Print Share Home

Cost Analysis Outside Manufacturing - Information for Evaluating Service Expenses


John Mothersole
John Mothersole, Manager, Cost Information Service, Standard & Poor's DRI, Washington, DC 202-383-3644,
Laura Hodges
Laura Hodges, Senior Associate, Cost Information Service, Standard & Poor's DRI, Washington, DC 202-383-3616,

85th Annual International Conference Proceedings - 2000 

Abstract. Getting a "good" or a "fair" price is often the principal measure of a buyer's performance. Yet trying to gather information on market prices or supplier costs is often quite difficult. This is especially true with regard to services, where every buy can be unique or customized. There are, however, a number of publicly available data sources that can be used to establish benchmarks for the comparative movement of prices and costs. This information can be useful in supplier negotiations, particularly in validating requested price increases. The trick is to focus your search so that the analysis is both efficient and effective. Here, a simple model of price formation can point you in the right direction. This paper presents a model of price change, uses it as a guide for price and cost analysis, and surveys data sources. In our survey, we pay particular attention to the sources of information on service related expenses.

Value of Price and Cost Analysis (PCA). The purpose of PCA is to ensure the price paid is both reasonable in terms of the market, and "fair" to the supplier over the longer-term. While a fair price for suppliers may seem an anathema to buyers, what we mean is the lowest price that ensures continuous supply at the required quality. In this context, the notion of a fair price seems reasonable, and fits in well with the economic theory of the firm where total revenue over the long term must at least equal total costs (including profit) for the firm to remain in business.

As important as PCA is in price determination, it also has a strategic application in purchasing organizations as a means of isolating and eliminating items of unnecessary or excessive cost. This is the role benchmarking plays, with PCA as the critical component in any benchmarking program.

A Simple PCA Model -- Think About Price Change. In an ideal world, collecting price and cost information for analysis would be relatively straightforward. Indeed, for many raw commodities, prices are readily available. But more often, what is being purchased is unique. This is almost always true with respect to services. Here the effort to establish a "fair" price for reference can be extremely difficult, time consuming and expensive.

But given the unique nature of almost all service buys, the price at any particular moment is probably less important than its movement. Knowing the general price trend, or uncovering what is driving prices or costs, is often more valuable in supplier negotiations. Moreover, focusing on price trends has the inherent advantage of being able to draw on a wealth of information, making the whole exercise easier. With an initial investment, any organization can assemble a powerful tool that can be used to minimize the impact of price increases, take advantage of price declines, or better understand margins.

We can begin by thinking about price change in a simple conceptual framework. Prices at any point in time are based on market conditions; however, over time, assuming a reasonable amount of competition, we would expect prices to follow a path dictated by the underlying movement in costs (see figure 1). Market activity helps to explain fluctuations about this base cost path. Statistically, our analysis shows that this straightforward notion of price change can yield surprising results. Anywhere from 50% to 80% of the movement in prices can be accounted for by the change in costs alone.

Given the dominant role that costs play in price behavior, the question now becomes how to best analyze costs? The literature on cost analysis within the purchasing function is extensive and provides a number of different avenues to pursue. Costs may be broken into direct and indirect costs, variable or fixed, with different aggregations to block expenses by manufacturing costs, general and administrative expenses, and/or shipping. All are designed to derive a total selling price, or in our analysis, determine how the final selling or transaction price is likely to move and why.

We have found that, in terms of explaining price movement, analyzing variable costs or production/manufacturing costs (purchased materials, direct labor, and energy - see figure 2), generally yields the best results. Fixed or indirect expenses, although important in gaining a complete picture of an industry's or firm's cost structure, often provide little help in explaining price movement. By definition, they tend to remain constant, or change only slowly over time.

How Does This Apply to Services? Fine, so we have established a simple model of price change, but it seems to apply most readily to manufacturing. True, it does. However, if we remember that most services, broadly defined, are extremely labor intensive, our notion of variable costs as the primary driver behind price movement applies to service costs. This can be a little misleading because some so-called service industries now employ more capital per worker than some manufacturing industries.

Nevertheless, a disproportionate share of the cost for many services is labor. Think of engineering services, legal services, information processing, health care, advertising and public relations, property management, personnel supply services, and travel agencies - all are primarily labor. By focusing on compensation costs then, we can get a good handle on what drives costs for many services. Other service industries like transportation and warehousing, or telecommunications have a profile more like traditional manufacturing with a similar cost distribution.

Data Sources. With a simple model of price change highlighting the importance of variable costs, where can analysts go for hard information on these kinds of expenses for specific industries? Service costs, even though we make a case that they are primarily labor, are less tangible and, therefore, have traditionally been harder to collect information on. Fortunately, there is recognition that, with the growing orientation of the economy towards services, data collection for this sector must be improved.

Two statistical programs are addressing this economic transformation, and are providing a roadmap to other data sources that are already rich in detail for the wage and compensation expenses that are important for tracking service costs. All are publicly available, and collectively provide a wealth of information. First is the new North American Industry Classification System (NAICS), which updates the older Standard Industry Classification (SIC) system. The goal of the NAICS is not only to integrate the collection of economic data across North America, but also to provide greater detail for service sector industries. The second is the Producer Price Index (PPI) Program. Since the late 1980s, this program has gradually expanded to include more comprehensive coverage of service costs. A fairly good collection of service sector data, with sufficient history, now exists for time series analysis. These databases and others are reviewed below.

The Bureau of Labor Statistics (BLS), a division of the Department of Labor, collects most of the publicly available information on prices and labor costs. In particular, the BLS collects labor cost information under the Average Hourly Earnings and Employment Cost Index programs while the Producer Price Index and Consumer Price Index programs collects pricing information. Data from these programs can be downloaded via the BLS website (

The Average Hourly Earnings (AHE) data is published as nominal dollars per hour and is based on reports of gross payrolls and corresponding paid hours for production workers, construction workers, and nonsupervisory workers. AHE series are calculated by dividing aggregate industry payrolls by corresponding aggregate industry man-hours. These figures exclude benefit costs, but include overtime, shift premiums and paid leave. Industry information for the AHE's can be as detailed as the 4-digit Standard Industrial Classification level.

The Employment Cost Index (ECI) is a quarterly measure of change in the cost of labor defined as compensation per employee hour worked. Although the self-employed, owners-managers, and unpaid family workers are excluded from coverage, its coverage is broader than that of the AHEs. The ECI is also considered a better measure of labor costs because it is a fixed-weight index at the occupational level and thus eliminates the effects of employment shifts among occupations. However, the occupational detail under the ECI program is lacking.

The Producer Price Index (PPI) measures monthly average changes in selling prices received by domestic producers for their output. The import prices are not included in the PPI universe. PPIs are currently classified under the Standard Industrial Classification (SIC) coding system and provide comparability with a wide range of industry-based data for other economic phenomena, including productivity, production, shipments, employment, wages and earnings. Although there are certainly more coverage of manufacturing than service industries, the coverage of service industries (SIC 40+) now totals more than 1,000 and is growing each year. Price indexes are now available for: Transportation services (rail, trucking, air and water), warehousing, telecommunications, insurance, travel expenses (air, hotels, rental cars), software, personnel supply (temporary help), health services (hospitals and outpatient), legal and A&E services.

The Consumer Price Index (CPI) measures changes in the prices of goods and services bought by households. It is based on a sample of prices for food, clothing, shelter, fuels and transportation, medical services, and other goods and services that people buy for day-to-day living. While the CPI program contains a number of measures that look at different kinds of service costs, the program samples retail prices, and may therefore give a misleading picture of price movement for businesses.

The CPI has some regional price information at the aggregate level, such as all items, medical care, and transportation. However, until BLS's introduction of the National Compensation Survey in 1996, there was no detail occupational wage information at the regional level. The National Compensation Survey now provides wage (and eventually benefit) data for localities, geographic regions, and the U.S. as a whole. Currently, the data gives you a point-in-time estimate of the $/hour wage for a particular occupation in a certain location. The table below highlights the wage differential across the different metropolitan areas of computer systems analysts' wages.

Wages for Computer Systems
(average $/hour in selected cities, private industry)
EAST Wage Time of Survey
Atlanta, GA 27.03 January 1999
Boston, MA 31.76 October 1998
27.78 October 1998
Philadelphia, PA 29.36 February 1999
Hill, NC
28.71 July 1998
Rochester, NY 25.07 June 1998
Tampa-St. Petersburg-
Clearwater, FL
27.23 November 1998
27.30 August 1998

WEST Wage Time of Survey
Austin-San Marcos,TX 30.19 October 1998
Honolulu, HI 26.82 February 1999
Kansas City, MO-KS 27.21 August 1998
Phoenix-Mesa, AZ 32.02 November 1998
Portland-Salem, OR-WA 30.55 August 1998
San Diego, CA 31.42 November 1998
San Francisco, CA 35.17 July 1998
Seattle-Tacoma-Bremerton,WA 26.91 February 1999

Source: National Compensation Survey, Bureau of Labor Statistics

In addition to the price and labor cost information, the Census Bureau conducts an Economic Census across the each sector of the U.S. economy every five years. For more information or to download these reports, please see the Census Bureau website ( Information in the Economic Census is the critical piece in the puzzle when trying to form a picture of industry costs - whether for the manufacturing or service sector. Published information includes detailed examinations of cost by industry, providing the road map to the actual price and wage data that can then be used to track movement (see example in next section).

Do note that the industrial classification embedded in most of the above data sources will be changing over the next five years. The North American Industry Classification System (NAICS) was introduced in 1997 and updates the older Standard Industrial Classification (SIC) System to reflect the current set of industries within the U.S. economy. Please see the Census Bureau website for more information on implementation dates and a crosswalk between the existing SIC structure and the NAICS ( On balance, the new NAICS structure promises more comprehensive coverage of services

Putting the Data to Work. Now the hard part - putting all of this information together to help in your supplier negotiations. To walk through an example of our framework, we have chosen the Trucking and Courier Services (SIC 421) Industry. The first part of this exercise is to locate the Economic Census Survey. In our case, table 4 from 1997 Transportation Annual Survey contains information on the operating costs for the Trucking and Courier Services. From this, we gather what portion of total operating expenses is spent on payroll, energy and other expenses. For example, 20% of total operating expenses was spent on purchased transportation in 1997.

From here, we find a price and/or wage proxy for each of the operating expense categories using the above data sources. Since we know that changes in input costs are the main drivers of a price change, we download identified prices and wages from the BLS website and identify which category(ies) is responsible for the upward/downward pressure on prices. Supplied with this information, you are now a more informed buyer, and therefore, are in a better position to question any proposed price change for a given commodity and/or service.

Trucking Operating Costs
Share of
Price/Wage Change from Year-Ago
1998:4 1999:1 1999:2 1999:3
Trucking and Courier Services $184,793 100.0% 3.1 2.9 2.8 2.8
Payroll 52,438 28.4% 2.2 1.7 2.2 2.3
Benefits 27,590 14.9% 0.1 0.3 1.3 1.7
Purchased Fuel 13,777 7.5% -30.2 -26.8 8.0 34.1
Purchased Transportation
Leased Trucks with Driver 23,208 12.6% -0.5 -1.6 0.4 -1.0
Leased Trucks w/o Driver 4,940 2.7% -0.5 -1.6 0.4 -1.0
Railroads & Airlines 6,175 3.3% 0.3 0.3 1.2 1.2
Lease and Rental 3,148 1.7% 1.5 1.2 2.3 2.5
Insurance 5,717 3.1% -- -- 0.7 1.0
Maintenance and Repair
Parts 4,694 2.5% -0.4 -0.5 -0.6 -0.5
Repair 4,771 2.6% 3.3 3.5 3.4 3.3
Depreciation 9,289 5.0% 0.9 1.8 2.8 1.7
Other Operating Expenses 29,046 15.7% -2.9 -2.5 -0.9 1.1
Source: U.S. Census Bureau's Transportation Annual Survey: 1997, Bureau of Labor Statistics

Summary. Although scattered across a variety of different sources, information on service related expenses can be pieced together. Even a relatively simple model of price change can provide a framework to help pull these pieces together. The result is a more complete picture of service industry costs and a powerful tool in supplier negotiations or for evaluating the performance of any purchasing organization.

Figure not available in text-only version of this article.


Aljian, George Purchasing Handbook, McGraw-Hill, 1973

Earl, Paul, Mike Ramondi and John Mothersole "A Structural Approach to Price and Wage Modeling: Methodology and Data Sources", unpublished working paper, Standard & Poor's DRI

Graw, Leroy "Is Price Analysis a Lost Art?" 83rd Annual International Purchasing Conference Proceedings, National Association of Purchasing Managers, 1998

Leenders, Michiel R., and Harold Fearon, Purchasing and Supply Management, Boston: Irwin-McGraw-Hill, 1997

Schoen, Joseph "Benchtracking Purchasing Performance - A Management Tool", 82rd Annual International Purchasing Conference Proceedings, National Association of Purchasing Managers, 1997

Back to Top