NOTES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

See Jerry Cromwell and Peter Merrill, "Minority Business Perform-
ance and the Community Development Corporation," The Review of
Black Political Economy (Spring 1973) and Abt Associates, Inc., An
Evaluation of the Special Impact Program: Final Report, Vol. 2
(Cambridge, Mass.: 1973).

These quotes are taken respectively from: Edward D. Hollander et
al., The Future of Small Business (New York: Praeger, 1968), p. 43;
E. F. Schumacher, Small is Beautiful (New York: Harper and Row,
1973), p. 60; Robert Dahl, "Governing the Giant Corporation," in
Ralph Nader and Mark J. Green, Corporate Power in America (New
York: Grossman, 1973), p. 23; and Ralph Nader, "Introduction,
in Morton Mintz and Jerry S. Cohen, America, Inc. (New York;
Dial Press, 1971), p. 18.

Gustav Schachter and Edwin L. Dale, Jr., The Economist Looks at
Society (Lexington, Mass.: Xerox College Publishing, 1973), p. 218.

Geoffrey Faux, CDCs: New Hope for the Inner City (New York:
Twentieth Century Fund, 1971), p. 45.

Francis X. Sutton et al., The American Business Creed (Cambridge,
Mass.: Harvard University Press, 1962), p. 60.

E. K. Hunt and Howard Sherman, Economics (New York: Harper and Row, 1972), p. 152.

For a detailed discussion of these and related issues, see Barry A.
Stein, Size, Efficiency, and Community Enterprise (Cambridge,
Mass.: Center for Community Economic Development, 1974).

Ibid. p. 58.

For a general discussion of these techniques, see Stein, Size, Efficiency. For more detail, see W. G. Shepherd, "What Does the Survivor Technique Show about Economics of Scale?" Southern Economics Journal (July 1967); Joe S. Bain, Barriers to New Competition (Cambridge, Mass.: Harvard University Press, 1962); and Robin Marris and Adrian Wood, eds., The Corporate Economy (Cambridge, Mass.: Harvard University Press, 1971).

See Stein, Size, Efficiency. As to the statement about small firms, this may be seen as a general application of the law of diminishing returns: the larger the unit, the less are likely to be the marginal returns from any additional increment of scale.

In this context, "short term" means the time span in which it is not possible for industry, say, to react to a change in environment by setting up new facilities, but in which it instead makes do with minor adjustments to its existing stock and technology. The initiator technique essentially assumes that the time span over which comparisons are made is short term, in that sense. The

31-717 O-79-5

survivor technique, on the other hand, embodies an inherent conflict,
since it must posit a long enough time period to let real changes
in size develop, while at the same time assuming that the bulk of
the changes observed are not entirely due to new facilities.

12. The Bureau of the Census uses the term "establishment" throughout
the references mentioned here. An establishment is a location at
which a particular production process is carried out, and the
figures reported on employment, shipments, and the like are based
on that unit. If two distinct production processes occur in the same
physical facility, it is defined as two establishments, each of which
may be placed in and aggregated with a different industry. Thus a
single physical facility may be further subdivided by the Census
Bureau, even though to its owners it is a single plant. This is
extremely important because the assumption underlying our work,
and the Bureau's data, is that those separate processes could in
fact be physically separated with little or no loss in efficiency. This
hinges on the issue of agglomeration efficiencies, that is the extent
to which each gains economies by being located next to the other.
In general, these effects will be small for the case we are considering,
and that is certainly true if we look at ownership rather than location.
Our point is that essentially the economies from this source can be
gained as well by operating the two establishments under different
managements. This, in fact, is the principle of industrial parks.

13.

14.

15.

The Standard Industrial Classification system is a scheme used
almost universally (in the United States) for collecting and analyzing
data on business and industry. It defines industries either on the
basis of products manufactured, or of a well-defined process or
technology. The intent is to provide a meaningful way of aggregating
data, and although it has flaws, there is no better system in wide
usage. It also provides for more and more detailed categories by
aggregating and labeling data at different levels of generality, which
groupings are then usually referred to in terms of the number of
digits in the category. Thus, SIC 20 is the category "Food and
Kindred Products a two-digit level of agregation), SIC 201 is the
category "Meat Products," and SIC 2011 is the four-digit category
"Meat Packing Plants. ' For further details and a complete
explanation of the system, see the Standard Industrial Classification
System Manual, (Washington: Bureau of Management and Budget,
Government Printing Office, 1973).

In addition, we make the assumption that, over the time span of
interest, there are no plants that move two or more categories up
or down. That is, we assume that all moves that can be made by
a shift to the next category happened, whereas no shifts took place
across intervening categories. This is undoubtedly not true in all
cases, but this phenomenon is believed to be rare and its possible
occurrence is compensated for by the conservatism of the method
sufficiently so that it can be disregarded.

We believe, in other words, that there are production units where
the efficient size is five to nine employees, although this says
nothing about their importance.

Stein. Size, Efficiency, pp. 65-71.

See Mark B. Hodax, "The Impact of the Energy Crisis on CDC
Industrial Development" (Cambridge, Mass.: Center for Community
Economic Development, 1974), memorandum to OEO.

For additional perspective on this issue, see Barry A. Stein, The
Community Context of Economic Conversion (Cambridge, Mass.:
Center for Community Economic Development, 1971).

Even though the figures are for assets at book value and reflect
accounting conventions regarding depreciation, the details of
which may differ from company to company depending on which of
the acceptable alternatives are employed, the actual book value shown
is at least approximately equal to the real (market) value of those
assets in used condition. One ought to be able to acquire the needed
equipment and facilities at something like the book value, at least
in cases where the items are more or less conventional in industrial

use.

This approach is potentially useful to community organizations for
the development of ventures specifically oriented to the needs of the
local population. Ronald A. Cornman, who helped in the present
study by orienting our thinking in this direction and by discovering
sources of the required data, is continuing to work on that approach.
Preliminary memoranda are available from the Center for Community
Economic Development, Cambridge, Mass.

Fabian Linden, ed., Market Profiles of Consumer Products (New
York: The National Industrial Conference Board, 1966).

22. The specialization ratio, as defined by the Census Bureau, reflects the mean value across all plants in the category, regardless of size. If there is a systematic variation (nonrandom) of specialization with size of plant, then the average will be in error by some amount, perhaps even a large one. However, we assume that these errors are not likely to be great in general, because in most cases we expect that production of other than primary products is a function of technology (involving, for example, byproduct or waste use) and is thus not likely to vary by proportion. The actual value of errors from this source is of course unknown.

23.

24.

25.

Dun and Bradstreet, Inc., "Cost of Doing Business in 185 Lines"
(New York; 1971).

Department of Commerce, Input-Output Structure of the American Economy: 1963, (Washington, D. C.: Government Printing Office, 1969). These tables enable direct proportions to be read off, indicating the extent to which goods produced by any industrial sector (roughly at a four-digit level of aggregation) go directly to final demand.

The SICS Manual (see Note 13) details the products contained in any particular category. There are a number of cases in which the initiator technique, after deducting adjacent categorical changes as

26.

27.

usual, suggests that the plants put up are bimodally distributed.
That is, there is evidence that new plants were both small and
large, with little or nothing in between. This can easily be
explained by assuming that the products in the particular SIC
category fall into two classes or clusters of products, each class
requiring very different production technologies.

There is an ongoing and important debate as to whether or not the
capital-intensive technologies promoted by present development
and industrial structures are the best types of technology to use
in other circumstances, particularly where labor is in excess,
people are without adequate jobs and income, and industry is largely
owned outside the area. This is just as much a problem for
depressed parts of the United States as it is in the third world or
in less developed countries elsewhere. We believe that there are
more opportunities than are often seen to develop intermediate
(often called "appropriate") technologies that are both market-
efficient and better designed to provide more jobs and training for
local persons. CDCs in our view should carefully consider this
issue before automatically making the assumption that the highest
technology is the best for their purposes. For a good presentation
of this point of view, see Z. F. Schumacher, Small is Beautiful
(New York: Harper and Row, 1973).

A good starting place to find such sources would be the reference
book National Trade and Professional Associations of the United
States and Canada and Labor Unions published annually by Columbia
Books, Inc., Washington, D. C.