Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 166
O Measures of
O Productivity for
Companies
The preceding chapters have examined the conceptual issues and measure-
ment problems involved in measuring productivity for industries and
sectors of the economy. This chapter considers measuring productivity for
individual companies. It examines the special uses, data requirements, and
other problems associated with company-level measures.
One reason for considering measures of productivity for companies is to
elaborate the uses to which these measures are particularly well suited.
The use of productivity measures in finding particular ways to increase
productive efficiency is probably most effective at the company level. The
first section of this chapter describes how a company can improve its
performance by developing its own measurement system. It also briefly
describes general procedures and problems of establishing a measurement
system and presents a few examples.
Measures of productivity for companies and related ratios can also be
used in research. Explaining the reasons for differences in productivity
among companies helps in understanding the sources of productivity
growth. Estimates of productivity at the establishment level in particular
are needed to make estimates of economies of scale. The second section of
this chapter reviews research based on company records collected by the
Census Bureau. It also discusses how the confidentiality of company
records prevents private researchers from gaining access to the data
directly and so limits the scope and quality of this research.
Another reason for narrowing our focus to the company level is to
consider a special problem that arises from the lack of coordination of the
166
OCR for page 167
Measures of Productivity for Companies
167
different data collection agencies. Different agencies collect the basic data
on prices, output, hours, etc., that enter into the calculation of productivi-
ty ratios at all levels of the economy, from the establishment to the private
business sector. The third section discusses how productivity measures
could be improved by better coordination of the agencies that make up the
federal statistical system.
MEASUREMENT TO IMPROVE COMPANY PRODUCTIVITY
Since, in the long run, companies are basically concerned with
profitability, why should they be concerned with measuring their produc-
tivity? Profitability is the best overall indicator of company performance: it
measures the outcome of all management decisions about sales and
purchase prices, levels of investment and production, and innovation as
well as reflecting the underlying efficiency with which inputs are converted
into output. However, profits are often influenced by factors external to the
firm, such as shifts in demand or inflation, and so profits may rise or fall
for reasons beyond the control of management. Measures of productivity,
on the other hand, are measures of the efficiency with which resources are
used. They reflect directly management's efforts to improve profits and to
remain competitive by improving efficiency (Kendrick and Creamer 1961~.
How much a company can learn from a productivity measurement
system depends on how extensively it is carried out. Productivity can be
measured for each production facility within a company, for different
levels of operation, such as fabrication and distribution, and for various
divisional and organizational elements as well as for the entire company.
When measures are available for several units, comparisons can be made
that should reveal when one unit is less efficient than or not performing as
well as others. Comparisons of production facilities are particularly widely
used, but when data are also collected from support units such as sales
divisions, the company takes a broader view of possible places for
improvement in productivity.
Many companies limit their measurement to labor productivity, but a
complete set of multi-factor and partial productivity ratios is preferable,
since it is advantageous to save any input. A company may nevertheless
benefit greatly from even a single-factor measure such as labor productivi-
ty because the company may produce a labor-intensive product, or it may
want to compare its performance with average output per hour for its
industry as published by BES.i A company can also use productivity
measures to forecast labor requirements and capital investment needs
when expansion of output is planned, and in collective bargaining
OCR for page 168
168
REPORT OF THE PANEL
situations, to help in preparing estimates of the impact of a wage
settlement on costs.
Company measures of labor productivity may be used by regulatory
agencies to help determine by how much a company should be allowed to
increase its prices. Public utility commissions have considered the use of
productivity estimates in their "cost and efficiency adjustment" formulas
by which rates are set. For example, in 1973 the New Jersey Bell
Telephone Company was permitted to automatically pass certain cost
increases on to customers. Rates could be increased in accordance with the
rate of increase in average hourly earnings minus the past rate of increase
in output per employee-hour. This adjustment provides an incentive for
the utility company at least to equal its past productivity performance so
as to be able to recoup increases in unit labor costs. There could be even
more incentive for a utility or other regulated firm to improve its
productivity if the regulatory agency extended the productivity adjustment
to cover other inputs (Kendrick 1975~.
The Price Commission planned to use individual company measures for
this purpose during phase II of the Economic Stabilization Program, in the
early 1970s, but found that it could not rely on the rather hastily and
inaccurately prepared estimates it received.2
MEASUREMENT PROBLEMS AND DATA SOURCES
In preparing productivity measures, a company faces many of the same
data needs and measurement problems associated with aggregate mea-
sures. Since the conceptual and other measurement problems are discussed
elsewhere (see Kendrick and Creamer 1961), here we mention only briefly
the issues most likely to arise in constructing company measures. We then
discuss the sources of data particular to a company.
What kinds of problems are likely to arise in measuring output? One of
the first problems for a multi-product company is how to combine its
diverse products. There are two ways to measure multi-product output: by
deflated value or by a quantity index using base-year price or labor-
requirement weights. If a company settles on value of sales deflated by a
price index as a way to measure output, two other problems arise. First, it
must adjust sales for additions to and subtractions from inventories of
finished goods and from work in progress in order to calculate production
for a given time period and to relate that production to the inputs used
during the same period. Second, the value of inventories should be
corrected for the arbitrary valuation methods used in standard accounting
procedures, which often do not reflect market values.
A non-financial company should usually exclude from its output
OCR for page 169
Measures of Productivity for Companies
·69
measure in a productivity ratio the income earned from financial assets
because this income is not produced by labor input. However, the output
of support activities to production, such as food services, maintenance, or
force account construction should be included whenever the correspond-
ing labor input is included.
A number of problems are likely to arise in measuring inputs. The
company must first determine whether it will be better served by a labor or
a multi-factor productivity measure. If the multi-factor measure is chosen,
labor and capital inputs must be measured and combined; for a more
complete productivity accounting, materials inputs should also be includ-
ed. Labor input may be a simple count of hours, but even in this case, a
company should keep in mind that hours spent at the workplace are a
better measure of actual labor input than hours paid for (see Chapter 6~. A
slightly more complicated measure of labor input involves weighting hours
by skills. When combining labor hours with other factors, it is conceptual-
ly correct to use weighted hours. Materials input, like output, is usually a
multi-product measure and presents the same measurement problems that
output does, except that in the case of materials there may be more
components for which collecting data on prices poses a problem.
Capital input is usually the most difficult input to measure. The several
steps of constructing a capital input measure are outlined in Chapter 6 and
are not repeated here. However, it is worth emphasizing that a capital
input measure is substantially more representative of actual input if assets
are valued at replacement, not historic, costs and if depreciation is based
on a reasonable service life of assets, not on arbitrary accounting life.
After measuring individual inputs, the next task is to combine the factor
inputs into a single index. The most obvious way is to use factor-share
weights. There are other weighting schemes, but they are often more
complicated and raise such issues as what is the appropriate production
function, an issue that is incongruous at the company level.
A company can find much of the data needed to measure productivity in
its own cost-accounting, payroll, and other internal records. Supplementa-
ry data can be found in the forms that many companies, especially large
ones, send regularly to the Census Bureau, the BES, the IRS, and regulatory
agencies. The Annual Survey of Manufactures (ASM) contains data on
employment and hours and on costs of labor, capital, and materials inputs.
See Table 8-1 for examples of productivity ratios that can be constructed
with ASM data. There may be occasions when a company needs to collect
data especially for productivity measurement. For example, a special
collection program for prices might be established because the price data
may be widely scattered among numerous invoices. As a shortcut, the
company may go to external sources of data available from agencies that
OCR for page 170
170
- ~ ~ ~t cat ~ cry cat ~ cat 0 ~ ~ ~ 0 0 ~ cry
- , ~ ~ ~ ~ - , _ ~ ~ ~ ~ ~ ~ o ~ ~ ~ of
___~______~_ _~
,.
V)
:^ -
o o
~ z
,~)
_
~ ,
an
-
Cal
._
V3
o
Hi. ~
rho ~ _
to
._
Cad
~o
;^
.`
°
-t r
-
_ - O
_ ~
Cal _ ~ ~
C)
·_
-
C~
_ ~ ~
. . .
_ ~ ~
O ~ Via ~ ~ ~ ~ ~t _
- I 1 1 1 1 ~ ~ As' ') - ) ~ Cat ~ ho cat
00 0 ~ ~ ~ ~ ~ ~ ~
- , _ _ _ ~ r~ ~ ~ ~
~ ~ ~ ~ ~ O X ~ ~ ~ O ~ ~ ~ 00 ~ ~ C~ X ~
00 ') 00 X ~ ~ OC) ~ ~ ~ - 1 ') - 1 ~ ~ ~ ~ X O ~
c~ c~ ~ r~ ~ ~ _ c~ c~ r~ ~ - ' ~ ~ ~ ~ c~
00 ~ ~ ~ U) O ~ ~ ~ O0 ~0 X ~ X O
~o ~ ~ u~ ~ ~ ~ ~ u~ o ~ ~ c~ c~ ~ ~ ~ ~ ~ ~
cr~ ~ ~ ~ cr~ ~ ~ ~ c;- ~ G~ ~ C~
0 ~ ~ x ~ r~
0- r~
. . . . . . .
~ ~ ~ ~ ~ ~ c`]
- - ~
~ :>
- -
OCR for page 171
Measures of Productivity for Companies
171
produce economic statistics; for instance, it could use components of the
BES consumer or producer price indexes.
EXAMPLES
Here we present two examples of companies that have developed
productivity measurement systems. The first example is General Foods, a
major food processor. The measurement program of General Foods has
many of the features described in the preceding section. First, it has a
comprehensive set of productivity measures, including measures for labor,
capital, and materials, and a multi-factor measure that combines all three.
A measure of materials productivity is important for General Foods
because it is an intensive user of raw and packaging materials. Second,
General Foods has measures for several of its plants. Thus, it can compare
plant performances but with the awareness that productivity levels and
rates of change diner for plants producing different items. Third, General
Foods makes the program a company-wide effort, enlisting the participa-
tion of management, accountants, engineers, and plant-level employees.
A special feature in the construction of its productivity measures is the
use of costs of materials lost instead of total purchased materials.
Purchased materials include materials consumed in making a food item
and materials lost, for example, grain spilled in unloading freight cars.
General Foods concentrates on the materials lost about which something
can be done over a short span of time.
The second example is United Airlines, a major domestic air carrier.
The controller's office has the responsibility for labor productivity
measures for the entire company for major divisions. Labor productivity is
measured in terms of equivalent passengers boarded per employee, and
company-wide productivity is also measured in terms of revenue ton-miles
per employee. The United Airlines program deals mostly in physical
output measures, thus avoiding many of the problems associated with
deflated-value measures. Its measures are akin to an efficiency concept of
reducing labor input and are not designed for considering overall costs, as
are those of General Foods. United Airlines uses internally collected data
as well as the data provided to the Civil Aeronautics Board (CAB) to
calculate its ratios. It also uses CAB data to construct productivity
measures for other air carriers in order to have a basis for productivity
#,
comparison.
OCR for page 172
OCR for page 173
OCR for page 174
OCR for page 175
OCR for page 176
OCR for page 177
OCR for page 178
OCR for page 179
OCR for page 180
Representative terms from entire chapter:
productivity measures
172
at:
c
.
-
C)
Q
rat
4-
~o
Q
0 ~
.
~
cn
lo cnl
1
in
c
_
Q
0
cat
o
-
~s
cat
o
c:
it>
c,
v,
x
LL
v,
a)
m
r
o
IS
Cat
00
-0
.0~7,
fin
<^
o ti
.
o o
V)
CO
V'
c a,
.o
a)
>
V,
CO
n
~ ~n
1m ~
~}
c ~0 ~
1=o
I-Q
~r E
° 1
G |
o ~ c E
~ ~ IOQ~
1= 1
1 ~
1 ~
1 CJ,
Measures of Productivity for Companies
ASSISTANCE FOR COMPANIES
173
A company can get help in preparing productivity measures from
handbooks, from federal government agencies, or from productivity
centers. Measuring Company Productivity by Kendrick and Creamer,
although written in 1961, is still the most complete and informative
treatment of the topic. The federal government is active in promoting
productivity improvement at the company level. The U.S. Department of
Commerce has conducted seminars for companies in trade-impacted
industries to inform them about recent developments in production
technology. The National Center for Productivity and Quality of Working
Life was until 1978 the most active federal agency in encouraging and
assisting companies to develop measurement programs. This function of
the National Center has recently been transferred to the U.S. Department
of Labor.
In several countries there are private or public productivity centers that
advise companies on how to improve their productivity.4 These centers
often conduct a special type of program that allows a company to compare
its performance with other competing companies that produce a similar
product by similar techniques. Figure 8-1 illustrates the typical set of
evaluation ratios provided to participants in a company comparison
program. A company receives its own ratios and, in coded form, those for
the other participants. The center, or perhaps a trade association or an
accounting firsts, processes the company records.
All the ratios in Figure 8-1 are components or derivations of what most
companies think of as the key measure of performance, the rate of return
on assets ratio 1. If a company shows a poor performance on that key
ratio, it can follow the lines of Figure 8-1 to see if its labor costs as a
percentage of sales (ratio 10) or some other factor deviates unusually from
the average.
The company comparison program as it is usually handled provides
assessment mostly in terms of financial ratios and so functions as a means
of evaluating the short-term performance of a company vis-a-vis its
competition. For small- and medium-size firms, this kind of appraisal
program is probably sufficient. A large company, especially one with
unique production methods or a wide variety of products, may gain more
from an independent measurement program.
I74
REPORT OF THE PANEL
Recommendation 21. The Panel recommends that companies investi-
gate whether having measures of productivity would improve their
performance, planning, and evaluation. It encourages the U.S.
Department of Labor and the U.S. Department of Commerce to
continue to inform companies of the potential benefits of productivity
measurement programs.
RESEARCH ON COMPANY PRODUCTIVITY
Thus far we have discussed measures constructed by companies for the
purpose of improving their own productivity. Now we turn to data
collected from companies (or their establishments) by the Census Bureau
and examine how these data are used to study productivity change in
industries and in the economy. The Census Bureau data are collected in
the quinquennial economic censuses and by other special surveys and
reports for manufacturing, wholesale and retail trade, services, construc-
tion, and agriculture. Most productivity research has been limited to
manufacturing industries, probably because these data conform best to the
underlying concepts of productivity measurement.
From the study of productivity in manufacturing establishments, a great
deal could be learned about how and why productivity change is taking
place. Productivity change may result from the introduction of new
establishments with higher productivity and the abandonment of high-cost
marginal establishments. This result may often be associated with a shift in
the location of an industry, such as the movement of the textile industry
from the Northeast to the South. Similarly, productivity may increase
through expansion of establishments with high productivity and contrac-
tion of establishments with low productivity, without marked changes in
the productivity of individual establishments. Alternatively, productivity
increases may reflect changes within individual establishments from new
techniques or learning through experience. To understand what is taking
place, these different situations need to be distinguishable. A few studies
have tried to investigate some of these possibilities and have met with
varying degrees of success, often depending on the form in which the data
were made available to the researcher.
One approach to studying the causes of productivity change is to explain
differences in productivity levels among establishments. Klotz (1973)
attempted to analyze the causes of dispersion in labor productivity using
data from the 1967 Census of Manufactures. He tried to explain the
dispersion of value added per production-worker hour among establish-
ments in terms of the dispersion of other ratios easily formulated from
Census Bureau data, such as total gross assets per production worker, a
Measures of Productivity for Companies
175
crude proxy for the capital/labor ratio. His measure of dispersion for any
ratio was the difference between the quartile average and the industry
average, with all quartiles formed by ranking establishments by value
added per hour.
Table 8-1 shows the industry and quartile means for value added per
hour and a few other ratios for four selected industries. In all cases, the
specialization ratio is high so that productivity differences are not due to
substantial differences in the products manufactured by establishments.5
The higher value added per hour is associated in every industry example
with higher assets per hour, as expected from production function theory
(see Chapter 3~. Higher average production-worker wages are also
associated with higher productivity. This association suggests that plants
are more productive because they hire more skilled, and therefore more
expensive, labor. Finally, the more productive plants are in two cases
larger in terms of employment per establishment, so that economies of
scale may be affecting their productivity. Based on the associations
suggested in Table 8-1, Klotz tried to explain productivity dispersion
within 200 4-digit manufacturing industries for which the specialization
ratio was high, but he obtained no significant empirical results. A major
weakness in Klotz's study is the level of aggregation of the basic data.
Working with deciles or finer breakdowns might have improved his
empirical results, but that degree of detail would make confidential
company records public.
Another approach to the study of productivity change is to explain
movements in the productivity of establishments over time. In the 1960s
the Census Bureau created a longitudinal file of establishments that would
permit a time-series approach. However, it found many difficulties in
constructing this file, such as tracing an establishment that is classified in
different industries for different census years and dealing with the fact that
even large establishments do not report every year between censuses. The
result of the Census Bureau's effort was a data file for just three years,
1952-1954, which was the basis of very little research (Kallek 1975~.
One of the more successful productivity research projects using time-
series data on establishments is the one by Griliches (1980) relating R&D
input to productivity (described in Chapter 7~. Part of the success of that
project may be due to its more limited scope; it deals only with the large
manufacturing companies responsible for most of R&D expenditures.
There are two interesting points to note about the construction of the data
set for Griliches's research. First, the study was possible because two
micro-data sets held by the Census Bureau were merged (see Chapter 7~.
This suggests other mergers that may open up new areas of research.
Griliches has listed some possibilities: the merger of Census Bureau data
176
REPORT OF THE PANEL
with company records from the IRS, the Securities and Exchange
Commission, or the patent once.
Second, because of confidentiality restrictions, Griliches, like all private
researchers, did not have direct access to company records. Following
Griliches's instructions, the Census Bureau processed the basic data into
moment matrices. This procedure hampered the research in two ways:
first, it caused long delays 3 years lapsed from the conception of the
project to the delivery of the first matrices; second, the data in their
aggregated form were not flexible enough to allow other calculations, such
as the social rate of return to R&D.
Confidentiality restrictions have repercussions beyond their impact on
research of private scholars; they also affect the quality of the basic data
themselves, which are used in estimating price movements, GNP, employ-
ment, and productivity.
The issue of confidentiality of company records has two major aspects.
First, confidentiality of data as it relates to the invasion of privacy of
individuals has been a major issue over the last 15 years, as computerized
files have been developed. There has been concern that sensitive records
will be brought together to create dossiers on individuals without their
knowledge, and that these dossiers will be used against the subjects by
government agencies or by private businesses. This concern about the
invasion of privacy has not been confined to government records. The
computerized files of credit agencies, banks, schools, and the personnel
records of employers have all been the subject of controversy. In such a
climate, companies are naturally concerned about the confidentiality of
records that the government obtains from them.
The second aspect of the confidentiality of company records relates to
the procedures under which information is collected. In order to get
accurate, detailed, and valid information from companies about their
operation, collecting agencies such as the Census Bureau have found it
important to assure companies that the information provided will be kept
confidential and will not be given to regulatory or administrative agencies
concerned with such matters as tax collection, enforcement of antitrust
laws, or monitoring wage and price behavior. Even in situations for which
by law some reporting by companies is compulsory, collecting agencies
believe that without such a promise of confidentiality, companies would be
less cooperative and the information provided would be less complete.
Where the provision of information is voluntary, the promise of
confidentiality is even more important. As a result, almost all information
collected from companies by federal agencies is kept confidential under
rules that may be even stricter than those relating to data obtained from
individuals.
Measures of Productivity for Companies
INTERAGENCY COOPERATION
177
The data relevant to productivity measurement are scattered among many
different agencies. The Census Bureau conducts censuses arid annual
surveys covering such important information as shipments, inventories,
cost of materials, capital expenditures, employment, and employee-hours.
The Bureau of Labor Statistics has the main responsibility for collecting
price information, and it also collects information on wages, employment,
hours, and earnings. The Federal Reserve Board compiles a production
index from a wide variety of different sources. In addition to these sources,
administrative and regulatory agencies collect much related information:
the Internal Revenue Service obtains information in connection with
corporate tax returns; the Social Security Administration obtains informa-
tion from employers in connection with social security contributions; the
Securities and Exchange Commission obtains quarterly income statements
and balance sheets for all corporations listed on stock exchanges; and the
Federal Trade Commission obtains line-of-business information from
individual firms. There are also specialized sets of data collected by such
agencies as the Federal Communications Commission, the Interstate
Commerce Commission, the Civil Aeronautics Board, and the U.S.
Departments of Energy, Agriculture, Housing and Urban [Development,
Transportation, and the Interior.
Although in some instances data at the individual company level are
exchanged between agencies, this is the exception rather than the rule. The
Census Bureau has authority to obtain company and establishment records
from other federal agencies, and they use the records of the Internal
Revenue Service and the Social Security Administration. But reports to
the Census Bureau, by statute, cannot be given to other agencies. Among
the other agencies, little or no exchange takes place.
In some cases, interagency committees have been established to improve
the comparability of data among agencies. For example, there is a
"working group on the 1977 Census Production Indexes" composed of
representatives of the Census Bureau, the Bureau of Labor Statistics, the
Federal Reserve Board (FRB), and the Bureau of Economic Analysis. Each
of these agencies prepares estimates of output for many 4-digit manufac-
turing industries. The FRB estimates are published first as part of the
monthly industrial production index, followed by annual estimates
prepared by the BES and BEA and finally the Census Bureau quinquennial
estimates. Even after all agencies have benchmarked their estimates to the
Census Bureau data, discrepancies among the estimates may persist
because each agency has its own set of procedures, deflators, and weighting
schemes. For example, for recent benchmark years, considerable
178
REPORT OF THE PANEL
differences were found in the four output measures for the pulp and paper
industries (see Myers and Nakamura in this volume). Also, sizable
differences were found between the output measures of 50 manufacturing
industries prepared by the BEA and by the FRB (see Popkin in this volume).
These results suggest that agencies should develop a more integrated effort
to measure real output.
Recommendation 2a The Panel recommends that the relevant
agencies try to reconcile their different output measures that cover the
same industry or sector to improve the measures and to acquire a
better understanding of measurement problems associated with
weighting, deflation, and other procedures. This can be achieved by
strengthening the existing mechanisms in government that bring the
agencies together, such as committees formed by the Office of Federal
Statistical Policy and Standards.
A major problem with the comparability of the basic data has been that
different agencies assign the same establishments to different industry
classifications; as a consequence, aggregated data at the industry level are
not in fact comparable from agency to agency. The Census Bureau has
prepared a Standard Statistical Establishment List that indicates how it
classifies different establishments by industry, but this listing has not been
made available to other government agencies because it is felt that this
would constitute a breach of confidentiality. The classifications assigned by
the Census Bureau are based on its report, in which establishments list
their products and services. Unless a uniform establishment list can be
made available to all federal statistical agencies, standardization of
information collected by different agencies is not possible.6 The result of
the high degree of decentralization of federal statistics coupled with
confidentiality requirements has been that the integration of data needed
for productivity measurement has not been achieved even at the industry
level.
Recommendation 18. The Panel supports legislation that would allow
the Census Bureau to share with other federal statistical collection
agencies the Standard Statistical Establishment List, so that all those
agencies could sample from a common universe, making the basic
economic data more comparable.
The President's reorganization project includes a statistical component,
which is concerned with improving coordination among the statistical
Measures of Productivity for Companies
179
agencies. The objectives of the improved coordination include reducing the
reporting burden of companies and establishments, which may now report
approximately the same information to several different federal agencies,
and improving the utilization of existing information within the federal
government, within the context of ensuring confidentiality. To achieve
those objectives, procedures would have to be set up to make possible the
interchange of data among statistical agencies with full protection of
confidentiality. These would be similar to the procedures currently used
for transferring administrative data from IRS and the Social Security
Administration to the Census Bureau. Improved coordination would also
promote cooperation among statistical and regulatory agencies in design-
ing their statistical programs so that the data collected by different
agencies would be complementary rather than conflicting or redundant.
Under such conditions it should be possible to obtain production, labor
input, and capital input information that would permit the computation of
productivity measures at the establishment level.
Recommendation 19. The Panel stresses the urgency of finding a
solution to the problem of coordinating data collection and allowing
data interchange among the federal statistical collection agencies for
statistical and research purposes, in such a manner that the rights,
benefits, or privileges of individual respondents are not violated.
NOTES
1. A comparison of productivity in a single plant or company with the average
productivity in an industry may not be valid if the plant or company produces an
output mix very different from other plants or companies in the industry (see
Chapter 2~. Also, the individual plant or company and the industry productivity
measures should be prepared in the same way. (For an explanation of how to
construct individual plant productivity measures akin to the BES measures, see
Greenberg [19733.)
2. The discussion in Chapter 2 about the possible misuses of productivity
measures in setting wages and prices and in forecasting in connection with the
aggregate measures also applies to company-level measures.
3. For other case studies, see Kendrick and Creamer (1961) and National
Center for Productivity and Quality of Working Life (1976~.
4. The American Productivity Center, a private productivity center supported
by major corporations, has recently been established in Houston.
5. The specialization ratio is the ratio of industry primary product shipments
to total industry shipments. It should be kept in mind that even a high
180
REPORT OF THE PANEL
specialization ratio is consistent with a variable mix of primary products among
establishments.
6. Even though agencies may have their industry classification verified by the
Census Bureau, all differences in assignment of establishments cannot be resolved
in this way. Thus, there remains a need for sharing the Census Bureau list of
establishments. It should be noted that several state governments compile lists of
establishments in their states by industry and make them available to the public.
