In February 2016, a federal judge in Albuquerque, NM, approved a $1 billion settlement between the Obama administration and nearly 700 Native American tribes. This settlement ended a decades-long class action lawsuit over how the Bureau of Indian Affairs (BIA) had distributed aid to tribes since the mid-1970s.

This case came about because of some disagreements over how to measure costs. For over 150 years, the BIA was directly responsible for most of the health care, education, economic development, and other core services delivered on Native American reservations. But then, starting in the mid-1970s, it shifted its focus from direct service provision to helping tribes become self-sufficient. Instead of managing services, it redirected its resources toward training, technical assistance, and other efforts to help tribes launch and maintain their own services.

To make that transition, BIA re-classified many of its activities as “contract support costs.” This change was not just semantic. Funding for direct BIA-administered services is part of a regular federal budget appropriation. That appropriation was stable and predictable. By contrast, funding for support costs on federal government contracts is variable and is often subject to renegotiation. Perhaps not surprisingly, BIA spending declined steadily under this new capacity-building model.

Tribes across the US argued that by re-classifying many of BIA’s costs, the federal government gave itself permission to slash BIA’s budget without Congressional approval. The tribes alleged that this simple cost measurement maneuver allowed BIA to operate well outside its authority and inflict substantial harm on Native Americans around the country. BIA argued that the cost reclassification was a standard accounting change that had been happening across the federal government for decades. The case was ultimately settled for far less than the tribes requested. Still, the federal government did agree to re-classify contract support costs as direct service costs, for which federal funding is far more transparent and predictable.

This case illustrates the central point of this chapter. How we define and measure costs matters tremendously. In this instance, cost measurement is not just a technical exercise; it had real impacts on the lives of hundreds of thousands of Native Americans. The same is true for virtually all public services. How we define, measure, and plan for costs affects which services we deliver and how we deliver them.

What is Cost Analysis?

If you’ve ever flown on an airplane, there’s a good chance you know Boeing. The Boeing Company generates around $90 billion each year from selling thousands of airplanes to commercial and military customers around the world. It employs around 200,000 people, and it’s indirectly responsible for more than a million jobs through its suppliers, contractors, regulators, and others. Its main assembly line in Everett, WA, is housed in the largest building in the world, a colossal facility that covers nearly a half-trillion cubic feet. Boeing is, simply put, a massive enterprise.

And yet, Boeing’s managers know the exact cost of everything the company uses to produce its airplanes, every propeller, flap, seat belt, welder, computer programmer, and so forth. Moreover, they know how those costs would change if they produced more or fewer airplanes. They also know the price at which they sold that plane and the profit the company made on that sale. Boeing’s executives expect their managers to know this information in real time if the company is to remain profitable.

Cost accounting (also known as managerial accounting) is the process of creating information about costs to inform management decisions.

Managers need good information about costs to set prices, determine how much of a good or service to deliver, and manage costs in ways that make their organization more likely to achieve its mission. Managers in for-profit entities like Boeing have instant access to sophisticated cost information that would assist with those types of decisions. But managers in the public and non-profit sectors usually don’t. There are many reasons for this:

• Large parts of the public sector do not produce a “product,” but they deliver services like counseling juvenile offenders, protecting the environment, or housing people experiencing homelessness. Sometimes, we know the “unit” of production and can measure costs relative to that unit. In the case of counseling juvenile offenders, we might think about the cost per offender to provide those services. But for services without a clear “end user,” like environmental protection, this analysis is much more difficult.
• Most (usually around 80 percent) of the costs incurred by a typical public sector organization are related to people. A parole officer will see many different types of parolees. Some will demand a lot of attention and follow-up. Some will need next to none. Some parole officers are comfortable giving each case equal time and attention. Others are not. This type of variability in how and where people spend their time, and as a result, where labor costs are incurred, can make cost analysis quite difficult.
• Employees often work across multiple programs. A program manager at a non-profit organization might work across two programs funded by two different grants from two funding agencies. Unless that program manager allocates their time exactly equally across both programs – and that’s unlikely – we can’t know the exact cost of each program without a careful study of how and where that employee spends their time.
• Public services often share buildings, equipment, vehicles, and other costs. It isn’t easy to know the full cost of services without a system to track which staff and programs use exactly which resources.
• Good cost analysis has no natural political constituency. Careful cost analysis requires substantial investments in information technology, staff capacity, accounting information systems, and other resources. Most taxpayers and funders would rather see that money is spent on programs and services to help people in the short term rather than on information systems to analyze and plan for future costs.

These are just a few of the many barriers that prevent public organizations from acting more like Boeing, at least with respect to cost analysis.

And yet, good cost analysis is critical to public organizations. Public financial resources are finite, scarce, and becoming scarcer. Public managers must understand how and where they incur costs, how those costs will differ under different service delivery models, and whether that pattern of costs is consistent with their organization’s mission and objective. In this chapter, we introduce the core concepts of cost accounting and show how to apply those concepts to real management decisions.

At the outset, it is essential to distinguish between full-cost accounting and differential-cost accounting. Full cost accounting is the process of identifying the full cost of a good or service. Differential cost accounting – sometimes called marginal cost analysis – is the process of determining how the full cost of a good or service changes when we deliver more or less of it. Sound financial management requires careful attention to both.

Let us start with a simple example. Imagine a copying machine that is shared among three departments within the Environmental Health Department of a county government. Those three departments are:

1. Food Protection. This includes inspection and licensing of restaurants and other establishments that serve food. This program is designed to prevent outbreaks of food-borne diseases like E. coli, botulism, and Hepatitis A. Staff in this division make around 500 copies each day, mostly related to documenting restaurant inspections.
2. Animals and Pests. This includes animal control, rodent testing and control, and educational programs to promote pet safety and neutering/spaying. These programs are designed to prevent communicable diseases, including rabies, that vagrant animals most often spread. Staff in this division make around 250 copies each day, but that number can increase in the event of an outbreak of avian flu or other communicable disease.
3. Wastewater. The Wastewater Department is responsible for treating wastewater. Staff in this division issue water discharge permits to businesses and industrial operations and test water quality near wastewater discharge sites. These programs are necessary to prevent waterborne communicable diseases like cryptosporidium. Wastewater staff typically make around 100 copies daily but make up to 1,000 per day when processing complex industrial building permits. They process around six such permits each year.

As a manager, you would want to know what it costs to operate the copier and how those costs should be spread across the three departments. To put this question in the language of cost accounting, we want to know:

1. What is the full cost of operating the copier?
2. How should we allocate the costs of operating the copier across the three departments?

To answer these questions, we first need to know all of the different ways the copier incurs costs. A few come to mind immediately: paper and toner to make the copies, a lease or rental payment to take possession of the copier, and the occasional maintenance and repairs. A few might be less obvious: electricity to run the machine, space within a building to house the copier, and an office manager’s time to coordinate maintenance and repairs. We can observe many of these costs, but we’ll need to estimate or impute others.

The next question is how the departments should share these costs. Imagine, for instance, that they split those costs one-third for each department. This approach is simple, easy, and transparent. But what’s wrong with it? Each department makes a different number of copies, and each has a different workflow related to the copier. These departments also have different potential “economies of scale” for copying. Also, keep in mind that the Animals and Pests department will need “emergency” or “surge” capacity while the other two departments may not. So, if an even distribution is not the most appropriate, what is? With careful attention to cost accounting methods, we can begin to address these and other questions.

Full Cost Accounting

Measuring Full Cost: The Six-Step Method

To answer the question “what does this service cost?” cost accountants follow a six-step process. Each step of this process is driven by policies and procedures that are defined by an organization’s management:

To answer the question “what does this service cost?” cost accountants follow a six-step process. Each step of this process is driven by policies and procedures defined by an organization’s management.

1. Define the cost object. The cost object is the product or deliverable for which costs are measured. Service-oriented public organizations typically define cost objects in terms of the end user or recipient of a service. Examples include the cost to shelter a person experiencing housing insecurity for an evening, the cost per counseling session delivered to recovering substance abusers, the cost to place a family in affordable housing, and so forth.
2. Determine cost centers. A cost center is a part of an organization that incurs costs. It could be a program, a unit within a department, a department, a grant, a contract, or any other entity defined for cost accounting purposes. Generally, cost centers work best for homogeneous groupings of activities.
3. Distinguish between direct and indirect costs. Direct costs are connected to a specific cost center. In fact, they are often called “traceable costs.” Examples include salaries for staff who work entirely within a cost center, facilities and supplies used only by that cost center, training for cost center-specific staff, etc. Many public organizations further stipulate that a cost is direct to a cost center only if that center’s management can control it. Indirect costs apply to more than one cost center. They include shared facilities, general administration, payroll processing services, and information technology support. Some managers call them service center costs, internal service costs, or overhead costs because they are usually for support services provided within an organization.
4. Choose allocation bases for indirect costs. One of the main goals of full-cost accounting is to distribute indirect costs to cost centers. This follows from the logic that all direct costs require support from within the organization. An allocation basis is an observable metric we can use to measure the relationship between direct and indirect costs within a cost center. For example, a non-profit might allocate indirect costs according to the number of full-time equivalent (FTE) employees within a cost center or the percentage of the organization’s overall payroll earned by employees within that cost center. The full cost of any service is the sum of direct costs plus the unit’s share of indirect costs. 
5. Select an allocation method. There are two main methods to allocate or apportion indirect costs to cost centers. One simply calls indirect costs their own cost centers and plans accordingly. For instance, a non-profit could choose to call the executive director its own cost center. In that case, it would plan for and report the executive director’s salary, benefits, and other costs as a stand-alone entity rather than allocate those costs as an indirect cost to other direct service cost centers. A more common approach is to allocate by a denominator that is common to all the cost centers that incur a particular indirect cost (see below).
6. Attach costs to cost objects. One of the biggest challenges for public organizations is that cost objects are usually people, and no two people are alike. For instance, a parole officer might have 30 clients, but each requires a different amount of time, attention, and counseling. When the cost per client varies a lot, the cost accounting system should reflect those differences, usually by applying different overhead rates or percentages to different types of clients.

Let’s illustrate some of these concepts with the copier example. To begin, assume that the copier is a cost center. Services like copying, information technology, and payroll exist to serve clients within the organization, so they are called service centers. One of the goals of cost accounting is to allocate service center costs to mission centers that are more directly connected to the organization’s core programs and services. In this case, we can assume Food Inspection, Animals and Pests, and Wastewater are mission centers that will ultimately receive costs allocated from the copier service center. Given those assumptions about cost centers, we can assume the cost object for the copier service center is the cost per copy.

With those assumptions established, we can define direct and indirect costs for the copier service center. Direct costs include paper, toner, rental/lease fees for the copier, and machine maintenance. These costs are incurred exclusively by the copier. Electricity, building space, and the office manager’s time are indirect costs. They are incurred by the copier cost center and by other cost centers.

To illustrate, the table below lists some details on the copier’s costs.

INDIRECT COST ALLOCATION: COST DRIVERS AND ALLOCATION BASES

To find the full cost of the copier cost center, we’ll need to find some way to allocate its share of indirect costs to it. A good cost allocation scheme follows from a clear understanding of an organization’s cost drivers. A cost driver is a factor that affects the cost of an activity. A good cost driver is a reliably observable quantity that shares a consistent relationship with the indirect cost in question. Fortunately, we have an intuitive cost driver for the copier cost center: the number of copies.

Ideally, we can allocate indirect costs according to their key cost driver(s). An allocation basis is a cost driver common to all the cost centers that incur an indirect cost. For building space, for example, we might find the portion of the total building space that is occupied by the copier and allocate a proportionate share of the building space costs to the copier copy center.

For example, this county government allocates electricity costs to different cost centers per kilowatt hour (kWh). Sometimes it is feasible to measure electricity use with this level of precision, and sometimes it is not. Assume that the copier has an individualized meter measuring its electricity use.

This government allocates building space costs per square foot. This assumes it has a reasonably sophisticated way to measure how much space each cost center uses. Allocations by space can be contentious because not every unit uses space in quite the same way to accomplish its mission. For instance, most Food Protection staff spend most of their time in the field inspecting restaurants. They report to the office at the beginning and end of the day but infrequently during the day. This is quite different from the Animals and Pets center, where most of the staff spend most of their time in the office.

These figures also assume that the government allocates the office manager’s time to individual cost centers. The office manager can do this if they track the amount of time they spend on work related to each cost center. Some public organizations have systems, often based on a billable hours concept, similar to those used by professionals like lawyers or accountants. Many do not.

Also, note that the copier cost center does not receive overhead from other service centers. We don’t see, for example, that the copier center is allocated a portion of the county administrator’s salary, insurance expenses, or other organization-wide indirect costs. This is a policy choice. Some public organizations do not require service centers to receive overhead costs, mostly to keep down the rates they must charge their internal clients. Many state and local governments have budgeting rules that state programs – if independently financed or paid for with specific fees or charges rather than general fund resources – do not need to allocate their indirect costs or receive an indirect cost allocation.

That said, many public organizations allocate overhead to internal cost centers. In fact, when they do, they typically use the step-down method of allocating indirect costs. That is, they allocate organization-wide indirect costs to all cost centers first, then allocate service center costs, including their portion of the organization-wide indirect costs, to the mission centers.

With those assumptions in place, recall that:

• The Food Protection mission center averages 500 copies each day. Assuming 260 workdays/year, that is 130,000 copies (i.e., 500 copies x 260 days). In this case, the 130,000 copies are the relevant range or the amount of activity upon which our cost analysis is based. Assuming the Food Protection mission center would require twice as many copies, our per-unit costs and cost allocations would look quite different. Good cost analysis follows from clear, defensible assumptions about the relevant range of activity that will drive costs.
• The Animals and Pests mission center makes 250 copies each day but makes many more in the event of a communicable disease outbreak. Assuming no outbreak, that is 65,000 copies (i.e., 250 copies x 260 days).
• The Wastewater division makes 100 copies daily and up to 1,000 copies around six times yearly when processing complex permits. Let’s assume a typical surge in copies for a complex permit will last for five days. That would mean 230 typical days and 30 “surge days” (i.e., six permits x 5 days/permit). So total number of copies would be around 53,000 (i.e., typical day copies of 100 copies X 230 days + and surge days copies of 1,000 copies X 30 days).

From these figures, we can determine that the copier will make 248,000 copies each year (i.e., 130,000 copies + 65,000 copies + 53,000 copies).

If we divide the full annual cost of the copier by the number of copies ($21,380/248,000 copies), we arrive at a unit cost of $0.0862/copy (i.e., 8.62 cents per copy).

With those full costs established, we must ask how the Environmental Health Department should allocate the full costs of the copier cost center across the three mission center departments. Fortunately, this is easy to do because the copier cost center has a clear cost object (cost per copy), and each department/cost center measures the number of copies it makes. As a result, each department would be assigned copier center indirect costs at a rate of 8.62 cents/copy.

• Food protection would be assigned $11,206 (i.e., 130,000 copies X $.0862/copy).
• Animals and Pests would be assigned $5,603 (i.e., 65,000 copies X $.0862/copy).
• Wastewater would be assigned $4,569 (i.e., 53,000 copies X $.0862/copy).
  Due to a rounding difference, the total adds up to $21,378.

With an appropriate allocation basis, it is possible to allocate any indirect costs in a similar way.

This copier example also shows why the cost center and cost object are essential to the cost allocation process. For instance, imagine that the copier was defined not as one cost center but as separate cost centers for large copying jobs (say, more than 500 copies) and small copying jobs, or for color copies vs. black and white copies. This would also require different cost objects, such as the “cost per black and white copy” or “cost per color copy.” The cost per black and white copy would presumably be less than the cost per color copy, and the cost per copy for large print jobs would presumably be less than that for small jobs. Different cost centers, cost objects, and allocation methods can mean substantially different answers to the question, “What does a copy cost?”

One potential drawback of the step-down method is that it allows “double counting” or “cross-allocation” of service center costs to service centers already allocated to mission centers. For example, recall that the annual full cost of the copier service center was $21,380. That full cost incorporated the indirect costs of the office manager’s time to manage the copier. Under the step-down method, the cost of the office manager’s time is allocated to the copier cost center, and the copier cost center costs are then allocated across the mission centers. But what happens if the office manager makes copies? Under this indirect cost allocation scheme, the office manager’s copies would not be reflected in the total volume of copies made, and the office manager would not receive any of the copy center’s costs. As a result, the mission centers subsidize the office manager’s copying by absorbing a larger share of the copy center’s costs.

In this particular example, those subsidies are a negligible amount. But in many other scenarios, cross-allocation of service center costs can significantly impact the full cost of a good or service. For instance, imagine a non-profit organization with three mission centers, a service center for the executive director, and a human resources service center. The human resources service center spends most of its time interacting with the executive director, as is often the case in small non-profits. If this organization uses the typical step-down approach, and it first allocates the executive director’s costs to the other service centers, then the full costs of the three mission centers will include a sizable subsidy for the costs of the executive director-human resource center’s interactions.

To address this problem, many public organizations instead use the double-step-down method. After each service center/department’s costs have been allocated once, each center/department’s cost not included in the original allocation is totaled and allocated again. To illustrate, let’s return to the copy center-office manager example above. If this allocation were done with the double-step-down method, the office manager’s copies would be included in the total copy figure. The copy center would first allocate its costs to the mission centers, excluding the office manager’s copies. Then in a second step, the office manager’s share of the copying costs would be allocated to the mission centers separately. This double-step method minimizes the cross-allocation of service center costs. For more on cost allocation methods for governments, see several chapters in Zach Mohr, ed. (2016), Cost Accounting in Government: Theory and Applications (New York: Routledge); also see the chapter “Cost Accounting and Indirect Costs” in Dittenhoffer and Stepnick, eds. (2007), Applying Government Accounting Principles (Lexis-Nexus Publishing).

Indirect Cost Allocation: Indirect Cost Rates

Cost drivers and allocation bases work well when the service has a clear cost objective and a measurable unit of service. Most public organizations, as described above, don’t have this luxury. Many don’t deliver services with measurable outcomes. Most public organizations’ costs relate to personnel, and personnel costs are not distributed evenly across clients or cases. Moreover, a growing number of public services are delivered through partnerships and collaborations where it’s often unclear how costs are incurred and murkier how those costs ought to be allocated across the partner organizations. Traditional cost allocation methods often don’t work in the public sector for these and many other reasons. And yet, it’s still critically important to measure and properly account for full costs, especially indirect costs that can be difficult to measure.

To address these problems, many public organizations rely on indirect cost rates. An indirect cost rate is a ratio of indirect costs to direct costs. For instance, a city police department might determine that its indirect cost rate is 15 percent. That means that for every dollar of direct costs like police officer salaries and squad cars, it will incur 15 cents of payroll processing, insurance, procurement expenses, and other indirect costs.

Let’s illustrate this with a more detailed example. Surveys show that many local public health departments would like to offer more services related to hypertension outreach and management. Chronic health conditions like heart disease and diabetes are known to be related to high blood pressure, so better management of high blood pressure can affect public health in a substantial, positive way. However many citizens, especially those without health insurance, cannot access regular blood pressure screening and other services needed to identify and manage hypertension.

Say, for example, that Smallville County and Riverdale County would like to launch a new, shared hypertension prevention and management (HPM) program. Neither currently has a formal program in this area, but both offer services through a patchwork of partnerships with local non-profits. Smallville County has roughly twice the population of Riverdale County, and Riverdale’s per capita income and property values are 30 to 40 percent higher than Smallville’s.

What does it cost to deliver this service? As with most public health programs, the main costs will be related to personnel, namely public health nurses, outreach counselors, and nutritionists. The program will also require space and other overhead costs. The outreach and education components will require advertising, travel, and other costs. For a service-sharing arrangement to work, the two counties must decide how to share these costs.

Suppose the counties also agree in advance to share costs evenly. This approach is simple and straightforward. However, it ignores many of the program’s underlying cost drivers. Smallville has a much larger population than Riverdale, so more participants will probably come from Smallville. Simply splitting these costs “50-50″ means Riverdale will likely subsidize Smallville, an arrangement Riverdale’s leaders might find unacceptable.

So, what’s the alternative? Smallville could bill Riverdale for each Riverdale resident who participates in the program. They could use an allocation basis like population or assessed property values. A more cutting-edge scheme might be to share the costs according to the incidence of the chronic diseases the HPM program is designed to prevent. Each of these strategies demands a trade-off. Some are simpler but at the expense of fairness. Some require additional data that might result in an expensive cost measurement process or one that is not feasible. Others are more feasible but might place costs disproportionately on the population the program is designed to serve.

To begin, let’s assume Smallville will structure the new HPM as a cost center within the Health Behaviors division of its Public Health department. Let’s also assume that since HPM’s main “deliverable” will be blood pressure screening, it will define its unit cost as the cost per blood pressure screening performed.

Given those assumptions, Smallville’s budget analysts estimate that for the first year of operations, the HPM program will serve 400 clients, and its costs will include:

But the much more important question is how we account for the indirect costs and for the direct costs that are more difficult to observe. The lower part of this table outlines those costs. When we include the indirect labor and indirect non-labor costs, we see the full cost of the program increases to $785,997, or $1,965/client. Or, put differently, the full cost of the program increases by more than 30 percent if we include all the indirect costs in our estimate of the full costs. Recall that Smallville County plans to bill Riverdale County for its share of program costs. If Smallville bills only for the direct costs, it undercharges Riverdale by nearly 30 percent. That is why it is important to measure full costs, especially when pricing services or requesting reimbursements for expenses incurred.

What about potential indirect cost rates? According to these figures, 74 percent of the full cost to employ a nurse is direct labor costs, and 10 percent is direct non-labor costs (i.e., mileage, outreach, and other direct non-labor). It follows that the remaining 16 percent is indirect costs related, in a predictable way, to those direct costs. Each nurse and health counselor will be insured, have their payroll processed by the payroll office, occupy space, and so on. If those figures are predictable, we can assume the current indirect cost rate for nurses is 16 percent. For the health counselor, the direct costs are a bit higher at 76 percent. Direct non-labor costs are 11 percent, and indirect costs are 14 percent.

In practice, this means that in future budgets, the HPM program could assume that for every dollar it will spend on nurse salaries, it can expect to incur 16 cents of indirect costs, and for every dollar it will spend on health counselor salaries, it can expect to incur 14 cents of indirect costs (i.e., indirect cost rate).

Some organizations compute indirect cost rates based only on direct labor costs. In that case, the rate for nurses would be the indirect cost rate would be 22.6 percent – i.e., $10,456+$4,799)/$67,500. For healthcare counselors, the indirect cost rate would be 22.8 percent.

We can also consider an indirect cost rate for the entire HPM program. For that, we compare the total indirect costs to the total direct costs. Total indirect costs for the nurses are $114,413; for the counselor, $13,318. That means total indirect costs are $127,731. Total direct costs are $573,390 for the nurses and $84,876 for the counselor, for total direct costs of $658,266. The indirect cost rate would be 19.4 percent (i.e., $127,731/$658,266). Again, all these figures assume the HPM program serves 400 clients.

Information about indirect cost rates is relevant to many types of decisions. For instance:

• HPM staff might compare their indirect cost rate to the rates of other programs within Smallville County. If its rates are noticeably higher or lower, it might more carefully review its cost structure and how it manages its costs. If its rates seem grossly out of line with other units, it might request an additional review by Smallville County’s budget staff.
• The counties might use these rates when applying for federal or state grants or for support from philanthropic foundations to support the HPM program.
• The counties might eventually decide to contract out some or all of HPM’s operations to a non-profit healthcare provider. In that case, these rates would be a focal point for negotiating the per-client rate at which the counties would reimburse a prospective contractor.

• Other governments might review these rates as an initial indicator of whether they can afford their own HPM program.

Cost Sharing Alternatives

Traditional cost allocation works best when it’s possible to observe when and where all the costs are incurred. When that information is unavailable, as is often the case for partnership arrangements that span multiple organizations, there are several alternative ways to organize a cost allocation plan. To illustrate, assume the full annual cost of HPM was $800,000 and that Smallville County must bill Riverdale County for Riverdale County’s share of those costs.

• Equal share. Total costs are divided equally across all participating partners. This is more typical for informal arrangements. It’s also common for services where it is unclear who “receives” or “uses” the service or to observe all the relevant indirect costs predictably and consistently. In equal share approaches, one partner often subsidizes the other, sometimes unknowingly. In the HPM example, Smallville would keep $400,000 of the costs and bill Riverdale its equal share of $400,000.
• Per capita. Total costs are divided by the population proportion in each partner jurisdiction. This approach is good for services without an observable “client” or discrete individual services. It’s less useful when population size is not the best cost driver or when the populations involved are different on some key characteristic that might affect the utilization of the service in question. Per capita sharing is often the most transparent way to share costs. In the case of HPM, recall that Smallville’s population is 240,000 and Riverdale’s population is 160,000. In other words, 60 percent of the population served resides in Smallville and 40 percent resides in Riverdale. Under a per capita model, Smallville would bill Riverdale $320,000.
• Cost Plus Fixed Fee. Personnel costs are often step-fixed costs, and it can be challenging to know when those costs will “step up” at higher levels of service (more on step-fixed costs below). To account for that uncertainty, some cost allocation strategies call for non-weighted cost sharing plus some fixed periodic fees. The fee part of the plan is designed to buffer the sharing arrangement against the uncertainty surrounding step-fixed costs. For HPM, one potential application of this method would be for the counties to share costs per capita but for Smallville to receive an annual payment of $35,000 at the start of its fiscal year to compensate in advance should it need to hire an additional nurse during the year. The cost-plus fixed fee model can also be used where overhead costs (e.g., space, utilities, administration, and accounting) would be shared one way, and incremental costs (e.g., costs for lab work or medical supplies) are charged based on volume.
• Ability to Pay. Some cost allocation arrangements are designed to make a service available where citizens and clients are otherwise unable to pay for it. In these cases, allocating costs according to the ability to pay makes sense. We can approximate the ability to pay through assessed property values, median household income, or other measures of income or wealth. In the HPM example, consider the following scenario: Riverdale’s median household income is $50,000, and Smallville’s is $40,000. Riverdale has a smaller population but is wealthier. In this case, Riverdale’s median household income ratio to Smallville’s is 1.25 ($50,000/$40,000). This is commonly known as a wealth factor. Recall that an equal share allocation is $400,000 for each jurisdiction. If 50-50 share is adjusted by a wealth factor, Riverdale’s share of costs would be $400,000 X 1.25, or $500,000, while Smallville’s share would be $300,000.
• Prevalence. In this method, the parties share costs according to the prevalence of the public health problem the service is designed to address. In the HPM example, the partners could share the total program costs according to observed instances of diabetes or heart disease. The logic here is simple: diabetes and heart disease tell us something about the expected number of people with hypertension. If the prevalence of the disease is not known, the partners can use a proxy, like socioeconomic status, to project the anticipated need for services in each population. In the HPM example, Riverdale’s higher overall wealth suggests its residents are at lower risk for hypertension compared to Smallville residents. Sharing by prevalence adds substantial complexity because cost sharing is now based on data from a series of measurements unrelated to costs. In this case, those measurements are the incidence of disease or an indicator of socioeconomic status, which can be difficult to measure reliably, and other health-related behaviors like smoking or medication adherence. That said, this approach is especially good where population, property values, income, and other measures vary too much among sharing jurisdictions to offer meaningful comparisons. In this instance, we assume Smallville will have an estimated 12,740 cases of type 2 diabetes during the coming year, and Riverdale County will have an estimated 5,460 cases. This strategy considers each county’s share of the total incidence across both counties. According to that logic, 70 percent of the cases will be found in Smallville and 30 percent in Riverdale. Allocating costs this way leads to a share of $560,000 for Smallville and $240,000 for Riverdale. Some versions of prevalence also incorporate a moving average so that one community does not incur huge costs in a single year, and costs are recovered over time.
• Weighted formula. This plan addresses some of the big problems with the per capita sharing approach. For example, in a weighted formula approach, the participants might agree to share total costs according to a combination of population, median household income, usage, and other factors. By incorporating these other factors, the cost apportionment method will better reflect differences in fixed costs in urban vs. rural areas, differences in travel distances within each county, and a wide variety of other factors that may affect service delivery. For HPM, assume that Smallville and Riverdale decide to share costs according to a three-factor formula that incorporates population, ability to pay, and prevalence of type 2 diabetes. This formula reflects both counties’ shared understanding of the cost structure and cost drivers of the HPM program. The counties, realizing the difference in the prevalence of type 2 diabetes, agree to weigh that difference in prevalence more heavily in the cost-sharing formula. They agree to a three-factor formula where population accounts for 25 percent, prevalence is 50 percent, and ability to pay is 25 percent of the total costs allocated to each county. Recall that Smallville accounts for 60 percent of the population served by HPM, and Riverdale accounts for 40 percent. At the same time, Smallville accounts for 70 percent of the prevalence factor and Riverdale for 30 percent. We would apply that formula as follows:
  • Cheng County: $800,000 X ((.6 X .25) + (.7 X .5) + (.44 X .25)) = $488,000;
  • Duncombe County: $800,000 – $488,000 = $312,000

Cost Behavior

We know that what a service costs largely depends on how much of it we deliver. This is broadly known as cost behavior. Every cost type falls into one of three different cost behavior categories:

• Fixed costs do not change in response to the amount of service provided. In the HPM case, Smallville County owns some of its own blood pressure screening equipment, so the costs of acquiring equipment costs won’t change even if the HPM program delivers a lot more blood pressure screenings. There, however, is a caveat to this. It is reasonable to assume that fixed costs will not change within a relevant range. Relevant range frequently refers to capacity. If current demand exceeds capacity, we’ll need to incur additional costs to expand capacity (e.g., room size, equipment that is available, etc.).
• Step-Fixed Costs are fixed costs within a relevant range. Step-fixed costs are frequently personnel costs. A nurse, for example, can see eight patients per day, and a counselor can see five patients per day. They are paid a fixed rate, regardless of number of patients they see. That said, they will likely not have the time to see more patients than the current workload. Therefore, we need to hire an additional nurse if we expect to see more than eight patients daily and an additional counselor if we expect to see more than five patients daily. So, if nurse visits per day were 20, we would need to hire three nurses and have the excess capacity to see an additional four patients (i.e., three nurses x a maximum of eight patient visits = 24 visits per day).
• Variable Costs change directly in response to the amount of service provided. For the HPM program, this might include copies and other office supplies needed to process physician referrals and mileage required to travel to outreach sessions, among others.
• Mixed Costs (or Semi-Variable Costs) have both a fixed and variable component. Utilities and equipment lease agreements are frequently semi-variable, with a fixed monthly charge irrespective of use and rate per unit of service consumed. Returning to our example of the copier, the monthly rate is fixed and the per-page service fee would be variable.

The figure below illustrates the cost behavior concepts for a generic, hypothetical service. The horizontal axis is the quantity of service provided, and the vertical axis is the total cost. The horizontal line at $50 represents a fixed cost. It does not change, regardless of the level of service provided. The triangle-marked line identifies a variable cost. Here we see each additional unit of service increases the total cost by $10, and that change is constant from zero to three units of service. The line marked with squares shows a step-fixed cost. Here, the cost is fixed at $20 from zero to three units of service. Once we reach four units of service, that total cost steps up to $40, where it stays fixed until seven units of service.

It’s useful to think about a program or service with reference to these main cost behaviors. In fact, we can place most programs/services/organization units into one of six cost behavior categories. Those categories are outlined below, along with examples of each from typical non-profit organizations and government programs.

It’s immensely helpful to think about cost behavior when we have to make decisions about how to design and fund programs. Consider this simple example based on the previously mentioned HPM program.

HPM staff have some rough budget projections. Their program is expected to incur fixed costs of $800,000 and variable costs of $400 per client. The program has expanded a lot since it launched, and it now expects to serve 550 clients but could serve up to 600 with current staffing levels. Meanwhile, nearby Emerald County has offered to pay $750 per client to expand the program to include an additional 50 Emerald County residents. Should Smallville and Riverdale counties agree to partner with Emerald County on these terms?

HPM’s cost behavior is outlined in the table below. Given its projected fixed and variable costs, at 550 clients, the average per-client cost is $1,855. If HPM scales up to serve 600 clients, its average cost will decrease to $1,734 per client. However, that average cost of $1,734 is still much higher than the $750 per client that Emerald County offering. On the basis of “average” unit costs, this proposal is a definite “no-go” for Smallville and Riverdale.

However, keep in mind the relationship between fixed and variable costs. Recall that HPM staff have said they can add 50 more clients without taking on additional fixed costs. If that is true, then the new cost to add a client is only the additional variable cost. Put differently, the average cost of each client is $1,734, but the marginal cost, or the cost of a new client, is $400. If HPM is reimbursed $750 per client, the additional “profit” is $350. If HPM makes this decision “at the margin” or with reference only to the marginal cost, it should take the deal with Emerald County. This is a good example of a service with positive economies of scale; the marginal cost of each unit of service decreases as the volume of service delivered increases.

Of course, there are trade-offs here. At 600 clients, the HPM program will operate at full capacity. HPM staff will almost certainly have to spend less time with clients. This could lead to a decline in the quality of service and could even increase staff burnout and turnover. But if we look just at the marginal cost, it makes sense for Emerald County to join the program.

This example also illustrates the key concept of sunk costs. Many fixed costs are for capital items like equipment, land, and buildings that can be bought and sold. A public health department can, in concept at least, recover some of those costs by selling those capital items. However, HPM’s spending on employee salaries, training, insurance, and many other costs cannot be recovered. Those costs are sunk.

Some economists argue that sunk costs ought to be irrelevant to future decisions. In other words, at the margin, all that matters are the future, measurable, variable costs. Of course, this is difficult in practice. In the HPM case, scaling up to full capacity will mean additional stress on staff, and perhaps more importantly, it would mean giving up the opportunity to take on additional clients without taking on additional fixed costs. These costs are much harder to measure, but they are key components of decisions about cost sharing. The key takeaway here is that when considering a service-sharing arrangement, be sure to consider both the marginal costs and the opportunity costs.

Cost-Volume-Profit Analysis

So far, we have reviewed how public managers can identify the full cost of their services and how that full cost changes as they deliver more or less of a service. Those are crucial questions that all good managers can answer.

However, public managers must often confront a different question: What should we charge for this service? They are also routinely asked a corollary question: What volume of service should we deliver, given that service’s cost structure? To answer these questions, we turn to a particular set of concepts within differential cost accounting known as cost-volume-profit analysis (CVP). CVP is how an organization determines the volume of activity needed to achieve its profit or mission goal. It is the price it needs to charge to break even or make a profit or the cost limits that it must manage to achieve its profit or mission goal. CVP analysis is usually done for a particular program or service within an organization. The basic equation is:

Profit = Total Revenue – Total Costs

From this discussion so far, we also know that Total costs = Fixed costs + Variable costs. And since fixed costs are fixed, we can represent the cost equation as:

TC = a + bx

Where TC = total costs, a = fixed costs, and b = variable costs. We also know that total revenue is simply the price of service (p) times the volume of service delivered (x). That said, we can show the fundamental profit equation as:

Profit = px – (a + bx)

For any service, the break-even volume is the point at which total revenue (px) equals total costs (a + bx). To illustrate how we use this formula, let’s go back to the Environmental Health Department’s copier. Assume for the moment that the county government’s leadership wants to make copying more affordable, so it caps the price of copying at 7 cents per copy. At that price, how many copies must the copier center deliver each year to break even? In other words, what’s its annual break-even volume?

We know that the copier cost center’s fixed costs (a) include $6,000 for the machine rental, $900 for machine maintenance, and $1,500 for its space allocation. Let’s also assume electricity and the office manager’s time allocation are fixed costs of $180 and $100, respectively. So total annual fixed costs are $6,000 + $900 + $1,500 + $180 + $100, or $8,680.

Variable costs (b) are the largest cost items. Recall that last year the copier made 248,000 copies. Total paper costs were $10,000, so the per-copy cost for paper is ($10,000/248,000), or $.04/copy. Total printing cartridge costs were $2,700, so the per-copy cost for printing cartridges was ($2,700/248,000), or $.011/copy. These two variable costs together give us total variable costs of $.051/copy.

At break-even, profit = 0, so we can re-arrange the fundamental profit equation to px = a + bx. Since the price per copy is capped, per management’s policy, at $.07, we can then express this equation as:

$0.07x = $8,680 + $0.051x

To solve, we first subtract $0.051x from both sides, leaving us with $0.019x = $8,680. To solve for x, we divide both sides by $0.019, and we’re left with x = 456,842. In other words, at 7 cents per copy, the copier cost center’s break-even quantity is 456,842 copies. That’s nearly twice as many copies as it produced so far. Management might want to rethink this decision.

We can also express break-even quantity as a/(p-b). Arranged this way, the previous quantity is $8,680/($0.07 – $0.051), or $8,680/$0.019 = 456,842. The price minus variable costs is known as the contribution margin. In some ways, break-even quantity is finding how many units of volume we need to spread the contribution margin to arrive at break-even quantity.

Let’s ask a different question. Assume that the copier technician says the current copier is aging and will likely break down if asked to make more than 150,000 copies per year. At that volume, how should the copier cost center adjust its prices, so it continues to break even?

Here we re-arrange the formula as p = (a/x) + b. In other words, the break-even price is the fixed costs divided by the volume plus the variable cost per unit. For the copier cost center, this is ($8,680/150,000) + $0.051, or ($0.0579 + .051) = $0.1089. Put differently, if capped at 150,000 copies, the copier cost center would need to charge 10.89 cents per copy to break even. Once again, management should take a careful look before implementing this policy.

The copier cost center is a useful illustration but is also an outlier because most of its costs are variable. This is typical in for-profit manufacturing, logistics, and other industries but atypical among public organizations. For a more typical public organization example, let’s return to the HPM program.

Recall from earlier that HPM has estimated fixed costs (a) of $800,000 and variable costs (b) of $400/client. Say HPM’s management wants to keep the program relatively small to ensure a quality service, so it decides to limit its enrollment to 400 clients. At that volume (x), its break-even price is ($800,000/400) + $400, or $2,400/client. If management is willing to expand enrollment to 500 clients, the break-even price becomes ($800,000/500) + $400, or $2,000/client.

In another CVP scenario, management reports that governments and philanthropies will pay a maximum of $1,800 per client to participate in the HPM program. At that price, what is HPM’s break-even quantity? Once again, the formula is a/(p-b), or $800,000/ ($1,800-$400), or 571 clients. To break even at that comparatively low price and contribution margin ($1,800-$400, or $1,400), the program will need to serve substantially more clients than it has served recently.

The table below summarizes these various cost-volume-profit calculations.

CVP is a powerful tool that can directly illuminate many important management decisions. Most CVP today is done in spreadsheets to allow for maximum flexibility when exploring alternative cost scenarios, especially when analyzing services with step-fixed costs. That said, it’s still important to understand the basic concepts.

Cost Analysis and Flexible Budgeting

Once we’ve identified a program or service’s cost structure, we can take our first steps toward preparing a budget. A budget is simply a plan for what an organization wants to accomplish and the resources it will use to accomplish it. Budgets are important – for governments, an adopted budget is the law. It describes taxes and other resources the government intends to collect. Since taxpayers must pay those taxes, the government must spend those resources per the adopted or amended budget. Recall from the earlier discussion that the budget is one of citizens’ most important tools to hold the government accountable. Non-profits’ budgets don’t have the force of law, but they’re also critically important. Why? Because if a non-profit deviates substantially from its budget, its funders, and donors will question its ability to make good on its promises and accomplish its mission.

Budgets are made and organized in many ways (see the next chapter). Regardless of an organization’s overall budget process, effective cost accounting is paramount to effective budgeting and management, particularly at the program or service level. Public managers must know what portion of the costs they’re responsible for are fixed, variable, and step-fixed. They must also understand how different cost items connect to service delivery outputs and how their cost center is assigned indirect costs. And perhaps most important, they must understand how their program’s cost structure and cost behavior will change under different performance scenarios. That is why the best budgets are flexible budgets.

A flexible budget allows a manager to quickly and easily perform sensitivity analysis to explore how changes in key cost assumptions affect a program’s unit costs, total costs, and service-delivery volume.

Flexible budgeting follows a basic four-step process:

1. Identify types of output or activity. Sometimes those levels of activity correspond to cost centers. Sometimes they relate to programs or “lines of business” within cost centers. Sometimes they correspond to entire departments or divisions, especially if that department or division delivers a single good or service.
2. Collect cost and price data. For most budget-making, last year’s budget is the best source of data on costs and prices. Pay careful attention to past budget trends unless you’re tasked with developing a budget for a new service (and someday you will be!). Two basic types of costs are reflected in most public organizations’ budgets: operating and capital costs. Operating costs are incurred through the organization’s regular, year-to-year activities, like salaries, benefits, and training. Capital costs are related to capital assets like property, equipment, and buildings. Recall that the portion of a capital asset that is “used up” each year is reflected in the organization’s operating costs as depreciation expense.
3. Classify inputs. Every organization classifies its cost inputs a bit differently. For a simple budget, it’s appropriate to classify the spending inputs in broad objects of expenditure like “salaries” or “commodities” and to call the revenue inputs “service revenues,” “fees,” or some other appropriate broad label. Some organizations prepare budgets using the same chart of accounts used to prepare the basic financial statements. Others have much more detailed budget item codes to identify specific types of budget inputs. The key to a classification scheme is that it’s as consistent as possible. Once you make a budget with it, make every budget with it.
4. Develop a cost rate or unit cost for each key budget input. A cost rate measures the resources needed to staff a service, usually expressed as a unit of output. A unit cost is the cost to acquire or produce a unit of a good or service.

Personnel costs are the largest and most visible budget inputs for most public organizations. That’s why effective budgeting for public organizations starts and often ends, with careful attention to budgeting for personnel.

So, let’s illustrate this four-step process with a personnel budgeting example.

Cooks can prepare three meals per day for a maximum of 50 clients per day and drivers can deliver three meals per day for a maximum of 35 clients per day. How many cooks and drivers do you need to hire and what are the related personnel costs for providing three meals per day to 200 clients?

What about a scenario where most of the key input items are variable costs? Here we would use the unit cost method. The steps for that method are as follows:

1. Project the demand for the service
2. Estimate the resources consumed per unit of output
3. Determine the average cost per unit of output
4. Multiply these three items together to determine the appropriate budgeted amounts

To illustrate, let’s return to the HPM program and focus on the counselor’s travel costs. Recall that the counselor travels to deliver outreach programs on healthy eating and active lifestyles designed to prevent hypertension among Smallville and Riverdale County residents. Let’s assume that for the coming fiscal year, the number of programs will be a bit lower than in previous years but robust nonetheless:

HPM staff project 175 outreach programs

The average mileage per outreach program is 75 miles, and the rate per mile is $0.40.

175 programs X 75 miles per program X $0.40/mile = $5,250

In this case, the “unit cost” and the “object of expenditure” are the same– individual outreach programs. And once again, we’d ideally set this analysis up in a flexible spreadsheet-based budget.

BUDGET VARIANCE

The difference between actual results and budgeted results is known as a budget variance. Effective public managers understand when, where, and why variances happen. They also use insights from variances to identify inefficiencies and other current management concerns and how they make future budgets.

The simplest variance analysis approach is to compare budgeted expenses to actual expenses. For example, suppose the HPM program budgeted $759,000 for FY15, but actual spending for the year was $789,000. This would be a total variance of ($759,000/$789,000) = .962, or 96.2 percent. In other words, budgeted spending was only 96.2 percent of actual spending. This is a negative variance or unfavorable variance of 3.8 percent, and it directs HPM’s management toward a few important questions about HPM’s operations. Did HPM serve more clients than expected? Did it spend more than expected on variable costs like mileage or supplies? Was the contract for communications and outreach properly structured? Are its indirect cost allocations fair and consistent with its actual operations? To address these questions, of course, it would help to compute variances on individual line items (known as line-item variances). Larger organizations often compute department variances.

Revenue-side variances demand a different type of interpretation. Imagine, for instance, that HPM budgeted for revenues of $800,000 but collected $750,000. The variance here would be ($800,000/$750,000) = 1.0667, for a negative variance of 6.667 percent. Here, Smallville County’s HPM management should ask a different set of questions. For instance, is HPM collecting all its revenues? Is Riverdale County slow in reimbursing Smallville County?

Positive variances beg a different set of questions. For example, say HPM’s total spending was just $700,000, compared to its budgeted spending of $759,000. This variance is ($759,000/$700,000) = 1.084, for a positive variance of 8.4 percent. But what does positive mean here? Was management able to drive down costs and deliver the expected volume of service at lower overall costs? Or did it not deliver as much of the service as expected?

Cost Structure and Management Strategy

By now, it should be clear that the principles of cost behavior, differential cost analysis, and cost-sharing suggest various management strategies. One of the best approaches to manage costs and bolster profitability for organizations with mostly fixed costs is to “scale up.” Since fixed costs are fixed, one way to manage them is to spread them across the largest possible volume of service. However, for organizations with mostly variable costs, scaling up will simply increase variable costs. The better approach in that circumstance is to invest in new technology, procurement processes, or other strategies that can drive down variable costs. With the core principles of cost analysis established, it becomes clear how a public organization’s cost behavior might be most effectively managed.

It is also essential to consider how an organization’s financial management strategy is contingent on the parts of its cost structure it can control. In other words, how much of its costs are direct and under the control of its management, and how much of its costs are indirect and assigned to it by management? For example, organizations with mostly direct costs often find themselves “scrubbing expenses” because they have the authority and ability to affect those costs. However, organizatioXns with mostly indirect costs are more likely to partner with other organizations to leverage economies of scale, diversify their potential customer/client base, or access new technologies or processes they might not have the authority to pursue on their own.

When we consider these two main parts of an organization’s cost structure – direct vs. indirect and fixed vs. variable vs. step-fixed – we begin to see a set of attenuated management strategies. The table below lists some of those strategies.

For example, in the bottom left section, we see financial management strategies commonly employed by organizations with primarily direct, step-fixed costs. For these organizations, the key is managing workflow. As their volume increases, so do their fixed costs. In turn, many organizations of this type work to spread their fixed costs within whichever step their fixed costs currently stand. That means maximizing staff-to-client ratios at all times, including queuing up clients with waiting lists or staggered enrollments before hiring new staff. Since most of those costs are direct, organizations have the latitude to employ these strategies.

By contrast, organizations with mostly indirect costs may not have the ability to enact new enrollment management policies. They must also “manage the step,” but for them, the better strategy is to narrow or limit the scope of services or to find new clients that can lead to more predictable enrollment patterns. In short, managing direct step-fixed costs is quite different from managing indirect step-fixed costs.

This table is by no means a comprehensive list of strategies, but it does orient you toward the interesting and nuanced relationship between cost structure and financial management strategy.

RELATED CASE STUDIES

Case 5.1: Heartland Community College
Case 5.2: Caring for Coho’s Seniors
Case 5.3: City of Hamtramck Snow Removal
Case 5.4: The Cascadia Hearing School