Cases on Cost Analysis and Government Decisions

C h a p t e r 6

Measuring and Improving Efficiency

In recent years, local, state, and national governments have focused much effort toward making government more efficient. Efficiency is the “accomplishment of objectives at minimal cost.”1 In other words, effi- ciency is the amount of resources used to produce an output. Efficiency is frequently measured in terms of the cost for each unit of output. Cost analysis can be a powerful tool to improve the efficiency of government programs. In this chapter, we will introduce four tools for measuring and improving the efficiency of government programs:

• Unit cost analysis;

• Standard cost analysis;

• Intergovernmental unit cost comparisons; and

• Activity analysis.

UNIT COST ANALYSIS Unit cost is the cost per one unit of a good or service. Examples of unit cost include: the cost of refuse collection per household per year, the cost per tree planting, and the cost per arrest. Unit cost can be a useful tool to measure improvements in efficiency. For example, a park main- tenance crew might use the maintenance cost per acre to measure the ef- ficiency of new maintenance techniques. A very high unit cost might in-

81

dicate inefficiency, while a very low unit cost could indicate that insufficient personnel and equipment is being used.

How to Calculate Unit Costs. The unit cost is simply the total cost divided by the number of units of output. Several key points should guide a unit cost calculation. First, the costs included in the numerator should be those costs incurred to produce the units of output in the de- nominator. For example, if a village is calculating the cost per tree plant- ing, the costs of labor, materials, and equipment in the numerator should be for the specific trees counted in the denominator. Second, a unit cost benchmark that is used to evaluate managerial performance should only include those costs that the manager can control. Thus, it should not include allocations of costs that the manager has no control over. Third, the output used for the denominator should be a variable that most closely reflects the goals of the service. For example, if the goal is to maintain parkland, then cost per acre maintained should be used rather than cost per tree planted. Fourth, the units of output in the denom- inator should be comparable. For example, if the cost per tree planting is being compared across years, the average height of the trees planted in each year should be similar. Otherwise it is very difficult to deter- mine whether a change is unit cost is due to a difference in efficiency or tree size. Finally, when comparing unit costs over time, it is necessary to remove the effect of inflation from the cost data. (Chapter 5 shows how to adjust data for inflation.) Exhibit 6-1 shows the unit costs of the motor pool in Mercer County, New Jersey.2

How to Read Unit Cost Data. If not used carefully, unit cost data can be misleading. Although higher unit costs are often associated with inefficiency and high total costs, it is very possible that a program with higher unit costs is more efficient and has lower total costs. When inter- preting unit cost data, it is important to look behind the scenes to under- stand why unit costs have changed. The key questions to ask when in- terpreting unit cost data are:

1. Has there been a change in the quality or amount of service pro- vided per unit? In other words, is more or less service being pro-

82 ! Cost Analysis and Activity-Based Costing for Government

vided to each customer, or has there been a significant change in quality? If more service is being provided per unit, then an in- crease in unit costs may not indicate inefficiency, but the higher costs of providing more service. It is even possible that the ser- vice is more efficient if the same total amount of service is consol- idated into fewer units of service. For example, when a purchas- ing department consolidates purchase orders for repeated purchases into a blanket purchase order, the unit cost per pur- chase order will increase even though paperwork is reduced and efficiency is improved.3

2. Is another variable affecting unit costs? It is possible that a change in unit costs is due to another variable that is not in- cluded in the unit cost ratio. This variable could be a change in the environment that affects unit costs. For example, the unit cost of clearing snow may increase from one year to another due

Measuring and Improving Efficiency ! 83

Exhibit 6-1 ! An Example of Unit Costs

Cost Item Cost

Personnel Costs

Salary $626,404

Longevity $22,350

OT $55,000

Benefits (25%) $161,969

Total Personnel $865,723

Other Expenses

Parts, Tires, etc. $172,000

Special Services $99,000

Other $25,565

Total Other $296,565

Grand Total $1,162,288

Vehicles 489

Per Vehicle Maintenance Cost $2,377

to higher snowfall in the second year, not a decrease in efficiency.

A change in the clientele or usage of a program can also cause a change in unit costs. Exhibit 6-2 shows an example where a change in the participation rate in a recycling program has an opposite effect on two different unit cost measures.4 As the participation rate increases, cost per household shows an in- crease in unit costs (suggesting less efficiency), while the cost per ton shows a decrease in unit costs (suggesting more effi- ciency). The unit cost per household increases as the participa- tion rate increases because workers require more time to collect more material in each neighborhood and require more frequent trips to unload the truck at the recycling facility. The cost per ton decreases because it takes less time and less driving to collect each ton of recyclables because more houses are participating. Unit costs can be misleading because the output variable chosen for the denominator in the unit cost ratio does not give a com- plete picture of costs.

3. Are fixed costs a significant proportion of the total costs? and if so, What effect do the fixed costs have on unit costs? It is very

84 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 6-2 ! Two Unit Cost Measures that Show Different Pictures of Efficiency

important to consider the effect of fixed costs on unit costs be- cause unit costs can be misleading when they contain fixed costs. When a significant portion of the total costs are fixed costs, it is difficult to know whether a change in unit costs is due to a change in efficiency or a change in quantity. This is because both a change in efficiency and quantity can affect unit costs when there are significant fixed costs.

If there are no fixed costs (i.e., all costs are variable) unit costs provide a convenient way to compare the total cost at two different quantities. Without looking at the unit cost, it is diffi- cult to determine whether, for example, $14 for 8 gallons of gas- oline is better than $25 for 13 gallons. Unit costs make this com- parison simple: $1.75/gal. versus $1.92/gal. When all of the costs are variable costs, the unit cost shows clearly whether an increase in total costs is due to an increase in quantity or a de- crease in efficiency. Exhibit 6-3 illustrates this. The top diagram of this exhibit shows the total costs of two alternatives. One al- ternative is an increase in total costs from A to B1. The second al- ternative is an increase from A to B2. Although B2 has higher to- tal costs than B1, the unit costs shown in the bottom diagram reveal that this B2 is a better alternative because it has lower unit costs than B1.

If a unit cost only contains variable costs, then it can be a good measure of efficiency because the unit costs will change if there is a change in efficiency and remain the same if a change in total costs is due to a change in quantity. However, if a unit cost contains fixed costs, it is difficult to know whether a change in unit cost is due to a change in efficiency or quantity. Unit costs that contain significant fixed costs will tend to decline as the quantity increases because additional units bear a portion of the total cost. Exhibit 6-4 shows how an increase in unit costs could be due to either a change in efficiency or a change in quantity. This exhibit shows two alternatives with the same increase in unit costs. One alternative is an increase in unit costs from A to B1. The second alternative is an increase from A to B2. Although both alternatives have the same unit cost, the increase in unit costs to B1 is caused by a decrease in quantity, while the increase to B2 is caused by a decrease in efficiency.

Measuring and Improving Efficiency ! 85

Two ways of avoiding the misleading nature of unit costs that include fixed costs are to: 1) look at the change in total cost rather than the change in unit cost, and 2) exclude fixed costs and measure only the unit variable cost.

4. Is there excess capacity or the need to expand capacity in the near future? Excess or deficient capacity is also an important consideration when interpreting unit cost data. In some cases, high unit costs may be due to a recent expansion in capacity and may suggest excess capacity that is being underutilized, while unusually low unit costs may suggest the need to expand capac- ity in the near future. When capacity is expanded, unit costs will

86 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 6-3 ! Unit Costs and Variable Costs

experience a sharp increase and then decrease as the capacity is used up.

STANDARD COST ANALYSIS A standard cost is a unit cost in normal or optimal conditions.5 In other words, a standard cost is what the unit cost should be. In the government sector, standard costs are usually set by each individual government. The standard is based on either historical data or the judgment of indi- viduals familiar with the organization’s costs. Standard costing6 is ap- propriate for repetitive services, such as processing tax payments, in which the work performed is the same for each unit of service.

How to Do Standard Costing. In order to use standard costs to pin- point inefficiencies in a budget, first a standard cost is determined for each line item in the budget. Then, these standard costs are multiplied by the actual quantity of each line item to calculate a flexible budget. A flexible budget is a hypothetical budget that shows what the dollar value of each line item would be in an efficient scenario.7 Finally, the flexible budget is compared with the actual budget to pinpoint ineffi- ciencies in specific line items. Standard costing involves three basic steps:

1. Determine standard costs.

2. Calculate a flexible budget by multiplying the standard unit cost by the actual quantity of each line item.

Measuring and Improving Efficiency ! 87

Exhibit 6-4 ! Unit Costs and Fixed Costs

3. Use variance analysis to compare the actual and flexible budgets.

Exhibit 6-5 shows an example of a standard cost analysis for a garbage collection service.

Step 1: Determine Standard Costs. The first step in a standard cost- ing analysis is to determine standard costs for each line item in the bud- get. First, knowledgeable individuals estimate the amount of resources necessary to produce a level and quality of output that meets the orga- nization’s objectives. Second, the cost of each resource is estimated. Third, the total cost of each cost component is divided by the output to calculate a standard unit cost for each component.

There are three approaches to setting standard unit costs. One ap- proach is to make the historical average the standard. This approach is easy to determine, however it incorporates past inefficiencies and may demotivate employees to surpass the standard. A second approach is to make normal operating conditions the standard. This approach does not incorporate past inefficiencies; however, it requires more detailed analysis and more subjectivity as to what are considered normal condi- tions. A third approach is to make optimum conditions the standard. This approach moves the bar up to the highest level and incorporates no

88 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 6-5 ! How to Do Standard Costing

inefficiencies. The disadvantage, however, is that it might demotivate employees by setting a standard that is impossible to maintain. Exhibit 6-6 summarizes the strengths and weaknesses of the three approaches to setting standard costs.8

Step 2: Calculate a Flexible Budget. The second step is to calculate a flexible budget by multiplying the standard unit cost by the actual quantity of each line item.

Step 3: Do a Variance Analysis. The third step is to use variance analysis to compare the actual and flexible budgets. Variance analysis can be used to examine whether differences between the actual and the flexible budget are due primarily to one of the following:

1. Change in the quantity of output (i.e., the demand for the service);

2. Change in the quality or amount of service provided for each unit;

3. Change in efficiency (i.e., change in the amount of labor and ma- terials used for each unit of output); and/or

4. Change in prices (i.e., wage rates and material prices).

INTERGOVERNMENTAL UNIT COST COMPARISONS An alternative to benchmarking unit costs against an internally gener- ated standard is to compare unit costs with other governments who provide the same services. In an effort to improve the performance of

Measuring and Improving Efficiency ! 89

Exhibit 6-6 ! Approaches to Setting Standard Costs

Standard Costs Based On: Strengths Weaknesses

Historical averages • Uses actual performance data

• Easy to determine

• Incorporates past inefficiencies

• May demotivate employees to surpass standard

Normal operating conditions • Reflects current performance

• Past inefficiencies are not automatically included

• Requires detailed analysis • Requires subjective judg-

ment to define “normal” conditions

Optimum conditions • No past inefficiencies built into standard

• Lowest unit cost

• May demotivate employees by setting a standard im- possible to maintain

local government services, several programs have been developed to collect and compare performance measurement data and cost data across governments.

The International City/County Management Association’s Com- parative Performance Measurement Program assists more than 130 cit- ies and counties in the United States and Canada to collect and analyze performance and cost data. Governments that participate in the pro- gram can compare their performance and cost data to the full sample or a subset of comparable jurisdictions. (http://www.icma.org/performance)

The North Carolina Local Government Performance Measurement Project (NCBP) currently includes 14 cities in North Carolina. The NCBP was initiated in 1995 and is one of the most advanced compara- tive benchmarking projects. Appendix E includes a copy of the NCBP form used to collect cost accounting data. (http://ncinfo.iog.unc.edu/programs/perfmeas/)

The South Carolina Municipal Benchmarking Project was pat- terned after the NCBP and includes 19 cities. (http://www.iopa.sc.edu/eJournal/Benchmark2.htm)

The Kansas City Performance Measurement Pilot Project began in January 2000 and includes 11 cities and counties. (http://www.marc.org/performance/home.htm)

Ontario Municipal CAO’s Benchmarking Initiative (OMBI) was initiated in 2000. OMBI plans to integrate activity-based costing data with performance data from participating jurisdictions. (http://www.ombi.ca/index.asp)

The Municipal Performance Measurement Program (MPMP) col- lects performance and cost data from all local governments in Ontario, Canada. The program is an outgrowth of an earlier program initiated in 1996 by the Province of Ontario Ministry of Municipal Affairs and Housing. Unlike the other comparative performance measurement pro- grams, participation in MPMP is mandatory. (http://municipalbestpractices.ca/reference/index.asp)

Caveats When Comparing Unit Cost Data. The following caveats9

should be considered when comparing unit cost data across governments:

1. Differences in cost accounting can create problems when com- paring costs across governments. Some governments may in-

90 ! Cost Analysis and Activity-Based Costing for Government

clude only direct costs while others may include both direct and indirect costs. In addition, governments may differ in what indi- rect costs they include. For example, some governments may not include general government overhead, pension costs, and depreciation in their indirect costs. Further, governments may differ in the allocation bases that they use and the method that they use to allocate indirect costs. For example, some govern- ments may use a single-step allocation while others may use a step-down or reciprocal allocation. These problems can be avoided if all the participating governments follow a standard cost accounting methodology.

2. Governments may have acquired capital assets at considerably different times. If the governments both include significant de- preciation expenses in their costs, then a comparison of costs should be done carefully because inflation can distort a compar- ison of their depreciation expenses (see Chapter 5 for a more complete discussion).

3. Governments may differ in their geography, climate, cost of liv- ing, population density, and societal makeup. Each of these fac- tors can influence the cost of providing identical services. For example, garbage collection costs per household will likely be higher in a county with a low population density than in a county with a high population density because of the greater distance between houses.

4. Governments may differ in their level (i.e., state, county, munic- ipality) and type (e.g., general-purpose, special-purpose).

5. There may be significant differences between the same service performed in different governments. For example, one govern- ment may only pick up garbage at the curb, while another gov- ernment may pick up garbage at a resident’s backdoor (which requires more labor-hours). One way to avoid this problem is to note significant differences in service and divide the sample into two comparison groups: governments with curbside pickup and governments with backdoor pickup. However, this may not be possible if the differences become numerous and complex.

Measuring and Improving Efficiency ! 91

6. Output measures do not necessarily provide information on outcomes and quality of the service. Although one government may have a lower cost per unit of output for job training services than another government, its service may have lower quality and effectiveness than the comparison cities. In other words, its greater efficiency may have been at the expense of effectiveness and quality.

7. Unit cost comparisons do not show a government how it is per- forming compared to an ideal standard, but only show how it is performing relative to its peers. If comparable governments also possess the same inefficiencies, then a unit cost comparison will fail to spot areas of improvement.

ACTIVITY ANALYSIS A traditional approach to increasing efficiency is to cut budgets – reduc- ing the amount of labor and materials while maintaining the same level of output. While this approach will increase efficiency in the short run, in the long run, costs may return to previous levels as employees tire of the unusually fast pace. Another approach is to uncover the inefficient or unnecessary activities that cause the costs. If these activities can be made more efficient or eliminated, then efficiency can be improved per- manently. This second approach is called activity analysis.

Activity analysis (or value analysis) focuses on the activities that an organization performs. The first step in an activity analysis is to draw a detailed diagram of the steps involved in providing a service. This dia- gram is sometimes called a process flowchart. Each activity in the process is then classified as either a value-added or nonvalue-added activity. A value-added activity is an activity that increases the value of a service to the customer. A nonvalue-added activity is an activity that does not in- crease the value of a service to the customer. For a snow plowing ser- vice, a value-added activity would be the actual plowing of the streets. A nonvalue-added activity would be vehicle repair because the activity of repairing vehicles does not improve the clearing of snow. In other words, if vehicles were so reliable that vehicle repair could be elimi- nated, the loss of this nonvalue-added activity would not decrease the value of the snow plowing service. Since nonvalue-added activities consume resources, if they can be eliminated, resources can be saved

92 ! Cost Analysis and Activity-Based Costing for Government

without taking away from the goal of the service. Overall efficiency is improved by continually finding ways to eliminate nonvalue-added ac- tivities and improve the efficiency of value-added activities.10 Appen- dix F provides additional guidance on performing an activity analysis.

Endnotes

Measuring and Improving Efficiency ! 93

1. James E. Sorensen, Glyn W. Hanbery, and A. Ronald Kucic, “Managerial Accounting,” Budget Formulation & Execution, 2nd ed. (Athens, Georgia: Carl Vinson Institute of Government, 1996), p. 452.

2. This table is adapted from State of New Jersey, Department of Treasury, Achieving Excellence: A Guide for Local Officials and Taxpayers To Identify Cost Savings and Improve Local Services (October 1996), p. 57.

3. This example is adapted from Joseph T. Kelley, Costing Government Services: A Guide for Decision Making (Chicago, Illinois: Government Finance Officers Association, 1984), pp. 62-63.

4. This chart is adapted from Jim Morris and Wayne DeFeo, “Fully Understanding Costs,” Practical Recycling Economics (New Brunswick, New Jersey: Rutgers, the State University of New Jersey, 1999), p. 42.

5. This section on standard costing is based on Bruce R. Neumann, James D. Surer, and William N. Zelman, Financial Management: Concepts and Applications for Health Care Providers, 2nd ed. (Dubuque, Iowa: Kendall/Hunt Publishing Company, 1997) and Robert N. Anthony and David W. Young, Management Control in Nonprofit Organizations, 5th ed. (Burr Ridge, Illinois: Irwin Publishers, 1994), pp. 194-196.

6. The term “standard costing” is also used to refer to a “method of estimating the historical cost of a capital asset by establishing the average cost of obtaining the same or a similar asset at the time of acquisition.” Stephen J. Gauthier, Governmental Accounting, Auditing, and Financial Reporting: Using the GASB 34 Model (Chicago, Illinois: Government Finance Officers Association, 2001), p. 651.

7. Another use of the term flexible budget is a budget “which authorizes varying levels of spending depending on demand or revenues.” This is in contrast to a fixed budget, which establishes a fixed spending cap that cannot be exceeded without special authorization. Gauthier, Governmental Accounting, Auditing, and Financial Reporting, p. 304.

8. This table is adapted from Neumann, Surer, and Zelman, Financial Management, p. 332. 9. These caveats are based on H.M. Coombs and D.E. Jenkins, Public Sector Financial Management

(London, England: Chapman & Hall, 1994), pp. 25-28; Leon E. Hay and Earl R. Wilson, Accounting for Governmental and Nonprofit Entities (Chicago, Illinois: Irwin Publishers, 1995), pp. 591-592; and the GFOA Recommended Practice, “The Use of Trend Data and Comparative Data for Financial Analysis (2003).” This section includes some caveats not discussed in the GFOA Recommended Practice.

10. This section is based on Atkinson et al., Management Accounting (Upper Saddle River, New Jersey: Prentice Hall, 1997), pp. 62-69.

C h a p t e r 7

Setting Fees and Charges

This chapter begins by briefly looking at the advantages of fees and user charges.1 Following this is a review of the fee-setting process. The first step in setting fees is to determine whether a fee is appropriate for the service. The second step is to determine the primary purpose of the fee, because the purpose will determine the method that should be used to set the fee. For example, if the primary purpose of the fee is to recover the full cost of a service (or subsidize a percentage of the full cost), then one should determine either the direct and indirect costs or the fixed and variable costs of the service. However, if the primary purpose of the fee is to recover the cost of additional users of a service, then one should determine either the marginal costs or fixed and variable costs of the service. The final step after performing the cost analysis is to consider any significant market factors, political factors, and broad policy goals that may influence the amount of the fee.

ADVANTAGES OF FEES AND USER CHARGES Fees and user charges2 comprise about 16 percent of total state and local government revenues.3 Fees and user fees have a number of important advantages. They generate revenue and send important signals on user demand, they enable a government to control how a service is used and carry out public policy, and they provide a means of financing that is fairer than taxes.

95

Revenue. User fees can be an effective way to generate revenue. Fee increases are usually more politically popular than tax increases. Fees can also help a government to diversify its revenues so that the government is not overly reliant on particular sources of revenue. In ad- dition, user fees that cover the total cost of a service enable a govern- ment to expand a program without increasing taxes or the costs to indi- vidual users because the additional users pay for the expansion.

Information. In addition to revenue, fees also provide important information to the government on the demand for a service. For exam- ple, monthly parking spaces that are sold quickly at a high fee indicate that there is a high demand for parking. The demand for a service after a fee is set equal to the cost of a service can tell policy makers whether res- idents are willing to pay what it costs to provide a service.

Control How a Service or Resource is Used. Fees can also be used to control how customers use a service or resource. For example, even a small fee on water can prevent users from wasting this resource. Li- brary fines encourage users to return books to the library so that they can be shared by others.

It is also possible to use fees to smooth out peak demand for a ser- vice by charging higher fees during peak periods. This would encour- age customers to use a service at non-peak times and may reduce the need for long lines and extra staff during peak periods.

Carry Out Public Policy. Since user fees influence behavior, they can also be used to carry out public policy. For example, if a govern- ment wants to encourage its citizens to recycle more, it might charge a fee based on the volume of garbage that a resident puts at his or her curbside. If a government wants to decrease automobile emissions, it might lower bus fees and increase car registration fees.

Fairness. Another advantage of fees is that they are fairer than taxes in that those who use a service pay for it. If a service is financed by taxes, all of the taxpayers must pay for it whether they use it or not. In addition, it is possible for nonresidents to use the service even though they do not pay taxes to support it. Fees avoid both of these inequities. One disadvantage with fees, however, is that they may wipe out inten- tional subsidies for low-income residents, unless special adjustments are made.

96 ! Cost Analysis and Activity-Based Costing for Government

HOW TO SET FEES The process of setting fees involves answering four basic questions: (1) Is a fee appropriate for this service? (2) What is the primary purpose of the fee? (3) What is the full cost of the service? and (4) How do market factors, political factors, or broad policy goals affect the amount of the fee? The government’s policy on fees and charges should provide a guide to answering these questions.

Develop Policy on Fees and Charges. The Government Finance Officers Association (GFOA) and the National Advisory Council on State and Local Budgeting (NACSLB) both recommend that govern- ments develop a formal policy on setting fees and charges, which should have the following characteristics:

• Includes a requirement to review all fees and charges;

• Identifies how fees and charges will be set and what factors will be taken into account;

• States whether the government intends to set fees to recover the full cost of services. (The GFOA and NACSLB recommend that governments estimate the full cost of providing services and use this information when setting fees.);

• Describes instances in which the government may set a fee higher or lower than the full cost of a service and provide its ra- tionale for doing so;

• Sets a frequency for undertaking cost-of-service studies;

• Allows stakeholder input; and,

• Is made available to the public.

Appendix B provides the full text of the GFOA recommended practice on fees and charges.

Is a Fee Appropriate? Before starting a cost analysis, it is important to determine whether the service has specific characteristics that make it suitable for a fee. In general, a fee is appropriate when a service meets the following criteria:

1. Individuals who do not pay can be prevented from using the service.

2. The cost of administering the fee is relatively low.

3. Those who would pay the fee receive most of the benefit of the service. In addition to being fairer to payers, this criteria is also

Setting Fees and Charges ! 97

better for nonpayers. For example, if a fee is charged to students to cover the full cost of a public school, then nonpayers are at a disadvantage because they receive benefits from the quality of the school in their community, but they cannot protect this ben- efit through their payments. Thus, they risk losing this benefit if students move to another school.

4. The customer demand for the service is somewhat responsive to the fee. In the terminology of economics, the service should have some elasticity of demand. It is preferable that the demand is sensitive to the level of the fee so that the fee can be used to re- duce consumption or can be used to measure the value that resi- dents give to the service and thus provide information for re- source allocation decisions.

5. The use of a fee does not work against government policy goals. For example, a high fee on public transportation might counter- act environmental goals of reducing auto emissions.

What is the Primary Purpose of the Fee? It is important to identify the primary purpose of the fee because the purpose will determine the method that should be used to set the fee. Essentially, fees are imposed for one of two purposes: collecting revenue or influencing how resi- dents use government services. When the general purpose is to collect revenue, a government might attempt to set a fee to recover the full cost of the service, recover a portion of the full cost and subsidize the rest, or recover the cost of additional users. When the general purpose is to in- fluence how residents use government services, a government might attempt to set a fee to discourage waste, ration scarce resources, carry out a policy goal, or smooth out peak demand. In the following para- graphs we will show how to set a fee in each of these cases.

Recover the Full Cost of a Service. In many instances, the primary purpose of a fee is to recover the full cost of a service. A government would want to recover the full cost of a service when it wants to pass the full cost of the service on to the users of the service and does not want the service to be subsidized by taxpayers – a water utility and municipal golf course are good examples. Even if the government intends to subsi- dize the service, it is still useful to calculate the full cost so that policy makers know how much of a subsidy they are providing. There are two alternative methods of setting a fee to recover the full cost of a service.

98 ! Cost Analysis and Activity-Based Costing for Government

One method uses direct and indirect costs. A second method uses fixed and variable costs.

Method 1: Direct and Indirect Costs. One method of recovering the full cost of a service is to estimate its direct and indirect costs (see Chap- ter 3) and then divide the total cost by the number of units of service to calculate a total cost per unit of service (see Exhibit 7-1).

When setting a fee to recover the cost of a service, it is important that the fee does not just recover the full cost of the service someday, but when the government must meet its cash flow needs. Assume, for ex- ample, that a county government uses a ten-year bond to finance an air- port with a useful life of twenty years. Under accrual accounting, the county would depreciate (or apportion) the capital expense of the air- port over twenty years. If a fee was set to meet this depreciation ex- pense, the county would have insufficient revenue during the first ten years when it must make debt service payments, and excess revenue during the second ten years after the bonds have been paid. Thus, it is important to consider cash flow needs when setting fees.4

Method 2: Fixed and Variable Costs. Another method of recovering the full cost of a service is to estimate its fixed and variable costs (see Chapter 4) and then use a version of the breakeven formula5 to calculate the fee that should be charged per unit to recover the full cost (see Ex- hibit 7-2).

Setting Fees and Charges ! 99

Exhibit 7-1 ! Calculating Fees Using Direct and Indirect Costs

Exhibit 7-2 ! Calculating Fees Using Fixed and Variable Costs

The formula above might be used if a county government wanted to determine the fee that it should charge to recover the full costs of its building inspection program. If a county’s building inspection program currently has fixed costs of $300,000 per year, total variable costs of $70,000, and conducts an average of 4,000 building inspections per year, then it would use the formula below to calculate the user fee that it would have to charge for the program to break even. The variable cost per unit would be the total variable cost divided by the number of building inspections ($70,000 ÷ 4,000 = $17.50 per building inspection). This calculation shows that the county would have to charge a user fee of $92.50 per building inspection to cover the full cost of its building in- spections.

The advantage of estimating the full cost using fixed and variable costs is that it is easier to adjust the fee to changes in demand. For exam- ple, if a recreation center expects 300 new users next year, it can adjust the annual fee by simply adding 300 users into the formula. The flexibil- ity with the breakeven formula enables a government to ask “What if?” questions such as: Could the fee be lowered if we increased demand by 30 percent? What fee increase would be necessary to cover the increased fixed costs of an additional facility?, and How much higher should the fee be to residents who request backdoor (rather than curbside) garbage pickup? It is important to note that the breakeven formula is valid for a limited range of users or units. As the number of users/units becomes much larger or smaller, the fixed and variable costs may change. A dis- advantage of the breakeven formula is that fixed and variable costs can sometimes be more difficult to estimate than direct and indirect costs.

Recovering the Costs of Additional Users. In some instances, the purpose of a fee is to recover the additional costs caused by additional us- ers of a service, not the total cost of the service. This may be the case for a town that provides lawn-cutting service to a local school district. The town covers the total cost of mowing town property by using the gen- eral fund, but it wants to recover the additional costs of mowing school property by charging a fee. If the town wants to recover its additional

100 ! Cost Analysis and Activity-Based Costing for Government

labor, fuel, and maintenance costs, then it would set a fee equal to its marginal cost.

The marginal cost is the increase in costs caused by a particular ac- tion. For example, the marginal cost of increasing the frequency of street cleaning from once a month to twice a month would be the cost of the additional man-hours, fuel, etc. necessary to clean the streets a second time during the month. The marginal cost of extending pool hours from five p.m. to seven p.m. would be the costs of staffing and operating the pool during those two hours.

Estimating the marginal cost can be very simple, and in some cases, simply a matter of accurate record-keeping. To estimate the marginal cost of mowing school property, the town in the example above might simply keep a record of the amount of man-hours and fuel used to mow school property. Alternatively, the town might estimate the variable cost of mowing one acre of grass, and then multiply this by the number of acres of school property that it mowed. Exhibit 7-3 shows how to esti- mate the marginal cost using both of these methods.

Since the marginal cost is less than the full cost, using marginal costs to set fees would be appropriate when a government wants to en- courage the use of a service or otherwise subsidize a service. One ad- vantage of marginal costing is that it can keep a subsidy from increasing

Setting Fees and Charges ! 101

Exhibit 7-3 ! Setting a Fee to Recover Marginal Costs

Method 1: Record-keeping 1. Record the type and quantity of all of the resources used by an action. 2. Estimate the average unit cost of each resource. 3. Multiply the quantity of each resource by its average unit cost. 4. Sum the costs and add any additional fixed costs generated by the action.

Method 2: Estimating the Variable Cost 1. Estimate the unit variable cost of a service before the action is taken.

a. Record the type and quantity of all of the variable costs of a service and the amount of service provided at that cost.

b. Estimate the average unit cost of each resource. c. Multiply the quantity of each resource by its average unit cost. d. Sum the costs to calculate the total variable cost. e. Divide the total variable cost by the amount of service provided to calculate the unit vari-

able cost. 2. Estimate the amount of units used by the action. 3. Multiply the unit variable cost by the number of units. 4. Add any additional fixed costs generated by the action.

as additional customers use a service because each new customer pays for the additional costs that they generate. Another advantage is that marginal costing can be simpler to estimate than full costs because it does not require an allocation of indirect costs.

It may also be fairer in some cases not to allocate indirect costs. For example, if a grade school rents out its gym to community groups dur- ing after school hours, it may not be fair to allocate the school’s indirect costs to the community groups because the school would still incur these costs even if the gym was left vacant. In other words, allocating these fixed, indirect costs would cause the school to benefit at the ex- pense of the community groups. In addition, the amount of fee revenue involved would probably not justify the complexity involved in a cost allocation.

Influencing How Residents Use Government Services. Although all fees generate revenue, many times the primary purpose of a fee is not to collect revenue but to influence how residents use government services.6 For example, fees can be an effective tool to discourage resi- dents from wasting public resources such as water, parking spaces, and library books. Even a small fine on a library book can encourage users to return them on time so that others can use them.

In addition, fees can be used to carry out broad policy goals such as protecting the environment. For example, charging residents based on the volume of garbage that they throw away is an effective means to en- courage residents to recycle more and preserve landfill space. Rather than imposing a regulation, such as prohibiting lawn sprinkling during certain days of the week during the summer to reduce water use, a higher fee could be charged during summer months. In this instance, a fee carries out the policy goal and eliminates the need (and cost) of enforcement.

Finally, fees can actually reduce the cost of a service if a higher fee is charged during peak periods. For example, a higher fee might be charged for water usage during the summer months than during the winter months. It can be very expensive in terms of capital costs and staffing costs to provide excess capacity to handle the peak demand rather than the average demand for a service. A higher fee during peak periods will encourage users to spread their demand more evenly and reduce the need for excess capital capacity and staff.

102 ! Cost Analysis and Activity-Based Costing for Government

How Do Other Factors Affect the Amount of the Fee? The final step after performing a cost analysis is to consider any significant mar- ket factors, political factors, and broad policy goals that may influence the amount of the fee.

Market Factors. Unlike taxes, many types of fees are voluntary in the sense that users can avoid the fee by choosing not to use a service or by using less of it. Because of this, it is important to consider how the amount of the fee will affect the demand for the service. In general, as a fee increases, users will demand a smaller quantity of a service, and as a fee decreases, users will demand a larger quantity of a service.

In some cases, a fee can be set too high in an attempt to generate more revenue. It is important to realize that increasing a fee will not al- ways increase revenues and may sometimes decrease revenues. For ex- ample, if few residents are willing to pay the higher fee necessary to cover the total cost of a service, then increasing the fee may decrease revenues because fewer residents will purchase the service.

It is also possible for a fee to be set too low. For example, if a village government charges $25 per month for train station parking when most of its residents are willing to pay $50 per month, then those who want to purchase parking stickers will be inconvenienced with long lines and waiting lists.

It is important to realize that the sensitivity of consumer demand to changes in fees depends on the type of service being provided. For ex- ample, consumers are relatively insensitive to changes in fees for ser- vices that fulfill essential needs such as drinking water, but are very sensitive to changes in fees for services that are non essential, such as the use of recreational and cultural facilities.

Many fees operate in a marketplace in which users can choose to purchase the same service from another local government or a private organization. Thus, it is important to compare the amount of the fee to the fees and prices charged by other local governments and private organizations.

Political Factors. Although fees must be politically acceptable, po- litical factors should not be the primary consideration when setting fees.

Broad Policy Goals. In some cases, a fee might be intentionally set higher or lower than the actual cost of the service in order to accomplish policy objectives. For example, a lower fee might be charged to senior

Setting Fees and Charges ! 103

citizens or low-income families. If a government wants to encourage the use of public transportation and discourage use of automobiles, it might set a low fee on public transportation and a high fee on automo- bile permits.

If a service provides significant benefits to people who do not pay for the service, then the government should subsidize a portion of the cost and users should not be charged the full cost of the service because they do not receive the full benefit. For example, all the residents in a community may benefit in terms of higher property values from the community’s parks, however, only a fraction of the residents may actu- ally pay to use particular facilities in the parks. Since the residents who do not use the facilities still receive benefits from the existence of the park, then the government should not charge residents who use the park facilities a fee based on the full cost of the park.

Updating and Revising Fees. After a fee is set, governments should review and update it periodically to adjust for changes in infla- tion, market conditions, and government policy. In addition, govern-

104 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 7-4 ! The Fee-Setting Process

1. Develop formal policy on fees and charges. 2. Determine if a fee is appropriate for this service.

a. Individuals who do not pay can be prevented from using the service. b. The cost of administering the fee is relatively low. c. Those who would pay the fee receive most of the benefit of the service. d. The customer demand for the service is somewhat responsive to the fee. e. The use of a fee does not work against government policy goals.

3. Set the fee based on the primary purpose of the fee. a. Collect revenue

i) Recover the full cost of the service ii) Recover a portion of the cost and subsidize the rest iii) Recover the costs of additional users

b. Influence how the service is used i) Discourage waste ii) Carry out policy goals iii) Smooth out peak demand

4. Consider how market factors, political factors, or broad policy goals affect the amount of the fee. a. What prices do other governments or private providers charge for the same service? b. How will the public react to the fee? c. Should the fee be adjusted to support important policy goals?

5. Update and revise fee based on changes in inflation, market conditions and government policy.

ments should conduct a review of the cost analysis and assumptions that the fee is based on. The City of San Luis Obispo, California con- ducts a comprehensive cost of services analysis every five years and up- dates fees on an annual basis to adjust for inflation.

Endnotes

Setting Fees and Charges ! 105

1. This section draws frm two GFOA publications: Joseph T. Kelley, Costing Government Services: A Guide for Decision Making (Chicago, Illinois: Government Finance Officers Association, 1984) and John E. Petersen and Dennis Strachota, Local Government Finance: Concepts and Practices (Chicago, Illinois: Government Finance Officers Association, 1991), pp. 135-151.

2. It is worth noting some important distinctions between specific types of charges and fees. Utility charges are charges placed on goods and services such as water, sewer, and electricity, which are provided by the government, but might also be supplied by the private sector. User charges and fees are charges on goods and services, unlikely to be offered by the private sector, that provide some benefits to non payers (e.g., library charges, garbage collection, health service fees, etc.). Special assessment fees are mandatory fees charged to property owners in a specific area, which are used to support specific services for that area. The mandatory nature of assessment fees makes them distinct from other sorts of fees. License and permit fees differ from other types of fees because they are a mandatory payment for a privilege granted by the government such as a hunting license or a building permit rather than a payment for a service.

3. “State and Local Government Finances by Level of Government and by State: 1999-2000,” U.S. Census Bureau, Governments Division. The percentage was calculated by the author.

4. This point is made in the GFOA Recommended Practice, “Application of Full-Cost Accounting to Municipal Solid Waste Management Activities (1998).”

5. The breakeven formula is: Break-even number of users = (Fixed cost) / (Revenue per user – Variable cost per user).

6. This section is based on Kelley, Costing Government Services, pp. 38-39.

C h a p t e r 8

Make-versus-Buy Decisions

In the quest for more efficient and effective government services, many governments have turned to the private sector and other entities out- side the government to provide these services. One option is to contract out the service to a private-sector firm. Private firms provide many of the same services as local governments, including: building and street maintenance, garbage collection, and snow removal. Another option is to contract with nonprofit organizations for certain human services, such as job training, homeless assistance, and recreational activities. A third option is to contract with other governments to provide services.

So is a government better off contracting out its services? And if so, which services? This chapter shows how to perform a cost analysis to answer this question. The process of deciding whether to perform ser- vices in house or contract them out is commonly referred to as the make-versus-buy decision. This chapter will walk through the steps of a make-versus-buy cost analysis focusing primarily on the cost of the ser- vice. Although this chapter focuses on costs, there are other important considerations in a make-versus-buy decision including:

• Quality of the service;

• Government’s ability to control the service;

• Government’s ability to evaluate the service;

• Impact on government employees, customers, and taxpayers; and,

• Statutory and regulatory issues.1

107

Before walking through the steps of a make-versus-buy cost analy- sis, two points need emphasis. First, an evaluation of contracting out should use a differential cost perspective. Second, the cost comparison should cover a multi-year period and discount future cash flows.

Use a Differential Cost Perspective. The differential cost is the key cost analysis concept for evaluating the contracting out of a service. The differential cost shows how a decision to contract out will change a gov- ernment’s costs. It is crucial to look at the differential costs and not merely compare the total costs of the status quo to the total cost of a pri- vate contractor. The pitfall of comparing total costs is that they may in- clude fixed costs that cannot be saved if the service is contracted out. This could give the appearance that the government will incur less costs with a private contractor when it actually will incur more.2

For example, let us say that a private waste hauler offers to provide waste collection services to the City of Unionsville for $550,000 per year. Currently, the total cost of waste collection services is $750,000 per year. Thus, it appears that the city could save $200,000 per year by selecting the private hauler. However, a closer look at the city’s fixed costs re- veals that the city is committed to spending much of the $750,000 even if it switches to a private hauler. More than half of this amount is person- nel costs, which the city is committed to due to a no-layoff policy and the fact that the truck drivers perform other responsibilities. In addi- tion, the city is committed to $50,000 per year in debt service payments for the facilities used to store and maintain its garbage trucks.

Sunk Costs. A potential mistake in a make-versus-buy cost analy- sis is the inclusion of sunk costs. A sunk cost is a cost that has already oc- curred and will remain the same regardless of what decision is made. An example of a sunk cost is the cost of conducting a survey to deter- mine resident interest in an outdoor pool. After the survey has been conducted, the cost of the survey is a sunk cost because it will remain the same regardless whether a pool is built. Therefore, this cost should be ignored in an analysis.

To see how including sunk costs can lead to bad decisions, suppose a county government is considering outsourcing its warehouse func- tion to private suppliers who can maintain inventories of all of the county’s supplies and ship them overnight. One year earlier, the county had spent $500,000 in consulting services to develop a state-of-the-art inventory process. This process would be scrapped if the county de-

108 ! Cost Analysis and Activity-Based Costing for Government

cides to outsource this function. Opponents of the outsourcing plan ar- gue that the county shouldn’t outsource this function because it had just poured $500,000 into perfecting its current system. However, this $500,000 should not influence the decision because it cannot be recov- ered regardless of the decision that the county makes. Only the differen- tial cost of the two alternatives should influence the decision. If the dif- ferential cost of outsourcing the warehouse function is $340,000 less per year, including the sunk cost of the consulting services would lead the county to spend $340,000 per year more than it has to.

Opportunity Costs. Another important cost concept in make-versus-buy decisions is opportunity cost. Opportunity cost is the lost opportunity of using an asset or resource in a way other than the chosen alternative. For example, if a suburban government sells a pub- lic swimming pool to a private company to own and operate, the oppor- tunity cost would include the admittance fee revenue that would have been collected if the pool remained a public asset. Likewise, the oppor- tunity cost of not selling the pool would be the revenue from the sale of the pool.

Opportunity costs should be included in the cost analysis for a make-versus-buy decision. If opportunity costs are not included as costs, then some proposals may appear to be better just because they use existing government resources. To show how this can lead to bad decisions, suppose a small city government is considering two propos- als for providing a community fitness center. In the first proposal, the fitness center would be operated by city employees and would involve the use of a vacant, downtown, city-owned building (that could be sold for $2.3 million). In the second proposal, the fitness center would be op- erated by a private company, but would require the city to purchase va- cant land in a residential area and construct a new facility.

The top box in Exhibit 8-1 illustrates what the costs and benefits of both proposals would look like in the first year if the opportunity cost of using the city building were ignored. Note that the total cost of proposal A ($1.2 million) appears to be nearly $1 million less than the total cost of proposal B ($2.1 million).

The bottom box in Exhibit 8-1 illustrates what the costs and benefits look like when the opportunity cost is included. Since the building could be sold for $2.3 million, its value to the city government is $2.3 million. Therefore, using this building for a fitness center entails an op-

Make-versus-Buy Decisions ! 109

portunity cost of $2.3 million. Including this opportunity cost shows that the total cost of proposal A is $1.4 million more than proposal B.

To discover opportunity costs, first consider all of the government resources that are used by the proposed project – land, employee time, facilities, etc. Second, determine the value of each of these resources to the government, or the greatest benefit that the government would ob- tain by using each resource in another way.

Cover a Multi-year Period and Discount Future Cash Flows. A cost comparison should cover a multi year period such as three to five years. This is important for two reasons. First, a multi-year analysis is more likely to reveal whether contracting out will provide a long-term savings to the government. A government should experience much of the savings of contracting out in later years as leases and contracts ex- pire and many fixed costs become variable costs. Second, a multi-year contract should be more attractive to potential vendors – creating more competition.

In a multi-year analysis, future cash flows should be discounted to their present value. This insures that appropriate weight is given to fu- ture costs and benefits. Discounting is the process of converting a future value into its present value. See the end of Chapter 5 for an explanation of discounting.

110 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 8-1 ! Including Versus Ignoring Opportunity Cost

HOW TO DO A MAKE-VERSUS-BUY COST ANALYSIS A make-versus-buy cost analysis involves four basic steps:

Step 1: Define the service. Step 2: Calculate the in-house costs that could be avoided by con-

tracting out the service. Step 3: Calculate the total costs of contracting out. Step 4: Compare the costs that could be avoided with the costs of

contracting out. Exhibit 8-2 illustrates the steps in a make-versus-buy cost analysis.

To simplify the diagram, the analysis is shown for a single year. In ac- tual practice, the analysis should cover a three- to five-year period and steps 2b through 4 should be repeated for each year. The totals in each year should be discounted to their present value.

Step 1: Define the Service. The first step in a make versus buy cost analysis is to clearly define the government service that is being consid- ered for outsourcing. In other words, it is important to specify the qual- ity and quantity of the service and the output and outcomes that are expected. This is necessary so that there is an apples-to-apples compari- son between the service that the government currently provides and the services proposed by outside contractors. If a service is vaguely or in- correctly defined, then in-house costs may be higher than outside bid- ders simply because the government is providing more service than what is documented in the RFP.

When specifying the quality and quantity of the service, it is impor- tant to investigate whether government employees informally provide additional services to residents. For example, a parks and recreation de- partment may perform tree trimming for elderly residents upon re- quest, or may deliver woodchips free of charge.

It is also important to examine whether the resources (i.e., labor, fa- cilities, equipment, and material) used for the service are formally or in- formally shared with other government services. For example, garbage collection crews might assist with snow removal after a heavy snowfall, or may help the parks and recreation department clean up after a sum- mer festival. All of these details must be documented in the contract to provide an accurate cost comparison and to avoid disputes with ven- dors after the contract is signed.

Step 2: Calculate the In-House Costs that Could Be Avoided by Contracting Out the Service. The second step is to calculate the total

Make-versus-Buy Decisions ! 111

112 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 8-2 ! Make-Versus-Buy Cost Analysis

government costs that could be avoided or saved by contracting out the service. In order to determine the costs that would be saved, first, item- ize the full cost of the service, including all of the direct and indirect costs.3 Then, use this list of costs as a basis from which to determine the specific costs that would be saved if the service is contracted out. It is important to remember that many fixed costs – such as overhead costs – will remain the same even though the resources behind the costs are not being used. The Government Finance Officers Association (GFOA) rec- ommended practice on Measuring the Cost of Government Services (Ap- pendix A) defines the in-house costs that can be saved by contracting out as, “those costs that can be eliminated immediately” and “those costs that can be eliminated after a transition period.”

These cost estimates should be made on a multi-year basis (i.e., three to five years) and discounted to a present value. In order to treat inflation consistently, nominal costs should be used if a nominal dis- count rate is used, and real costs should be used if a real discount rate is used.4

Step 3: Calculate the Total Costs of Contracting Out. The third step is to calculate the total costs of contracting out the service. The costs of contracting out includes: the bid from the contractor, the govern- ment’s contract administration costs, and the government’s transition costs, minus the additional revenue generated from contracting out. These cost estimates should be discounted to their present value and made over the same time period as the cost savings in the previous step.

In order to be consistent throughout the analysis, only new costs should be counted, as opposed to costs that the government would in- cur anyway even if it decided to keep the service in-house. For example, if a government already has employees dedicated to contract adminis- tration, these costs should not be counted because contracting out would not generate any new costs. This is true as long as the contract administration can be performed with existing employees.

Contract Administration Costs. The government’s contract ad- ministration costs include all of the tasks necessary to select and man- age a vendor through the life of the contract. These tasks may include: reviewing and evaluating RFPs, writing and negotiating a contract, processing change orders and amendments to the contract, monitoring and evaluating vendor performance, dealing with disputes, and pro- cessing payments to the vendor.

Make-versus-Buy Decisions ! 113

There are two primary methods of estimating contract administra- tion costs: using informed judgment or adopting a standard formula.5 If an estimate is made based on informed judgment, a reasonable range for the estimate would be 10 percent to 20 percent of the amount of the contractor’s bid. Small contracts should be closer to the upper end of the range, while larger contracts should be near the lower end of the range. In actual practice, state and local governments estimate the costs of con- tract administration between 0 percent and 25 percent, depending on the government.

Another method of estimating contract administration costs is to adopt another government’s standard formula. The U.S. Office of Man- agement and Budget uses the formula shown in Exhibit 8-3 to estimate the number of government staff required to administer federal con- tracts.6 The State of Texas uses the formula shown in Exhibit 8-4 for the same purpose.7 In both of these formulas, the numbers in the first col- umn are the number of government employees that are currently used

114 ! Cost Analysis and Activity-Based Costing for Government

Exhibit 8-3 ! OMB Staffing Formula for Estimating the Cost of Government Contract Administration

Number of Government Employees Used to

Provide Service

Number of Staff Needed for Contract Administration

(FTEs)

10 or less 0.5

11 – 20 1

21 – 50 2

51 – 75 3

76 – 100 4

101 – 120 5

121 – 150 6

151 – 200 7

201 – 250 8

251 – 300 9

301 – 350 10

351 – 450 11

More than 450 2.5% of the number of government employees used

to provide service

to provide the service that may be contracted out. The second column shows the number of government staff that would be needed to admin- ister the contract if the service is contracted out.

Transition Costs. The government’s transition costs include all of the costs that the government incurs when it makes the transition to an outside contractor. These transition costs may include the various per- sonnel-related costs that result from laying off employees such as un- employment compensation, accrued vacation benefits, and severance pay.8 These transition costs may also include the preparation of govern- ment facilities and equipment for use by an outside contractor. Alterna- tively, the assets might be sold or disposed of – resulting in a positive or negative net salvage value. If the assets are rented, an early termination of the lease may involve additional costs.

Revenues from Contracting Out. Any additional revenues that a government collects (as a result of contracting out) should be sub- tracted from the costs of contracting out. In some cases, a decision to contract out a service may result in an increase in tax and fee revenues from the contractor. For example, a contractor may pay additional property taxes because they build a new facility within city limits. An- other source of revenue from contracting out is the sale of government assets that are no longer needed because a vendor uses its own assets to provide the service. For instance, a government might sell its garbage

Make-versus-Buy Decisions ! 115

Exhibit 8-4 ! State of Texas Staffing Formula for Estimating the Cost of Govern- ment Contract Administration

Number of Government Employees Used to

Provide Service

Number of Staff Needed for Contract Administration

(FTEs)

20 or less 1

21 – 42 2

43 – 65 3

66 – 91 4

92 – 119 5

120 – 150 6

More than 150 2% to 4% of the number of government employees used to

provide service

trucks if a vendor uses its own vehicles. Care should be taken to include only the additional revenues that are the result of contracting out.

Step 4: Compare the costs that could be avoided with the costs of contracting out. The final step is to calculate the difference between the costs that are saved and the costs that are generated by contracting out the service. On the one hand, many in-house costs are saved when a ser- vice is contracted out. On the other hand, many new costs are generated when an outside contractor provides a service. If the costs that are saved by contracting out are significantly higher than the costs of contracting out, then contracting out may make good financial sense.

Sensitivity Analysis. It is possible that the final result is based on assumptions that have a fair degree of uncertainty. As a countermea- sure, it is beneficial to conduct a sensitivity analysis on the final result. Sensitivity analysis tests how sensitive an analysis is to changes in the underlying assumptions of the analysis.9

Consideration of Costs and Benefits Not Included in the Analysis. After conducting a costs analysis, it is important to consider other costs and benefits that could not be quantified in the analysis. For example, there may be a significant difference in the quality of the service pro- vided by government and the private contractor. Or, it may be difficult to evaluate the quality of the service. If there is inadequate competition among contractors, the government may be subject to price increases or losses in service quality if it contracts out the service. In addition, the government may have a reduced ability to use the service to carry out other policy goals, such as helping the poor and reducing pollution. Further, there may be negative impacts on government employees, cus- tomers, and taxpayers. Finally, underutilized facilities, labor, and equipment during the transition period may entail opportunity costs to the government. To address these types of problems, some govern- ments require the cost savings of contracting out to be at least 10 percent greater than the costs of providing the service in house.10

Endnotes

116 ! Cost Analysis and Activity-Based Costing for Government

1. Practice 6.1, Develop Programs and Evaluate Delivery Mechanisms, NACSLB. 2. This point is made in the GFOA Recommended Practice, “Application of Full-Cost Accounting

to Municipal Solid Waste Management Activities (1998).” See Appendix C for the full text of the practice.

3. See Chapter 3 for guidance on estimating the full cost a government service. 4. See the end of Chapter 5 for a discussion of discounting to the present value.

Make-versus-Buy Decisions ! 117

5. This discussion on estimating the costs of contract administration is based on Lawrence Martin, How To Compare Costs Between In-House and Contracted Services (Los Angeles, California: Reason Foundation, March 1993).

6. This exhibit is adapted from: OMB Circular No. A-76 (Revised), Executive Office of the President, Office of Management and Budget, 2003, p. 50.

7. This exhibit is adapted from: Office of the Texas State Auditor, “Guide to Implementing the Competitive Cost Review Program” (Austin, Texas: 1992).

8. There is not agreement on whether these costs should be included in a cost analysis. The Treasury Board of Canada recommends that severance pay and vacation benefits should not be included in transition costs. (“Stretching the Tax Dollar: Make or Buy,” Treasury Board of Canada Secretariart, Innovative and Quality Services, Financial and Information Management Branch.)

9. There are three main methods of conducting a sensitivity analysis. One method is to recalculate the result of an analysis in a pessimistic, expected, and optimistic scenario. A second method is to recalculate the result of an analysis many times by testing each uncertain assumption over a wide range of values. A third method is to calculate a probability distribution for the result of an analysis. For an additional discussion of sensitivity analysis, see R. Gregory Michel, Decision Tools for Budgetary Analysis (Chicago, Illinois: Government Finance Officers Association, 2001), pp. 82-87.

10. The City of Cincinnati, Ohio, the State of Texas, and the federal government have all established 10 percent thresholds.

C h a p t e r 9

Changing the Level of Service

Every budget season, local governments must decide whether to ex- pand or reduce funding to various programs. If budgets are tight, a gov- ernment may want to determine what the cost savings would be if grass cutting were reduced from four times per month to twice a month. If burglaries increased dramatically, a city government may want to know what the additional costs would be if the city added five new offi- cers to the force. Each of these scenarios involves an analysis of the cost (or savings) of increasing or decreasing the level of service. Three differ- ent approaches can be used to perform this analysis:

• Marginal cost analysis;

• Cost behavior approach; and

• Average costing.

MARGINAL COST ANALYSIS Marginal cost measures the increase or decrease in cost due to a change in the level of service. Marginal cost focuses only on the change in cost rather than the total cost. If an expansion or reduction is being consid- ered, it is not necessary to estimate the total cost of the service but only those costs that will change with the proposed change in service. For ex- ample, the marginal cost of increasing the frequency of trash collection

119

from once a week to twice a week would be the cost of the additional man-hours, fuel, etc. necessary to pick up trash a second time during the week. The marginal cost of extending library hours from 8 p.m. to 10 p.m. would be the costs of staffing and operating the library during those two hours.

How to Calculate the Marginal Cost. To estimate the marginal cost of an increase in service, first, itemize all of the additional resources required by the increase (e.g., 350 labor-hours, 120 gallons of fuel, $6,000 in fees, etc.). Then, convert each resource to a dollar amount by estimating the unit cost of each resource and multiply the amount of each resource by its unit cost. Finally, sum the costs of each resource and add or subtract any change in fixed costs caused by the increase in service. (The fixed cost will remain the same unless the increase in quantity requires the government to increase its capacity by building larger facilities, purchasing additional vehicles, etc.)

COST BEHAVIOR APPROACH The cost behavior approach uses the concepts of fixed and variable cost to determine the relationship between the output and total cost of a ser- vice. Once this relationship is known, it is relatively easy to predict the total cost at different levels of output.

The typical relationship between the output and total cost of a ser- vice can be shown graphically by a sloped line with total cost on the Y axis and output on the X axis. The Y-intercept of the line is the fixed cost and the slope of the line is unit variable cost. If you remember your high school geometry, a line can be represented by the equation: Y = m(X) + b. Following this format, the equation of a line for cost analysis pur- poses is: total cost = (unit variable cost) x (output) + fixed cost. Exhibit 9-1 shows the relationship between output and total cost in the form of an equation and a chart.

To estimate the cost of increasing the level of service using the cost behavior approach, first, estimate the fixed cost and unit variable cost of the service to generate the equation in Exhibit 9-1. (Chapter 4 provides several methods of estimating fixed and variable costs.) Then, substi- tute the new level of output into the equation to calculate the new total cost at that level of output.

120 ! Cost Analysis and Activity-Based Costing for Government

AVERAGE COSTING Another method of estimating the costs of an expansion or decrease in service is average costing. Average costing uses the current unit cost to estimate what the total cost would be at a new level of output.

To use average costing, first determine the average unit cost of a service by dividing the total cost by the number of units of service (e.g., $200,000 / 10,000 residents = $20 per resident). Then, estimate the ex- pected increase in units (e.g., 1,000 additional residents). Finally, multi- ply the average unit cost of the service by the expected increase or de- crease in units (e.g., $20 per resident x 1,000 new residents = $20,000).

Average Costing Can Distort the Actual Change in Costs. Aver- age costing is a much simpler method than calculating the marginal cost. However, when a service has significant fixed costs, average cost- ing may provide a misleading picture of the actual change in costs. In essence, average costing assumes that all costs behave like variable costs, such that each additional unit of service will increase the total cost by the average unit cost. The problem is that fixed costs cause costs to change in ways that are different from the average unit cost. For exam- ple, if a government has deficient capital capacity, 1,000 additional resi- dents may cause a government to make substantial capital expendi- tures such as building a new fire station. In this case, the government

Changing the Level of Service ! 121

Exhibit 9-1 ! Cost Behavior Equation and Chart

would experience a jump in total costs much higher than the average unit costs. Thus, average costing would underestimate the cost of adding 1,000 new residents.

If a government has excess capital capacity and fixed costs make up a high proportion of the total costs, the total cost may change little when 1,000 new residents are added because the new residents use the excess capacity at no cost. In the case of excess capacity, average costing would overestimate the cost of adding 1,000 new residents. As a general rule, average costing should not be used if fixed costs make up a significant portion of total costs.

Exhibit 9-2 illustrates how average costing can sometimes distort the actual change in costs when there is an increase in the level of ser- vice. The first chart shows an example in which average costing under- estimates the total cost when there is deficient capital capacity. In this example, a budget analyst uses average costing to predict the total cost of a service if the number of residents increases from 1,500 to 2,000. The estimate is based on the average cost at 1,500 residents, which is just be- low the threshold at which the city requires a substantial increase in capital capacity. The second chart shows an example in which average costing overestimates the total cost when there is excess capital capac- ity. In this example, a budget analyst uses the average cost at 500 resi- dents to predict the total cost of a service if the number of residents in- creases from 500 to 1,500. In both of these examples, fixed costs make up a high proportion of the total costs.

122 ! Cost Analysis and Activity-Based Costing for Government

Changing the Level of Service ! 123

Exhibit 9-2 ! Marginal Costing Versus Average Costing