”wiufiifl 1‘ ~.;L?Er“"$w“rvl‘3‘?5‘l‘“
u “‘9. ‘ at?“ Q“ '
. A 1‘4 ”"-
5: .252 ’ Jug'ffiifiaéi
a .153: g .. 5.1%??? . ’
aéékvnw.” ‘ Nerf-'5‘” “a"“t‘c‘w UV
@394.” : ‘ (2-“:th 13b“ hw;4}'51§33"‘
;., . w a n' A .,.-/ a“ J”
QIE'J'JJ 1” , , ‘15-‘45 f it?“ r-Lcigmgd‘

figfl'y . ,~ I‘.»- l.
mu, . .. any”
J; ‘vfl

(& ' AA

1U: 5"??T#‘I'L;
. ..é§it'¢"'“
‘ fig?“
asfi
- Iffia'm' ; "

3F. ‘5‘:
"F“ ‘>
r. ~r-‘r ‘
e}, 32:34?
. a

‘ -. 5,

‘itlr M M;

E; :w:
”-3. ‘1‘?“

33L
“ . “‘13. :;.,
.. fix ' ' "s " “ wibfliw
‘, .‘w‘g‘ ‘ . . ‘ . '

‘Fz v
' r n

1‘“: '.
239:2?“

3f:

’43-
my?

aw

w“
. .x4.‘.~1
.1». 1.‘

‘ ‘13"9"

 

m

i
fit?

 

MICHIGAN STATE UNI VE

iUlLIHl; rm MI 1m umnumu L

301409 8580 4..

pf. 11—. L‘BRARY
Michigan State

Unlversi Y

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

HUMAN CAPITAL AS A REGIONAL FACTOR
IN THE GROWTH OF SMALL FIRMS

presented by

Joan Margaret Kendall

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Geography

 

Major pro v

Date March 20, 1995

 

MS U i: an Affirmative Action/ Equal Opportunity Institution 0-12771

4‘ Av ___.fi‘..—~—-‘_-_L

fiarig-~———~—~_ ~—_. ‘ -—-— —,_-—~hf h... .—

PLAcE ll RETURN BOX to mouthi- ohookom from your mood.
To AVOID FINES Mom on or baton dot. duo.

DATE DUE DATE DUE DATE DUE

 

DEC 17 39%
a ‘ . __ . ‘
- 1:31:91.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jl | |

MSUIoAnAfflmdMAoflaVEwOppommRylmmon
W

 

m1

 

HUMAN CAPITAL AS A REGIONAL FACTOR
IN THE GROWTH OF SMALL FIRMS

BY

Joan Margaret Kendall

A DISSERTATION
Submitted tO'
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Geography
1995

ABSTRACT

HUMAN CAPITAL AS A REGIONAL FACTOR
IN THE GROWTH OF SMALL FIRMS

BY

Joan Margaret Kendall

Small firms are accounting for an increasing share of
new employment, but growth in the number of small firms is
not occurring equally over space. Since small firms are
less likely to make explicit location decisions, regional
variation in the number of such firms is less likely to be
explained by cost-related factors traditionally thought to
be most important in attracting firms to a region. Regional
factors which influence new firm formation and survival
rates, such as a region’s human capital, are expected to be
more important to growth in small firms. Regional variation
in the distributions of small firms and of several human
capital variables are examined using MSA-level data, and
regression analysis is used to measure the impact of human
capital on small firm growth. This relationship is examined
for two different firm-size categories and for four
industrial sectors. In all but one of the sectors examined,
human capital variables are found to be significant

predictors of growth in the smallest category of firms.

ACKNOWLEDGEMENTS

I would like to thank the Department of Geography for
the financial support I received, the faculty and staff for
their guidance and support, and my fellow students for their
friendship and moral support. In particular I would like to
express my appreciation to the members of my committee:
Bruce Pigozzi, Gary Manson, and Assefa Mehretu from the
Department of Geography, and Karen Roberts from Labor and
Industrial Relations; and also to Judy Olsen, for
encouraging me to pursue this degree. And last, but
definitely not least, I thank my entire family, whose
support and faith in me contributed significantly to my

success in completing this degree.

iii

TABLE OF CONTENTS

L I ST OF TABLE 8 O O O C O O O I O O O O O 0

LIST OF FIGURES O O O O O O O O O O O O 0

CHAPTER

I. INTRODUCTION . . . .'. . . . . . . . .
II. SMALL FIRMS AND REGIONAL GROWTH . . .
Definition . . . . . . . . . . . . . .
Contribution to Jo Growth . . . . . .
Regional Patterns in Small Firm Growth
Components of Small Firm Growth . . .
New Firm Formation . . . . . . . . .
Small Firm Survival . . . . . . . . .
III. HUMAN CAPITAL AND DEVELOPMENT . . . .
Definition 0 O O I O O O O O O O O O 0
.Measuring Human Capital . . . . . . .
.Human Capital as a Regional Factor . .
Regional Patterns of Human Capital . .
IV. RESEARCH PROBLEM AND HYPOTHESES . . .
Which Human Capital.Measures? . . . .
Which Size Small Firms?. . . . . . . .
Which Industries? . . . . . . . . . .
Industrial Structure . . . . . . . . .
Regional Implications . . . . . . . .
v. ANALYSIS 0 O O O O O C O O O O O O O 0
Data 0 O O O O O O O O O O O O O O O 0
.Methodology . . . . . . . . . . . . .
Hypotheses Testing . . . . . . . . . .

VI 0 RESULTS 0 O O O O O O O O C

General Hypotheses Testing
Sector Specific Analysis .
Regional Patterns . . . .
City-Size Comparison . . .
Residuals Analysis . . . .

iv

vi

. vii

VII. CONCLUSION . . . . . .

Human Capital as a Regional
the

Sectoral Differences in

.Human Capital . . .
Regional Differences .
Conclusions . . . . . .
Future Research . . . .

APPENDIX A . . . . . . . . . .
APPENDIX B O O O O O O O O O O

BIBLIOGRAPHY . . . . . .-. . .

Factor
Impact of

105
105

108
110
113
116
120
123

133

1.
2.

3.
4.

5.
6.
7.

TABLES

Regression Results: Equation 1 . . . . . . . . .

Percent Change in Number of Establishments by

Census Region .

Human Capital by Census Region . .

Percent Distribution of Wage and Salary Workers
by Selected Occupation and Industry, 1988 and

Projected to 2000

Hypotheses Testing .

Regression Results: Equations 2 and 3 . .

Regression Results: Equation

Differences)

Regression Results: Equation

Differences.

vi

5

0 OS

(Regional

(MBA Size

12
40

59
79
84

92

95

FIGURES

1. Percent Change in Total Establishments . . . . . . . . 15
2. Percent Change in Manufacturing Firms . . . . . . . . 17

3. Percent Change in Finance, Insurance, and Real
Estate Firms 0 O O O O O I O O O O O O O O O O I O 1 9

4. Percent Change in Service Firms . . . . . . . . . . . 20
5. Percent Change in Business Service Firms . . . . . . . 22
6. Percent of 25+ Population With a College Degree . . . 42
7. Percent of the Population with a High School Diploma . 44
8. Local Per Pupil Education Expenditures . . . . . . . . 45
9. Percent of the Labor Force in Management Occupations . 47
10. Residuals from Equation 8 . . . . . . . . . . . . . . 97
11. Residuals from Equation 9 . . . . . . . . . . . . . . 99
12. Residuals from Equation 10 and 11 . . . . . . . . . 101

13. Residuals from Equation 12 and 13 . . . . . . . . . 103

vii

CHAPTER I

INTRODUCTION

The decade of the 1980’s saw small-business dominated
industries' share of total employment steadily increasing
(U.S. Small Business Administration, 1989), and currently
small firms are responsible for an increasing share of net
new employment (Erdevig 1986; Loveman, Piore, and
Sengenberger 1990; U.S. Small Business Administration 1989;
White and Osterman 1991). Some of this increase in small
firm employment can be accounted for by the shift to
services (since service firms tend to be smaller than
manufacturing firms), but evidence also indicates that small
firm growth not related to inter-industry shifts is
occurring in all major sectors of the economy (Mardsen 1990,
Piore 1990). Much of this growth in small firms is
attributed to the rise of flexible specialization and
assciated downsizing, vertical disintegration, and increased
outsourcing (Loveman et.a1. 1990), all responses to changing
global competitiveness. According to Spindler and Forrester
(1993), such economic changes are related to the increased
premium placed on the skills and education of the labor
force; and Reich (1991) emphasizes that in an increasingly
global economy, where most factors of production are more
mobile than labor, it is the skills and abilities embodied
in a nation’s laborforce that are the key to its future

productivity and competitiveness. These ideas suggest that

l

the success of any business, but particularly the small
firm, will increasingly be dependent upon its human capital.
Does this also translate into a more immediate and direct
link, at the regional level, between human capital and the
above noted growth in small firms? This question is the
focus of this dissertation.

Research related to regional variation in the growth of
entrepreneurship and small firms suggests that regional
factors do play an important role in the growth of such
firms (Beyers, Christopherson, Erickson, Gibson, Hewings,
Malecki, McConnell, Rees 1990; Erdevig 1986), but regional
factors which explain variation specifically in small firm
growth have not been identified. Although factors which
impact regional growth in general would certainly also
explain some of the differential growth in small firms,
Watts (1987) suggests that because small firms do not
usually make explicit location decisions, cost factors such
as local tax and utility rates, which are usually emphasized
when the focus is upon attracting industry, are likely to be
relatively less important to the location of small firms.
This is not to imply that location is not important to the
success of small firms, only that locational adoption is
more likely to be associated with small firms, while
locational adaption, the ”... rational selection of an
optimal location for a firm," is more characteristic of
larger firms (Berry, Conkling, and Ray (1993, p. 296). This

suggests that regional characteristics which foster the

creation and success of small firms, factors such as the
region’s industrial structure and its human capital, are
likely to be more important determinants of small firm

location.

The Growing Importance of Human Capital

Although according to Kiker (1966) a few 19th century
economists such as Fisher and von Thunen did argue for
treating people as capital, it is primarily since the middle
of this century that leading economists, including Theodore
Schultz (1971, 1990) and Lester Thurow (1970), began to
acknowledge the quality of human input as a significant
factor in economic growth. And even recently, says Schultz
(1990, p.3), ”Growth models either omit or underrate the
increases in income from investments in human capital." For
the most part, because the quality of labor was generally
less important in primary and secondary industries, the
importance of human capital to economic growth and
development has focused on the quantity, rather than the
quality, of the laborforce (Salamon, 1991). As post-
industrial economies evolve, the emphasis shifts to labor
force quality, the key to productivity and the focus of

human capital theory. As Thurow (1970) has stated,

Economists cannot determine the sources of economic
growth if they measure labor as a homogenous
commodity...Increases in the productive capacities of
labor play an important role. Increases in human
skills, talents, and knowledge are of primary
importance. Measuring labor in terms of human capital
focuses attention on this source of economic growth.
Labor is no longer regarded as a homogeneous, fixed
commodity, but as a commodity that may be expanded and
improved (p. 11).

Packer (1991), acknowledging the relationship between
productivity and economic growth, focuses on the
relationship between education and productivity. He sees
the need for increasing productivity as a response to both
demographic and economic change. The primary demographic
issue centers on the need for greater laborforce
productivity resulting from the increased dependency ratio
which will occur as the baby-boomers began to retire. At
the same time, he points out (p. 45) that economic changes,
the shift away from manufacturing, the technological
revolution, and changing competitive environment, ”...have
outpaced the change in our education and training
institutions and methods."

There is no question that in advanced industrial
economies industrial restructuring and changing global
competitiveness are increasing the demand for skilled labor
and for more responsive production systems (Beyers, et.a1.
1990; Howell and Wolff 1991; Lever 1985; Spenner 1988;
Spindler and Forrester 1993; Storper and Walker 1983).
Although traditional location theory emphasized transport

costs (Webber 1984), as the economic base shifts away from

manufacturing, transport costs decline in relative

importance, and productivity is increasingly related to
human resources (Beyers, et.al. 1990; Leven 1985; Lever
1985; Storper and Walker 1983). According to Leven, as the
economy shifts from goods to services, and as the
information content of output increases, regional variations
in levels of technological knowledge become a factor in
regional development. As Berry, Conkling, and Ray (1991, p.
298) explain:

Location theory can no longer confine itself to

identifying the specific location that optimizes some

requirement of an economic maximizer. Instead, it must
seek to understand the processes that enable an
entrepreneur to start up a small business, to survive
and to grow, as well as the processes that cause firms,
large or small, to fail and to exit.

Schultz (1990) focuses on the role of the entrepreneur;
he views productivity as increasingly related to education,
not only because education increases skills, but because it
enhances entrepreneurship, which he sees as essential to
restoring the disequilibria which results from a modernizing
economy. It is the relationship between education and
entrepreneurship, he says, which explains why the proportion
of farmers with college degrees is increasing. The
importance of entrepreneurship is also stressed by Salamon
(1991), citing documentation of a strong relationship
between education and the adoption of innovations, a
characteristic associated with entrepreneurship (Clark
1985). Also associated with entrepreneurship is new firm

formation. Firm formation is receiving increasing attention

because it is thought to explain much of the regional

variation in the growth of small firms, whose share of new
employment, as noted above, has been increasing.

The purpose of this dissertation is to examine spatial
variations in the role of human capital in small firm
growth. Human capital will be measured by education and
occupational background, and their relationship to small
firm growth will be examined both for different size small
firms and for different industrial sectors, over both
Metropolitan Statistical Areas and census regions. This
dissertation will differ from.much of the traditional
research on the role of human capital in economic
development in the following ways: 1) its focus on human
capital at the regional, rather than national, level; 2) its
inclusion of entrepreneurship as a regional factor, as
opposed to examining characteristics of individual firm
founders; 3) its specific focus on small firms, and 4) the
use of establishment data.

The dissertation will begin with an examination of
small firm growth, with emphasis on the components of growth
and the location of small firms. This will be followed by a
discussion of human capital as a factor in economic
development, particularly as such growth relates to small
firms, consideration of how best to measure human capital
within the context of regional development, and an
examination of geographic patterns in the distribution of

human capital.

CHAPTER II

SMALL FIRMS AND REGIONAL GROWTH

There is a considerable literature related to the
increased contribution of small firms to employment growth,
and there are reasons to expect that regional variations in
this phenomenon may be related to geographic differences in
human capital. The first step in examining the possibility
of such a relationship is to settle upon a precise
definition of small firms, examine actual patterns of small
firm growth over space, and then take a closer look at the
components of small firm growth and the factors involved in

their increasing contribution to job growth.

Definition

Exactly what is meant by a small business is not clear,
and according to Bannock (1981, p.25), "It troubles many
people that there should be doubt about exactly what a small
firm is." Most literature relating to small firms deals
with the problem by simply avoiding any specific definition.
The Small Business Administration (1989) indicates that
small firms are sometimes defined as those with fewer than
500 employees, and sometimes defined as only those with
under 100 employees. The SBA then offers a breakdown used
by all federal agencies in publishing business data, and one
which is consistent with that developed by the Office of
Management and Budget: under 20 employees, very small; 20-99

employees, small; 100-499 employees, medium; and over 500

employees, large. Size, however, is not the only criteria
to be considered. Bannock suggests that small firms are
defined less by their absolute size and more by other
factors, the essential characteristic being that a small
firm is managed personally by the person who owns it. He
also points out that "small" is relative to market share and
to industry, and that even when industry is taken into
consideration, definitions vary widely and tend to be

somewhat arbitrary.

Contribution to JOb Growth

Small firms’ increased share of employment was noted in
the introduction, but no evidence was offered to indicate
how much impact such firms have on employment. The impact
of change in the number of small firms in each sector on
employment in that sector is examined for each sector by a
simple regression of employment growth on small firm change
as follows:
{1} EMPi - a + biSFi + ei

where: EMPi = percent change in total employment in
sector 1

SF1 = percent change in number of small firms,
sector i from 1983-1988

Data used correspond with the breakdown (discussed
earlier in this chapter) which the Small Business
Administration indicates is used by federal agencies
publishing business data for very small and small
establishments (1-19 employees and 20-100 employees,

respectively).

The results of this analysis are seen in Table 1. All
coefficients are positive and significant at P=.01, and
while Rz's are not large, they clearly indicate that small
firm change has a positive impact on employment in all
sectors. This impact appears to be greater in the Finance,
Insurance, and Real Estate and the Service sectors than in
the Manufacturing or Business Services sectors. However, it
appears that for the Finance, Insurance, and Real Estate and
the Business Service sectors, very small firms have a
greater impact, while in the Manufacturing sector Small

firms have the greatest impact; in the Service sector, the

TABLE 1. Regression Coefficients: Equation 1
(Impact of Change in Small Finns on Employment Growth)

 

Sector 3 B t-value R2
NmemmWMQ
Very Small 0.052 0.301 5.538 0.087
Small 0.044 0.380 7.288 0.142
F. l.R.E.
Very Small 0.108 0.582 12.676 0.338
Small 0.154 0.352 8.670 0.121
Swvmms
Very Small 0.129 0.587 12.843 0.342
Small 0.155 0.588 12.895 0.344
BwamsSawmes
Very Small 0.349 0.426 8.339 0.179
Small 0.383 0.359 6.787 0.126

 

10

impact of both size firms appears to be equal. For Business
Services, also, very small firms appear to have a greater
impact than those size 20-99.

Increased small firm share of employment is explained
by a variety of factors relating to restructuring, changing
technology, and increasing competitiveness. The shift to
services, because they are generally smaller establishments,
accounts for some increase in the relative number of small
firms; however evidence indicates that more of this increase
can be accounted for by within-sector change than by inter-
industry shifts (Mardsen 1990, Piore 1990). And, in fact,
the Small Business Administration (1989) indicates that
current growth in small firms represents a change from
historical trends. The SBA states,

Small businesses are generating relatively more of the

job growth in traditionally large-business-dominated

manufacturing, but relatively less of the growth in
retail trade and services, industries generally

dominated by small firms. (p.15)

Much of the trend toward smaller firms is thought to
result from vertical disintegration, as rapidly changing
technologies increase the need to contract out more
specialized functions if firms are to remain competitive
(Loveman, et.al. 1990). Also in the interests of
competitiveness, firms are increasingly relying on smaller
cores of full-time, permanent employees and depending more
upon contingent labor, both to minimize costs and to foster

flexibility (Abraham 1990). According to Bannock (1981) and

White, et.al. (1988), it is this increased flexibility which

11

allows small firms to be able to respond to change more
quickly, providing them with a key advantage over large
firms. Thus, it is not difficult to see why, in the context
of today’s economy, small firms might be growing faster than
larger firms. Where they are growing faster is a more
difficult question to answer, although it seems reasonable
to expect that areas with greater access to information
which would be likely to increase awareness of the need for
greater flexibility and also in the ability to achieve it
would be more likely to show strong small firm growth. The
search for an answer to why small firms are growing faster
in some places than in other will begin in the following
section with an examination of regional patterns in the

growth of small firms.

Regional Patterns in Small Firm Growth

Small firm data are the same as used in the previous
section (equation 1). Growth in the number of small firms
is measured as the percent change in the number of small
firms from.1983-1988. ”All sectors” in Table 2 refers not
just to the sum of the four sectors shown here, but
represents total establishments. Starting with MSA data
which has been aggregated to the census region level, Table
2 shows the percent change in the number of firms for
selected sectors. It appears that for very small firms
(Table 2a), over all sectors and for each of the individual

sectors except Manufacturing, the strongest growth is found

12

in the Southeast. Even in Manufacturing, growth in the
Southeast is considerably higher than in all but the
Mountain region. The Midwest and Southwest regions,
particularly the Southwest, had the worst rates of small
firms growth; in these regions, only the Service sector
displayed "reasonable" growth rates.

In the 20-99 employee category, (Table 2b), the
Southeast region again displays the highest overall growth
in firms; however, much of this aggregate growth must have
occurred in sectors not included in the analysis, since this
region not only does not have the highest rates of growth in
any of the individual sectors, but has a very low growth

rate in the Business Service sector (SIC 73). The Mountain

TABLE 2. Percent Change in Number of Establishments by Census Region

 

FRmbn
Firm Size/Sector NE SE MW SW MT WEST
a) 1-19 Employees
All sectors 14.39 19.61 10.41 2.79 17.08 15.82
Manufacturing -3.01 11.43 -0.15 0.59 15.54 3.21
F.l.R.E. 12.58 18.57 2.10 4.71 11.24 12.87
Services 18.46 26.81 17.83 15.63 26.64 20.25
Business Sew 3.43 22.85 6.40 -4.59 7.67 2.05
0) 20-99 employees
All sectors 23.94 39.03 23.82 6.18 25.13 30.75
Manufacturing -3.23 13.12 12.05 -1.41 20.59 14.56
F.l.R.E. 15.58 15.28 11.18 11.19 16.60 14.91
Services 39.13 48.42 35.99 23.41 43.12 48.89
Business Serv 13.26 5.95 18.42 -1.82 19.25 23.95

 

13

and Western regions both show greater increases in small
firms in most sectors, while the Southwest has the smallest
rates of change (even negative in two sectors). The Midwest
and Northeast regions fare better than the Southwest,
although the Northeast does poorly in Manufacturing.

In general, the figures in Table 2 indicate that there
is considerable spatial variation in the growth of both very
small and small Manufacturing firms. The same is true for
very small Finance, Insurance, and Real Estate firms,
although small firms in this sector vary less than those in
other sectors. The sector which reflects the most
consistency in growth rates, for both firm-size categories,
is the Service sector; however, Business Service firms, a
subset of Services, displays the most variation over space.
Table 2 indicates that in the "all sectors” category, small
firms with 20-99 employees grew much faster in the Midwest,
Mountain and Western regions than did the very small firms.
This was not the case in other regions. In both the
Northeast and Southwest regions, negative growth in
Manufacturing was greater in the larger firms, and in the
Southeast, growth in the larger firms was slower in both
Finance, Insurance, and Real Estate and Business Service
firms. These aggregate census region data, while helpful in
determining general patterns, are likely to hide
considerable within-region variation, as reflected in maps

of MSA level data.

14

Sectoral changes in the number of firms over
Metropolitan Statistical Areas are mapped in Figures 1
through 5. An examination of Figure 1 shows that over all
sectors (for both small and very small establishments), most
areas of the Southeast do exhibit strong growth, with the
strongest growth concentrated in Florida, Atlanta, and parts
of North Carolina. In fact there is a fairly strong
bicoastal pattern. The midwest and south central/southwest
show the least growth, while actual decline in small firms
seems concentrated in a vertical belt beginning south of
Minneapolis and running all the way to the Gulf. Even the
traditional "rust belt" exhibits small firm growth in many
MSA’s.

For all sectors, the major differences between very
small firms (Figure 1a) and small firms (Figure 1b) are in
the Northeast and the Great Lakes areas. In the Northeast,
very small firms appear to be more successful than small
firms, at least in relative terms. In contrast, in the
Great Lakes area, specifically in MSA’s around the Detroit
area, small firms are growing at a faster rate than very
small firms. (Note: the classes are not the same on both
maps; the range of change for small firms is considerably

higher than for very small firms).1

 

1Because the range of change, both between very small and
small firms and among sectors, was so great, it was not
possible to use an absolute scale for all maps and have a
meaningful result; thus it was decided to use categories
which produced equal numbers of cases, i.e. divided MSA’s
into lower, middle, and upper thirds.

15

a) CHANGE IN VERY SMALL EST ABUSHMENTS

 

 

 

  

2 Choc-90
E] -1534 16 7.62

7.62 16 16.13
- 16.1310 181.00

 

b) CHANGE IN SMALL ESTABLISHMENTS

 

 

 

1 V10 ‘
0 4 .‘ a: Change
a .l D -26.92 16 19.51
‘ ‘1 19.51 to 30.68
‘ 1' - 30.68 to 246.13

 

Figure 1. Percent Change in Total Establishments

16

For individual sectors, the maps of changes over MSA’s
revealed the considerable variation within census regions
that did not appear in Table 2. For example, when MSA
changes in the number of very small Manufacturing firms over
MSA’s are examined, while Table 2a indicated negative growth
in very small Manufacturing firms in both the Midwest and
Northeast, Figure 2a shows many MSA’s in both of these
regions to have high levels of growth in very small
manufacturing firms, along with areas of negative growth.

For very small Manufacturing firms, the southeast, and
particularly Florida, along with MSA’s in Colorado, southern
Arizona and New Mexico, experienced the greatest growth.
Many areas of strong growth also occurred along the west
coast as well as in New England. While there does not
appear as strong a bicoastal pattern among Manufacturing
firms (compared to all firms), the areas of greatest decline
are concentrated in the middle of the country.

Small Manufacturing firms along the East coast (both
north and south) appear to be growing much more slowly than
very small firms (Figure 2b). On the other hand, small
Manufacturing firms fared better in the Great Lakes area,
the upper Midwest, and on the West coast than did very small
firms. For example, Detroit and Saginaw, Michigan; Duluth,
Minnesota; and Wausau, Wisconsin all show strong growth in
small Manufacturing firms but only weak or negative growth
in very small Manufacturing firms.

In contrast to Manufacturing firms, growth in very

17

a) CHANGE IN VERY SMALL MANUFACTURING FIRMS

 

 

 

XChonqe

[:I -31.65 16 0.70
[3 0.7016 10.23

- 10.23 16 57.15

 

b) CHANGE IN SMALL MANUFACTURING FIRMS

 

 

 

71 Change
[:] -500016 3.13

I I 3.1310 16.96

- 16.9610 157.15

 

Figure 2. Percent Change in Manufacturing Firms

18

small Finance, Insurance, and Real Estate establishments,
which appeared positive across all regions in Table 2a,
exhibits a strong bicoastal pattern (see Figure 3a), with
the exception of the Northwest. In the Midwest, where very
small firm growth is the weakest in this sector, the Ann
Arbor MSA, along with Kenosha, Wisconsin, two MSA’s in the
Chicago area, Columbus, Ohio, and Columbia, Missouri, stand
out. Much of the remainder of the Midwest shows actual
decline in Finance, Insurance, and Real Estate. In the
Southwest, there is also considerable decline, although two
isolated areas of strong growth are Little Rock, Arkansas
and Austin, TX.

The greatest difference in growth between very small
and small Finance, Insurance, and Real Estate firms is the
lack of the bicoastal pattern for the small firms (Figure
3b). Growth in these firms shows the least amount of
geographic concentration of any of the sectors of either
size category.

Figure 4a maps the growth in very small Service firms,
which, next to small Finance, Insurance, and Real Estate
firms, appear to be the least concentrated, with the
exception of the Southeast. In general, the Service sector
shows strong growth in the Southeast and in the Washington
D.C. area, and parts of the Mountain and Western regions;
and although the Northeast region did not experience the
strong growth (compared to other regions), this was the

Northeast’s strongest sector. In contrast, small Service

19

a) CHANGE IN VERY SMALL F.l.R.E. FIRMS

 

 

  
 
 

[:3 -2120 16 1.62
[:3 1.62 16 14.58
- 14.58 to 60.63

 

b) CHANGE IN SMALL F.l.R.E.FIRMS

 

 

 

” 3 a»
CI 1 g Q; a
4 M u
. X Change
[:I -51.1116 3.17
E] 3.1716 20.16
- 20.16 to 12501

Figure 3. Percent Change in Finance, Insurance, and Real Estate Firms

20

a) CHANGE IN VERY SMALL SERVICE FIRMS

 

  
 
 
 

b) CHANGE IN SMALL SERVICE FIRMS

 

 

Figure 4. Percent Change in Service Firms

 

SM

[3 -44.0610 15.56

E] 15.5810 23.02

- 23.0216 51.67 1
j

 

 

\

7: Change

[:1 -19.151o 31.34
31.34 to 47.47

- 47.47 to 205.45
#4

21

firms (Figure 4b) exhibited a stronger bicoastal pattern
than did very small firms.

Growth in Business Service firms, a subsector of
Services, was quite different for the two size categories.
According to Table 2, in the Northeast, growth in very small
Business Services firms was much weaker than in Services.
However, comparing Figures 4a and 5a, this does not appear
to be the case for very small firms. These firms,
particularly in the New England area, appeared much stronger
in Business Services growth than in Services. It is likely
that the strong growth in Services in the Washington D.C.
area is responsible for the aggregate strength reflected in
the Table 2 figures. In general, in contrast to Service
firms, growth in the number of very small Business Service
firms is much greater in the Southeast and in Florida
(Figure 5a) than in any other region. There was much
negative growth in this sector in the Southwest and even in
the Northwest, with the exception of the Seattle area, This
is somewhat surprising, given the research of Beyers (1990)
and Beyers and Alvine (1985) on the growth of producer
services in this area. However, since this research focused
on firms exporting services (which are generally larger than
those providing services for local consumption), perhaps a
decline in the number of very small firms in this sector is
the reciprocal of an increase in small Busines Service
firms, which both Table 2 and in Figure 5b show to be much

higher in the Western region. In general, the maps of small

22

a) CHANGE IN VERY SMALL BUSINESS SERVICE FIRMS

 

 

 

   

 
 

S Change
(3 -33.3316 3.45

3.4516 17.12
- 17.1216 100.01

 

b) CHANGE IN SMALL BUSINESS SERVICE FIRMS

 

 

    

2 Change
D -57.14 10 9.36
C) 9.36 16 37.50

- 37.5016 400.01

 

Figure 5. Percent Change in Business Service Firms

23

firm growth show considerable variation within regions which
is not evident in the figures in Table 2.

An interesting phenomena that is seen when firm growth
in all sectors is examined is the apparent spatial
polarization by sector. Although many MSA’s, such as
Atlanta, exhibit strong growth in all sectors, there are
numerous MSA’s which fall into the highest growth categories
in some sectors and the lowest in others. For example,
Pueblo, CO shows stronger growth in very small Manufacturing
firms than does adjacent Colorado Spring, yet Colorado
Springs is experiencing very strong growth in Services while
Pueblo shows negative growth in this sector. The same is
true for adjacent Minneapolis and St. Cloud, MN.

To understand why small businesses are growing faster
in some regions than in others, it is necessary to examine

more closely the components of small firm growth.

Components of Small Firm Growth

Change in the number of firms, like population change,
results from birth, deaths, and net migration, i.e. firm
formations, minus business failures, plus net relocations
(including branch plants). It also occurs as a result of
both inter-industry shifts and downsizing within industries.
Since both downsizing and the shift to services are fairly
universal phenomena resulting from restructuring and
competitive pressures, change attributed to these factors
might be expected to vary more by industry than by region.

If this is true, it would be reasonable to assume that,

24

within each industry, most of the regional change in the
number of small firms will result from births, failures, and
net relocations.

Of these three major factors, relocations are expected
to account for less of the variation in small firm growth
than are births and survival rates since, according to Watts
(1987), small firms do not usually make explicit location
decisions. Relocation is more likely to be a factor in
regional small firm growth as a result of branch plant
location. While not dismissing relocation as accounting for
some of the change in small firms, more regional variation
will be accounted for by differences in firm formation rates
and failure rates. Thus the role of human capital in these

two processes will be considered more closely.

New Firm Formation

A considerable amount of literature examines variation
in new firm formation rates at the regional level (Ashcroft,
Love, and Malloy 1991; Bartik 1989; Carlton 1979, 1983;
Gould and Keeble 1984; Lloyd and Mason 1984; Moyes and
Westhead 1990; Schmenner, Huber, and Cook 1987; Watts 1987).
Watts’1987 review of this literature summarized regional
characteristics affecting new firm formation into three
categories: 1) industrial mix, 2) occupational and social
characteristics, and 3) plant size structure. It is the
second category which includes human capital; variables
examined by studies in this category are educational

attainment, occupational mix, age structure, unemployment

25

rates, income, savings, and homeownership. In the analyses
cited above not all of these variables are related to new
firm formation, and some, such as age structure, show
inconsistent results. Others are problematic for a variety
of reasons. For example, as Horiba and Kirkpatrick (1979)
point out, income cannot be used as a surrogate for skill
since male earnings exceed female earnings significantly
when female levels of education are equal or ever greater
than males. Further, although income might be associated
with human capital insofar as it may enhance natural ability
and education, it is more commonly thought of as resulting
from education and abilty. The two measures most commonly
associated with new firm formation are educational
attainment and occupational background, particularly
managerial experience. Of these two variables, education
appears to be most universal.

Although occupational mix is expected to vary more at
the regional scale than at the national scale, in the above
mentioned regional level analyses, educational attainment is
more consistently related to new firm formation than is
occupational background. In fact Bartik (1989), in a state
level analysis which used the variable percent scientists
and engineers as a measure of occupational background (and
using data which allowed him to distinguish between firm
starts and branch plants), found this variable not to be

significantly related to firm starts, although both percent

26

high school graduates and levels of public school spending
were related.

When the location process, as opposed to the creation
process, is considered, higher levels of education appear to
be less significant than other laborforce skills. For
example, in the case of a branch plant location, the
location decision is made by someone outside of the region,
and thus the entrpreneurial aspect of this location decision
would not be a characteristic of the region receiving the
firm; in this case, the region’s laborforce skills would be
more important than factors related to entrepreneurship.
Schmenner, Huber and Cook (1987), in a study relating to the
location decisions of new.manufacturing plants, found that
in general lower educational levels were more important; the
percent high school graduates was positively related only to
the location of manufacturing plants characterized by new
product engineering.

These analyses may indicate that occupational skills,
while important to growth in general, and often significant
factors in a location decision (depending upon the type of
industry involved), are less important in firm formation,
i.e. occupational structure may be more important to
attracting than to creating firms. Much of the creation
process, as part of what is considered the entrepreneurial
factor, is probably not industry-specific, except insofar as
a founder’s industry background oftens determines what type

of firm is started. Education, as opposed to technical

27

training, is both more general and more transferrable.
Insofar as firm formation is dependent upon general vs.
specific knowledge, it will be more strongly related to
education than to occupational background.

Numerous variables related to education are found in
the literature, some measuring educational attainment (years
of schooling, percent high school grads, percent college
grads, etc.). Others are more qualitative, such as
pupil/teacher ratios, dollars spent on education, and
standardized scores, all measures of local school systems.
Generally, all of these education-related variables tend to
be associated with firm formation to some degree, although
not consistently over space, due to the mobility of human
capital.

A study by McNamara, Kriesel, and Deaton (1988)
suggests that variables which measure the quality of
education (pupil-teacher ratios, dollars spent), are less
likely to impact local economic development than are
education attainment variables which measure the quality of
the laborforce (percent with college degree). They
categorize the former as flow variables, and the latter as
stock variables and argue that since the output of a local
areas’s education system does not necessarily remain in the
region (and often does not in rural or depressed areas where
employment is not available), an educational flow variable
is not the best measurement if the focus is on regional

characteristics, such as those which might encourage firm

28

formation. On the other hand, if the focus is on a region’s
potential, e.g. a corporation is considering locating a
branch plant in an area and needs to consider if the school
system can meet its labor needs, then a flow measurement
might be more appropriate. Warner's (1989) comparison of
the impact of education and other human capital variables on
economic growth to cost-minimizing factors, used one measure
of each: percentage of the population with more than 16
years of schooling (stock), and pupil/teacher ratios (flow).
In a regression analysis using data from the 44 Metropolitan
Statistical Areas in the southeast, both variables were

significant predictors of growth (at alpha 8 .05).

Small Firm Survival

Since a high percentage of small firms fail within a
short period of time, regional factors which encourage firm
survival are also quite important. Many of the factors
associated with new firm fOrmation will also be related to
firm survival, but not all. For example, research has
documented instances where firm formation rates increase
with unemployment, the explanation being that the firm
founder had no other opportunity for employment. However,
such a person is not necessarily going to have the education
and skills required to successfully manage a small firm.
.Also, as Bruderl (1992) mentions, if it is unemployment
‘which has motivated someone to start a business, there has
(probably not been time to adequately plan, look for and

«evaluate the best opportunities, or get the best advice.

29

Income also might be less important to the operation of a
firm than during the start-up phase, although Lloyd and
Mason (1984) found that lack of personal capital did tend to
result in chronic undercapitalization of new firms,
resulting in low rates of growth. Which characteristics are
more important to firm survival?

According to Bruderl (1992), contrasting human capital
theory with organizational ecology theory, research relating
to organizational failure indicates that characteristics of
the founder are the key to success; he describes successful
individuals as coordinators, risk-takers, and innovators,
pointing to managerial incompetence and lack of relevant
experience as factors in failure. Human capital not only
increases chances of success after a firm is set up, as it
affects productivity, efficiency, and results in higher
profits, but has an impact befbre. Banks more likely to
loan money to individuals with more education and
experience, and such individuals are better able to get
relevant information, and make good decisions. Other
factors which he indicates are also relevant to firm
survival include prior self-employment, parental self-
employment, and "leadership experience." Although a
regional scale analysis cannot assess characteristics of
individuals, to some extent such characteristics can be
taken into consideration through surrogates. For example, a
variables such as "percent in management“ might reflect

leadership experience.

30

In summary, considerable spatial variation in the
growth of small firms is evident. Most of this variation is
expected to be related to differences in small firm
formation and survival, and only to a lesser extent to
differences in those regional factors which might be related
to the attraction of new firms. Because firm formation and
survival have been shown to be associated with human
capital, it is expected that regional variation in small
firm growth is related to regional variations in human
capital. The following chapter will take up the questions
related to human capital: how it can/should be measured, how
it relates to economic development at the regional level,

and how it is distributed over space.

CHAPTER III

HUMAN CAPITAL AND DEVELOPMENT

Before developing hypotheses relating "human capital"
to small firm growth, it is first necessary to determine
exactly how human capital is to be defined, i.e. which
variables might be used to measure human capital, and to
consider which of these variables will best operationalize
the model to be proposed. This will also require a more in
depth discussion of how human capital functions at the
regional level and an examination of how how human capital

varies from region to region.

Definition

Although the term human capital may seem somewhat
vague, there is general agreement as to its meaning, and it
has changed little over time. Two-hundred years ago, in The
Wealth of Nations, Adam Smith (1961) defined human capital
as "... the acquired and useful abilities of all the
inhabitants of the society..." and asserted that a nation’s
human capital was an important part of its wealth. In 1962,
Weisbrod identified health, learning, and location
(migration) the principal types of human capital investment;
Thurow (1970, p.1) defined human capital as "...an
individual’s productive skills, talents, and knowledge," and
Salamon (1991, p. 9) describes human capital as ”...the
size, productive capabilities, or useful life of the work

force..." These characteristics go beyond simply labor

31

32

quantity; they reflect labor productivity and
entrepreneurship (organization and management skills), often
considered a separate factor of production (O'Farrell 1986).
As an input to production, human capital has implications
for both the attraction and the creation of industry. To
the extent that labor productivity and entrepreneurship both
vary over space, human capital is a factor in the location
of economic activity and can provide a regional comparative
advantage. Although a return to investment in human capital
is not disputed, quantifying human capital in order to

estimate this return is more problematic.

.Measuring Human Capital

According to Spindler and Forrester (1993, p. 34),
"...the link between education and increased productivity is
generally accepted." Three measures of human capital stock
commonly used in either growth or new firm location studies
are identified by McNamara, Kriesel, and Deaton (1988) as:
1) number of persons 25 years of age or more having a
college degree, 2) median years of schooling, and 3)
percentage of adults with a high school education. Most
studies tend to ignore health investment, and rely on formal
education as the measure human capital (Denison 1962, Glomm
and Ravikuman 1992). At the theoretical level, there is the
problem that some of the expenditure associated with
improving the well-being and abilities of human beings also
represents consumption, not just investment (Salamon 1991).

And, as Salamon points out, although this may be somewhat

33

true of investment in education and training, it is far more
true of other forms of human capital investment, such as
investment in health care. Another reason for the focus on
education, as Becker (1962) pointed out, is that in
developed economies, earnings are far more strongly related
to education than to physical ability and strength; thus
investment in education is seen as more directly related to
development than investment in health. Not only is the
return on investment in healthcare more difficult to
quantify, but it may take longer to be realized.

According to Parnes (1984), this concentration on
education further tends to be limited specifically to skills
and abilities that have required some investment to acquire
and that are in demand in the labor market. This specific
focus on formal education is due in part to the fact that
differences in natural ability and experience are more
difficult to quantify. Also, as Schultz (1991) points out,
at least within large populations the distribution of
inherent, as opposed to acquired, abilities probably does
not vary significantly.

McNamara, Kriesel, and Deaton (1988) suggest that one
of the problems with demonstrating the relationship between
education and economic development may be a result of the
failure of much of this research to distinguish between
human capital stocks and flows. Most of these studies have
used various measurements of educational attainment, which,

according to McNamara, Kriesel, and Deaton, would be

34

considered human capital stock, since they measure the
existing level of education, the levels necessary to support
the existing economic structure. In contrast, the authors
argue, per pupil expenditures, the percent of teenagers in
high school, and standardized test scores measure human
capital flows, since they reflect marginal change in
educational attainment. They appear to suggest that perhaps
flow measures would be more important to attracting
industry, since a relocating firm’s concern is with
laborforce potential, skills which will not outmigrate if
appropriate jobs become available. A study which appears to
support McNamara, Kriesel, and Deaton's position is that of
Killian and Parker (1991). Results of this study, which
compared education variables in both metro and nonmetro
areas, indicated that although educational attainment was a
significant factor in employment change at the metro level,
it was not at the nonmetro level. In the rural areas, the
initial job mix was a better predictor of employment change
than educational attainment, a result which they attributed
to the problem of the outmigration of human capital from
areas of low job opportunity.

The fact that labor is mobile, that human capital,
unlike fixed capital, does not have to remain at the
location of the investment, raises another issue related to
assessing the impact of investment in education,
particularly in rural areas. If appropriate job

opportunities do not exist in a region, a condition more

35

likely in rural areas, increasing the education or skills of
the laborforce may only result in the outmigration of this
capital, as reflected in Killian and Parker’s (1991)
results.

With respect to the determining the economic impacts at
the regional scale, the human capital argument depends to
some extent on the assumed immobility of the laborforce.
And, although it is true that labor is more mobile at the
regional level than at the national or international scale,
according to the Committee for Economic Development (1987),
both economic and demographic factors are contributing to
declining workforce mobility. Watts (1987) also questions
the mobility of labor, suggesting that the assumption that
labor follows jobs needs to be examined more closely. To
the extent that labor is not as mobile as it once was (or
was assumed), the possibility of human capital as a factor

in local economic growth increases.

Human Capital as a Regional Factor

At the national scale, the role of human capital in
development is well documented (Psacharopoulos and Woodhall
1985, Schultz 1971, Thurow 1970, Weisbrod 1962). At the
individual level, the return to investment in education, in
the form of future earnings, has also been verified (Schultz
1971). Its importance at the regional level, however, at
least until recently, has not been emphasized (Beyers,

Christopherson, Erickson, Gibson, Hewings, Malecki,

36

McConnell, and Rees 1990; Haider 1992). Spindler and
Forrester (1993) indicate that human capital development
policies focus on national economic growth, and according to
McNamara, Kriesel, and Deaton (1988, p.61), "...limited
progress has been made in isolating the spatial impacts of
specific local investment in human capital on economic
development."

Although evidence linking human capital to development
at the regional scale may be less than conclusive, there has
recently been increased interest in human capital as a
factor which might provide a local competitive advantage.
Haider (1992. p.127), discussing what he calls place
advantage, suggests that with the globalization of markets,
local competitive advantages change more frequently, and
"...people and their know-how and knowledge have become more
important than places and things." In other words, in a
scenario where, increasingly, comparative advantages
associated with relative location and resources are
changing, it is the human capital which can provide
stability to a region.

Clarke and Gaile (1992) indicate that, in addition to
global economic trends, cutbacks in federal economic
development programs are resulting in what they refer to as
the "new centrality of locality," which emphasizes place-
specific attributes as competitive advantages. Although in
the past such place-specific advantages might have centered

on a region’s natural resources or its industrial base,

37

today it is often the type of labor force skills availabale
in a region that determines what type of economic activity
exists in the region. As O'hUallachain (1991, p.73) says,
”Differences between places are based less on specialization
by sector and more on intrasectoral specialization by type
of labor process."

One manifestation of human capital that has received
little consideration as a regional attribute is
entrepreneurship. Although economists have recently
stressed the importance of this factor in responding to the
changing economic conditions (Schultz 1990, Salamon 1991),
Leven (1985, p.576) states that at the regional level,
"...there are really no definitive studies of payoffs of
improved human capital to the investing region or of
policies to promote entrepreneurialism per se." Haider
(1992, p.128) also mentions the lack of study relating to
this factor, indicating that "...we have not yet discovered
why some places are more entrepreneurial than others.” And
Erdivig (1986) suggests that accounting for the spatial
variation in high-tech industries will require deemphasizing
global and corporate factors and focusing on regional
factors likely to foster entrepreneurship.

Thompson (1965) suggested that the lack of knowledge
relating to the role of entrepreneurship may be related to
the difficulty in defining and quantifying this factor.

According to Gillis, Perkins, Roemer, and Snodgrass (1987),

38

the concept of the entrepreneur, as developed by
Schumpeter, was

... someone who had the imagination to see the

potential for profit from the innovation, the

initiative to carry out the task of introducing the
innovation, and a willingness to take a calculated risk
that the effort might fail and lead to a loss rather

than a profit (p. 26).

This definition sounds very much like that of a firm
founder, which is generally what the term is associated
with. Recent studies relating human capital to regional
economic growth (Beyers, Johnson, and Stranahan 1987; Lloyd
and Mason 1984; O’Farrell and Hitchens 1989) have tended to
examine characteristics of the individual entrepreneur, as
opposed to regional characteristics. Many of the studies
which have taken a regional approach have focused on human
capital investment as a rural economic development strategy
and much of this research, as discussed earlier, has not
found strong evidence for a return to investment in
education (Killian and Parker 1991, McNamara, Kriesel, and
Deaton 1988). However, regional approaches which have
considered the impact of human capital at the metropolitan
level suggest that it is a factor in growth.

One such study (Warner 1989), an examination of growth
in per capita income, compared the impact of a human capital
strategy to a more traditional cost-minimizing strategy.
The human capital strategy focused on quality of the labor
force, quality of public goods, and quality of life as

determinants of growth. Variables used in the human capital

analysis were pupil-teacher ratios, percentage of the

39

population with a college degree, and a quality-of-life
variable; the sample consisted of forty-four Metropolitan
Statistical Areas. Warner’s results indicate that these
human capital variables provide a better explanation for
economic growth than do measures relating to cost-
minimization. His findings suggest that more economic
growth occurs as a result of local area firm formation and
growth than from business relocations or branch plant
openings, the targets of cost-minimization strategies. (He
also suggests that this was not necessarily always the case,
that in the past ten years, changes in technology and in the
world economy have lessened the importance of cost factors
and increased the importance of the education and skill of
the laborforce.) Further support for the idea that human
capital is a significant factor at the regional level is
provided by Rauch (1993), whose MSA-level study concluded
that increases in factor productivity were related to the
average level of formal schooling.

As discussed in the introduction, treating human
capital as a regional factor assumes that significant
differences in the distribution of human capital across
regions do exist (and that regional differences in rates of
small firm growth also exist and are related to differences
in human capital). Thus, before turning to the discussion
of small firms, the spatial distributions of human capital

will now be examined.

Reg;

ca;

mea

ti.
ad

of

VE

d!

40

Regional Patterns of Human Capital

As suggested in the above section on measuring human
capitall, formal education is the most widely accepted
measure of human capital. Thus, the following examination
of the distribution of human capital by Metropolitan
Statistical Areas will use three variables related to
education: the percent of the population age 25 and older
with a college degree (COLLGRD), the percent of the same
population with only a high school education (HSGRD), and
the level of local spending on education (EDEXP). In
addition, a measure of occupational background, the percent
of the labor force in management occupations (PCMGT) is
included. (A thorough rationale for the selection of all
variables to be used in this study, along with more precise
definitions, is presented in the chapter V.)

In Table 3, MSA-level data are aggregated by census
region. These figures suggest that the MSA’s in the
Mountain and Western regions (followed by the Northeast

region) are the more human capital rich, having both the

highest levels of college graduates and the highest levels

TABLE 3. Human Capital by Census Region

 

_R_e_gllJn COLLGRD(%) HSGRD(%) PCMGT(%) EDEXP(%)
Northeast 18.68 35.49 1 1 .36 46.70
Southeast 15.62 32.68 10.70 44.60
Midwest 16.53 37.51 10.39 46.20
Southwest 18.10 30.94 10.92 49.70
Mountain 21.03 35.11 12.15 45.10
Western 19.81 32.32 1 1 .97 43.60

 

41

in management occupations. The Southeast region appears
most deficient in human capital, while the Southwest and
Midwest regions fall somewhere in between, but with very
different human capital profiles. While both have
approximately the same proportions of the laborforce in
management occupations, the Midwest has fewer college
graduates but more high school graduates than the
Southwest. Interestingly, the level of local per capita
education spending does not appear to be related to other
measures of human capital; i.e. it is highest in the
Southwest, not one of the stronger human capital regions,
and lowest in the West, which has relatively high levels of
human capital. This may indicate that MSA variations in
local per capita expenditures on education vary considerably
as a result of state level differences in methods and levels
of funding.

MSA patterns in human capital are mapped in Figures 6-
9, and all of the four variables exhibit distinct variations
across MSA’s, but not all reflect broad regional
differences. Figure 6, mapping percent college graduates,
shows only a few MSA’s in the highest category; they were
widely scattered, geographically, and are, generally,
college towns (the Washington DC area; Ann Arbor, MI;
Madison, WI; Bloomington, IN; Columbia, MO; Iowa City, IA;
Bryan/College Station, TX; Santa Fe, NM; Lawrence, KS; and
Boulder, CO). The most noticeable absence of college

graduates is in MSAs in the southeast.

42

 

owns 0. modn I
moon 6. 8.3
OWN». 3 mmdp D
co:o_5aoa +3 .0 R

 

02on owe—EU e E3 core—soon +3 05 mo 8083 .9 0.5»:—

 

 

43

In contrast, a more definite regional pattern emerges
in Figure 7, with the high proportion of high school
graduates in Midwest region MSA's quite evident. Higher
percentages of high school graduates in the parts of the
northeast are also reflected in this map, although the
almost equally high percent in the Mountain region
(indicated by Table 3) does not appear as pervasive in
Figure 7. Since this variable measures those with only a
high school diploma (It doesn’t include college graduates),
it was expected that the pattern would to some extent be the
inverse of the college graduate pattern (R = -.359); i.e. an
area might scoring low on HSGRD precisely because it has a
high percentage of college graduates. This appears to be
true for most areas. On the other hand, an area with low
scores on both HSGRD and COLLGRD would be an area of very
low human capital. It is in many areas of the south that
this pattern is most noticeable.

As mentioned above, the deficiency in formal education
in the south does not appear to be related to the percent of
local budgets spent on education. No discernable spatial
pattern appears to exist with respect to this variable
(Figure 8). This might reinforce the idea that perhaps most
of the variation in local dollars spend on education is a
result of state-level differences in the amount of state
funding is available to local schools. On the other hand,
the range displayed by this variable is not that great;

MSA’s in the lowest region averaged 43.6% while those in the

44

 

Ban 9 was. I
met 3 on? m..w...........a
our... 3 KS D

5.6.366 +3 .6 u

 

 

 

Essa 626m .3: a as. 85.38 +2 65 a6 E83 s 2:9...

       

 

. 1‘ fl...
' V “Fawrl‘sx. . l’ '6
any a S.

no... ‘oi‘

 

th<DQ<mG wwwjoo

45

 

KS 6. one. I

on? 2 8.? ....

8.? 2 8.0 D
00:98 N

 

Cowvsm .80.:0 .xb moose—55%..“— aouaosem =9:— aom .83 .a 0.5»:—

 

 

zo_._.<o:om zO 02.02wdw 1200..

46

highest averaged 49.7%. Figure 8 does reveal intraregional
differences which do not show up as clearly in Table 3. For
example, there is a distinct difference in the percent of
local budgets spent on education in MSA’s in the northern
part of the Western region (Oregon and Washington) compared
to those in most of California.

The percent of the laborforce in management
occupations, seen in Table 3 to be highest in the Northeast,
Mountain, and Western regions, does not appear in Figure 9
as a strong regional pattern. This map looks a lot like
that of college graduates, but with somewhat more of a
bicoastal pattern, and not as many southern MSA’s in the
lowest categories. One area where the percent in management
is much greater than percent college graduates is Florida.

To summarize, generally the east and west coasts and
the Mountain region appear most human capital rich, and the
deep south, especially the south central area, appears to be
human capital poor, with low levels of human capital in all
categories. Although this descriptive analysis does not
address the issue of whether there is a relationship between
human capital and small firm growth, it does serve to verify
that there are great differences in levels of human capital
over space and, thus the possibility that human capital may
account for some of the spatial variation of small firm
growth is reasonable to consider.

The next chapter will consider the theoretical links

between human capital and small firm growth, along with what

47

 

3.6. 6. .98. I
9.2 6. 9.6 ......M._..,
mam 6. 66 _III_

emcee-U N

 

 

823980 «costume—32 5 020m 3&3 05 Lo E083 .a 0.53..—

 

 

hzw§m0<z<5 2. RED... mOmS .....O w0<hzmommd

48

has been seen of the spatial distribution of these
phenomena. Based upon both of these considerations,
specific hypotheses concerning the relationship between
human capital and the formation and survival of small firms

will be developed.

CHAPTER IV

RESEARCH PROBLEM AND HYPOTHESES

Although the literature reviewed in the previous two
chapters suggests a relationship between human capital and
the growth of small firms, much of this discussion has
focused on separately examining human capital and small
firms as factors in economic growth. Where these two
factors come together is in the context of the changing
global economic environment, an environment which is
increasingly dependent upon innovation, entrepreneurial
skill, flexible production processes, and responsiveness to
change. These are characteristics which are associated with
both higher levels of human capital and with smaller firms
which, because of their smaller size, are often at an

f/advantage in being able to quickly respond to market changes

\\and changes in technology (Bannock 1981, White et. a1. 1988).
To compete successfully in today’ s increasingly global,
increasingly high tech economy requires both better access
to information and greater ability to use such information,
and it seems likely that individual entrepreneurs with
higher educational attainment are more likely to satisfy
such requirements.

The role of human capital inwsmall firm growth has
spatial implications both because the quality of human
capital varies over space and as a result of what Pred

(1977) refers to as spatial bias in the availability of

49

50

specialized information. According to Fred,
...all economic and locational actors - whether
individual entrepreneurs, business firms and
corporations, or government organizations - can choose
only from alternatives of which they have become
aware... The specialized information pertaining to
costs, prices, supply, demand, and technological and
other matters which influences the decisions underlying
city-system growth and development is virtually never
universally available.(p.20)
Thus it can be argued that the education and experience,
human capital, which help to transcend spatial bias in
information availability become even more critical for the
small firm, since small firms are generally more constrained
in their locational choices than larger firms.

The central questions that this dissertation will
attempt to answer are: 1) does human capital play a
significant role in the growth of small firms at the local
level, 2) are there industry-specific differences in the
impact of of human capital, and 3) are there regional

differences in the impact of human capital? Hypotheses

designed to examine these questions will now be discussed.

Which Human Capital Measures?

Both human capital theory and research suggest formal
education as the best measure of human capital. Thus the
first five hypotheses relate to various measures of
education and its impact on small firms in general. They are
not industry-specific.

The first education-related variable to be examined is

percent of the workforce, age 25 and over, with a college

51

degree (COLLGRD). Even though, for many industries, the
skill requirements of the laborforce do not dictate a
college degree, the basis of the argument relating to higher
education is that both the formation and the successful
management of a small business requires a certain level of
entrepreneurial skill, regardless of the type of business;
this skill is expected to be strongly related to higher
education. Thus it is hypothesized that:

Hl: At the regional level, small firm growth is
positively related to educational stock, as
.measured by the percent of the laborforce with a
college degree.

Although it has been shown that the existence of a
large proportion the labor force with a high school
education is often a regional factor considered by firms
relocating or opening branch plants, small firms seldom
relocate or open branch plants. Relocation, therefore, is
expected to account for less of the change in small firms
than are firm formation and survival. Thus, percent of a
local laborforce with a high school education (HSGRD) is
not expected to be a significant factor in small firm
growth. The following hypothesis will be tested:

HZ: The percent of the laborforce with a high school
diploma will not be a predictor of small firm
growth.

Both of the preceding hypotheses are related to

educational attainment, rather than educational quality,
i.e. stock vs. flow measures, according to McNamara,

Kriesel, and Deaton (1988). According to McNamara, et.a1.,

flow measures are also relevant to economic growth insofar

52

as the quality of local education is an indication of the
potential skill of the local labor force and therefore
likely to account for some small firm growth resulting from
relocations. Local educational quality is also, according
to Warner (1989), a reflection of the general quality of
life in a metropolitan area, a characteristic considered
conducive to economic growth. Since research has indicated
that per pupil expenditure on education is strongly related
to standardized scores and thus a valid surrogate for output
quality (McNamara, et.al. 1988), local per pupil expenditure
(EDEXP) is the variable used as a qualitative measure. The
hypothesis is:

H3: Regional variation in small firm growth is

positively related to regional variation in local
‘per pupil expenditures.

Educational attainment variables are expected to be of
greater importance to small firm formation and survival than
qualitative measures. Since qualitative (flow) measures do
not reflect how much of the local school system’s output
remains in the area, they are considered to be more
important to firm attraction than firm creation. According
to the argument put forth by McNamara, Kriesel, and Deaton,
qualititive, flow measures of education represent the
incremental additions to human capital that a location-
seeking firm might expect from the local workforce, after an
initial lag period, if employment opportunities were
available. In contrast, stock measures such as the

educational attainment of the region’s laborforce would be

53

better indicators of the actual, as opposed to potential,
level of entrepreneurial skill. Since the
creation/management component of small firm growth is
expected to be more important than the attraction component,
educational attainment is expected to be a better predictor
of growth than a qualitative measure. Thus, the following
hypothesis is added:

H4: The percent of the laborforce with a college degree
will be a greater predictor of small firm growth
than will local per pupil education expenditures.

The percent of the laborforce with a high school

diploma is, like the percent with a college degree,
considered a measure of educational attainment, rather than
of quality; however, like the educational quality measure,
it is expected to be more important to attracting industry
than to creating industry. It must also be acknowledged
that the labor needs of some industries can best be
satisfied by high school graduates, and thus the percent of
the population with a high school diploma might be a better
predictor of growth for such industries. For a large,
labor-intensive firm, seeking to locate a branch plant this
might be an important consideration. However, for small
firms, which are not being managed from outside the area,
percent of college graduates, the variable thought to be
most strongly associated with firm formation and survival,
is expected to be a stronger factor than the variable

measuring general laborforce skills. Thus, the following

54

hypothesis is suggested:

H5: The percent of the laborfOrce with a college degree
will be a greater predictor of small firm growth
than will the percent of the laborfbrce with a high
school education.

To briefly summarize these hypotheses relating to
education, in the aggregate analysis, the impact of a
college education is expected to be a greater factor in
small firm growth than the other educational measures. The
primary reason for this expectation is that for small firms,
the creation and management components of growth are thought
to outweigh other laborforce considerations.

Watts’ 1987 review of relevant literature suggests that
occupational characteristics are a significant factor in
firm births, and this literature suggests that occupational
background is probably a greater factor in small firm
survival than in small firm births. Although growth in any
size firms depends not only upon firm births but also upon
survival, because the failure rate of small firms is much
greater than that of larger enterprises, factors influencing
survival are expected to be particularly important to small
firms. According to Bruderl (1992), small businesses are
more likely to succeed when their founders/managers have
prior management experience. This leads to a sixth
hypothesis:

H6: The growth of small firms is positively related to

the percent of the laborfOrce in management
occupations.

As discussed previously, empirical evidence suggests

that occupational background, while important to firm

55

survival, is not as strong a factor in new formation as
educational attainment. This could indicate either that
firm formation is a more important component of small firm
growth than is survival, or that a formal education is more
important to successful management than is actual experience
in management. Regardless of the explanation, since
education seems to be the more transcendent factor, the
following hypothesis will be tested:

H7: The percent of the laborforce with a college degree
will be a greater predictor of small firm growth
than will the percent of the laborforce in
management occupations.

Which Size Small Firms

Beyers, et.al. (1987) found that among producer service
firms, founders of export-oriented firms were even more
likely to have a least a bachelor’s degree. This makes
sense insofar as managing an export-oriented industry
requires increased knowlege concerning distant markets and
more understanding of nonlocal economic conditions.
Location theory would tend to reinforce this as it indicates
that since basic industries are less constrained by the need
to locate near their markets, the role of other locational
factors, such as the need for skilled labor, may be
relatively more important. Unfortunately, it would be
difficult, within the framework of the present research, to
test the hypothesis that the impact of a college education
is greater for basic than for nonbasic firms. However, such

an hypothesis does suggest something with respect to which

56

size category of small firms might be most impacted by
education. If one assumes that larger firms are more likely
to be export-oriented, for such firms, education is probably
more important than for very small firms. Also supporting
this hypothesis is the observation made by Acs and Audretsch
(1989) that

...small firms are more inhibited than their larger

counterparts from entering industries which have a high

technological environment and where human capital plays

an important role (p. 468).

Based upon these considerations, it is hypothesized that:

H8: The impact of education (specifically a college

degree) will be greater fer the 20-99 size category
than for the <20 category.
Which Industries?

For small firms, for the reasons discussed at the
beginning of this chapter, it is not expected that industry-
specific differences in human capital based upon the labor
needs of such firms will be as important as for large firms.
The impact of both a college education and management
background are expected to transcend industry differences to
a great extent, since the success of small firms in any
industry is expected to be primarily related to their
founders’ and managers’ entrepreneurial abilities. Thus for
all industries, small firm growth is expected to be more
strongly associated with a college education than with any
of the other human capital variables.

Considering industrial structure as a factor in the

impact of human capital is somewhat problematic insofar as

57

industrial structure may not be as important a dimension of
variation as it once was. As discussed earlier, and
supported by O’hUallachain (1991), there is likely to be
more variations among places based upon labor processes than
upon industrial sectors. However, having emphasized this,
it is conceded that some arguments for industry-specific
variations based upon differential labor needs are possible.

Even among small firms the laborforce requirements of

 

u...-.._._—.. _.—m -MHH-__.... _

industries vary, and thus the relative importance of certain

~-.-_

 

human capital variables will differ from industry to

-— Haw-—

industrf:L For industries which are more information

‘M -c~_,.
/ ‘* *-qn..~~.n_.1~....mflm.. h, ...-...... ...... “W ”cm m‘- M'”‘“ ”5" -- “g

intensive, or more dependent upon advanced.technolegy; the

 

 

relatixe importance of a college degree, compared to the
other human capital variables will probably be greater.
Thus, although a college education is hypothesized to be the
most important human capital variable, the relative
importance of human capital variables is expected to show
some variation by industry.

One indication of industry-specific distinctions in

human capital needs is provided by the Bureau of Labor

Statistics’ occupational employment projections summarized

\x. -..t——-"“' "E" ‘LHW'H

......

by Silvestri and Lukasiewicz (1989). For the three
occupational categories most dependent upon higher education
and technical skills (executive, administrative, and
managerial; professional specialty; and technical and
related support occupations), Silvestri and Lukasiewicz

show, for the United States, both the share of employment by

\

58

sector for 1988 and the anticipated share in the year 2000.
These data are summarized in Table 4. According to their
figures, service industries have a higher demand for
professional and technical skills than for executive,
administrative, and management skills. Since it is the
former category that has the greatest higher education needs
(Silvestri and Lukasiewicz), for service industries
educational attainment would be expected to be more
important than management background. Assuming, therefore,
that the laborforce-related educational needs of small firms
in a given sector are similar to those of that sector as a
whole, these findings would suggest that

,Hg/flThe relative impact of a college education on small

firm growth will be greater for service industries,
than fOr.manufacturing.

\‘w

 

“w

W—M m—ouh .m- —

e survey results of Beyers, Johnsen, and Stranahan

'/‘(1987) provide another indication of human capital needs.

'I
I/

When they compared the employment composition of producer
service firms to that of manufacturing firms, they found
that manufacturing firms.had.only_ll% of their workforce in
professignal’angmtgghpical occupations (those requiring the
highest level of education), compared to 43% for producer

servicewiizmax. They also found that the majority of the

\\\p£oducer service firms indicated that most new employees

. M”--.
were recruited from the local area, implying that local

_ - -v-' Vin-"1...,

 

educational "stock" is important. The findings of Howell

w w. Mflm-v ’-

and Wolff (1991) also support the implications of the

 

research of Beyers, Johnsen, and Stranahan. Howell and

59

TABLE 4. Percent Distribution of Wage and Salary Workers by Selected
Occupation and Industry, 1988 and Projected to 2000

 

Occupation
EXIADIMGR PROF/SP TECH

Industry 1988 2000 1988 2000 1088 2000
Goods Producing 8.8 9.5 5.1 5.8 2.6 2.8
Mining 11.9 12.0 7.8 8.2 3.1 3.1
Manufacturing 8.6 9.5 6.3 7.4 3.2 3.7
Services Producing 10.1 10.4 14.7 15.3 3.8 4.3

Transportation and

Communication 7.9 8.4 4.4 5.1 4.3 4.7
Retail 8.3 8.6 1.2 1.3 2.0 2.0
Wholesale 11.6 11.5 1.5 1.8 2.0 2.4
F.l.R.E 22.5 23.8 2.8 3.4 1.9 2.2
Sewices 8.2 8.6 28.1 28.0 6.2 6.9

 

Source: Slvoctn,GoorgeandJohn Lukasiewicz. 1991. Projectlomoi WSW,
1988-20W.Buuudhbw8btbflcoMaflflyLabochvlsw.m1989

60

Wolff show that educational attainment is highly correlated

with both cognitive and interactive skills, and employment

”MM-um.—

growth in jobs requiring these skills is greater for service

industries, particularly advanced services, than for good-

«a

producing industries. These findings would support the

~-.‘_

hypothesis that:

H10: For small business service firms, the percent of
the workforce with a college degree would be of
greater relative importance than fbr other small
service firms.

The correlative of hypotheses 9 and 10 might be as

follows:

H11: For small manufacturing and service firms, the
percent of the workforce with a high school
diploma will be of greater relative importance
than for other sectors.

The data in Table 4 also suggest that for the Finance,

Insurance, and Real Estate sectors, management background

h...‘

might be more important than education. Based upon this

“woe—-

 

information, it might be hypothesized that

312: The relative importance of management background,
compared to education, will be greater for
finance, insurance, and real estate industries
than for other industries.

 

IndustrialTStrugture_fl

\kAlthough a region’s industrial structure is not a
measure of human capital, the literature indicates that the
growth in firms in a given sector is, to a considerable
extent, related to that sector’s existing share of the
region’s activity (Watts 1985). This is due to the fact

that new firm founders are more likely to start and succeed

61

in a business in the same sector in which they have
experience. It may also be a reflection of agglomeration
economies. Since some of the growth in firms in a
particular sector is expected to be explained by that
sector’s share of all firms, sector-specific analyses will
control for this factor by including a variable measuring

the sectgrns share of total employment. The associated
hypothesis is:

H13: Growth of small firms in sector "j" will be
positively related to sector j’s share of the
region’s establishments.

Regional Implications

The descriptive analyses relating to the spatial
distribution of human capital and of small firm change,
discussed in Chapters 2 and 3, were presented, in part, to
verify that both do vary considerably over space. If this
analysis revealed a strong coincidence between the
geographic distributions of human capital and small firm
growth, regional hypotheses relating to the impact of human
capital would be suggested. However, what was most obvious
in Figures 6-9, the maps of human capital variables, is that
only high school graduates (Figure 7), the one variable
hypothesized not to impact small firm growth, shows any
strong geographic concentration (at the regional scale).
The other three variables, percent college graduates,
percent of labor force in management occupations, and local

per pupil expenditures on education, are fairly evenly

distributed over space (particularly the first two, which

62

are expected to be most significant). In contrast, Figures
1-5 (maps of small firm growth) indicate that areas of
relatively strong small firm growth do exist, particularly
in the Southeast, an area of low human capital. The absence
of human capital in regions of strong small firm growth may
suggest that the link between human capital and small firms
is weak (or overshadowed by other factors) in these areas.
On the other hand, it could also be argued that stronger
small firm growth in the face of minimal levels of human
capital (such as in the Southeast) points to the increased
impact of human capital in such areas (more bang for the
buck). Since the first interpretation would be inconsistent
with the overall thesis that human capital is a significant
factor in small firm growth, it is the second explanation
that will be pursued. Since the Southeast region is the
most human capital poor, yet the region displaying the most
consistent growth in small firms, the following hypothesis
is proposed:

H14: The impact of human capital will be greater in the
Southeast than in other areas of the country.

The lack of other obvious hypotheses suggested by the
descriptive analysis is actually wholly consistent with the
general thesis of this research, i.e. that the impact of
human capital is relatively universal. For the same reasons
that human capital’s impact is expected to transcend
industry differences (as discussed in the previous section)

it is also expected to be fairly consistent over space.

63

Thus, while not abjuring the hypothesis just proposed, a
more general hypothesis will also be offered:

H15: The impact of human capital will vary little among
census regions.

Although the descriptive data offer scant reason for
further speculation, theoretical considerations may provide
the basis for further geographic hypotheses. Product cycle
theory suggests that stages in the development of products
influence their location insofar as the labor needs of these
stages vary. In earlier stages, when skilled labor and
urbanization economies are more important, location in
metropolitan/core areas is favored; in the more mature
stages, when production is more routine and less dependent
upon innovation and skilled labor, firms tend to locate in
peripheral and nonmetropolitan areas. Since small firm
growth is expected to be significantly related to new firm
formation, the greatest geographic variation in the
importance of human capital will be along the urban-rural
continuum, a dimension of variation that cannot be examined
in this analysis since only metropolitan areas are included.
There may, however, be reasons to expect differences in the
impact of human capital based upon MSA size.

It is expected that human capital, i.e. education and
management skills, becomes more important as economies shift
to advanced services and become increasingly global, as
industries depend more upon advanced technologies and as
production becomes more information intensive. This would

imply that areas characterized by advanced service

64

industries and technologies, areas which function as hubs in
information networks, etc. would be more human capital
dependent. The direction of this relationship, however, may
be two-way, i.e. human capital might also be dependent upon
such areas, or at least its impact might be enhanced by the
urbanization economies associated with larger metropolitan
areas. Further, it might be argued that large MSA’s and
small MSA/s are different, even if not extreme, points along
the rural-urban continuum. The theoretical principles
discussed above might then suggest that the impact of human
capital would be greater in larger metropolitan areas than
in smaller MSA’s. Thus the following hypothesis is
proposed:

H16: The impact of human capital will be greater in
large MSA’s than in small MSA’s

As already discussed, a firm’s labor needs are also
related to the type of industry, although perhaps not as
much as in the past (O’hUallachain 1991). Nevertheless, in
the aggregate (vs. industry-specific) analysis, some
regional patterns in the importance of human capital may
result from variation in regional industrial structure. For
example, a college education might be expected to be more
important in areas of the country associated with high tech
and advanced service industries. However, because certain
regions of the country rely more on human capital because of
the demands of their industrial structure does not
necessarily mean that the impact of this human capital is

greater, only that there is likely to be more of it. In

65

other words, an area whose economy is based upon agriculture
will undoubtedly have a smaller proportion of its population
with a college degree than one whose economy centers on
advanced services; but assuming both areas are experiencing
growth in some types of small firms, is there any a priori
reason why the impact of college education on the growth of
these firms would be less in the agricultural region?

Other theoretical considerations have not suggested
further explanations related to regional differences in the
impact of human capital on small firm growth. Although the
importance of amenities has been emphasized as a locational
factor (Stafford 1974), particularly for corporate
headquarters and industries dependent upon attracting a
highly educated laborforce, since small firms do not usually
attract labor from outside the local area, the amenity
factor is not likely to be a significant in their growth.

More recently, the locational implications of
laborforce requirements have been related to not just
laborforce skills, but to social aspects of production such
as capital/labor relations (Clark 1988). Thus much of the
growth in the South has been attributed to industry’s need
for cheap labor and its desire to minimize capital/labor
conflict, neither of which are usually associated with areas
of high human capital.

In summary, it seems that there are not many reasons
for expecting regional differences in the impact of human

capital, at least not within the purview of this study. The

66

focus of this analysis is on the type of human capital and
firms for which variation is expected to be greater along
the urban-rural continuum than at the interregional level.
Thus, as hypothesized above, with the exception of the
Southeast region, little interregional variation in the
impact of human capital is expected. Further speculation
concerning regional patterns will follow evaluation of the

results of the analysis.

CHAPTER V

ANALYSIS

This study will employ regression analysis to measure
the impact of human capital on the growth of small firms,
across both industry sectors and census regions, and
regional patterns will be further analysed by examining
residuals from the regression analysis. This chapter will
identify the data and explain the methodology for testing

the hypotheses developed in the previous chapter.

Data

The unit of measurement will be the Metropolitan
Statistical Area (MSA). The study will include all MSA’s
and Primary Metropolitan Statistical Areas (PMSA’s) in the
coterminous United States, with the exception of the 26 New
England MSA’s. MSA’s in New England, unlike those in the
other census regions, are not defined in terms of counties.
Since the data being used for the dependent variable (change
in small establishments) is only available by county, the 16
New England County Metropolitan Areas (NECMA’s), which are
defined by county boundaries, will be used in place of the
MSA’s in this region. Altogether, the study will include
317 metropolitan areas (see Appendix B for complete listing
of MSA/county equivalents). Although many studies of
regional growth use counties to measure variables, MSA’s are

thought to better represent functional economic areas.

67

68

Counties are simply political units, whereas, according to
the Statistical Abstract (1989, p.908), the concept of an
MSA “...is one of a large population nucleus, together with
adjacent communities which have a high degree of economic
and social integration with that nucleus.”

One problem with the use of MSA/s that does not occur
when counties are used is that MSA.definitions change over
time. Since the dependent variables will be calculated as a
rate of change over time, the spatial units of measurement
must be the same at both ends of the time period used.
Therefore, the most current MSA definitions (1988) will be
used for both the beginning and ending time periods. The
reason for this is that, assuming growth in small firms
reflects a growing MSA, such growth might be underestimated
if the earlier MSA definition were used. Of course it is
also true that if the number of small firms is declining,
use of the more recent MSA definition might bias the results
in the opposite direction, but it is assumed that this is
the more unlikely error, since the MSA is less likely to
have grown if the number of establishments are declining.

For all variables, county-level data will be aggregated
to the MSA-level before transformations are calculated. For
example, to determine the percent change in small Service
firms for the Lansing, Michigan MSA, the number of small
Service firms in Ingham, Eaton, and Clinton counties (the
three counties making up the Lansing MSA) in 1983, will be

totaled; the number of small Service firms in these counties

69

in 1989 will also be totaled. These aggregate figures will
then be used to calculate the percent change in these firms
for the Lansing MSA.

The dependent variable will be the percent change (from
1983-1988) in the number of small establishments by sector
and by size category. (This time period is dictated by data
availability.) The source of these data is the U.S. Census’
County Business Patterns. It is the expectation that this
net growth figure will reflect both the creation and the
survival of small firms. Firm birth data, such as the Dun
and Bradstreet data used by many studies focusing only on
firm formation, give no indication as to the success of
these firms. As pointed out by Acs and Audretsch (1989),
such data do not account for business failures. Since a
large percent of new small firms fail within a short time
(Bruderl 1992), the percent change in the number in small
firms over time is a net figure, which reflects both firm
creation and survival rates, is a better indicator of the
contribution small firms make to the economy.

The use of establishment-level data as a measure of
small-firm based economic growth is dictated by the
unavailability of employment data by firm size
classifications. However, although it is more common to use
employment data to measure economic growth, the use of
establishments is not without some advantage. As White and
Osterman (1991, p.242) point out, ”The number of jobs in a

particular location is more important than an aggregated

70

number of jobs that a firm may have at multiple locations."
Whether or not an establishment is part of a larger
enterprise, it is counted as a separate entity because it
has an independent location and thus is important for local
development.

Two establishments size categories (based upon number
of employees), as defined by the Small Business
Administration, will be examined. They are: very small (1-
19 employees) and small (20-99 employees). The four
industry sectors included are: Manufacturing; Finance,
Insurance, and Real Estate; Services; and Business Services
(SIC 73), a subdivision of services.

Data for the independent variables, will come from the
KU.S. Census. Since the hypothesized relationship between
these variables and the dependent variable is causal, and
some lag time is generally assumed between cause and effect,
1980 data will be used for the independent variables. The
rationale for the use of these variables has been discussed
in previous chapters and so will not be repeated. As
mentioned earlier in this chapter, all data are measured by
county and then aggregated into MSA’s, with the exception of
local per pupil expenditures, which was available only by

MSA. The variables are:

71
* Percent of the population 25 years and older with a
college degree (COLLGRD)

* Percent of the population 25 years and older with
only a high school education2 (HSGRD)

* Percent of the local expenditures for education
(EDEXP)

* Percent of the laborforce in management occupations
(PCMGT)

* Percent of the laborforce employed in each of the
four industries listed above (INDSHR)

In relation to the "local expenditures for education"
variable, it should be noted here that the map of this
variable (Figure 3), while reflecting considerable within-
state variation (see, for example, Florida), also shows
entire states with much lower level of local spending, e.g.
California. This suggests that some of the state-to-state
variation in local spending on education may be related to
differences in state funding for education, and thus it may
not be a very good indicator of total spending on education.
Nevertheless, because considerable within-state variation
was also exhibited, it was decided to include this variable

in the analysis.

 

zThe reason for using a figure which does not include high
school graduates who also have a college degree is explained
later in the chapter, in a discussion of colinearity.

72

.Methodology
Regression analysis will be used to test all
hypotheses. The generic model, used to test the first six

hypotheses, is as follows:

{2} CSF = a + b1COLLGRD + bZHSGRAD + b3PPEXP + b4MGT + e

where:
CSF = percent change in the number of small
firms
COLLGRD = percent of 25+ age group with a college

degree

HSGRAD = percent of 25+ age group with a high
school diploma (only)

PPEXP percent of local expenditures for

education

MGT = percent of laborforce employed in
management occupations

e - error term; a, b1, b , b3, and b4 are

parameters to be estimated

\

When the equation is run for specific sectors, another
independent variable, the industry share of total

employment, will be included, producing the following:

{3} CSFi = a + b1COLLGRD + b2HSGRAD + b3PPEXP + b4MGT +

where: INDSH = industry i’s share of total employment

This equation will be modified by the addition of a
dummy variable to reflect growing vs. declining areas. It
will be recalled that an underlying assumption of this

research is that areas where the number of small firms are

73

increasing are areas whose economies are growing, i.e. that
there is a positive relationship between overall growth and
growth in small firms. However, it must be considered that
some relative increase in the number of small-firm.might be
accounted for by general economic decline in a region, i.e.
as employment in larger firms decreases, these firms shift
into smaller size categories. If the relative number of
small firms is increasing in both growing and declining
regions, it is possible that, statistically, there will be
either no relationship or a negative relationship between
overall growth and the growth of small firms, a situation
likely to confound the analysis since these two types of
regions would be unlikely to have similar human capital
profiles. Two approaches to dealing with this problem are
considered.

The first approach would be to test for a relationship
between growth in total employment in a given sector and
growth in the number of small firms in that sector. This
has been done (regression analysis in Chapter 2), and the
results are shown in Table 1. Although the relationships
are positive for all sectors, indicating that small firm
growth is occurring in growing region (not as a result of
decline), because coefficients are not strong, a further
test is proposed.

To determine if increases in small firms are occurring
in declining regions, some means of distinguishing between

growing and declining regions might be useful. This could

74

be accomplished using a dummy variable. A dummy variable
will allow the impact of human capital in regions which are
experiencing growth to be distinguished from the impact in
regions experiencing decline. Values of the dummy variable
will be based upon a region’s employment growth relative to
the nation’s employment growth. This measure can be made
industry-specific for testing hypotheses relating to
specific sectors. Growing regions will be those in which
the regional rate of change in total employment for industry
"j” is equal to or greater than the national rate of change
for this industry; declining regions will be those where
this rate is less than the national. In other words, a
location quotient will be calculated as follows:

L9 = Eij/Euj

where: Eij 8 percent change in employment in region i,
sector j

ENj = percent change in employment in the nation
in sector j

The sector-specific model would then be as follows:

{4} csri = a + b1COLLGRD + bZHSGRAD + b3ppexp + b4MGT +
b5INDSH + bGD + b7(COLLGRD*D) + B8(HSGRAD*D) +

b9(PPEXP*D) + B10(MGT*D) + b11(INDSH*D) + ei

growth in total employment and growth in small firms is

where:

CSFi

COLLGRD

HSGRAD

PPEXP

MGT

INDSH

D

ei

If analysis

indicates that the relationship between

75

percent change in the number of small
firms in industry i

percent of 25+ age group with a college

degree

percent of 25+ age group with a high

school diploma

per capita expenditures on education

percent of laborforce employed in
management occupations

industry i’s share of total employment

dummy variable to distinguish growing

regions from declining regions

0 = growing
1 a declining

error term

positive, and thus there is no need to use dummy variables

to distinguish growing from declining regions, equation (5)

will incorporate five regional dummy variables (plus an

intercept dummy) to distinguish the six census regions; the

default will be the Northeast region.

equation is:

{5}

CSFi = a + b18E + bZMW + b3sw + b4MT + b5WST + bGHC

The resulting

b7(se*acj) + b8(MW*HCj) + bg(sw*ch) +

b10(MT*HCj ) '1' b11(WS*HC

j)

+ ei

j'l'

where:

CSFi

SE

MW

SW

MT
WST
HC
SEHC-
MWHC

SWHC-

U U. U

MTHC

U.

WSHCj

76

percent change in

in industry i

intercept dummy =

intercept dummy =

intercept dummy =

intercept dummy =

intercept dummy =

human capital
slope dummy a
slope dummy 8
slope dummy =
slope dummy =

slope dummy -

Equations incorporating

the number of small firms

1

#1 ha 1H #-

if Southeast Region
if Midwest Region
if Southwest Region
if Mountain Region

if Western Region

variable j

SE*HC-

3
* .
MW HCJ

SE*HC-

J

MT*HC'

J

* 0
WS HCJ

regional dummy variables will

be run separately for each independent variable (since to do

all four human capital variables would have required 48

independent variaables) and, for each sector, will be run

only for those variables which were significant in the

equations {2} or {3}.

Finally, differences in MSA size will be tested by

incorporating dummy variables into the equation {3} to

distinguish large MSA’s (defined as MSA’s with a labor force

greater than 150,000) from small MSA’s. This equation will

be:

{6} CSFi =

a + bICOLLGRD + banGRAD + bBPPEXP + b4MGT +

bSINDSHi + b6D + b7(D*COLLGRD) + b8(D*HSGRD) +

b9(D*EDEXP) + b10(D*MGT) + b11(D*INDSHi) + e

77

where: D - dummy variable to distinguish MSA size
0 = labor force < 150,000
1 - labor force 2 150,000

The possibility of colinearity among the independent
variables was considered. In a preliminary investigation
using the 84 MSA’s of the North Central census region,
multiple regression output showed tolerances between COLLGRD
and PCMGT to be .9010183, clearly not an indication of
colinearity. In the current analysis (n = 317), correlation
between COLLGRD and PPEXP was examined using both Pearson
and Spearman tests; the results indicated correlation
coefficients of -.401 and -.418, respectively, indicating
some degree of colinearity, but not enough to seriously
compromise the analysis. Correlation between HSGRAD and
PPEXP was much lower (.044). A serious colinearity problem
did exist, however, between COLLGRD and HSGRAD, Pearson's
and Spearman’s being .765 and .707, respectively. Since
these two variables are already percentages, the problem
cannot be solved with a transformation. Likewise, putting
the four human capital variables through a principal
components analysis to produce two or three orthogonal
variables would defeat the purpose of including both
educational variables to begin with (the purpose being to
discrimate between the impact of different levels of
education). One of the reasons why these two variables are
collinear is that the HSGRAD figure includes COLLGRD
(assuming all college graduates are also high school

graduates); when COLLGRD is subtracted out of HSGRAD,

78

leaving a figure that represents the number of persons with
only a high school education, this figure should not (and
does not) exhibit the same amount of colinearity with

COLLGRD.

Hypotheses Testing

All hypotheses, along with the criteria for their
acceptance, are shown in Table 5. The first eight
hypotheses will be tested using equation {2}, since these
hypotheses are not industry-specific. Acceptance or
rejection of hypotheses {4}, {5} and {7} will depend upon
comparing coefficients within equations; thus, if more than
one of the coefficients in the equation is significant, it
will be necessary to determine if the coefficients are
significantly different from each other (as opposed to
significantly different from zero). SinCe each coefficient
is theoretically the mean of a normal distribution with
standard deviation equal to the standard error of the b-
value, a difference of means t-test can be conducted.
However, since the samples from which the estimates are
derived are not independent, the standard error of the
difference between means for correlated groups is used for
the denominator (Kendall and Pigozzi 1994, Runyan and Haber
1980):
{7} le-x2 ” 52x1 + 82x2 ' 2rsxlsx2

where "r" is the correlation between the two samples.

79

TABLE 5. Hypotheses Testing

HYPOTHESIS
1 Regional variation in small firm growth is positively related
Ito regional variation in educational stock, as measured by

the percent of the laborforce with a college degree

2 The percent of the laborforce with a high school diploma
will not be a predictor of small firm growth

3 Regional variation in small firm grow is positively related to

regional variation in local per pupil expenditures

4 the percent of the laborforce with a college degree will be a
greater predictor of small firm growth than will local per
pupil expenditures

5 The percent of the laborforce with a college degree will be a
greater predictor of small firm growth than will the percent of
the laborforce with a high school education

6 The growth of small firms is positively related to the percent
of the laborforce in management occupations

7 The percent of the laborforce with a college degree will be a
greater predictor of small firm growth than will the percent of
the laborforce in management occupations

8 The impact of human capital will be greater for the
20-99 size category than for the 1-19 category

9 The relative impact of a college education on small firm
growth will be greater for service industries than for
manufacturing

10 For small business service firms. the percent of the labor-
force with a college degree will be of greater relative
importance than for other small service firms

EQUATION #

2

TEST
accept if b1 is positive

and significant

reject it b2 is positive
and significant

accept if b3 is positive

and significant

accept if b1 > b3 and
significant

accept it b1 > b2 and

significant

accept it b4 is positive

and significant

accept if b1 > b4 and

is significant

81 (small firms) > 81

(very small firms)

81 (service firms) > 81

(manufacturing firms)

81 (business service

firms) > 81 (services)

80

 

TABLE 5 (cont'd)
HYPOTHESIS EQUATION # TEST
11 The percent of the laborforce with a high school diploma will 3 82 (manufacturing
be of greater relative importance for small manufacturing firms) > 82 (other
and service firms than for other sectors sectors)
12 The relative importance of management background. com- 3 B4 (F.l.R.E. firms) > 84

pared to education. will be greater for finance, insurance.
and real estate firms than for other firms

13 Growth of small firms in sector "j" will be positively related to
sector j's share of the regions establishments

14 The impact of human capital will be greater in the Southeast
region than in other regions

15 The impact of human capital will vary little among census
regions other than the Southeast

16 The impact of human capital will be greater for large MSA's
than for smaller MSA's
b = coefficient

8 = beta value

(other sectors)
accept if b5 is positive
and significant

accept if b7 is positive

and significant

reject if b8-b11 are

positive and significant

accept it b7-b11 are po-
sitive and significant

81

Hypotheses 9-13 are industry-specific and thus require
separate equations. These hypotheses will be tested by
comparing beta coefficients within and across equations.
Since all variables are expressed in comparable units
(percentages), such comparisons will be reliable. These
hypotheses do not take the form of asserting, for example,
that b1 in one equation is greater than b1 in a different
equation; rather, they state that the importance of b1
relative to other coefficients in the same equation is
greater than the importance of b1 in a different equation
(relative to other coefficients in that equation). For
example, such an hypothesis might be that when the dependent
variable is the change in manufacturing firms, b1 is more
important relative to other b's than when the dependent
variable is the change in service firms. Firm size

comparisons will be made in the same manner.

CHAPTER VI

RESULTS

The first issue to be resolved in this analysis has to
do with the need for equation {4}, in which dummy variables
were incorporated to distinguish growing regions from
declining regions. It will be recalled that a potential
problem with the basic model was that the model assumes
growth in the number of small firms is occurring in growing
regions, not simply resulting from downsizing in declining
regions. In reality, however, some growth in the number of
small establishments might result from shrinking employment
of larger firms in declining regions. To clarify which
scenario better explains the growth of small firms, two
approaches were suggested. First, it is assumed that if
most of the increase in small firms were the result of
downsizing there would be either no relationship or a
negative relationship between employment growth and the
growth of small firms in a given sector. The relationship
between growth in employment and growth in the number of
small firms was tested and the results (Table 1) indicated
that in all four sectors employment growth and small firm
growth were positively correlated. Thus it is assumed that
growth in the number of small firms is not primarily the
resulting of downsizing occurring in declining regions.
Equation {4} was also run, and the results (not reported
here, since none of the dummy variables were significant)

confirmed that there were no significant differences in the

82

83

impact of human capital between growing and declining
regions. This analysis has, therefore, proceeded on the
assumption that most growth in the number of small

establishments is related to growth, not decline.

General Hypotheses Testing

The first eight hypotheses are tested with respect to
both the aggregate and sector-specific growth in
establishments. In this section, only the results of the
aggregate analysis will be presented.

Hypotheses 1-7 are tested using equation {2}. These
hypotheses all examine the impact of the four human capital
variables on the change in the number of small and very
small establishments in all sectors. In general, it was
hypothesized that all variables except HSGRD will be
significant, but that the percent of the population with a
college degree will have the greatest impact. The results
indicate that while PCMGT was significant for very small
firms, none of the education variables were significant in
the growth of aggregate small firms of either size category
(Table 6). Thus, for all sectors, hypotheses 1, 3, and 4
must be rejected. Hypothesis two, which stated that HSGRD
would not impact the growth of small firms, can be accepted
since this variable was not significant. Hypothesis six,
that the growth of small firms is positively related to the

percent of the laborforce in management occupations, can

84

TABLE 6. Regression Results: Equations 2 and 3

 

Sector a COLLGR HSGRD EDEXP PCMGT INDSHR ADJ R2
All Sectors
1-19 employees -15.409 0.186 0.143 0.064 1.733 0.050
(1 .014) (.900) (.603) (3.136)*
20-99 employees -10.288 .175 .293 .039 .912 0.052
(.559) (1.081) (.215) (.965)
Manufacturing
1-19 employees -7.264 0.162 0.006 0.139 -0.027 -0.058 0.023
(2.218)” (.091) (2.389)“ (-.360) (-.91 1)
20—99 employees 15.737 -0.001 0.05 -0.081 -0.002 -0.009 0.000
(-0.013) (.803) (-1.376) (-.029) (-.144)
F.l.R.E.
1-19 employees 3.838 0.022 -0.153 0.011 0.287 -0.11 0.090
(.312) (-2.617)*‘ (.186) (3.897)” (-1.706)
' 20-99 employees 14.532 -0.038 -0.042 -0.109 0.21 -0.077 0.033
(-.525) (-.693) (-1 .862)” (2.763)“ (-1.165)
Services
1-19 employees 9.961 0.127 -0.043 0.028 ‘ 0.214 -0.098 0.073
(1 .755)* (-.712) (.488) (3.025)“ (-1 .591)
20—99 employees 24.933 0.191 -0.014 -0.011 0.075 -0.033 0.042
(2.600)” (-.223) (-.182) (1.041) (-.532)
Business Services
1-19 employees 22.389 -0.014 0.038 -0.028 -0.0482 -0.154 0.025
(-.186) (.634) (-.469) (-.603) (-2.077)**
20-99 employees 2.371 0.084 0.081 0.077 -0.058 -0.214 0.062
(.244) (1 .353) (1 .319) (-.750) (2945)“

 

NOTE: For individual sectors. coefficients are beta-value

‘significant at .05
“significant at .01
(t-values)

85

also be accepted for the very small size category, since
PCMGT was positive and significant.

Hypothesis seven, which stated that the percent of the
laborforce with a college degree would be a greater
predictor of growth than the percent of the laborforce in
management must be rejected, since COLLGRD is not
significant. The last hypothesis tested at the aggregate
level (hypothesis 8), that education would be a greater
factor in small firms than in very small firms must also be
rejected since neither of the education variables was
significant for either size firms.

Although this analysis is perhaps not very helpful for
determining the relative importance of various educational
measures of human capital to small firm growth at the
aggregate level, it does suggest that occupational
background may be more important than suggested by much of
the literature reviewed. It also serves to confirm that
colinearity among the independent variables is not a serious
problem. Regression output shows the following tolerances
for the four independent variables: COLLGRD - .59432; HSGRD
- .86046; EDEXP - .93283; PCMGT - .63697.

Sector-specific.Analysis

Results of sector specific analysis, tested with
equation {3}, are also shown in Table 6. The first
hypothesis relates to the importance of college education

and can be accepted for both the Manufacturing and Services

86

sectors. College education is significant for very small
Manufacturing firms and for both size categories of Service
firms. It is not, however, significant for either the
Business Services or for Finance, Insurance, and Real Estate
sectors.

The second hypothesis, that high school education would
not be important to small firm growth, is also accepted. It
is not significant for the Manufacturing, Services, and
Business Services sectors, and is negatively related to the
growth of very small in the Finance, Insurance, and Real
Estate sector. Hypothesis three, that local per pupil
expenditures would be positively related to growth, can be
accepted for very small Manufacturing firms. This variable
(EDEXP) was also significant but negatively related to
growth in small Finance, Insurance, and Real Estate firms.

Hypothesis four, which stated that college education
would be a greater predictor of growth in small firms than
local per pupil education expenditures, can be accepted for
both size Service firms (since EDEXP was not significant).
It can also be accepted for very small Manufacturing firms
since the beta associated with COLLGRD is greater than the
beta associated with EDEXP, and the t-test conducted (using
equation {7}) indicates that this difference is
statistically significant (at .05). Generally, COLLGRD was
significant in three sector/size categories, whereas EDEXP
was only in one. Hypothesis five, that the percent of the

labor force with a college degree would have a greater

87

impact on firm growth than the percent with a high school
diploma is accepted since HSGRD was not positive and/or
significant for any of the sectors.

Hypothesis six stated that growth in small firms would
be positively related to the percent of the laborforce in
management occupations. This was accepted at the aggregate
level and can also be accepted for two of the four sectors
examined: very small Service firms and Finance, Insurance,
and Real Estate firms (both size categories). In addition,
the seventh hypothesis, that COLLGRD would have a greater
impact than PCMGT, can be accepted for very small
Manufacturing establishments and for small Service firms.
For very small Manufacturing firms, this hypothesis can be
accepted since PCMGT is not significant. For both very
small Service establishments and both categories of Finance,
Insurance, and Real Estate firms, PCMGT has a greater impact
than COLLGRD. For the very small Service firms, this
difference is significant (at .99).

The hypothesis relating to firm size, which stated that
human capital would have a greater impact in the 20-99
employee category than in the 1-19 size category, can be
accepted for Services. This is the only sector in which
COLLGRD was significant for both size categories, and the
beta value is not only greater for small firms than for very
small firms, but for small firms it is the only variable
that is significant (while for very small Service firms,

PCMGT is also significant). However, in general, it should

88

be noted that this variable is significant for very small
size firms in two sectors, while for small size firms it is
significant only in one sector. Also, while no hypothesis
relating to firm size was suggested for the PCMGT, it is
perhaps worth noting that in the Finance, Insurance, and
Real Estate sector, where this variable was significant for
both size categories, its impact was greater in the smaller
of the two size categories (based upon a comparison of beta
values), and for the Service sector, it was significant only
for very small firms.

Hypotheses nine, that the relative impact of a college
education on small firm growth will be greater for Service
firms than for Manufacturing firms can be accepted since
COLLGRD is significant for both size categories of Service
firms but only for very small Manufacturing firms. For
small firms, COLLGRD is significant only for Services, not
for Manufacturing. However, for the 1-19 category, the beta
value associated with COLLGRD for Manufacturing firms is
higher than the corresponding beta value for Services firms.

Hypothesis ten states that college education would be
more important in the growth of small Business Service firms
than for other Service firms; this must be rejected.
Although COLLGRAD is positive and significant for both size
categories of Service firms, for Business Service firms it
is significant for neither. This is probably the most
surprising result, since Business Service firms employ a

syreater percentage of people with college degrees than do

89

Service firms as a whole. Hypothesis eleven, that the
percent of the population with a high school degree might be
more important to the growth of small Manufacturing and
Service firms than to firms in other sectors, must be
rejected. HSGRD is not positively related to the growth of
small firms in any of the sectors examined.

The importance of management background to the growth
in Finance, Insurance, and Real Estate firms is strongly
suggested by the coefficients associated with PCMGT for both
size categories. Therefore, hypothesis twelve, which states
that for this sector, management background will not only be
more important than college education for but will be more
important for the Finance, Insurance, and Real Estate sector
than for other industries can definitely be accepted. PCMGT
coefficients for both size categories are positive and
significant, whereas those associated with COLLGRD are
neither. Furthermore, the only other sector in which PCMGT
is significant is Services, and only for the 1-19 size
category.

A variable measuring each sector’s share of total
employment (INDSHR) was included in the sector-specific
model primarily as a control, but the associated hypothesis
(13) was that growth in small firms in a sector would be
positively related to that sector’s share of employment.

For Manufacturing, Services, and larger Finance, Insurance,
and Real Estate establishments, INDSHR is not significant.

For Business Services (both size categories) this variable

90

is significant but negative. Thus hypothesis thirteen is
rejected.

To summarize these results, it can be said that human
capital variables are generally more important for the
growth of very small firms than for small firms. The
importance of college education and management background
are generally substantiated, while high school education in
itself does not appear to positively impact the growth of
small firms, not even in Manufacturing, a sector which
employs a greater proportion of people with only high school
level education than the other sectors. In general, for the
sector-specific analysis, ten of the first twelve hypotheses
can be either accepted universally or at least for some
sectors. Only hypotheses ten and eleven must be rejected
completely. The following section will present results of

the regional analysis.

Regional Patterns

The last three hypotheses relate to regional patterns.
Three techniques were employed to determine if regional
patterns in the impact of human capital exist: 1) the use of
regional dummy variables in the basic regression analysis,
and 2) the incorporation of a dummy variable to test for
differences in MSA size, and 3) examination of residuals
from regression. (Residuals were saved from equations which

included only the coefficients which were significant in the

91

original analysis). These residuals were standardized and
mapped (see Figures 10-14).

The equation incorporating regional dummy variables
(equation 5) was run separately for each independent
variable (since to do all four human capital variables would
have required resulted in 48 independent variables).
Results of these equations are shown in Table 7, which
indicates only significant coefficients for this entire set
of regressions (see appendix, Table 1, for complete results
of these regressions). In the aggregate model (all
sectors), for very small firms, PCMGT (the only significant
human capital variable in the original analysis) was
significant in none of the six regions. On the other hand,
for these same firms, COLLGRD, which was not significant in
the basic analysis, was positive and significant for all
regions.

When the impact on very small Manufacturing firms was
examined with regional dummy variables included, COLLGRD,
which was positive and significant overall (in original
analysis), is positive and significant only in the
Southeast. For these firms, however, EDEXP is positive and
significant for all regions.

For the Finance, Insurance, and Real Estate firms
sector, with the dummy variables included, a positive
relationship between PCMGT and growth in very small firms is
seen in the all but the Southwest and Mountain regions. In

both the Southwest and Mountain regions the impact of PCMGT

TABLE 7. Regression Results: Equation 5 (Regional Differences)
Note: only significant coefficients shown

Sector
All Sectors:
Northeast

Midwest

Mountain

Western
Manufacturing:

Northeast

Western
F . I . R . E.:

Northeast
Southeast
Midwest
Southwed
Mountain
Western

Services:
Northeast
Southeast
Mldmst
Southwest
Mountain
Western

Business Services:

Northmst
Southeast
Midwest
Southwest
Mountain
Western

'significent at .05
“significant at .01

COLLGRD

0.879 “
0.879 “
0.879 "
0.879 “
0.879 "
0.879 “

0.841 ‘

0.779 “
0.779 "
0.779 “
0.779 “
0.779 “
0.779 "

Very Small Firms

HSGRD EDEXP

0.527 ”
0.527 “
0.527 ”

0.5 “
0.527 “

0&2 ‘
0&2 '
-0.153 '
0&2 '
0&2 °
0&2 '

PCMGT

2.91 8
2.91 8
2.91 8
-0181
-2.347
2.91 8

1 .041
1 .041
1 .041
1 .041
1 .041
1 .041

 

COLLGRD

1.&1 '

Small Firms
HSGRD

4851 '

EDEXP

4.481
1 .738
0.579
-0.285
-1 .627
4.461

PCMGT

93

is negative and significant, particularly in the Mountain
region.

When regional dummies variables are added to the
equation testing the impact of COLLGRD and PCMGT on very
small Service firms, no regional differences appear (both
variables are significant in all regions). In contrast, for
small Service firms, the impact of COLLGRD (which, in the
original analysis was a stronger predictor of growth than
for very small firms), is positive and significant only in
the Western region.

Neither size category of Business Service firms was
positively related to any of the human capital variables in
the original analysis, but when dummy variables are added
EDEXP appear significant in some regions for both size
categories. For the very small firms, PCMGT is postive in
all regions but the Midwest region, where it is significant
but negative. For small Business Service firms, EDEXP is
positive and significant in the Northeast, Southeast,
Midwest, and Western regions; and it is negative and
significant in the Southwest and Mountain regions (the same
two regions where PCMGT is negative for Finance, Insurance,
and Real Estate). In addition to the positive impact of
EDEXP, one other difference occurs for the larger Business
Service firms; HSGRD, not significant otherwise, is
negative and significant in the Southeast region.

Based upon the above results, hypothesis fourteen,

which states that the impact of human capital will be

94

greater in the Southeast than in other regions, can be
accepted, for very small firms in the Manufacturing sector.
Likewise, hypothesis fifteen, that the impact of human
capital will vary little among regions other than the
Southeast, can also be accepted, particularly with respect

to the impact of college graduates.

City-Size Comparison

This section will examine the impact of human capital
variables on the growth of small firms for two different
size MSA’s. It was hypothesized, based upon both product
cycle theory and upon the general shift to a more advanced
service economy, that larger MSA’s would be likely to have a
greater proportion of industries requiring higher levels of
human capital than smaller MSA’s, and in addition these
larger MSA’s would be more likely to embody urbanization
economies which would increase the effectiveness of human
capital. Equation {6}, incorporating dummy variables (both
slope and intercept) to distinguish MSA's with a laborforce
greater than 150,000 from smaller MSA’s, was employed to
determine if city size was a factor in the impact of human
capital on small firm growth.

Almost no differences were found between large and
small MSA’s (see Table 8). The dummy variables were not
significant for equations which examined all sectors,
Manufacturing firms, or Service firms. In the Finance,

Insurance, and Real Estate sector, HSGRD, which is

95

TABLE 8. Regression Coefficients: Equation 6 (MSA Size Differences)

Sector
All Sectors:
1-19 Employees
- small MSA's
- large MSA's
20—99 Employees
- small MSA's
- large MSA's
Manufacturing:
1-19 Employees
- small MSA's
- large MSA's
20-99 Employees
- small MSA's
- large MSA's
F.l.R.E.:
1-19 Employees
- small MSA's
- large MSA's
20-99 Employees
- small MSA's
- large MSA's
Services:
1-19 Employees
- small MSA's
- large MSA's
20-99 Employees
- small MSA's
- large MSA's
Business Services:
1-19 Employees
- small MSA's
- large MSA's
20-99 Employees
- small MSA's
- large MSA's

-24. 358'

0.428 “
0.428 “

0.254 "
0.254 '

0.679 "
0.679 ’

COLLGRD HSGRD

-0.279 '
-0.997 *

-1.089 '

1.217 *
1.217 "

EDEXP

0.218 ‘
0.218 "

-0.378 '
~0378 ”

0.829 "
0.829 '

PCMGT

2.289 “
2.289 “

1.745 ”
1.745 ”

2.285 '

2.285 '

0.944 '
0.944 ‘

-6.881 ”
-6.881 “

R2

0.054

0.018

0.087

0.02

0.068

0.039

0.015

0.03

F-RATIO

2.947

0.477

1.633

0.524

4.29

1.709

3.487

2.389

1.515

2.04

 

‘slgnlficant at .05
“significant at .01

96

negatively related to growth in very small firms, appears
even more negative in the larger MSA’s. In the Business
Service sector, one change related to city size was noted;
for very small firms, in the larger MSA’s, HSGRD is negative
and significant. However, it is interesting to note that
although for this sector, none of the human capital
variables were significant without the dummies (equation
{3}), and while the larger size-category of Business Service
firms displays no differences by MSA size, when these dummy
variables are incorporated both HSGRD and EDEXP are positive
and significant (at .05); PCMGT is negative and significant
(at .00).

In general, it appears that the impact of human capital
is not related to MSA size; thus the final hypothesis must
be rejected. The only variable to vary in impact by MSA
size was HSGRD (not significant in earlier analyses) and in

both cases the impact was negative.

Residuals Analysis

As noted above, residuals were saved from equations
which included only the variables which were found to be
significant in the original analysis. Thus, for the
aggregate analysis (all sectors), the equation would be:

{8} CSF - a + blPCMGT + e

vs
where CSFvs = the change in very small firms
Residuals from this equation are shown in Figure 10.

The geographic pattern of these residuals is definitely

97

 

8.9 2 one I

 

one o. 5.0: I
5.0: 3 88.. D

m.o=o_mom pofiptoocoam

 

a 888m see .323 .2 use...

 

 

WEN..."— Jr_<_)_w >mw> ZO Luau—200 ".0 LID/$05:

98

different from the distribution of those in management
occupations seen in Figure 9. For example, in several MSA’s
which were among those with the highest percentage of
managers, residuals were either negative or less than +.5
(e.g. Orange County; Washington DC; Denver; Springfield,
Illinois; Boise, Idaho), indicating that in these MSA’s a
wealth of managers did not translate into increases in small
firms. On the other hand, there were also MSA’s with few
managers, but in which a strong relationship between
management and small firm growth was evident (Las Cruces,
NM; St. Cloud, MN). In general, the relationship between
management and the increase in small firms seems stronger in
the Northeast and the Southeast.

For very small Manufacturing firms, residuals were
saved from an equation regressing the change in very small
Manufacturing firms on both COLLGRD and EDEXP. This
equation is:

{9} CSF

V8 3 a + bICOLLGRD + szDEXP + e

where: CSFvs = percent change in very small
Manufacturing firms

The residuals from this equation (Figure 11) do not exhibit
the strong east coast concentration seen in Figure 10. The
impact of these two human capital variables on very small
Manufacturing firms is weaker in the Northeast but somewhat
stronger on the West coast (Figure 11). The impact of these
variables in the Southeast and in some Texas and Mountain

regions also appears somewhat stronger.

99

5 8:83 age assess .: 2:»...—

 

82 2 one I
omd 0. .mdi

5.0: 9 coal D

«.03—”zoom psnmpLOpcoam

 

 

 

 

WEN—.0 mv.Z_KDFO<u_DZ<_2 Jdgw >mw> r QXNDN 026. Own—01.400 00 F0<05=

100

The equation used to examine residuals for very small
Finance, Insurance, and Real Estate firms is:

{10} CSF = a + b1HSGRD + bZPCMGT + e

vs
where CSFvs = change in very small Finance,
Insurance, and Real Estate Firms
For these very small firms, the impact of PCMGT appears
greater in the Northeast than in the Northwest (Figure 12a),
and is also strong in the Southeast and somewhat stronger in
more Midwest MSA’s than for either of the previous patterns
of residuals. In most of the North Central part of the
country PCMGT appears to have a negative impact on growth in
this sector, despite the fact that most of the MSA/s in this
area fall into the ”middle" category with respect to their
percent in Management occupations.
Residuals were also saved for small Finance, Insurance,

and Real Estate firms, from the equation below.
{11} CSF8 a a + blEDEXP + b2PCMGT + e

where: CSF8 - percent change in small Finance, Insurance,

and Real Estate firms
Results are shown in Figure 12b. Residuals for these firms
are definitely weaker in all areas than those associated
with very small Finance, Insurance, and Real Estate firms.
Both PCMGT and COLLGRD impacted the growth in very
small Service establishments. Thus equation used was:
{12} CSFvs a a + b1COLLGRD + bzPCMGT + e
where: CSFvs - percent change in very small Service

firms

101

.) IMPACT OF HSGRD AND PCMGT ON VERY SMALL F.l.R.E FIRMS

 

   
 
 
    

 

Standardized Residuals
[3 -5.00 to ~05:

-0.sr to 0.50
- 0.50 to 13.00

 

b) IMPACT OF EDEXP AND PCMGT ON SMALL F.l.R.E FIRMS

 

   

 

fl is ..
m at: LU [b J». fl 1 an"
23' (a: 0 4‘
I; Dal? . £530"
IE3
all u .‘fih‘fi 6.} Gagfifigfi
a .. . .9“
' o 0 (1 Q0 \
£13184 3 RI! Q38
9 . 0*"- 7 V}: G ‘a Standardized Residual!
A l - . o - - i
‘ J 45).: :0 0::
~ “I; - 050101100

 

Figure 12. Residuals from Equation {10} and {11}

102

The impact of these variables in the Northeast does not
appear as strong as the impact of human capital variables on
growth in Finance, Insurance, and Real Estate firms (Figure
13a), but seems greater than for Manufacturing firms (Figure
11). In the West, impacts appear somewhat weaker than in
the East, and were also weaker than the impacts of human
capital variables on both Manufacturing and Finance,
Insurance, and Real Estate in this region. In the
Southcentral area very small Service firms showed a slightly
stronger response to these variables other sectors.

For small service firms, PCMTG was not significant, but
COLLGRD was, thus residuals were saved from the following
equation:

{13} CSF8 - a + b1COLLGRD + e

where: CSF8 - percent change in small Service firms
Compared to the very small Service firms, the impact of
human capital on small Service firms was much stronger on
the West coast, but somewhat weaker in the Southeast and the
Midwest (Figure 13b).

In summary, although statistically there were few
regional differences in the impact of human capital, these
maps do show some broad regional patterns and also indicate
some interesting patterns for individual MSA’s. For
example, Ann Arbor, Michigan, scores high in the impact of
human capital variables in the aggregate analysis, as well
as for Service and Finance, Insurance, and Real Estate

firms; yet for Manufacturing firms, the impact is negative.

103

I) IMPACT OF COLLGRD AND PCMGT ON VERY SMALL SERVICE FIRMS

 

 

 

Standardized Residuals
C] -s.00 to -0.51

[3 -0.51 to 0.50
‘7 - 0.5010 13.00

 

b) IMPACT OF COLLGRD ON SMALL SERVICE FIRMS

 

 

Standardized Residuals

[3 -5.00 to —o.sr
[:1 -0.51 to 0.50

- 0.50 to 13.00

 

Figure 13. Residuals from Equation {12} and {13}

104

Colorado Springs likewise has high residuals for both
Manufacturing the and Finance, Insurance, and Real Estate
sectors, but negative residuals for Service firms.

Residuals in other MSA’s, Jacksonville, North Carolina,
Charlotte, South Carolina, and many in Florida, are strong
in all sectors. The implications of these results, as
well as the other findings reported in this chapter, will be

discussed in the next chapter.

CHAPTER VII

CONCLUSION

The questions this dissertation set out to answer are
1) is human capital a factor in the growth of small firms at
the regional level, 2) assuming that it is, which human
capital variables are most important, 3) are there sectoral
differences in the impact of human capital, and 4) are there
regional differences in the impact of human capital. The
findings of this dissertation with respect to each of these

questions will be summarized and discussed.

Human Capital as a Regional Factor

The first of these questions can be answered in the
affirmative. One or more of these human capital variables
is significant in all sectors examined except Business
Services, and even in this sector, when regional dummy
variables are incorporated, two human capital variables are
seen to impact small firm growth in some regions.

Which types of human capital are most important varies,
depending both upon the sector and upon the size of firms,
although generally human capital appears to play a greater
role in the growth of very small establishments. .All human
capital variables included were significant for one of the
size categories of at least one sector. And the two
variables argued to measure human capital stock (based upon
McNamara, Kriesel, and Deaton), percent of college graduates

and percent of the local labor force in management

105

106

occupations, were (as hypothesized) more important than
local education expenditures and high school education, the
two which measured human capital flow. It will be recalled
that the rationale for this hypothesis was that flow
variables would be more important to relocating firms (which
presumably would be influenced more by indications of an
area’s potential for supplying labor), whereas measures of
human capital stock would play a greater role in firm
farmation, which was expected to account for more of the
variation in small firm growth. And, in fact, one of these
flow variables, HSGRD, is not positively related to growth
in any of the sectors examined; it is even negatively
related to the growth of very small Finance, Insurance, and
Real Estate firms.

The other variable assumed to be a measure of human
capital flow was EDEXP; this variable was positive and
significant only for very small Manufacturing firms (and, in
some regions, for Business Service firms). The explanation
for this might be related to this variable’s greater impact
in attracting firms, if it can be argued that in the
Manufacturing sector, more branch plant location occurs than
in non-manufacturing industries. However, as pointed out
earlier, this variable also reflects state-level variation
in educational funding, which makes its interpretation
difficult. Overall, the stronger showing of the two
measures of human capital stock reinforce the notion that

for small firm growth, the creation and management

107

components outweigh laborforce considerations. With respect
to formal education, the percent of the population with a
college degree is definitely the most important factor
affecting small firm growth.

Because the literature suggested that occupational
background is less important to new firm formation than is
formal education, the second measure of human capital stock,
the percent of the laborforce in management occupations
(PCMGT) was not expected to be as important as formal
education. However, this variables had the greatest impact
on aggregate growth of very small firms, the greatest impact
on growth of both size firms in the Finance, Insurance, and
Real Estate sector, and the strongest impact on very small
Service firms. When regional dummy variables were added,
PCMGT also appears positive and significant for Business
Services in most regions. The rationale underlying the
general hypothesis relating management background to small
firm growth had less to do with the firm farmation component
and more to do with firm survival. The fact that empirical
evidence suggested formal education would be a greater
predictor of growth in the number of small firms than would
occupational background may be related to two factors: the
size of the firms studied and the time period covered by
such research. With respect to size, since small firms have
a much greater failure rate than do large firms, PCMGT,
because it is considered a more important factor in firm

survival, might be a greater factor in the growth of small

108

firms than large firms. The time period is relevant since
much of the literature referred to above predated the
1980's; the stronger than expected showing of occupational
background in this study may reflect the increasingly
competitive environment during this period, an environment

in which firm survival rates are generally lower.

Sectoral Differences in the Impact of Human Capital

The third general question considered by this research
relates to sectoral differences in the impact of human
capital. It was hypothesized that for Manufacturing firms,
the impact of COLLGRD would be less than for Service firms,
and this was true overall, since this variable was not
significant for both size Manufacturing firms. However, for
very small firms, the impact of COLLGRD was greater in the
Manufacturing sector. The variable measuring educational
quality, EDEXP, was also significant for Manufacturing
firms. For both of these variables, the impact in this
sector was greater for very small firms. Although HSGRD was
not expected to be significant overall, it was expected to
be more important for the Manufacturing sector than for
other sectors. However, it was not significant in any of
the analyses.

For the Finance, Insurance, and Real Estate sector, the
importance of occupational background to firms in both size
categories was substantiated. This was expected, based upon

the high percentage of the administrators and managers in

109

this sector, and its lower requirement for professional and
technical labor, which more often requires college
education. All analyses (basic, regional, MSA size)
indicated that impacts of PCMGT in this sector were greater
for very small size firms.

As discussed above, based upon employment figures which
indicate that service industries have a greater demand for
skills which are dependent upon college education than for
management skills and which assume that the laborforce needs
of small firms will be the same as those of larger firms, it
was hypothesized that for the Service sector, COLLGRD have a
greater impact than for Manufacturing. This hypothesis was
accepted, for both size Service firms, but for very small
firms the impact of PCMGT was even greater. In contrast,
the impact of COLLGRD was was greater for small firms than
for very small firms. For Services, there were no
significant differences in human capital impacts by region
or MSA size.

For Business Service firms, the lack of relationship
with any of the human capital variables was very surprising,
particularly since the survey of Beyers, Johnsen, and
Stranahan (1987) found that producer service firms had 43%
of their workforce in occupations that require the highest
levels of education. When regional dummy variables were
added, EDEXP did appear to positively impact the growth of
very small firms everywhere except in the Midwest and it

also positively impacted small firm growth in all but the

110

Southwest and Mountain regions. The MSA size analysis also
indicated positive relationships between small Business
Services firms and both HSGRD and EDEXP. In general, it
appears that human capital’s impact in the Business Service

sector is greater for small than for very small firms.

Regional Differences

Finally, this dissertation addressed the question of
regional differences in the impact of human capital. With
respect to its distribution, human capital appears to be
greatest in MSA’s in the Mountain and Western regions,
followed by the Northeast region; the Southeast region
appears most deficient in human capital. In contrast, it is
in this most human capital poor region that the strongest
growth in very small establishments in all sectors is
occurring. While this might suggest a lack of relationship
between human capital and small firm growth, this does not
appear to be the case. In fact, the Southeast provides the
only instance of a human capital variable having a stronger
impact than in other regions (the impact of COLLGRD on the
growth of very small Manufacturing firms).

Regression results revealed few regional differences in
the relationship between human capital and small firm
growth. The lack of strong regional differences in the
impact of human capital, despite the obvious differences in
rates of small firm growth, suggests that even when macro-

level{factors influence growth in a general area, the human

111

capital factor may influence more specifically where this
growth occurs. The high levels of growth experienced by the
human capital poor Southeast region during the 1980’s is a
good example. The economic boom occurring in this region in
the 1980’s has been hypothesized to be related to the
opening of branch plants and relocation of firms to this
area to take advantage of lower taxes and its low-paid, non-
union laborforce, in a sense, its low (at least inexpensive)
human capital. And certainly much of this growth can be
attributed to such relocations. However, within this
decision to relocate to the south for cost-related reasons,
the more specific location within the south was probably
related to human capital in these areas.

Further, it is likely that much of the development of
these smaller firms was triggered by the relocation of the
larger firms in the area, since it is unlikely that many
firms in the 1-19 employee category would have relocated to
the area. While the larger firms’ location may have been
more related to cost factors than to human capital (except
perhaps negatively), the related growth of smaller firms in
the region, those that are most likely to have "developed in
place" vs. relocating, is more dependent upon having a local
comparative advantage in human capital.

The only other strong regional differences noted were
in the Southwest and Mountain regions. For both of these
regions, PCMGT was negatively associated with very small

firm growth in Finance, Insurance, and Real Estate and EDEXP

112

was negatively related to small firm growth in Business
Services. This is particularly mystifying since the
Mountain region, where the impact of PCMGT is most negative,
has the highest percent of its labor force in Management
occupations. This suggests that perhaps the relationship
between these two variables is not linear beyond a certain
point, i.e. there is a point where additional levels of
management began to have a negative effect, particularly if
this disproportionate number of managers also reflects a
dearth of other occupations. For very small Business
Service firms in the Midwest, PCMGT is also negatively
related to growth, but this region also has the lowest
percent of its labor force in management occupations.

At the subregional scale, although some variation in
the impact of human capital was evident, based upon the
residuals analysis, differences based upon MSA size were not
evident. The only sector in which the impact of human
capital varied by MSA size was the Business Service Sector,
in which the impact of HSGRD was more negative for large
MSA’s. The failure of urbanization economies associated
with the larger MSA’s to enhance the impact of human capital
can probably be explained in the light of recent trends in
urban growth, which see the smaller metropolitan areas and
exurban areas growing faster than many of the larger MSA’s.

The fact that the impact of human capital is
significant, even in areas where levels of human capital are

not strong, reinforces the idea that it is a universal

113

factor in growth. It will be recalled that in the section
where hypotheses were developed, regional differences were
expected to be minimal; as stated there, if human capital is
truly a universal factor which transcends variations in
other factors related to growth, its impact should not vary
significantly by region. It is more likely that other
growth-related factors will vary, but will be augmented by

human capital.

Conclusion

Human capital theory and research have always pointed
to human capital, particularly formal education, as a
significant factor in development, but generally at the
national level. Consideration of the role of human capital
in regional economic growth has been less common. Today,
however, advanced economies are undergoing transformation
both in industrial structure and in the type of technologies
used in all industries; in addition, the increased
globalization of economies means that greater productivity
is required to be competitive. At the same time that these
changes are occurring, in most developed countries,
increased productivity is also necessary due to falling
birth rates. As both economic and demographic factors
demand increased productivity, the education and skills of
the laborforce become increasingly important (Spindler and
Forrester 1993). Despite this need for a more educated and

skilled labor force, the federal government role in economic

114

development has diminished; thus, regional development
initiatives have focused more on local sources of
comparative advantage (Rich 1992) and placed more emphasis
on entrepreneurial factors (Clark and Gaile 1992), factors
which are associated with firm formation and survival.

The same economic environment that appears to enhance
the value of human capital also gives an advantage to small
firms. According to Bannock (1981),

The principal economic importance of small firms lies

in their responsiveness to change and since change is

what is required if economic growth is to be resumed,
it is desirable that more rather than fewer resources

should be channelled into small business (p.8).
Learning more about the role of human capital in the growth
of such firms will help to identify exactly which resources
are to be channelled into small business and where they
should be directed.

The question of which types of human capital are most
important to small firm growth is important at the local
level. This research appears to indicate, at least in the
relatively short term, that human capital stock, i.e. the
existing level of human capital, is more important to a
region than is its potential human capital, as measured by
qualitative (flow) variables. This would suggest that
regional strategies should focus more on attracting and
retaining educated individuals than on improving local
educational quality. However, since this study used as its

qualitative measure the percent of the local budget spent on

education, an amount which would vary from state to state

115

because of differences in state funding for education,
further research would be necessary before such a conclusion
could be reached. Even if such a conclusion were found to
be justified for the local area, if individual regional
strategies focused only on attracting human capital, rather
than creating it, for the nation as a whole, the result
would quite possibly be only a redistribution of human
capital, not an increase. This speaks to the question of
who should pay for upgrading the education and skills of the
laborforce. Local areas, like individual firms, often have
less incentive to invest in human capital which may not
remain in the area, particularly if it can be attracted from
elsewhere. Thus a case might be made for federal and/or
state government subsidies for education and training.

The increasing importance of an educated, skilled labor
force is occurring at a time when local areas are
experiencing cutbacks in federal funds for education and
training, fewer individuals can afford the cost of higher
education, and fewer firms can afford (or are willing, when
skills involved are transferrable) to invest in training
(Gaspersz and vanVoorden 1987). A case for more federal
support in developing human capital resources would seem to
be suggested by Warner's (1989) argument (relating to the

importance of human capital in promoting local economic

116

growth):
\

The human capital strategy, which focuses on raising
labor productivity, has the potential to increase the
growth path of the national economy and therefore raise
the overall standard of living. However, the cost
'minimization approach is far more likely to result in
relocation of resources within the economy and not
stimulate national economic growth. (p. 396)

Future Research

Although this research has indicated a relationship
between some types of human capital and small firm growth,
results say nothing about how this effect occurs. Does the
number of small firms increase as a result of increased new
firm formation or because of a decrease in firm failures?
This question could only be answered by utilizing a dataset
which breaks down establishment change by firm starts and
firm failures.

A variation on the question of how human capital
impacts small firm growth is to consider whether it is
primarily the human capital embodied in the firm’s founder
that is most important, or whether it is also necessary (for
the survival of firms) to have a workforce which is
characterized by high levels of human capital? As seen in
the above results, growth in small firms in a sector whose
laborforce had the highest percent of college educated
employees (Business Services, according to Beyers et.a1.,
1987), was not significantly related to human capital. A
possible way to resolve this question would be to survey new

small firms, to determine both their laborforce needs, with

117

respect to educational levels, and the educational and
occupational background of their founders.

The time period covered by this study was dictated by
data availability and the desire to use the most recent data
available. However, it was a time period characterized by
considerable restructuring and the continued shift to a
global economy, a time period during which some regions
experienced rapid growth while others stagnated. As a
Washington Post series, focusing on the economic outlook for
the 1990's, stated (1990, p. 8), ”In a way, there is no
American economy, but a collection of regional economies
that rise and fall to distinctive rhythms." Perhaps a
longer time period, which would be more likely to capture
long term trends, as opposed to shorter fluctuations in
growth, would be more appropriate. It might also indicate,
contrary to what the results of this study suggest, that
over the long term, investment in human capital potential,
i.e. educational quality, is more important to a local area
than increasing the level of human capital stock. A study
which focused on a smaller, more cohesive, region over a
longer time period, might also produce more definitive
results.

Possibly the most interesting finding of this study was
the existence of significant sectoral variation in the
impact of human capital. Further investigation of such
differences is certainly called for. This might take the

approach, as discussed earlier, of considering different

118

functional forms. Also, incorporating a greater number of
less general sectors, would probably be illuminating,
particularly if such a study were confined to a more
homogeneous region.

Although this study has made only a small contributin
to the question of how human capital influences economic
development at the regional scale, it has shown that human
capital is a factor at this scale. At a time when, as Reich
(1991) points out, the concept of a national economy is
losing its validity, the fortunes of regional economies are
more closely tied to the education and skills of their labor
forces than to the economic success of the state in which
they happen to be located.

Schumacher (1973), discussing education, which he
considered to be "the greatest resource,” pointed out that
throughout history, not only have civilizations flourished
in all parts of the world, but even when they have declined
and perished, new civilizations have developed in the very
same spaces, raising the question of how the necessary
resources for development have been reconstituted. His
explanation is that these resources were not simply
material. He says,

All history - as well as all current experience -

points to the fact that it is man, not nature, who

provides the primary resource: that the key factor of
all economic development comes out of the mind of man

(p. 72).
At a time when economies are in the process of

significant transformations, are increasingly threatened by

119

global competition, and when the competitive advantage
associated with material, place-specific resources is
diminishing, it is the resources embodied in the mind of man
that will determine which regions flourish and which

decline.

APPENDIX A

A311:

 

$00me San—mg

120

< X_Dzm&n_<

98. a
Re:
26

an:

26
63. E
.823-

25

3.3
mumd
>>2

6.3
5:
>22
Ea. E
«and.
>22

ASP...
98. F-

98;

9.8.3
08.0.
mm

83. :
8”:
mm
28.3
«and.
mm

6:;

8m.»
mm

€83

85. _. 7

mm

5. s 28:29...
8. s 2858...
98$ $93
:23 .83.?
520.. ..

gators n 28.5; .5288

Round no». 3
End .3. «.0 w-
mxmam a
683 6a.;
«8.0. 98.0
010.300 w

w>OuEOIO u o_ooto> Eoucoooo

End 683
L8... as.»
amodoo a
93.: 68s
8N... 8a....

620.. a

map—.010 u o_noto> E02800

0:5“. :36 ba> - n ease-Em "Essex c2329: .p mans.

121

28.9.:

 

3. s 282299.
8. a 2855.».
SAN: sens-e an: $8.3 :Be 845 65.2 fiance 29.: 68.3 2483 68.3
Bud tho. 28.? Sec- bzd- and. 83. .28 «8.». v8.8 818 .82. 82.:-
Nm 908m; 903:2 axmomgm “58222 508% 52, :2 26 >22 mm 503 a
m>meIo n 030.2; E03800
$83 on: an: 2 61 2:3 68... .85 so: $8.3 68.2 62.: 68.:
83 5.3 83- 83- 86 83 82.2- 9.32 2.2.. 83.- 83. .23.. N89
mm 5.2.0.2, 52:2 5226 5.222 52mm 52, 5 26 >22 mm 5.20.. a
68.2 oboe 28.3 68.3 sci fine 2.1.8.. 233 2.8: 8%.: $5.9 88.:
«No Reo- vomd. unme- Snd- 82.0. B: E.” 88.. Ibo 83 :23 .33
mm .3083 ....an $0030. .3822 3.680 52, E2 25 >22 mm aces-.00 .

w>mmw010 u o_no_._m> Eoccoooo

8.328. p 39.:

122

28.9...

 

6. .0 58:9...
8. a 2.8220...
Fun-T L808. :83. 2.03.. .881 5.0.. Lemma. .803. LBS. been. 2.8.... Lamas.
5... Rod. 2.8.0. 0:? 80.0- .4.2..- 0008 08.08 80.8. 08.0: 80.8. 8: 3.0.00.-
9. 0.0802. 90002.2 90002.0 0.08;: 0.00000 82. :2 .26 >22 mm 9000 a
60.... 2.3.... $8... $8... .CSN. $8.. $0.. 98.. 50.. .88... 2.02.... 208...
.8... 83 Q...- 300- 80.n- 288. R08. 83.. 2.0.8 :18 50.2.. EN 000.00-
02 0200202, 9.00:5 80050 000012,: 0200200 2.02. E2 30 2,2 00 0000... .
wwmwro u 030...; E03800
.280... 65.-. am... 68... Ea. $3.... 88... .55.. :83. 88.. 85.. LBS.
8a... .00.. 05.0- 33 2.0.0 33 08.8- 80.0- 02.3- 88. 08.0 .8... 2.0.8
mm .3802, ....an 3.6030 3.00222 .2800 E03 :2 >20 >22 mm 9.0.300 0

8.0005 n 0.0025 2.020000

05...... =0Em - 0 20:05am "8.2001 cease-50¢ d 39:.

APPENDIX B

APPENDIX B

TABLE B-1. MSA/County Configuration

M§A
Abilene, TX

Akron. OH

Albany, GA

Albany. NY

Albuquerque. NM
Alexandria, LA
Allentown, PA

Altoona, PA
Amarillo, TX

Anaheim. CA
Anderson. IN
Anderson, SC
Ann Arbor. MI
Anniston. AL
Appleton, Wi

Ashville, NC
Athens, GA

Atlanta, GA

QM!!! M
Taylor Atlanta (cont'd)

Portage

Summit

Daugherty

Lee

Albany

Greene

Montgomery

Rensaleer

Saratoga

Schnectady

Bemalllo

Rapides

Carbon

Lehigh

Northhampton

Warren

Blair Atlantic City. NJ
Potter
Randall
Orange
Madison
Anderson
Washtenaw
Calhoun
Calumet
Outagamie
Winnebago
Buncombe Bakersfield, CA
Clarke Baltimore. MD
Jackson

Madison

Ooonee

Barrow

Augusta. GA

Aurora/Elgin. IL

Austin, TX

123

Count!
Butts

Cherokee
Clayton
Cobb
Coweta
DeKalb
Douglas
Fayette
Forsyth
Fuhon
Gwlnnett
Henry
Newton
Paulding
Rockdale
Spaiding
Walton
Atlantic
Cape May
Columbia
McDuffie
Richmond
Aiken
Kane
Kendall
Hays
Travis
Williamson
Kern
Anne Amndel
Baltimore County
Canoll
Hartford
Howard

TABLE B-1 (cont'd)

M§A
Baltimore (cont'd)

Bangor. ME
Boston,MA

Baton Rouge, LA

Battle Creek. MI
Beaumont,TX

Beaver County, PA
Bellingham, WA
Benton Harbor, MI

Bergen-Passaic, NJ

Billings, MT
BlloxllGulfport. MS

Binghamton, NY

Birmingham, AL

Bismark. ND

Bloomington, IN

Blooming-Normal. IL

Boise City, ID
Boulder, CO
Bradenton, FL

9.931!!!

Queen Anne
Baltimore City
Penobscot
Essex
Middlesex
Norfolk
Plymouth
Suffolk
Ascension
East Baton Rouge
Livingston
West Baton Rouge
Calhoun
Hardin
Jefferson
Orange
Beaver
Whatcom
Berrien
Bergen
Passalc
Yellowstone
Hancock
Harrison
Broome
Troga
Blount
Jefferson
St. Clair
Shelby
Walker
Burlelgh
Morton
Monroe
McLean
Ada

Boulder
Manatee

124

M§A 99mm
Brazoria. TX Brazoria
Bremerton, WA Kitsap
Bridgeport. CT Fairfield
Brownsville, TX Cameron
Bryan/College Sta. TX Brazos
Buffalo, NY Erie
Burlington, NC Almance
Burlington. VT Chittenden
Grand isle
Canton, OH Carroll
Stark
Caspar, WY Natrona
Cedar Rapids, IA Linn
Champalgn/Urbana, IL Champalgn
Charleston. SC Berkely
Charleston
Dorchester
Charteston. WV Kanawha
Putnam
Chartotte. NC Cabamus
Gaston
Lincoln
Mecklenburg
Rowan
Union
York, SC
Charlottesville, VA Ablemarte
Fluvanna
Greene
Charlottesville
Chattanooga. TN Hamilton
Marion
Sequatchie
Catoosa. GA
Dade, GA
Walker. GA
Cheyenne, WY Laramie
Chicago, IL Cook

Table B-1 (cont'd)

M§A
Chicago, IL (cont'd)

Chico, CA
Cincinnati, OH

Clarksville. TN

Cleveland. OH

Colorado Springs, CO
Columbia, MO
Columbia. SC

Columbus, GA

Columbus, OH

Corpus Christi. TX
Cumberland, MD

Dallas, TX

125

Count!
DuPage

McHenry
Butte
Clerrnont
Hamilton
Warren
Boone. KY
Campbell. KY
Kenton, KY
Dearbom, IN
Montgomery
Christian, KY
Cuyahoga
Geauga
Lake

Medina

El Paso
Boone
Lexington
Richland
Chattahooche
Muscogee
Russell. AL
Delaware

Fairfield
Franklin
Licking
Madison
Pickaway
Union
Nueces

San Patricio
Allegany
Mineral, WV
Collin
Dallas
Denton

Ellis

M§A
Dallas, TX (cont'd)

Danville, VA

Davenport, IA

929.01!

Kaufman
Rockwell
Pittsylvania
Danville City
Scott

Henry. IL
Rock Island, IL

Dayton/Springfield OH Clark

Daytona Beach. FL
Decatur. AL

Decatur, IL
Denver, CO

Des Moines, IA

Detroit, MI

Dothan, AL

Dubuque, IA
Duluth. MN

Eau Claire, WI

El Paso, TX

Greene
Miami
Montgomery
Vqusia
Lawrence
Morgan
Macon
Adams
Araphoe
Denver
Douglas
Jefferson
Dallas
Polk
Wanen
Lapeer
Livingston
Macomb
Monroe
Oakland
St. Clair
Wayne
Dale
Houston
Dubuque
St. Louis
Douglas. WI
Chippewa
Eau Claire

El Paso

TABLE B-1 (cont'd)

M§A
Elkhart, IN

Elmira, NY
Enid, OK
Erie, PA
Eugene, OR
Evansville, IN

Fargo, ND

Fayetteville, NC
Fayetteviile, AR

Flint, MI
Florence, AL

Florence, SC

Fort Collins, CO
Fort Lauderdale, FL
Fort Meyers, FL

Fort Pierce, FL

Fort Smith, AR

Fort Walton Beach FL

Fort Wayne, IN

Fort Worth, TX

Fresno, CA
Gadsen, AL
Gainsville, FL

Galveston, TX

Gary/Hammond, IN

£99111!
Elkhart

Chemung
Garfield

Erie

Lane

Posey
Vanderburgh
Warrlck
Henderson, KY
Cass

Clay, MN
Cumberland
Washington
Genesee
Colbert
LauderdaIe
Florence
Larimer
Broward

Lee

Martin

St. Lucie
Crawford
Sebastian
Sequoyah, OK
Okaloosa
Allen
DeKaIb
Whitley
Johnson
Parker
Tarrant
Fresno
Etowah
Alachua
Bradford
Galveston

Lake

126

M
Gary, IN (cont'd)
Glens Falls, NY

Grand Forks, ND
Grand Rapids, MI

Great Falls, MT
Greeley, CO

Green Bay, WI
Greensboro, NC

Greenville, SC

Hagerstown, MD
Hamilton, OH
Harrisburg, PA

Hartford, CT

Hickory, NC

Houma, LA

Houston, TX

991101!
Porter

Warren
Washington
Grand Forks
Kent
Ottawa
Cascade
Weld

Brown
Davidson
Davie
Forsyth
Guilford
Randolph
Stokes
Yadlom
Greenville
Pickins
Spartanburg
Washington
Butler

Cumberland
Dauphin
Lebanon
Perry
Hartford
Middlesex
Tolland
Alexander
Burke
Catawba
Lafourche
Tenebonne
Fort Bend
Harris
Liberty
Montgomery

Waller

TABLE B-1 (cont'd)

MSA
Huntington, WV

Huntsville, AL
Indianapolis, IN

Iowa City, IA
Jackson, MI
Mackson, MS

Jackson, TN
Jacksonville, FL

Jacksonville, NC
Jamestown, NY
Janesville, WI
Jersey City, NJ
Johnson City, TN

Johnstown, PA

9mm
Cabell
Wayne
Boyd, KY
Carter
Greenup
Lawrence, OH
Madison
Boone
Hamilton
Hancock
Hendricks
Johnson
Marion
Morgan
Shelby
Johnson
Jackson
Hinds
Madison
Rankin

Madison
Clay
Duval
Nassau

St. Johns
Onslow
Chatauqua
Rock
Hudson
Carter
Hawkins
Sullivan
Unicoi
Washington
Scott, VA

Washington, VA

Bristol, VA
Cambn‘a

127

MSA

Johnstown (cont'd)
Joilet, IL

Joplin, MO

Kalamazoo, MI
Kankakee, IL
Kansas City, MO

Kenosha, WI
Killeen, TX

Knoxville, TN

Kokomo, IN

LaCrosse, WI
Lafayette, LA

Lafayette, IN
Lake Charles, LA
Lake County, IL
Lakeland, FL
Lancaster, PA

Lansing, MI

M
Somerset
Grundy
Will
Jasper
Newton
Kalamazoo
Kankakee
Cass

Clay
Jackson
Lafayette
Platte

Ray
Johnson, KS
Leavenworth, KS
Miami
Wyandotte
Kenosha
Bell
Coryell
Anderson
Blount
Grainger
Jefferson
Knox
Sevier
Union
Howard
Tipton
LaCrosse
Lafayette
St. Martin
Tippecanoe
Calcasieu
Lake

Polk
Lancaster

Clinton

TABLE B-1 (cont'd)

M§A
Lansing (cont'd)

Laredo, TX

Las Cruces, NM
Las Vegas, NV
Lawrence, KS
Lawton, OK
Lewiston,ME
Lexington, KY

Lima, OH

Lincoln, NE
Little Rock, AR

Longview, TX
Lorain/Elyria, OH

Los Angeles, CA
Louisville, KY

Lubbock, TX
LynChburg, VA

Macon, GA

92ml!
Eaton
Ingham
Webb
Dona Ana
Clark
Douglas
Comanche
Androsooggin
Bourbon
Clark
Fayette
Jessamine
Scott
Woodford
Allen
Auglaize
Lancaster
Faulkner
Lonoke
Pulaski
Saline
Gregg
Harrison
Lorain

Los Angeles
Bulitt
Jefferson
Oldham
Shelby
Clark, IN
Floyd, IN
Harrison, IN
Lubbock
Amherst
Campbell
Lynchburg
Bibb
Houston

128

MSA

Macon (cont'd)

Madison, WI
Manchester, NH
Mansfield, OH
McAIIen, TX
Medford, OR
Melbourne, FL
Memphis, TN

Merced, CA
Miami, FL
Middlesex, NJ

Midland, TX
Milwaukee, WI

Minneapolis, MN

Mobile, AL

Modesto, CA
Monmouth, NJ

M
Jones
Peach
Dane
Hillsborough
RIchIand
Hidalgo
Jackson
Brevard
Shelby
Tipton
Crittendon, AR
DeSoto, MS
Merced
Dade
Hunterdon
Middlesex
Somerset
Midland
Milwukee
Ozaukee
Washington

Waukesha
Anoka
Carver
Chisago
Dakota
Hennepin
Isanti
Ramsey
Scott
Washington
Wright

St. Croix, WI
Baldwin
Mobile
Stanislaus
Monmouth

0068"

TABLE 84 (cont'd)

LEA
Monroe, LA

Montgomery, AL

Muncie, IN
Muskegon, MI
Naplels, FL
Nashville, TN

Nassau, NY
New Bedford, MA
New Haven, CT

New London, CT
New Orleans, LA

New York, NY

Newark, NJ

M
Ouachita
Autauga
Elmore
Montgomery
Delaware
Muskegon
Collier
Cheatham
Davidson
Dickson
Robertson
Rutherford
Sumner
Williamson
Wilson
Nassau
Suffolk
Bristol

New Haven
New London
Jefferson

Orleans

St. Bernard
St. Charles
St. John

St. Tammany
Bronx

Kings

New York
Putnam
Queens
Richmond
Rockland
Westchester
Essex
Monis
Sussex

Union

129

MSA
Niagara Falls, NY

Norfolk, VA

Oakland, CA

Ocala, FL
Odessa, TX

Oklahoma City, OK

Olympia, WA
Omaha, NE

Orange County, NY

Orlando, FL

Owensboro, KY

Oxnard/Ventura, CA

Panama City, FL
Parkersburg, WV

Pascagoula, MS

$2211.91!
Niagara
Gloucester
James City
York
Chesapeake
Hampton
Newport News
Norfolk
Poquoson
Portsmouth
Suffolk
Virginia Beach
Williamsburg
Alameda
Contra Costa
Marion

Ector
Canadian
Cleveland
Logan
McLain

Oklahoma
Pottawattami
Thurston
Douglas

$8er
Washington
Pottawattami, IA
Orange

Orange

Osceola
Seminole
Daness

Ventura

Bay

Wood
Washington, OH

Jackson

TABLE B-1 (cont'd)

M§A

Pensacola, FL

Peoria, IL

Philadelphia, PA

Phoenix, AZ
Pine Bluff, AR
Pittsburgh, PA

Pittsfleld, MA

Portland, ME
Portland, OR

Portsmouth, NH

Poughkeepsie, NY
Providence, RI

Provo-Orem, UT
Pueblo, CO
Racine, WI
Raleigh, NC

29mm
Escambia
Santa Rosa
Peoria
Tazewell
Woodford
Bucks
Chester
Delaware
Montgomery
Philadelphia
Burlington, NJ
Camden, NJ
Gloucester, NJ
Maricopa
Jefferson
Allegheny
Fayette
Washington
Westmorland
Berkshire

Cumberland
Clackamas
Multnomah
Washington
Yamhill
Rockingham
Strafford
Duchess
Bristol

Kent
Providence
Washington
Utah
Pueblo
Racine .
Durham
Franklin

Orange

130

M§A

Raleigh (cont'd)
Rapid City, SD
Reading, PA
Redding, CA
Reno, NV
Richland, WA

Richmond, VA

Riverside, CA

ROBDOKC, VA

Rochester, MN
Orchester, NY

Rockford, iL

Sacramento, CA

QM!!!

Wake
Pennington
Berks

Shasta
Washoe
Benton
Franklin
Charles City
Chesterfield
Dinwiddie
Goochland
Hanover
Henrico

New Kent
Powhatan
Prince George
Colonial Heights
Hopewell
Petersburg Ct.
Richmond City
Riverside

San Bemadino
Botetourt
Roanoke City
Roanoke County
Salem City
Olmsted
Livingston
Monroe
Ontario
Orleans
Wayne

Boone
Winnebago

El Dorado
Placer
Sacramento

Yolo

TABLE B-1 (cont'd)

M§A
Saginaw, MI

St. Cloud, MN

St. Joseph, MO
St. Louis, MO

Salem, OR

Salinas, CA
Salt Lake City, UT

San Angelo, TX
San Antonio, TX

San Diego, CA
San Francisco, Ca

San Jose, Ca
Santa Barbara, CA
Santa Cruz, CA
Santa Fe, NM

Santa Rosa, CA
Sarasota, FL

99ml!

Bay

Midland
Saginaw
Benton
Sherbume
Steams
Buchanan
Franklin
Jefferson

St. Charles
St. Louis County
St. Louis City
Clinton, IL
Jersey, IL
Madison, IL
Monroe, II
St. Clair, IL
Marion

Polk

Montery
Davis

Salt Lake
Weber

Tom Green
Bexar

Comal
Guadalupe
San Diego
Marin

San Francisco
San Mateo
Santa Clara
Santa Barbara
Santa Cruz
Los Alamos

Santa Fe
Sonoma

Sarasota

131

MSA
Savannah, GA

Scranton, PA

Seattle, WA

Sharon, PA
Sheboygan, WI
Sherman, TX
Shreveport, LA

Sioux City, lA

Sioux Falls, SD
South Bend,IN
Spokane, WA
Springfield, IL

Springfield, MA

Springfield, MO

State College, PA

Steubenville, OH

Stockton, CA
Syracuse, NY

Tacoma, WA
Tallahassee, FL

Tampa, FL

9.91mi!
Chatham

Effingham
Columbia
Lackawanna
Luzeme
Monroe
Wyoming
King
Snohomish
Mercer
Sheboygan
Grayson
Bossier
Caddo
Woodbury
Dakota, NE
Minnehaha
St. Joseph
Spokane
Menard
Sangamon

Hampden
Hampshire
Christian
Greene
Centre
Jefferson
Brooke, WV
Hancock, WV
San Joaquin
Madison
Onondaga
Oswego
Pierce
Gadsden
Leon
Hernando

Hillsborough

TABLE B-1 (cont'd)

M
Tampa (cont'd)

Terre Haute, IN

Texarkana, TX

Toledo, OH

Topeka, KS
Trenton, NJ
Tuscon, AZ
Tulsa, OK

Tuscaloosa, AL
Tyler, TX
Utica, NY

Vallejo, CA

Vancouver, WA
Victoria, TX
Vineland, NJ
\frsalia, CA
Waco, TX
Washington, DC

132

Eau—nix M§A
Pasco DC (cont'd)
Pnellas
Clay
\frgo
Bowie
Miller, AR

' Fulton Waterloo, IA
Lucas
Wood Wausau, WI
Shawnee West Palm Beach,FL
Mercer Wheeling, WV
Pima
Creek.
Osage Wichita, KS
Rogers
Tulsa
Wagoner Wriliamsport, PA
Tuscaloosa Wilmington, DE
Smith
Herkimer
Oneida Wilmington, NC
Napa Worcester, MA
Solano Yakima, WA
Clark York, NY
Vrcton'a
Cumberland Youngstown, OH
Tulare
McLennan Yuba City, CA
Washington
Calvert, MD
Charles, MD
Frederick, MD

Montgomery, MD
Prince George, MD
Arlington, VA
Fairfax, VA
Loudoun, VA

Prince William, VA

Coung
Stafford, VA

Alexandria, VA
Fairfax City, VA
Falls Church, VA
Manassas City, VA
Manassas Park, VA
Black Hawk
Bremer

Marathon

Palm Beach
Marshall

Ohil

Belmont, OH
Butler

Harvey

Sedgwick
Lycoming

New Castle

Salem, NJ

Cecil, MD

New Hanover

Worcester
Yakima
Adams
York
Mahoning
Trumbull
Sutter
Yuba

BIBLIOGRAPHY

BIBLIOGRAPHY

Abraham, Katharine G. 1990. Restructuring the Employment
Relationship: The Growth of Market-Mediated Work
Arrangements. New Developments in the Labor Market
(edited by K.G. Abraham and R.B. McKersie). Cambridge:
The MIT Press.

Acs, Zoltan J. and David B. Audretsch. 1989. Births and Firm
Size. Southern Economic JOurnal 56 2:467-475.

Ashcroft, Brian, James H. Love, and Eleanor Malloy. 1991.
New Firm Formation in the British Counties with Special
Reference to Scotland. Regional Studies 25 5:395-409.

Bannock, Graham. 1981. The Economics of Small Firms: Return
from the Wilderness. Oxford:Basi1 Blackwell.

Barkley, David L. and Sylvain Hinschberger. 1992. Industrial
Restructuring: Implications for the Decentralization of
Manufacturing to Nonmetropolitan Areas. Economic
Development Quarterly 6 1:64-79.

Bartik, Tomothy J. 1989. Small Business Start-Ups in the
United States: Estimates of the Effects of

Characteristics of States. Southern Economic qurnal
55:1004-1018.

Becker, Gary S. 1962. Investment in Human Capital: A
Theoretical Analysis. The JOurnal of Political Economy
70 (supplement):9-49.

Bengtsson, Jarl. 1991. Human Resource Development. Futures
December:1085-1106.

Berry, Brian J.L. 1970. The Geography of the United States
in the Year 2000. Ekistics 171 (February):339-351.

Berry, Brian J.L., Edgar C. Conkling, and D. Michael Ray.
1993. The Global Economy. Englewood Cliffs, NJ:
Prentice Hall.

Beyers, William, Erik Johnsen and Harriet Stranahan. 1987.
Education and Economic Development. Commentary,
Winter:14-l7.

Beyers, W.B., Susan Christopherson, Rodney A. Erickson, Lay
J. Gibson, Geoffrey G.D. Hewings, Edward J. Malecki,
James E. McConnell, and John Rees. 1990. Industrial
Geography in Geography in America, Gary Gaile and Cort
Wilmott (eds.). Columbus: Merrill Publishing Company,
pp 290-314.

133

134

Bruderl, Josef, Peter Preisendorfer, and Rolf Ziegler. 1992.
Survival Chances of Newly Founded Business
Organizations. American Sociological Review 57
April:227-242.

Carlton, Dennis W. 1979. Why New Firms Locate Where They Do:
An Econometric Model. Interregional.MOvements and
Regional Growth, Wm. C. Wheaton (ed.). Washington DC:
The Urban Institute.

1983. The Location and Employment Choices of New
Firms: An Econometric Model with Discrete and
Continuous Endogenous Variables. Review of Economics
and Statistics 65:440-449.

 

Carnevale, Anthony P. 1991..America and the New.Economy.
U.S. Department of Labor, Employment and Training
Administration.

Clark, Gordon. 1985. Innovation Diffusion: Contemporary
Geographical Aproaches, CATMOG No. 40, Norwich: Regency
House.

Clark, Gordon L. 1988. Regional Development and Policy:
Emerging Themes of Research on Labour, the Firm, and
Location. Progress in Human Geography 12 4:549-559.

Clarke, Susan E. and Gary L. Gaile. 1991. The Next Wave:
Postfederal Local Economic Development Strategies.
Economic Development Quarterly 6 2:187-198.

Committee for Economic Development. 1987. Wbrk and Change:
Labor.Market.Adjustment Policies in a Competitive
WOrld. New York City.

Denison, Edward F. 1962. Education, Economic Growth, and
Gaps in Information. The Journal of.Political Economy
70:124-128.

Erdevig, Eleanor H. 1986. Small Business, Big Job Growth.
FRB Chicago Economic Perspectives, Nov/Dec:15-24.

Gaspersz, J.B.R. and W. VanVoorden. 1987. Spatial Aspects of
Internal Labour Markets Tijdschrift 78 5:359-365.

Gertler, Meric S. 1988. The Limits to Flexibility: Comments
on the Post-Fordist Vision of Production and Its
Geography. 186 Transactions N.S. 13:419-432.

Gould, Andrew and David Keeble. 1984. New Firms and Rural

Industrialization in East Anglia Regional Studies 18
3:189-201.

13S

Griliches, Zvi. 1970. Notes on the Role of Education in
Production Functions and Growth Accounting in
Education, Income, and Human Capital (edited by W. Lee
Hansen). New York: Columbia University Press.

Gudgin, G. and S. Fothergill. 1984. Geographical Variation
in the Rate of Formation of New Manufacturing Firms.
Regional Studies 18 3:203-206.

Haider, Donald. 1992. Place Wars: New Realities of the
19908. Economic Development Quarterly 6 2:127-134.

Hepworth, M.E., A.E. Green, and A.E. Gillespie. 1987. The
Spatial Division of Information Labour in Great
Britain. Environment and Planning.A 19:793-806.

Hirschhorn, Larry. 1988. Computers and Jobs: Services and
the New Mode of Production in The Impact of
Technological Change on Employment and Growth (edited
by Richard M. Cyert and David C. Mowery). Cambridge:
Ballinger Publishing Company.

Horiba, Yutaka and Rickey C. Kirkpatrick. 1979. Labor
Skills, Human Capital, and the Pattern of U.S.
Interregional Trade. Interregional.Mbvements and
Regional Growth.

Howell, David R. and Edward N. Wolff. 1991. Trends in the
Growth and Distribution of skills in the U.S.
Workplace, 1960-1985. Industrial and Labor Relations
Review 44 3:486-502.

Imrie, R. F. 1989. Industrial Restructuring, Labour, and
Locality: the Case of the British Pottery Industry.
Environment and Planning A 21: 3- 26.

Kendall, J. and B. Pigozzi. 1994. Nonemployment Income and
the Economic Base of Michigan Counties: 1959- 1986.
Growth and Change 25 1: 51- 74.

Kenen, Peter B. 1970. Skills, Human Capital, and Comparative
Advantage in Education, Income, and HUman Capital
(edited by W. Lee Hansen). New York: Columbia
University Press.

Kent, Calvin A. 1984. The Rediscovery of the Entrepreneur in
The Environment fer Entrepreneurship (edited by Calvin
A. Kent). Lexington: Lexington Books.

Kiker, B. F. 1966. The Historical Roots of the Concept of
Human Capital. JOurnal of Political Economy 74: 481- 499.

136

Killian, Molly S. and Timothy S. Parker. 1991. Education and
Local Employment Growth in a Changing Economy.
Education and Rural Economic Development: Strategies
for the 1990’s. U.S. Department of Agriculture.

Lambody, J.G. 1986. Locational Decisions and Regional
Structure. Human Behaviour in Geographical Space,
(J.B.P. Paelinck, ed.). Brookfield, VT: Gower
Publishing Co.

Leone, Robert A. 1972. The Role of Data Availability in
Intrametropolitan Workplace Location Studies. Annals of
Economic and Social.Measurement 1 2:171-182.

Leven, Charles, L. 1985. Regional Development Analysis and
Policy. Journal of Regional Science 25 4:569-592.

Lever, W.F. 1985. Theory and Methodology in Industrial
Geography. Progress in Industrial Geography, Michael
Pacione (ed.), Dover, NH: Croom Helm.

Lloyd, P.E. and C.M. Mason. 1984. Spatial Variations in New
Firm Formation in the United Kingdom: Comparative
Evidence from Merseyside, Greater Manchester and South
Hampshire. Regional Studies 18 3:207-220.

Loveman, Gary W., Michael J. Piore, and Werner Sengenberger.
1990. The Evolving Role of Small Business and Some
Implications for Employment and Training Policy. New
Developments in the Labor Market, Katharine G. Abraham
and Robert B. McKersie, eds. Cambridge: The MIT Press.

Marsden, David. 1990. Small Firms and Labor Markets in the
U.K. The Reemergence of Small Enterprise: Industrial
Restructuring in Industrialized Countries
(Sengenberger, Werner; Loveman, Gary; and Piorem
Michael; eds.). Geneva: International Institute for
Labour Studies.

McNamara, Kevin T., Warren P. Kriesel and Brady J. Deaton.
1988. Human Capital Stock and Flow and Economic Growth
Analysis:Note. Growth and Change 19 1:61-66.

Milne, S. 1990. New Forms of Manufacturing and Their Spatial
Implications: the UK Electronic Consumer Goods
Industry. Environment and Planning A 22:211-232.

Moyes, A. and P. Westhead. 1990. Environments for New Firm
Formation in Great Britain. Regional Studies 24 2:123-
136.

137

Noyelle, Thierry J. 1983. The Implications of Industry
Restructuring for Spatial Organization in the United
States in Regional Analysis and the New International
Division of Labor (edited by F. Moulaert and P.W.
Salinas), 113-134. Boston: Kluwer Nijhoff Publishing.

O’Farrell, Patrick N. 1986. Entrepreneurship and Regional

Development: Some Conceptual Issues. Regional Studies
20 6:565-574.

O’Farrell, P. N. and D. M. W. N. Hitchens. 1989. The
Competitiveness and Performance of Small Manufacturing
Firms: an Analysis of Matched Pairs in Scotland and
England. Environment and Planning.A 21: 1241- 1263.

O’hUallachain, Breandan. 1991. Industrial Geography.
Progress in Human Geography 15 1:73-80.

1991. The Location and Growth of Business and
Professional Services in American Metropolitan Areas,
1976-1986. Annals of the Association of American
Geographers 81 2:254-270.

 

Packer, Arnold H. 1991. The Demographic and Economic
Imperatives, in Human Capital and America’s Future
(edited by David W. Hornbeck and Lester M. Salamon.
Baltimore: The Johns Hopkins Press.

Piore, Michael J. 1990. The Changing Role of Small Business
in the U.S. Economy. The Reemergence of Small
Enterprise: Industrial Restructuring in Industrialized
Countries (Sengenberger, Werner; Loveman, Gary; and

Piore, Michael; eds.). Geneva: International Institute
for Labour Studies.

Rauch, James E. 1993. Productivity Gains from Geographic
Concentration of Human Capital: Evidence from the
Cities. Journal of urban Economics 34: 380- 400.

Reich, Robert B. 1991. The work of Nations. New York: Albert
A. Knopf.

Rich, Michael J. 1992. Local Economic Development in the

1980’s: Editors Introduction. Economic Development
Quarterly 6 2:148-149.

Runyon, R. P. and A. Haber. 1980. Fundamentals of Behavioral
Statistics. Reading, Massachuestts: Addison-Wesley
Publishing Company, 4th edition.

Salamon, Lester M. 1991. Overview: Why Human Capital? Why
Now? in Human Capital and.America’ s Future (edited by
David w. Hornbeck and Lester M. Salamon). Baltimore:
The Johns Hopkins Press.

138

Sass, Steven A. 1990. The U.S. Professional Sector: 1950-
1988. New England Economic Review Jan./Feb.:37-51.

Schmenner, Roger W., Joel C. Huber, and Randall L. Cook.
1987. Geographic Differences and the Location of New

Manufacturing Facilities. Journal of urban Economics
21:83-104.

Schultz, Theodore W. 1971. Investment in Human Capital: The
role of Education and of Research. New York: The Free
Press.

Schultz, Theodore W. 1990. Restoring Economic Equilibrium.
Cambridge: Basil Blackwell, Inc.

Schumacher, E.F. 1973. Small Is Beautiful. New York: Harper
and Row, Publishers, Inc.

Schwartz, Alex. 1987. Dun and Bradstreet's Market
Identifiers and Economic Geography. Professional
Geographer 39 2:166-172.

Scott, A.J. 1988. New Industrial Spaces: Flexible Production
Organization and Regional Development in North.America
and Western Europe, London: Pion Limited.

Silvestri, George and John Lukasiewicz. 1989. Projections of
Occupational Employment, 1988-2000. BLS Monthly Labor
Review, NOvember, D1-D24:37-60.

Smith, Adam. 1961. The wealth of.Nations, Representatives
Selections, edited by Bruce Mazlish. Indianapolis: The
Bobbs-Merrill Co.

Smith, David. 1971. Industrial Location. New York: Johy
Wiley and Sons, Inc.

Spenner, Kenneth I. 1988. Technological Change, Skill
Requirements, and Education: The Case for Uncertainty
in The Impact of Technological Change on Employment and
Growth (edited by Richard M. Cyert and David C.
Mowery). Cambridge: Ballinger Publishing Company.

Spindler, Charles J. and John P. Forrester (1993). Economic
Development Policy: Explaining Policy Preferences Among
Competing Models. urban.Affairs Quarterly 29 1:28-53.

Stafford, Howard. 1974. The Anatomy of the Location
Decision: Content Analysis of Case Studies in Spatial
Perspectives on Industrial Organization and Decision-
.Making (edited by P.E. Hamilton. New York: John Wiley.

139

Storey, D.J. and S. Johnson. 1987. Job Generation and Labour
.Market Change, London: The Macmillan Press, Ltd.

Storper, Michael and Richard Walker. 1983. The Theory of
Labour and the Theory of Location. International
Journal of Urban and Regional Research 1 1:1-41.

Thompson, Wilbur R. 1965. A Preface to urban Economics, The
John Hopkins Press.

Thurow, Lester C. 1970. Investment in Human Capital.
Belmont, CA: Wadsworth Publishing Company, Inc.

Unknown author. Across the U.S., a Test of Economic Mettle,
in The washington Post National weekly.Edition, Feb.
12-18, 1990, 8-12.

U.S. Department of Commerce, Bureau of the Census. 1980.
General Social and Economic Characteristics of the
Population Tables 175, Washington DC: U.S. Government
Printing Office.

U.S. Department of Commerce, Bureau of the Census. 1980.
County Business Patterns Table 1B and 2. Washington DC:
U.S. Government Printing Office.

U.S. Department of Commerce, Bureau of the Census. 1988.
County Business Patterns Table 2, Washington DC: U.S.
Government Printing Office.

U.S. Department of Labor, Commission on Workforce Quality
and Labor Market Efficiency. 1989. Investing in People,
A Strategy to Address America’s Workarce Crisis,
Washington D.C.

U.S. Small Business Administration. 1989. The State of Small
Business: A Report of the President, Washington: United
States Government Printing Office.

Warner, Paul D. 1989. Alternative Strategies for Economic
Development: Evidence from Southern Metropolitan Areas.
urban.Affairs Quarterly 24 3:389-411.

Watts, H.D. 1987. Industrial Geography, New York: John Wiley
and Sons, Inc.

Webber, Michael J. 1984. Industrial Location, Beverly Hills:
Sage Publications.

Weisbrod, Burton A. 1926. Education and Investment in Human
Capital. Journal of Political Economy 70
(supplement):106-123.

140

White, M., H.J. Braczyk, A. Ghobadian, and J. Niebuhr. 1988.
Small Firms’ Innovation: Why Regions Differ. London:
Pinter Publishers Ltd.

White, Sammis B. amd Jeffrey D. Osterman. 1991. Is
Employment Growth Really Coming from Small
Establishments? Economic Development Quarterly 5 3:241-
257.

Woodward, Douglas P. 1992. Locational Determinants of
Japanese Manufacturing Start-ups in the United States.
Southern Economic Journal 58 3:690-708.

ICHIGRN STATE UNI V.

llllllllmillill!”llllllHilllllllllllllllllllll