk of equity capital). Similarly the expected return on stock market does not depend on how the investors finance their investments. Of course for individual investor, the expected return changes if he uses more financial leverage (debt) with his investment. This affects however only return on his own capital (equity) but not the return for the whole investment. Changing leverage changes always the return and risk of equity and debt capital but it can not influence the underlying expected return of the whole investment. It merely allocates the risk and return in a new manner.

Practical performance reporting with EVA requires a certain procedure how WACC is calculated when solvency ratio changes. The following three examples demonstrate this problem through three different procedures to calculate WACC with different solvency ratios. Examples do not include the tax shield of debt in order to keep things simple.

Suppose that the cost of equity is 15% and the cost of debt is 5%. The target (and normal) solvency ratio of the company is 40%. How WACC can be calculated when solvency ratio is 30% and 50%?

1. Alternative: If we calculate WACC strictly according to financial theory, the costs of equity and debt have to be changed each time the solvency changes. This procedure might be too difficult in practical performance measurement.

2. Alternative: We can calculate WACC each time with the actual solvency ratio and with the same estimated costs of equity and debt. Then WACC changes always according to solvency ratio and thus the result is not in line with financial theory.

3. Alternative: We can calculate WACC each time with the target solvency ratio no matter what the actual solvency is. This procedure produces a result in line with financial theory and additionally it is quite simple.

The following figures will clarify these three procedures.

Figure 3 Alternative 1: WACC is calculated according to financial theory: costs of equity and debt change but WACC remains the same when solvency changes.

Figure 4 Alternative 2: WACC is calculated with actual solvency and fixed costs of debt and equity. WACC thus changes according to solvency. This procedure contradicts with financial theory.

Figure 5 Alternative 3: WACC is calculated with target solvency. WACC remains the same although solvency changes.

As presented earlier the alternatives 1 and 3 are in line with financial theory: WACC can not be decreased simply by replacing expensive equity capital with cheap debt capital, because solvency affects also the risk level of both equity and debt capital. Therefore alternative 2 (using actual solvency and fixed debt and equity costs) contradicts with the financial theory and seems to be out question. Alternative 1 (changing costs of equity and debt) is best in line with financial theory, but it requires that equity and debt costs should be scaled according to the prevailing solvency ratio. In practice it is therefore too complicated and time-consuming. Alternative 3 (WACC calculated with target solvency ratio) appears to be the best alternative since it is both simple and in accordance with the theory for essential parts. This method does not recognize that costs of equity and debt increase with leverage but on the other hand usually only the average cost of capital (WACC) is of importance. Academics and other experts like Stewart (1990, p.85-89), Löyttyniemi (1996) and Rappaport (1986, p.56) strongly recommend the use of target solvency in calculating WACC and EVA.

Optimal capital structure

The above examples ignored the different tax treatment of debt and equity and some other details. In reality the increased tax shield from debt decreases WACC somewhat when leverage increases. Therefore increasing leverage might decrease WACC slightly. On the other hand if leverage increases too much then the increased probability of bankruptcy and the costs attached to it increase WACC (Copeland&Weston 1992, p.498-499). These bankruptcy costs increase rapidly when solvency decreases from its already low level. Therefore low solvency levels are avoided although mathematically low solvency levels are as good as high since increased return should compensate the increased risk. Correspondingly bankruptcy costs increase at very moderate rate when the solvency decreases from high level. So it does not have very big difference with bankruptcy costs if company's solvency is 90% or 50%. However this would mean a some kind of change in WACC since it affects the tax shield from debt financing. Therefore very high solvency levels are avoided. Although no completely satisfactory financial theory has yet been found to explain the existence of optimal capital structure, casual empiricism suggests that firms behave as though it does exist (Copeland&Weston 1992, p.536). The changes in WACC are nevertheless quite small if the solvency changes moderately and near its optimal level. If e.g. the solvency ratio of an industrial company changes between 40% - 50% it probably has very small impacts on average cost of capital. That is because only changes in tax shield and changes in expected bankruptcy costs affect WACC and their effects are to the opposite directions. This optimal capital structure does change from one business field to another. E.g. real estate companies have on average very low solvency, normal industrial companies have moderate solvency and rapidly growing high-tech companies have high solvency. These different solvency levels reflect the differences in operational risk levels. E.g. real estate companies have very smooth operational cash flow (rents) so they tolerate more financial risk without too high bankruptcy costs. This should be remembered also with business unit controlling. The SBUs with low operational risk might have more financial leverage, lower solvency, than other SBUs.

The calculating of WACC (and EVA) in business controlling should take into account that there is some kind of optimal capital structure. Hence it is not desirable that the solvency ratios of SBUs would differ from it very radically. E.g. if solvency ratio is a SBU is very high, then the tax shield of debt is unused and shareholders will suffer. Furthermore high solvency ratio means besides low risk also low return on equity. However shareholders usually want high return from their investments and tolerate the higher risk level - otherwise they would have invested in bonds instead of in equity /stock market. SBU managers in turn prefer high solvency to low solvency because it is easily to operate with high solvency. High solvency enables the company to do the investments easily without asking for equity capital from parent company. Furthermore high solvency gives more discretion since operational cash flow does not go to fixed interest payments. The companies with high solvency are often referred as companies with "strong balance sheet" or as "healthy" companies. Already these expressions reveal that high solvency is a favorable thing.

If EVA reporting and bonus systems are based on fixed WACC (like example 3 on the previous pages) then EVA does not decrease no matter what the solvency ratio is. In this kind of situation SBU mangers will maintain high solvency and they are not willing to give up excess equity. Group management can of course always force the SBUs to decrease their solvency, but it is quite undesirable situation and leads to lengthy discussions of the "right" solvency level in business units. Better solution would be a controlling or incentive system that steers to optimal (target) solvency. This controlling system can not define WACC as it is in the reality. First of all it is very difficult or impossible to develop a formula that would define the right WACC at each solvency level and secondly this formula would be difficult to use and utmost difficult to communicate throughout organization. Therefore we have to resort to some kind of simplifying procedure. One possible solution would of course be to alter the alternative 3, which calculates WACC with target solvency ratio and fixed costs of equity and debt. WACC could e.g. be raised with 0,5 percentage points or more every time when solvency rises 5 percentage points from its target figure.

Another possible solution might be to turn back to the "incorrect" alternative 2 which calculates WACC with actual solvency and fixed cost of equity and debt. This procedure could be complemented with some sanctions if the solvency falls too low. This kind of system ensures that SBUs give up their excess equity and they would also understand better that equity is costly capital. This procedure includes however one major steering failure.

A steering failure in using actual solvency ratio

The weighted average cost of capital (WACC) should be a factor steering all the capital expenditures and investments. All the investments producing a return above WACC (NPV positive investments) should be executed and all the investments producing a return below WACC should be rejected. In practice this vital condition does not come true if the EVA control uses actual instead of target solvency ratio in calculating WACC. This is because the high solvency ratio enables the company to make quite big investments solely with debt capital. In order to increase EVA, these investments have only to produce a return more than the cost of capital.

EXAMPLE of steering failure when WACC is defined with fixed costs of equity and debt and with actual solvency ratio (alternative 3)

Let us assume that the cost of debt is 5%, the cost of equity 15%, solvency ratio 50% and the beginning capital 100 (equity 50, debt 50). So the current WACC is 10% = 50%*5% + 50%*15%11. Let the current return on investment be 11% (operating profit 11). Thus EVA is (ROI - WACC)*CAPITAL = (11%-10%)*100 = 1

The company faces an investment, which requires 25 of capital and offers a return of 6%. The current solvency ratio allows the whole investment to be financed with debt. If the investment were executed, the new capital base would be 125 (equity 50, debt 75). The new operating profit is 11 + 6%*25 = 12,5 and thus return on investment is 10%. WACC would in this new situation be somewhat lower: 0,4*15+0,6*5% = 9% (leverage would change and it would affect WACC). Thus the new EVA would be: (ROI - WACC)*CAPITAL = (10%-9%)*125 = 1,25

When using the actual solvency ratio, EVA might increase with investments producing less than WACC as the above example demonstrates. The increase in EVA is due to mixing operating and financing decisions. The capital resource affects EVA calculated with actual solvency. EVA is simply operating profit minus capital costs and if the investment is financed solely with debt, then the capital costs will only increase with the additional cost of debt. This pattern enables that EVA of a SBU and thus also the management bonuses might increase with accepting investment projects producing less than WACC. This holds only with short-term and with excess solvency. In the long run the company has to use debt and equity with target proportions and with already low solvency it can not solely stick to debt financing. However, the problem must not be underestimated because sometimes people tend to operate in short-term focus and SBUs face often the situation of excess solvency. Reserves and accumulated excess depreciation increase solvency although the net profit of a SBU was divided out as group contribution or dividends.

3.1.4 The essence of defining the capital costs accurately

Historically the ROI-targets are set for SBUs according to their current performance level. If the current performance is good then the ROI-target is also high and visa versa. The board of directors in parent company probably wants to include this kind of pattern also in EVA controlling. This method, defining capital costs according to current performance level and not according to the estimated opportunity cost of capital, is however against the principles of EVA. As presented earlier, the whole meaning with EVA is in establishing a capital cost based on risk-adjusted opportunity cost and that way assuring that the capital is in efficient use.

If the capital costs are set too low, it automatically allows the inefficiency of capital. With too high capital costs the SBUs will ignore some value creating investment opportunities (assuming that the incentive system is efficient). Assume that the cost of capital is set to be 16% and the true opportunity cost of capital is 13%. The SBU will ignore all the investment opportunities producing return between 13% and 16% although they all would improve the position of shareholders. The capital flows to parent company, which is unable to produce a return above 13%. Furthermore, normally most investment possibilities offer a return near the capital costs because we operate in a competitive world. So in this case investments which produce 14% are far more common than investments producing 17%.

With ROI-control the high capital costs are grounded with maintaining current high discipline. If the hurdle rate is decreased, this discipline and the current high profitability are likely to decline. This might be worse than ignoring some good investment opportunities. With EVA approach things are however different. Imposing capital costs well below current rate of return does not leave possibilities to decline current good profitability without decreasing EVA. That is because EVA is the absolute amount of capital the company generates above the capital cost. With EVA control the capital costs should be estimated as objectively as possible. If and when the company wants to set challenging targets they ought to introduce them as high EVA targets, not as high capital costs. There will always be business units with high profitability or with EVA figures biased upwards because of depreciated assets. The EVA targets for those business units should of course be far above zero.

Neither the distortions nor wrong periodizing of EVA should be taken into account in setting the capital costs. That is mainly because the cost of capital is used in estimating investment opportunities with NPV calculations and in this context there are no distortions. Hence modified WACC would cause harm with investment calculations. The distortions in performance measurement can be taken into account in EVA figures, if the company is able to estimate their impact. With steady capital spending program and common asset structure these distortions are usually immaterial.

3.2 EVA in Bonus systems
As discussed earlier EVA might be somewhat distorted because of inflation or periodized unevenly inside different years because of flat depreciation schedules. Furthermore it has been presented that these imperfections are exactly the same as problems with accounting rate of return (commonly ROI). If ROI were an accurate estimate of the true underlying return of an enterprise, then EVA would also be an accurate estimate of the excess return to shareholders in absolute terms. In the normal cases, i.e. with relatively stable investment schedule, normal asset structure and reasonable investment horizon ROI can be sufficiently accurate estimate of the true rate of return. Thereby also EVA is accurate enough in estimating the excess return to shareholders in absolute terms. If that is not the case, then ROI and EVA can be adjusted to sufficiently accurate measures with some modifications. Following paragraphs outline first what kind of bonus base EVA would be in the normal case (with no material errors). What the problems of ROI/EVA mean from the viewpoint of bonus systems will be discussed after that.

3.2.1 Arguments for using EVA in bonus systems
If EVA is zero, the shareholders have earned a sufficient rate of return on their capital. Many Finnish companies have earned negative EVA in the long run (Veranen&Junnila 1997). The shareholders of these companies had for sure been better of if the companies have earned positive EVA even though some part of it would have been paid out to company's managers or employees. The idea of EVA bonuses is that if management can be paid some bonuses, the shareholders have always earned higher return on their capital than they can expect. This kind of bonus system is usually beneficial both to management and the shareholders, because the performance level is likely to rise after introducing EVA bonus system (Wallace 1997). Motivating bonus system normally encourages managers to exceed the normal performance level and even after the payment of the management's bonuses, the return to shareholders is more than it would have been without the bonus system. With well designed bonus plan, the higher the bonuses that are paid, the better it is for the shareholders. EVA bonus paid is far from a cost to shareholders, because it is often a share in the discretionary value created.

Objective target level

With EVA bonus system the target performance level is very objective. Bonus can be paid e.g. according to some percentage of positive EVA or according to some percentage of improved (positive) EVA. Traditionally bonuses are often subjective, because they are based on the negotiated budgets. The managers negotiating their budgets in turn have incentive to sandbag i.e. to underestimate their potential performance level. That is because revealing the real potential would mean smaller bonus. With objective and unlimited bonus level, the SBU managers have an incentive to maximize performance and value instead of sandbagging their potential and wasting time and effort in managing earnings and the expectations of corporate office. The following citations describe the benefits of objective bonus system:

"Tying incentive compensation to EVA rather than to budget helped streamline the SPX budgeting and planning processes. "No more fussing around in the fall for months, messing around with a huge planning documents and worrying about sandbagging and things like that. It is gone," says Chuck Bowman, director of financial planning and analysis." (Kroll1997, p.109)

"...Instead of having budgets drive bonuses, the bonus system ought to drive the budgets."(Stewart 1990, p. 243)

Agency problems: spending free cash flow

EVA bonus systems are also good in decreasing agency problems. Management of a subsidiary wants usually to invest in their business as much as headquarters allows. Not many SBU management teams conclude voluntarily that they do not have any enough good investment projects and thus it is better to give the period's free cash flow out as dividends. Even if there would not be enough good investment projects, the subsidiaries would like to keep the excess capital in their balance sheet as liquid assets (or invest it in not-so-good projects). With powerful (enough motivating and rewarding) EVA-based bonus system the management is aimed to avoid this kind of behavior. That is because all capital producing a return less than WACC decreases their bonuses. If the incentives are tied to the change of EVA, excess capital in current assets or overinvestments in mature businesses can do a lot of harm to bonuses. Wallace (1997, p.15-16) presents strong empirical evidence showing that after introducing residual income based bonus system, managers avoid investments producing less than WACC. It applies also more generally, that because EVA measures the ultimate aim of any company, EVA-based bonus systems unite the interest of group management and shareholders or the interest of group and SBU managers.

Paying managers for performance with EVA-based bonus system

Private entrepreneurs, the managers of their own firm get paid just as they make money. Some successful entrepreneurs get rich and there is no set limit to their income level. Corporate managers can often make huge improvements to the wealth of shareholders since they have large amount of capital under management. However mangers are not paid accordingly i.e. with the line of the shareholder wealth increases. Even a little improvement in the capital efficiency might imply a big improvement in shareholder wealth in absolute measures. Some of this kind of big improvement should be paid out to managers in order to motivate managers to top achievements and in order to pay according to performance. In practice that would mean paying more than currently to good performers and less to bad performers.

A method to link the growth of productivity to payroll with line-workers

EVA might also be suitable to uniting the interests of the management/owners and ordinary employees. There has traditionally been an ever-lasting battle between employees and employers. It has led to the rise of strong trade unions and in some cases fruitless, frustrating and wealth-destroying strikes. The problem is that companies' profits are likely to increase due productivity growth and employees want always to get their share of the increased profit. The employees do not however exactly know what kind of share they could have. They also feel that they are always underpaid compared to the salaries of management and profits of shareholders. Therefore the demand for wage increases are often oversized. Economists tell that wage increases must not be over the growth of productivity (but that is hard to tell to employees, because even economists do not always agree on the definition of the productivity). So the raises often go over the growth of productivity and make profitability and capital efficiency too low from the viewpoint of owners, which in turn decrease employment. Hence the final sufferers are besides shareholders normally also the employees, no matter of employees' original intention.

EVA-based bonus system might be a way to pay employees according to the change in productivity. If part of the positive EVA is always handed over to employees they might be able to realize the connection between company's productivity (profitability) and their own payroll. In the last resort, the customers and the productivity pay the payroll and not the owners. EVA bonuses could also bring some elasticity in the payroll of workers. When the state of the market is good the employees get bonuses. When the state of economy is not so good there are no bonuses since there are no positive EVA. Good bonuses could prevent, if negotiated that way, some oversized wage increase demands. On the other hand, if wage increases go over the growth of productivity it is not possible to reach positive EVA and then there would be no bonus. In that sense the rise of the EVA-based bonuses follows quite well the increase in company's productivity.

With ordinary employees it might be difficult to tie the bonus plan to their own achievements because they can not contribute EVA materially or at least not in a measurable way. It is neither recommendable to tie the bonus to long run EVA, because it makes the link between company's profit and employees' payroll less visible.

3.2.2 Characteristics of feasible EVA-based bonus system
Noticing long run, EVA-bonus bank

The bonuses for corporate managers should always be tied to long run EVA because short term EVA can sometimes be manipulated upwards to the cost of long run EVA. The long run can be incorporated into EVA-based bonuses e.g. by "banking" the bonuses. This would mean that when EVA is good the managers earn a certain percentage (or other derivative) of it, but the bonus should not be paid out to them entirely. E.g. only one third of the bonus should be paid out to managers and the rest, two thirds, should be put in a bonus bank. In the following year managers earn again a certain bonus and this bonus is also put in the bank and then managers are paid one third of the bonuses in the bank. Each year the earned bonuses increase the balance in bank and managers are paid one third (or what ever the percentage is) of the accumulated bonuses. If the periodic EVA based bonus is negative, then the bonus put in the bank is negative and it decreases the balance already earned. This exposures the managers partly to the risk the shareholders are used to bear. At the same time it gives golden handcuffs to the good performers (with big positive balance in the bank) and encourages the bad performers (with negative balance in the bank) to leave the company. Stewart has presented the idea of bonus bank in his book (1991, p. 241)

There are of course some problems in calculating the bonuses in the long run when all of the key employees do not occupy the same post for many years. For managerial level this should be however done. Accumulated bonus from the current post has to follow a manager to the next post as long as it is inside the group. Retirement in the normal sequence should not affect the bonuses earned but other kind of leavings should erase the positive balance.

Consistency with bonus system

EVA-based bonuses should be consistent from year to year. If management in some SBU earns big bonuses along with outstanding results, the bonus system should not be altered in order to reduce these bonuses in the future. Of course fundamental errors with bonus systems can or should be corrected but big bonuses per se are not a sign of these kinds of errors. On the contrary big bonuses are a sign of well functioning bonus system which creates incremental return for shareholders.

Generous bonus from good performance

According to professors Michael J. Jensen from Harvard Business School and Kevin J. Murphy from University of Chicago the biggest problem with top management salaries is that managers are currently paid like bureaucrats rather than like value maximizing entrepreneurs (Jensen&Murphy 1990, p.1). They also state that traditional bonus systems produce far too small incentives for good performers and guarantee too big compensation for mediocre performers (1990, p.3). Corporate managers have often a lot of capital under their control. Because the stakes are so high, the potential increase in corporate performance and the potential gains to shareholders are great. The professors argued than even though the press often wonder the top management salaries (in the United States), those salaries are certainly not too big on average. Paying the top management in a more rational manner would eventually mean paying them according to achievements and with good performers that means paying them more than currently. (Jensen&Murphy1990, p.4).

If the shareholders want that the bonus system has desirable effects, the bonuses ought to be motivating. Positive EVA, if reached, can and should be truly rewarding meaning that top performers get big bonuses. If however a SBU operates already at positive EVA, the hurdle should be raised and bonus would follow only after the target EVA is exceeded. Other possibility is again to tie bonuses to the change of EVA. Of course no management team should be rewarded due to current positive EVA. That would not motivate and it would certainly be wasting shareholders' money.

The bonuses should not be capped, nor should they have diminishing marginal return. It would certainly not motivate the mangers to reach for stars if the bonus is limited to certain amount of money. Certainly the shareholders would not like to have their profits to be limited to certain level either. If the EVA goes to incredible figures, the bonuses should follow. Bonus system should neither have limits, nor would it be wise to make the bonuses raise with decreasing rate if certain EVA-target is exceeded, because it works in the same way with motivation. The bonus system should deviate from linear only if it has increasing marginal return: the bigger EVA the bigger bonus percentage. This kind of bonus system really motivates the managers to reach the stars. Furthermore it would be a good way to reduce the problem of ROI overestimating the true rate of return under inflation.

Changes of EVA more important than absolute values

As presented earlier changes in EVA tie more closely to share prices than absolute values. That is possible because the changes of EVA are not as likely to be subject to accounting distortions etc. as absolute values. Stewart (1993, p.13) suggests therefore that management rewards are tied to year-to-year changes in EVA instead of absolute values. A bonus system based on changes in EVA emphasizes the focus on continuous improvement.

Changes in EVA are in some occasions the only objective way to define bonuses. That is because in a company operating at positive EVA there is no sense paying bonuses based on that "already earned" EVA. Instead the bonuses can be based on year-to-year changes of EVA. If EVA is currently 100 and increases to 120, then the bonus base can be that incremental 20.

In order not to cut from the shareholders expected return, the rewards based on EVA changes should be paid only when the EVA is positive. If bonuses are paid according to changes of EVA in a situation where EVA is negative (but improving), the bonus system loses one of its essential characteristics. That is: EVA bonuses are never away from the expected return to shareholders.

Always tie to the current situation of the SBU

EVA-based bonuses should always be tied to the current situation of the SBU in question. Unit's business life cycle should affect the goals and thus also the bonus system of a SBU. Some unit might have mature line of business with strong positive periodic EVA and thus imminent danger of wasting the ample free cash flow via overinvestments in mature business. Other unit in turn might have plenty of profitable investment opportunities and good prospects of long term EVA although weak current EVA. The bonus system ought to be formulated so that it does not fight against strategic goals. Sometimes it might be even recommendable not to use any EVA-based bonus system. If EVA do not fit in SBU's current situation it should be left out.

3.2.3 The impacts of EVA's imperfections to bonus system
Accounting distortions from inflation and historical values

Because the true rate of return differs often somewhat from the accounting rate of return, also EVA can differ from the true Economic Value Added. This problem might often be insignificant and therefore ignored. This is the case especially if current assets make up considerable part of total assets or if the investment horizon of the company is relatively short. If however these distortions have material effects in EVA, there are at least a couple of ways to circumvent the problem. Firstly according the assets structure, inflation rate and investment horizon the extent of this problem can be estimated and hence the bonuses can be tied to the estimated target EVA, which corresponds the zero EVA in real terms. Second possibility is of course always to tie bonuses to the periodic changes in EVA instead of absolute values. The distortions in these periodic changes are so insignificant that they can be ignored. Third solution would be bonus system with increasing marginal compensation. Bonus percentage can be small at low levels of EVA but increasing when EVA increases. This kind of pattern decreases the effects of inflation biasing EVA upwards. Furthermore it has other appealing features like more motivation to top achievements.

The problem of wrong periodizing

If a company has depreciated almost all of its fixed assets, it might have - prior adjustments - big positive EVA even though the business would on average and in the long-run produce unsatisfactory true rate of return. In a similar fashion, if a company has a lot of undepreciated new assets in its balance sheet, it might show negative EVA even if the business would be quite profitable in the long run. Often businesses have steady growth and hence the above problems are luckily quite rare. If however EVA is unevenly periodized it has to be taken into account with bonus systems.

The problem of wrong periodizing can appear with different time horizons. The problem might be either chronic or temporary. If the problem is only temporary and will become even in couple of years, then merely emphasizing long run with the bonuses would solve problem. An example of chronic wrong periodizing would be e.g. old paper mill where the initial and massive investments in the factory are already depreciated totally. Another example is telecommunications operator that continues to invest in infrastructure and keeps making very small accounting profit or even loss in the near future. For both companies ordinary EVA bonus is clearly unsuitable and would not steer the operations correctly. The former (old paper mill) can however use EVA bonus system or actually that kind of bonus system is quite suitable for it. One solution for bonus system for this mill is to take only the change of EVA as bonus base. That way the managers of the mill have an incentive to drag out as much free cash flow as they possibly can. That is also the best way to operate from the viewpoint of the shareholders. The former company (teleoperator investing heavily) is probably unsuitable for any kind of EVA bonuses. Tying bonuses to changes in EVA would not work because EVA can be increased simply by decreasing long-term investments.

The problem of wrong periodizing in bonus systems can also be prevented by directly deferring some or all capital cost for some major investment. This is especially practical if the problem arises from one or two major new investments.

Although EVA has some imperfections they seldom outdo the benefits of EVA-based bonuses. Even in situations where EVA or bonus system is not adjusted to these imperfections the change in the approach and the behavior of management is great. If the SBU managers know how to operate in order to enhance shareholder value and they are also motivated to act accordingly because of good bonus systems, then some minor estimating errors with EVA figures do not matter. The most essential thing with EVA is however the fundamental change to adopt some kind of Shareholder value -approach. Empirical research (Wallace 1997) and plenty of examples (e.g. Gee 1997, p.7; Kroll 1997, p.109; Martin 1996, p.173) support the argument that adopting EVA or any Residual income based compensation plan benefits the shareholders.

3.2.4 Possible EVA-based bonus plans

This section seeks to summon up the discussion about the implications of EVA's imperfections to bonus systems. This is done by presenting some possible bonus patterns in different kind of companies.

Example 1

A typical industrial company has both new and old assets and is growing steadily. The company operates currently at small negative EVA (on average). EVA based bonus plan should be constructed so than it encourages reaching positive EVA and improving the performance continuously. Plan should also discourage making NPV negative investments.

Possible bonus system

Amount of bonus earned for each year: 	

Absolute EVA * Z% + Periodic change in EVA * 5 * Z%

The amount of bonus will be put in bonus bank every year. The bonus paid is 1/4 of the current balance in the bank. Change of EVA will increase bonus only if the EVA is positive. Improving negative EVA does not bring any bonus unless EVA increases above zero.

Example 2

An old business unit produces positive EVA all the time. The good profitability is however partly due to the fact that the company's fixed assets are mainly depreciated. Thus the capital costs are very small and the accounting rate of return overestimates the true rate of return.

And:

Example 3

A recently acquired business unit operates at a new business area. For some reason the current profitability is very good even though the unit has mainly new assets.

Possible bonus systems (for Examples 2 and 3)

Amount of bonus earned for each year: 	

Change in EVA * Y%

Or

Amount of bonus earned for each year: 	

(EVA - Target EVA) * Z% + Change in EVA * 5 * Z%

The amount of bonus will be put in bonus bank every year. The bonus paid is 1/4 of the current balance in the bank. "Target EVA" can be e.g. current EVA or current EVA plus something or some other EVA level determined by group managers.

Example 4

A new business unit operates at a new business area. In order to succeed in the future the unit has to make heavy investments. Furthermore the made investments are expected to produce big positive cash flows only in the coming years. Hence currently the unit operates at a very small net profit and at big negative EVA.

Possible bonus systems

EVA based bonus system is not suitable for this unit because of its growth phase and long investment horizon.

3.3 Implementing EVA control inside organization

Implementing EVA is and at least should be more than just adding one line in the monthly profit report. EVA affects the way capital is viewed and therefore it might be some kind of change in management's attitude. Of course this depends on how shareholder value focused the management and the company has been in the past. While implementing EVA represents some kind of change in organization, it should be implemented with care in order to achieve understanding and commitment.

Understand and tailor to your company

It is vital that group level managers gain first thorough understanding from the characteristics of the concept, how these characteristics affect controlling and above all, in what kind of situation the SBUs are currently from the viewpoint of these characteristics. Before implementing EVA to any SBU, the group management ought to assess whether the business units are currently cash flow generators in mature businesses or companies in rapidly growing businesses. This assessment should absolutely include careful estimation of the relative age and structure of the assets in order to know whether the current accounting rate of return is over- or underestimating the true rate of return. Only thereafter can the concept be properly tailored to the unique situation of each individual business unit. The group level managers ought also to know how to support the strategic goals of a SBU with EVA and how to create value with EVA in this individual SBU. According to John Shiely, the CEO of Briggs & Stratton Corp, "Adopting EVA simply as a performance measurement metric, in the absence of some ideas as to how you're going to create value, isn't going to get you anywhere." (Kroll 1997, p.109).

Gaining understanding and commitment at SBU level important

At the SBU level gaining understanding and commitment are also the most important issues. First task is to get the support of all the managers, not only of the managing director and treasurer but also of directors of production, marketing, sales etc. This is achieved with intense and thorough training. For managerial level attaining heavy commitment can be facilitated very much by introducing good incentive plan based on EVA.

Gaining commitment of the middle managers and other employees below the top management of a business unit is also important. Training and some kind of EVA-based compensation plans should also be considered with these target groups.

Other things to remember in implementing EVA

Keeping EVA simple is also viewed as an important feature in successful implementation (Gressle 1996). In principle EVA is a simple concept and like that it should be also offered to business units. In some cases it is even possible to simplify the current complex periodic profit reports by excluding some insignificant ratios. EVA summons up some important aspects of finance and value creation. Thus EVA might also give profound financial understanding to some operating people (sales, production) not familiar with these issues and confused about current great number of different financial measures.