Hurdle Rates for Screening Capital Expenditure Proposals Author(s): Eugene F. Brigham Source: Financial Management, Vol. 4, No.

3 (Autumn, 1975), pp. 17-26 Published by: Blackwell Publishing on behalf of the Financial Management Association International Stable URL: http://www.jstor.org/stable/3665186 . Accessed: 30/06/2011 06:04
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=black. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Blackwell Publishing and Financial Management Association International are collaborating with JSTOR to digitize, preserve and extend access to Financial Management.

http://www.jstor.org

HURDLE CAPITAL

RATES FOR SCREENING EXPENDITURE PROPOSALS

EUGENEF. BRIGHAM

Dr. Brigham, Professor of Finance and Director of the Public Utility Research Center, University of Florida, is author and coauthor of a number of books and many articles in finance.

Several of my colleagues and I are in the process of writing a series of financial management cases to illustrate what might be called "standard textbook procedures." While finding firms to illustratemost aspects of finance was not difficult, we were initially unable to find even one firm that used different hurdle rates for different projects or adjusted them as often as we expected. This raised some ratherfundamentalquestions: Are there difficulties with the hurdle ratesrecommended in the academicliteraturethat renderthem impractical? Are businessmen missing an opportunity to improve their operations? Or, had we simply examined an unrepresentativeset of firms?This paperreportsthe results of an investigationof these questions.

Capital Budgeting in Academe

Although many basic issues in the theories of capital budgeting and cost of capital are still unresolved,most academicians would accept the following statements: (1) In general, discounted cash flow (DCF) techniques should be used to evaluatecapitalexpenditureproposals. Autumn 1975

Further, the net present value (NPV) method is the best of the DCF criteria.(2) The discount rate used to calculate a project's NPV should reflect the project's risk (however measured). A large, multi-divisional, multiproduct firm will undertake projects with differing degrees of risk. Thus, diversified firms should ordinarily use a number of different hurdle rates in their capital budgetingprocess. (3) A firm's capital budgetinghurdle rates should be based on its average cost of capital, which depends on (a) capital market conditions (reflected in part by the level of interest rates) and (b) the amount of capital the firm plans to raise during the budget period. Since capital market conditions are do volatile, and since firms' capital requirements change from time to time, firms would be expected to make relativelyfrequenthurdlerate revisions.

Obtaining the Data

As in earlierresearch[1, 3,4, 8, 9, 11, 12, 13, 14] on capital budgetingdecisions, a questionnairewas used in the presentstudy. Uniquely, however,we simultaneously constructedboth a questionnaireand in 45 pages a pair 17

of hypothetical cases that detailed how a firm might apply textbook principles (e.g., the capital asset pricing model) to the problem of setting hurdle rates. The case explained how we thought hurdle rates should be set, while the questionnaireaskedhow it was actually done. Next, we discussed the case with the financial staffs of severalcorporationsand used it in MBAclassesin five different universities.By special arrangement,financial officers of such firms as Armco Steel and 3M examined the case, participated in the MBA classes, and helped clarify points and make the case more realistic. Also, after each class the corporatepeople were asked to complete the questionnaire. We next used the case at Duke Universityin a 3-week executive programdevoted to capital budgeting. All the participants had a responsibility for capital budgeting, and at the conclusion of the course completed the questionnaires and discussed them with the author. Participants were from both industrial companies (14 persons), on which this study concentrates, and utility companies,which will be examined separatelyin a later report. To further broaden our coverage, the case and the questionnaire were sent to a group of executives who had recently completed a University of Illinois program and who were (1) known to be interestedin the problem addressedin the case and (2) probably willing to take the time to provide thoughtful answersto the questionnaire. Of 20 companies polled, 14 completed and returnedthe questionnaires. In total, data were obtained on the 33 firms listed in Appendix A. This is certainly not a randomsample-the companies are all quite large, and the fact that they participated in university programscould mean that they are more in sympathy with "academictechniques"than the averagefirm in their industry and size group. Howof ever, even if the sample firms are not representative at present, the data may well indicate the fuindustry ture directionof capitalbudgeting.

Exhibit 1. Screening Techniques Used in Capital Budgeting

Question: Does your company use a discounted cash flow (DCF) analysis for at least some prospectivecapital expenditures?Please check the method or methods that are used. Number 24 16 26 23 6 Percent 74% 48 78 70 18 94

Payback Accountingrate of return(ARR) Internalrate of return(IRR) Net presentvalue (NPV) Benefit/costratio, or profitability index (PI) EitherIRR, NPV, or PI Total companies

31 33

Use of DCF Techniques

As noted above, academiciansexpect firms to use DCF selection criteria,especially the NPV method, and Exhibit 1 indicates that the samplefirms overwhelmingly meet this expectation. Actually, the sample companies tend to use a number of different screening criteria. In many cases, cash flow data are punched on cards, fed into a computer, and severalcriteriaare generated as output. Severalinterestingpoints can be made on the basis of the interviews,the questionnairedata, and marginalcomments on the questionnaires.First, one firm uses only the payback; another firm uses only the payback and 18

the ARR; and the other 31 companies all use one or more of the DCF criteria. These 31 firms, if they also calculate the payback, use it (1) as a risk index for large projects or (2) to screen relatively smallreplacementinvestments. On this point, one company indicatedthat its procedures manual makes the following statement: "DCF calculations are not requiredto justify replacement investments with a cost of less than $1,000 provided (1) the payback is 3 years or less and (2) the economic life of the project is 6 years or longer."(This company's capital budgeting system is not computerized.) Second, six firms calculated an IRR but not a NPV; 4 firms calculated a NPV but no IRR; and 18 firms calculated both. The questionnaire did not ask for this information, but several respondents did comment that greater weight is given to NPV than to IRR; none of the 18 companies indicated that the reverse was true. Finally, six firms calculate the benefit/cost ratio or profitability index (PI). Four of the six also calculate both NPV and IRR; a fifth calculatesNPV; and the sixth uses only PI for project screening.Three of the companies that use PI with another DCF criterion were interviewed; these executives suggested that little weight would be given to PI should it conflict with NPV. On balance, our sample companies are using DCF to a somewhat greaterextent than earlierstudieswould indicate, and they give greaterweight to NPV as opposed to IRR. This could be becausethe samplesare different, or it could simply reflect the fact that our data are newer, and firms are moving towardDCF in generaland NPV in particular.

Use of Multiple Hurdle Rates

Academic literaturesuggeststhat risk-adjusted hurdle rates should be used to screenprojects,so, except in the Financial Management

as unlikely event that all projects are equally risky, firms the cash flow projectionsare regarded being relatively should use more than one hurdle rate. To test for use of risky. Thus, implicitly, but in an unspecified manner, differentialhurdle rates are being applied.Second, intermultiple rates, we asked the question given in Exhibit 2. Of the 31 firms that use DCF and therefore require views established the fact that financial people engaged hurdle rates, almost half use one rate to evaluate all in project evaluation want very much to increase the projects throughout the corporation. The remainder weight given to quantitativerisk analysis-they want to categorize projects in some manner and use different measure risk in some manner and to incorporateit exhurdlerates for different investmentcategories. plicitly into the decision process. Exhibit 2. Use of Multiple Hurdle Rates in Capital Budgeting
Question: Some companiesestablishone hurdle rate and use it throughout the corporation. Others estimate different rates for different units (e.g., subsidiaries, divisions,or product lines), or use differentrates for different types of projects (e.g., replacement, expansion of old product lines, expansion into new products, overseas vs. domestic). Still other companies use different costs of capital for different individual projects depending on projects' riskiness or other characteristics.Which statement or statements are applicableto your company?

How Are Hurdle Rates Calculated?

We calculateonly one hurdle rate and use it to screenall projectsthroughoutthe corporation. We calculatedifferenthurdlerates for different organizational units of the company: 1. Subsidiaries 2. Divisions 3. Productlines 4. Domestic vs. overseas

Academicians generally agree that a firm's hurdle rates should be based on its cost of capital. There is controversyover the measurementof the cost of capital, but there is agreement that a risk-adjustedweighted averagecost of capital should be used to screenproposed investments. As Exhibit 3 shows, 10% of the 31 firms that use a DCF capital budgetingcriterion and thus require a hurdle or discount rate use their historical rate of return on investment;61%use a hurdlerate based on Number Percent* the weighted average cost of capital; and 29% checked the "other" category and provided a separateexplanation of their procedures.
15 48%

Exhibit 3. Basis for EstablishingHurdle Rates

Question: How do you establish your basic hurdle rate or rates?

5 3 3 5 14**

16% 10 Ourhurdlerates are based on historic 10 rates of returnon investment,e.g., 16 assets. operatingearnings/operating 45%**Pleaseexplain. Ourhurdlerates are based on our cost of capital,e.g., weightedaverageof our cost of debt and equity. Please 35% explain. Other (pleasespecify):

Number Percent

10%

We use differenthurdlerates for different types of investments(e.g., replacement; expansion;old product lines; expansion: new products). We attempt to evaluatethe riskinessof individualprojects (as opposed to units of company or types of investmentsas describedabove) and vary hurdle rates dependingupon projectrisk.

11

19 9 31

61 29 100%

7 31

*These percentagesrelate to the 31 firms that use DCF and thereforerequirehurdlerates. **The numbersand percentagesdo not add up because some units. firms use different hurdlerates for severalorganizational

Severalcomments are appropriate. First, althoughthe firms are not using multiple hurdle rates as exsample tensively as theory suggests, every firm interviewed noted that judgment plays an important role in capital budgeting. If management is confident of a project's projected cash flows, the project may be accepted even though it fails to meet the DCF criterion;conversely, a project that does meet the criterion may be rejected if
Autumn 1975

Academiciansregardthe use of historical rates of return to screen projects as being most inappropriateand not in the interest of maximizing equity value. Comments on several of the questionnairesprovide clues as to why historical returns are calculated and why they might be used as screeningrates. First, ratesof returnon investment are widely used to evaluate divisionalmanagers-firms recognize that, other things held constant, a high rate of returnon investmentis desirable.Further, if a successful firm earns, say, 15%after taxes on assets, and if it accepts no project with an expected return of less than 15%, then it will maintain or improve its historical return. However, the firm is clearly not maximizing stockholderwealth if the historicalrate of return happensto be different from the currentcost of capital.
19

The companies that use a cost of capital screening rate provided some interestinginsights into practicalattempts to measure the cost of capital. First, 29 of 31 companies use balance sheet figures (book weights) to calculate the weighted averagecost of capital.It was not always clear if the weights representedthe actual book value figures at the time the cost of capital was calculated or a target capital structure. At least some firms use a target as opposed to the actual capital structure. Second, the companies all use the after-tax cost of new debt. Most concentrate on long-term debt, but several use an average of long and short-termcosts. Further, executives have as much trouble estimatingthe cost of equity as academiciansdo, and many of the questionnaire respondentsand intervieweeswere quite candidin admittingthat their estimates rely heavily on judgment. As to quantitative measures, most use the D/P + g formulation, with g being estimated on the basis of past earnings growth. Two companies specifically indicated their use of the capital asset pricing model (CAPM)approach, although both indicated that they also use D/P + g. Finally, a number of the companies calculate an averagecost of capital, then increaseit for use as a screeningrate on profitable investmentsin order to offset zero (or negative) returns on environmentaland other non-earninginvestments.In other words, if a company estimates its averagecost of capitalto be 12%,but it must allocate 20%of its capitalbudget to non-revenue producing projects, then it might use 15%as its screening rate (12%/.8 = 15%). Although this procedure seems reasonable at first glance, it could present problems.What is happeningis similarto the sunk cost problem.There are two types of environmentalexpenditures: (1) those associated with new plants, where the analysis presumablyalready includes pollution control costs, and (2) those associated with bringing existing plants up to standard. For the first type, a regularcost of capitalhurdlerate should be used. For the second type, the correct analysisinvolves comparingthe cost of the environmentalinvestment to the entire profit contribution of the plant. The existing plant cost is, for purposes of this analysis, a sunk cost. If a company can raise capital at a cost of 12%, it should take all investmentsyielding anythingmore than 12%, even if it cannot earn sufficiently more than 12% to cover zero or negative return projects. Of course, if the firm thinks that plants currentlybeing built will require subsequent pollution abatement expenditures, it can use an inflated hurdle rate to screen projects,but it would be better to estimate pollution costs and build them into the analysis. It is impossible to generalizeabout the "other" category in Exhibit 3 except to state that these companies, while using a DCF acceptancecriterion,arenot attempting to screen on the basis of a cost of capital.Generally,

the companies made such statements as: "Our hurdle rate is set duringour strategicplanningprocess.It varies among operating units depending on economic conditions that affect specific product lines. Basically,we are concerned with what is possible and in getting rates of return that will meet our corporategrowth and profitability targets." We attempted without much success to pin down such statements in the interviews. While the intervieweesgenerallystated that a good deal of thought goes into the hurdle rate specification, those in the "other" category all agreed that the final rate is somewhat arbitrary, that it is not basedon the firms'cost and of capital.

How Often Are Hurdle Rates Changed?

Hurdle rates should reflect both capital market conditions and the firm's internal situation, including the amount of capitalit plans to raise.Given that conditions change fairly often, responses to the question in Exhibit 4 shows that firmsrevisehurdleratesless frequently than one might expect. Based on the interviewees' comments and notes accompanyingthe questionnaires, most of the companies that reviseless than once a year do not have a system for reviewingtheir hurdle rates. The "depends on conditions" companies stated that they revise rates to reflect product and capital market conditions, with revisions generally occurringless than once a year. Exhibit 4. Frequency of HurdleRate Revisions
Question: How often do you changeyour hurdlerate? Number Percent Annually Morethan once a year (please specify) Less than once a year (pleasespecify) Dependson conditions (pleasespecify) 4 5 12 10 31 13% 16 39 32 100%

Do Hurdle Rates Reflect the Firms' Volumeof Financing?

Because of transactionscosts, market imperfections, taxes, and other factors, most academiciansthink the marginalcost of capital schedule (MCC)facing a firm is upward sloping, at least beyond some volume of financing.(See Exhibit 5.) In theory (and in at least some firms), the budgeting process goes something like this: (1) The treasurerestimates the MCCschedule. (2) The

20

Financial Management

director of capital budgeting obtains from the operating divisions the approximate dollar volume of acceptable projects at different hurdle rates (e.g., $5 million at a 20% screening rate, $6 million at a 19%rate, etc.). (3) The projected, or ex ante, MCC and investment opportunity schedules (IOS) are plotted on a graph such as that shown in Exhibit 5. (4) The percentage interest rate (k) at which the two schedules intersect is used as the hurdle rate;if the ex post MCCand IOS schedules are reasonably close to the ex ante schedules, the use of this hurdle rate will produce an optimal capital budget. If the investment opportunity schedule is anywhere to the left of IOS2, then shifts in IOS, and consequently in the amount of funds raised (Q), have no effect on the MCC; hence, such shifts should not be reflected in the firm's hurdle rate. However, if the IOS is to the right of IOS2, then year-to-year shifts will affect the MCCand should be reflected in the hurdle rate. Exhibit 5. Relationship Between Hurdle Rates, Investment Opportunities, and the MarginalCost of Capital Schedule
Interest Rate(%)

fossil plants, it will need only $1 billion. In analyzingthe alternatives,the company has been using the same discount rate. Nuclear appears to have an advantage(the present value of total capital and operating costs are lower for nuclear), but the treasureris very concerned about the effect raisingthis volume of capital will have on his cost of capital.Discussionwith managementindicates that the treasurer can estimate a MCC curve in which he is reasonablyconfident, and that the nuclear/ fossil analysis will be rerun with higher discount rates applied to the nuclear plant. (The company is also considering the relative riskiness of the nuclear and fossil plants, and may also use risk-adjustedrates in the analysis, but at this point the risk differentialsare not clear. It is interesting, though, that the company is thinking seriously of shifting from a single to multiple hurdlerates.) Exhibit 6. Responsivenessof Hurdle Rates to Amount of CapitalRaised
Question: Does your hurdle rate reflect the amount of capital you plan to raise, e.g., if preliminary forecastsindicate a need to raise a relatively large amount of capital vis-a-vispast years, would you tend to use a higherrate to screenprojects? Number Percent Ourhurdlerate does reflect the amount of capitalwe plan to raise (or that is available). Ourhurdlerate does not reflect the amount of capitalwe plan to raise (or that is available).

17

55%

k1

14 31

45

100%

Q3

Q2

Q1

$ Raised Budget

During Period

Weights Givento Quantitative Analysis by Project Type

Two important points began to emerge in our early interviews. First, different types of projects have systematicallydifferent risk characteristics, and, second, these differences affect the relative weight given to formal quantitative analysis versus "judgment." A project's riskiness can be separatedinto two components-uncertainty about how much it will cost to buy or build a particularfacility and make it operative,and uncertainty about the cash flows the project will generate over its expected life. For many routine replacementsand small cost-reducinginvestments, both the initial cost and the cash flow stream can be estimated with a relativelyhigh degree of accuracy. On the other hand, building a new plant that utilizes a new technology to produce a new product will involve a great deal of uncertainty with regard to both the initial outlay and the annual cash flows. Executives recognize these differences, and they 21

MCC = Marginal IOS = Investment

Cost

of

Capital Schedule

Opportunity

Exhibit 6 shows that over half the firms consider financial needs when setting hurdle rates. The 45% that do not could, of course, simply be in the MCCrangeto the left of IOS2, but it was evident from the interviews that some of the companieshave simply never thought in a formal way of the relationshipbetween funds raised and capital budgetinghurdle rates. One particularlydramatic example is a utility company with assets of about $1.5 billion (1975) consideringthe constructionof nuclear or fossil generatingplants. If it "goes nuclear,"it will have to spend an additional $2 billionby 1985;if it builds Autumn 1975

account for them in two ways. First, projects are classified by type (replacement,expansion, etc.), and different screening procedures are employed for different types of projects. Second, where the cash flow data are relatively reliable, a great deal of weight is given to the DCF calculations,and where the cash flow data are uncertain, more weight is given to "judgment."In other words, if management is confident of the cash flow data, a project with a positive NPV or with IRR greater than MCCwill be accepted without much ado, but if it does not have such confidence, projects with positive NPV's may be rejected and negative ones accepted because of "corporate strategy" or other nonquantified reasons. To obtain information on the firms' use of investment project classifications, and also on the extent to which they rely on quantitative analysis as opposed to judgment, we asked the question given in Exhibit 7. Twenty-one of the 33 executives were able to answer

this question; the other 12 either indicated that they simply did not know how the final decisions were reached, or that their companiesdo not classify projects in the mannerindicatedin the question. The exhibit can best be understoodif it is first consideredin a somewhat artificialway. Suppose a firm makes 100 investmentsof equal size (say $1,000) in each of the 4 categoriesshown in columns 1 to 4. Focusingon column 1, we see that on average27 of the 100 replacement and modernization projects are accepted primarilyon the basis of quantitative data, e.g., NPV > 0, IRR > cost of capital, or payback < minimum acceptable payback; 36 projects are decided largely on the basis of the quantitativedata; in 33 cases most weight is given to "judgment"and the decision maker's "gut feel"; while 4 projects are decided almost entirely on the basis of judgment and gut feel. Moving to column 2, expanded production of existing product lines, quantitative analysis plays an even more important role: 85%of the projects in this category are

Exhibit 7. WeightsGivento QuantitativeAnalysisby ProjectType

Question: Most companies consider qualitative and quantitative factors when making accept-rejectdecisions on capitalprojects.However,these two types of factors are given different emphasisin different companies,and a given companymay use different weights for different types of expenditures.For example,a high weight might be given to NPV, IRR, or payback for plant modernization that decisions,but less weight for new productdecisions. Recognizing judgment is involved in answeringthis question, please indicate the approximate weight your company places on qualitativeversus quantitativeanalysisin each of the followingcategories. Approximate weightgiven to formal economicanalysis in the acceptrejectdecision of Percentage $ amountof capital budgetin each category or Replacement modernization (1) 75-100%of weight placedon formal economicanalysis 50-75%of weight placedon formal economicanalysis 25-50%of weight placedon formal economicanalysis 0-25%of weight placedon formal economicanalysis TOTAL Other Total Expansionof Expansion into new existing capital capital products expenditures* budget products (2) (3) (4) (5)

27%

35%

29%

28%

36

50

47

44

33

14

22

36

20

4 100%

1 100%

2 100%

55 100%

8 100%

*Includedhere might be pollution control expendituresand other non-incomeproducing and projects,R&Dexpenditures, the like.

22

Financial Management

decided primarily on the basis of formal economic analysis. Somewhat less weight is given to quantitative factors when new products are being considered,while in the "other" categoryrelativelylittle weight is given to the screeningcriteria-judgmentis the dominant factor. The total capital budget is a weighted averageof these 4 categories, and column 5 indicates that, overall, about 70% of the averagecompany's capital expenditures are based primarilyon quantitativeanalysis. The logic behind these findingsrelatesboth to quality of data and to cost/benefit considerations.Replacement and modernizationdecisionsinvolve relativelygood data, making quantitativeanalysisfeasible, but these decisions are often repetitive, involve numerous small projects, and are frequently too obvious to warrant refined analysis. These two forces offset one another, and the results are reflected in column 1 of the exhibit. There is some indication that category 1, "replacementor modernization," should have been divided into "replacement: cost reduction"and "replacement: maintenance." Had this separationbeen made, it is likely that relatively heavy weight would be given to quantitativeanalysisfor cost-reducing investments, but little weight to maintenance investments,which are often too obvious to warrant detailed analysis. The data inputs are generallyof lower quality for expansion investments,but the projects are larger,making formal analysismore worthwhile.So, on balance, formal analysis is more important for expansion than for replacement decisions. Investments in the "other" category (pollution, R&D, new home office building)generally produce no direct revenues upon which to base a NPV, IRR, or payback,or are so fraughtwith uncertainty as to make a DCF type analysisunrealistic.If reliable data are available,these projects can be decided on the basis of choosing the system that will do the job with the lowest present value of future costs; this is the DCF criteriaon which utility companiesbase their investment analysis. (See Brighamand Pettway [3] .)

Exhibit 8. Trends in CapitalBudgetingMethodology

Question: Various researchershave observed a trend toward greater quantification of the financial decision process. Has your company tended to give greateremphasis to quantitative versus qualitative factors in recent years? If so, is this trend likely to continue? Yes In general,we rely more heavily on quantitativeanalysis(in the financialarea) than we did five to ten years ago. DCF project selection criteria are given more weight than was true five to ten years ago. Is your companylikely to move in the following directions: Towardheavieruse of DCF? Towardmore quantification of projectrisk (e.g., simulation, sensitivityanalysis, etc.)? Towardgreateruse of multiple hurdle rates? Towardmore frequentrevisions of hurdle rates? 73 27 No

92%

8%

88

12

96 50 50

4 50 50

What Trends in Capital Budgeting Methodologydo the Companies Anticipate ?

To obtain the companies' views on future project selection procedures,we asked the questions givenin Exhibit 8. Consistent with trends shown in earlier studies, companies today rely more heavily on quantitative analysis in general, and DCF in particular,than in the past, and they expect this trend to continue. Most companies that do not expect to use DCF more heavily in the future generally indicated that they are already screeningmost projectsby DCF criteria. More strikingis the interest in quantifying risk: 96% of the questionnairerespondentsexpect their companies Autumn 1975

to move in this direction, and in interviewsexecutives continually stressed (1) that they do recognize risk differences among projects and (2) that they want to measure these differences in some manner and incorporate them into the decision process in a non-arbitrary manner. Based on these reactions, one might anticipate a major movement toward formal risk analysis in capital budgeting,at least for largerprojects. Although the questionnairedid not ask for forecasts of the analytical methods firms think will be used to quantify risk in the future, comments both on the questionnairesand in interviewsdid providesome information on this point. First, several respondents mentioned the possibility of using Hertz-typesimulation [7]. In a recent article, in the Winter1974 issue of Financial Management, Hastie [6] concluded that Hertz-type analysis is generally not operationally useful, and that "its use will probably continue to decline." Hastie strongly advocated a second approach,sensitivityanalysis, which involvesestimatinga project'sprofit sensitivity to changesin key input variables.Nothing in our survey could be interpreted as either confirming or refuting Hastie's conclusions.However,based on discussionswith respondents,it seems likely that a third approach-"optimistic, pessimistic, most likely" analysis-which lies 23

somewherebetween Hertz's full scale probabilisticsimulation and sensitivityanalysis,may well be the dominant method of risk analysisin the future. (Some preliminary ideas on this issue are presented in Barry,Brigham,and Crum [2].) Regardlessof how risk is quantified, half of the respondents feel that their companies will use multiple hurdle rates more extensively in the future, and half also expect to make more frequentrate revisions.In the case of multiple hurdle rates, this low percentage is somewhat inconsistentwith the movement toward quantifying risk.If companiesdo develop project risk indices, they will probably use this informationin setting hurdle rates. Thus, we can expect an increasingemphasis on quantifying risk to be accompaniedby an increaseduse of multiple hurdle rates.

Respondents'Reactions to the Study

A draft version of the paper was sent to each questionnairerespondent,along with a requestfor comments. The commentsvariedconsiderably.The "multiplehurdle rate" companies seemed generally to agree with the thrust of the paper, but the "single hurdle rate" companies gave some very cogent argumentsagainst more quantitativerisk analysis in general and againstmultiple hurdle rates in particular.The following excerpts are offered partly for the insights they offer and partly to offset any bias this papermight exhibit: 1. "Most respondents(including this one) indicated that in the future they expect their firms to place heavier reliance on quantificationof risk in their evaluationprocedures. Assuming most were individuals like myself, they may be primarilyindicating their hopes or goals, with continuing acceptance of new analytical methods by operatingpersonnel dependent on how good an educationaljob they do." 2. 'The whole financial analysis programof a business is to support, aid, and assist managementin exercising its business judgment. The analytical techniques used by those of us who work in this area must be understandableand meaningful to those people who have the authority to make and bear the responsibility for operatingdecisions. "Most of our important capital expenditureshave a time cycle of about five years. Generally,we can predict quite accurately how much capital funds will be required to accomplish a given program.Our degree of accuracy goes down when we estimate how much the finished product will cost at different points in time (even ignoring inflation). The cash proceeds are even more difficult to estimate because they are even more beyond our control. How many will the marketabsorb, and at what price? What will be the reaction of direct 24

competition? What about indirect competition or other technology that may come into the marketplace? "We believe that decision makers should be given as much usable informationas possible regarding risks,but it may sometimes be more harmful than good to push quantificationtoo far. We make an effort to explain all important assumptions,and to provide "what if" calculations where risksappearto be significant." 3. "Like many efforts in life, the desire to achieveis often not sufficient to overcome a basic lack of ability. By lack of ability, I refernot to lack of an understanding of the methods but to a lack of confidence in the quality of the necessary inputs. It is my experience (and confirmed by a good many of my peers) that it is damned tough to obtain reliable data for a single determination of return, let alone obtain a set of data for probabilistic risk analysis. To do the latter requires the capital budgeter to be a highly skilled psychologist and interviewer as well as an extremely competent teacher. Few of us possess these skills. The marketingman, engineer, purchasing agent, etc., must be taught and convinced that a method produces improvedresults before he will accept it and participateeffectively in its use." 4. "As to establishingvarious hurdle rates, I submit two arguments on why this is not a more widespread practice. First, we'd like to do it, but franklywe do not know how to meaningfullyapply known cost of capital methods to each businessin our mix. All the approaches require approximations. Like allocating corporate expenses, these approximations create controversy and, therefore, cast heavy doubts on the parts of the various groupexecutives. "Second, I'm not convinced that in an inflationary, volatile world we know how to definitivelymeasurecost of capital. Compoundingthe problem of an imprecisely measured cost of capital, with a wide range, is the fact that project returnshave largestandarderrors.This situation tends to minimize the need (1) to establishdifferent hurdle rates for different projects and (2) to alter hurdle ratesmore frequentlythan once a year." As indicated above, these are some of the more negative (but perhapsrealistic)statements.Otherrespondents are moving toward more risk quantificationand toward multiplehurdle rates.

Conclusions
As noted at the outset, the primarypurpose of this study is to determine to what extent businesses'capital budgeting practices are consistent with the academic literature, and if major inconsistencies appear, to suggest modifications in either what businessmen do or academiciansteach, or both. Based on a series of interviews, and a questionnaire survey of 33 firms, we reachedseveralconclusions. Financial Management

First, in general,the DCF methodology recommended in the academicliteratureappearsto be useful in a practical sense, and sophisticatedtechniques for quantifying risk analysis are also gaining ground. Business firms do have trouble obtaining the data to implement some academic procedures,but the task seems not to be impossible, and at least some firms are approachingthe "textbook case." Second, while some of the companiesare almost completely consistent with the academicliterature,most are not. In particular,although 94% of the samplefirmsuse DCF methodology, only 62%use a hurdlerate based on the cost of capital, only 53%use more than one hurdle rate (in spite of admitted risk differentials among projects), and less than half change their hurdle rates even once a year. Finally, the inconsistencies are caused by two primary factors. First, some of the techniquesnow being advocated in the academicliteratureare relatively new, and not all of the operating personnel in the firms are "checkedout" on these techniques.The respondentsare,

however, generally receptive to most of the new ideas, and over time many of the inconsistencieswill probably diminish. The second cause of the inconsistencieshas to do with the difficulty of measuringthe averagecost of capital, project risk, and appropriaterisk-adjusteddiscount rates. But here again, the sample companies are generally trying to overcome these measurementproblems, and assumingthey can get a better fix on the basic data, one can anticipate a further reduction of the observed inconsistencies. Data problems may, however, keep Hertz-type simulation from ever being widely used in industry. In closing, we should reiterate that these conclusions are based on a sampleof 33 large,relativelysophisticated firms, so no inferences can be drawnabout the practices of industrial firms in general and small business in particular. However, to the extent that firms such as the ones in the sampletend to lead others, the data may well indicate the directions in which capital budgeting is moving.

REFERENCES
1. Moustafa Abdelsamad, A Guide to Capital Expendi-

AmericanManageture Analysis, New York, AMACOM, ment Association, 1973. 2. C. B. Barry,E. F. Brigham,and R. L. Crum,"Perspectives on Risk Analysis in Capital Budgeting,"paperpresented at Southern Finance Association Meeting, New Orleans,November, 1975. 3. Eugene F. Brighamand RichardH. Pettway, "Capital Budgetingby Utilities," FinancialManagement (Autumn 1973).
4. George A. Christy, Capital Budgeting-Current Practices and Their Efficiency, Eugene, Oregon, Bureau of Business & Economic Research, University of Oregon,

ton, Indiana,Bureauof Business Research,IndianaUniversity, 1961. 9. Donald F. Istvan, "The Economic Evaluationof Capital Expenditures," The Journal of Business (1961).

10. J. R. Lyon, "Using MultipleCutoff Rates for Capital Investment," Journal of Petroleum Technology (July

1975). 11. Thomas P. Klammer, "EmpiricalEvidence of the Adoption of Sophisticated Capital Budgeting Techniques," Journal of Business (July 1972), pp. 387397. 12. Terry J. Nolan and Frederick A. Banda, "An Empirical Study of the CapitalInvestmentDecision Making Process in Selected Ohio Companies in 1971-Part I,"
Akron Business and Economic Review (Spring 1972), pp. 10-16; Part II, Akron Business and Economic Review (June 1972).

1966. 5. Gordon Donaldson, "StrategicHurdleRates for Capital Investment," Harvard Business Review (March/April

1972). 6. K. LarryHastie, "One Businessman's View of Capital

Budgeting," Financial Management (Winter 1974).

13. Norman P. Pflomn, "ManagingCapital Expenditures," Studies in Business Policy, 107, New York, The

7. David B. Hertz, "Risk Analysis in Capital Investment," Harvard Business Review (January 1964). 8. Donald F. Istvan, Capital Expenditure Decisions: How They Are Made in Large Corporations, BloomingAutumn 1975

NationalIndustrialConferenceBoard, 1963.
14. George Terborgh, Business Investment Management,

A MAPIStudy and Manual,Washington, D.C., Machinery and Allied Products Institute and Council for Technological Advancement,1967. 25

Appendix. Sample companies Companies Agway,Inc. AlliedChemical Corporation ArmcoSteel ClarkEquipment of Container Corporation America Crown-Zellerbach Deere& Company Inc. Esmark, F. S. Services,Inc. Mills General Goodrich Gulf Oil HanesCorp. IBM Industries KeystoneConsolidated Maytag Company & Minnesota Mining Mfg.Co. Ward Montgomery SCM Corp. Sears,Roebuck& Co. SignodeCorp. SouthernRailwayCo. Shell Oil Staley Mfg.Co. StanleyWorks Oil Standard of Indiana SybornCorp. Tenneco TRW,Inc. Unio Oil Oil Universal Products Weyerhaeuser Williams Companies 1974 Assetsin Millions $ 456 1,968 2,542 1,006 862 1,269 2,022 1,266 122 1,117 1,647 12,503 183 14,027 196 123 2,841 2,151 647 11,339 247 2,024 6,129 285 278 8,915 419 6,402 1,698 3,459 443 2,327 1,286

26

Financial Management