Content uploaded by Dustin Chambers
Author content
All content in this area was uploaded by Dustin Chambers on Aug 17, 2017
Content may be subject to copyright.
1
A Moving Target:
The Geographic Evolution of Silicon Valley, 1953-1990
Stephen Adams
Department of Management and Marketing, Salisbury University
E-mail: SBAdams@salisbury.edu
Dustin Chambers
Department of Economics and Finance, Salisbury University
E-mail: DLChambers@salisbury.edu
Michael Schultz
CDM Smith
E-mail: SchultzMD@cdmsmith.com
Abstract
!
This article provides an empirical examination of high tech firm location data from 1953
to 1990 to show a dramatic shift in geographic center of what is now called Silicon Valley.
Universities (most notably Stanford), venture capital and law firms acted as magnets for
divisions of established firms and local start-ups. These institutions combined with the
Santa Clara County’s available land to pull the high tech region’s epicenter southeastward
from San Francisco, an early source of investment capital and legal expertise. These
findings add another element (spatial change) for consideration in explaining the evolution
of industry clusters.
Keywords: High-tech Clusters; Industrial Location; Geography; Regional: Silicon Valley
JEL Classification: N92, L20, R12
2
1. Introduction: Locating Silicon Valley
1.1 The Problem
In the 1950s, economist Walter Isard bemoaned the tendency of economists to view the
world as a “wonderland of no spatial dimensions.”1 In the second half of the twentieth
century, Isard and his peers helped shape the field of regional science. Placing economic
variables in the context of location led to path-breaking scholarship in the 1990s
regarding the region as a source of competitive advantage.2 A centerpiece of this
scholarship was Silicon Valley, where regional networks and economies external to the
individual firm contrasted to vertically integrated enterprise of the second industrial
revolution.3 One of the subsequent challenges facing scholars of regional studies was
articulated by Wal and Boschma: “. . . most studies analyze clusters from a static
perspective, while questions like where clusters initially emerge, and why and how
clusters and the advantages associated to them change over time are largely ignored.”4 In
short, a wonderland of spatial dimensions without change. This paper looks at spatial
change over time in Silicon Valley, “the master cluster.”5
Silicon Valley is widely hailed as the world’s pre-eminent high-tech,
entrepreneurial region. The Valley has become the primary symbol of the knowledge
economy, in which industrial location is based less on access to raw materials and
transportation nodes than on access to brain power. From electronics and computers to
video games and social media, developments in the Valley have changed the way the
world lives, works, and plays. Little wonder that so many regions, with names like
Silicon Forest, Silicon Plain, Silicon Fen, and Silicon Wadi, have attempted to replicate
3
the Valley’s success.6 Those who seek innovation, entrepreneurship, and regional
economic development increasingly ask: How did Silicon Valley come to be?7
This article will show the importance of also asking: Where is Silicon Valley, and
where has it been? The most common answer to that question has been that the world’s
foremost high-tech region grew up on the northern end of Santa Clara County. The 1980s
brought the first spate of books about Silicon Valley: The New Alchemists (Hanson),
Silicon Valley Fever (Rogers and Larsen) and The Big Score (Malone).8 The three books
capture a Valley that gave the appearance of a hermetically sealed region. Malone, for
instance, notes ‘more than any industry in history, it is a self-contained, living city.’9
Rogers and Larsen suggest ‘Almost all of Silicon Valley lies in Santa Clara County.’10 In
1983, that definition made perfect sense because, as shown by Schmeider, that was where
the high-tech action was.11 Therefore, much of the subsequent literature on the Valley
has portrayed San Francisco as external (and always having been external) to this high-
tech region, and of the Valley as having a fixed center which changed little as start-ups
proliferated and the cluster expanded.12 An exception, published in 2015, is a comparison
between the San Francisco Bay Area (not just the San Francisco Peninsula) and Greater
Los Angeles, but Storper, et al emphasize developments within defined metropolitan
areas rather than spatial changes over time.13
Data we gathered shows that from 1960 to 1980, the geographic center of the
region’s tech activity (defined as aerospace, computers, electronics, and
telecommunications) shifted more than 11 miles southeast (out of a total of 48 miles
along the Peninsula between San Francisco and San Jose). In important ways, the time
period ending in the 1980s represented a departure from the recent past: Venture capital
4
and legal activities for the Valley were now based in the Valley (as opposed to the period
ending in 1960), most of the Valley’s enterprise involved commercial and consumer
products (as opposed to the period ending in 1965), and most of the Valley’s employees
worked for firms headquartered there (as opposed to the period ending in the mid-
1970s).14 Amongst all of that change, a static view of the Valley’s center would hold
sway in much subsequent literature and popular perception of the region.
Focusing on one place yields quite a story. Early in the 20th century, Santa Clara
County was largely orchards and farmland. Thanks to the promotion efforts of the
Southern Pacific Railroad at the turn of the twentieth century, the region became known
as the Valley of Heart’s Delight. Little wonder that many have penned stories with titles
such as ‘From the Valley of Heart’s Delight to Silicon Valley’ or ‘From Orchards to Hard
Drives.’15 Malone elegizes, ‘What had been the Valley of Heart’s Delight had been
bulldozed and paved out of existence.’16 Rogers and Larsen write that Silicon Valley ‘in
1950 was the prune capital of America. . . Today the fruit trees have disappeared.’17
Saxenian notes that the Valley ‘remained an agricultural region as late as the 1940s,
famous primarily for its apricot and walnut orchards.’18 O’Mara writes “over the second
half of the twentieth century, this region evolved from a primarily agricultural landscape
far away from the centers of industry and capital to ‘Silicon Valley’. . . ‘the ultimate post-
industrial city.’”19
Indeed, in just a few decades, Santa Clara County went from predominately
orchards and farm workers to industrial parks and high-tech employees.20 That truth,
combined with the accepted wisdom about the region’s boundaries, has led to a major
historical misconception—one that this article seeks to correct.
5
If Silicon Valley remains a spatially static story, then what the world sees is an
economic miracle: a rapid transformation from apricots to chips, from agricultural to
post-industrial—and with a competitive advantage to boot. Saxenian writes,
‘Unhampered by the constraints imposed by pre-existing industrial traditions, the
region’s founders created a distinctive technological community.’21 The twentieth century
transformation of the ‘Valley of Heart’s Delight’ from orchards to high-tech campuses
provides encouragement to developing economies harboring hopes of jumping directly
from an agrarian economy to a knowledge economy. You, too, can create something from
nothing. This article will show that this is not Silicon Valley’s full story.
The remainder of section 1 will describe developments that caused the center of
tech activity to move further from San Francisco during the second half of the twentieth
century. Section 2 discusses the high-tech firm location data analyzed in this paper.
Section 3 explores the cluster patterns in the data and tests for a southeastern shift in the
Peninsula’s high-tech firms. Section 4 concludes with implications of this study.
1.2 Activity beyond the Valley
Even amidst orchards, companies in the Valley had access to key urban resources,
repurposed from previous industrial activity. An utter focus on Santa Clara County
ignores the extent to which during the period 1909-1970, San Francisco played a major
role in Valley developments. By the early 20th Century, San Francisco had become the
leading financial center of the American West, due to an accumulation of capital from the
region’s extractive and agricultural industries.22 Distance from eastern equipment
suppliers and associated shipping difficulty and cost necessitated western innovation and
6
provided competitive protection for western manufacturers.23 Consequently, California
was the second-ranked state in the U.S. in patents per capita in both 1910 and 1920,
requiring the services of an increasing number of patent attorneys.24 Prior to development
of the region’s electronics industry, a financial and legal ecosystem was already present
in San Francisco.
From the beginning, the tech region that became known as Silicon Valley was
really two places. For its two decades in the Bay Area (1909-1931), the Valley’s first
major high-tech firm, the Federal Telegraph Company, was multilocational, with
manufacturing and research in Palo Alto, and administration and marketing (aided by
local legal and financial expertise) in San Francisco.25 A division of labor between San
Mateo County--the next county south of San Francisco--and Santa Clara County on the
one hand (manufacturing and R&D) and San Francisco on the other (finance and law)
would exist until the 1970s.26 The nexus of high-tech enterprise gradually headed south,
attracted—directly and indirectly--by Stanford University and available land nearby, but
still tethered to San Francisco’s financial resources and legal expertise.27
Prime examples of this southward movement were electronics pioneers Charles
Litton and Russell Varian. Litton was a San Francisco native and neighbor of the founder
of Moorhead Laboratories. Otis Moorhead helped spark Litton’s interest in radio and
tubes. Litton earned a graduate degree at Stanford, and after two years at Bell Telephone
Laboratories made transmission tubes at the Federal Telegraph Company from 1927 until
1931. After Federal’s departure from the Bay Area, Litton established Litton Engineering
Laboratories, a maker of vacuum tube manufacturing machinery, in San Mateo County.
7
In the mid-1940s, he established Litton Industries, a manufacturing enterprise that would
become a defense giant, in San Carlos.28
Russell Varian also attended Stanford, earning undergraduate and graduate
physics degrees in the 1920s, and then worked for San Francisco television inventor Philo
Farnsworth. After a stint in the east, Varian joined his brother Sigurd as ‘research
associates’ in the Stanford physics department. Working with Professor William Hansen,
they developed the klystron tube, which would have important applications in radar.
Beginning in 1938, the Varians worked at Stanford University as part of an agreement
with the Sperry Gyroscope Company. In 1940, the Stanford-based group moved to
Sperry’s Long Island facility, and then returned to campus after the war. In 1948, they
established Varian Associates in San Carlos, near Litton’s manufacturing plant.29
Other high-tech firms founded prior to 1950 settled in the San Mateo County
cities of San Bruno, San Carlos and Redwood City. In addition to Varian and Litton, such
was the case for telecommunications firms Eitel-McCullough and Dalmo Victor, as well
as Ampex, developers of professional audio and video equipment. Indeed, of the region’s
five firms involved in major financial events (initial public offerings or as a significant
acquisition target) in the 1950s, only Hewlett-Packard (a maker of instrumentation) was
from Santa Clara County. The others were closer to San Francisco, based in San Mateo
County.30
The draw of Stanford University, which attracted the region’s first major tech
firm in 1909, intensified in the 1950s. In addition to the university’s role as a developer
of brains, Stanford became a developer of land as well. The university’s budget problems
of the 1930s and 1940s combined with concerns about property taxes led to discussions
8
about putting “idle lands” to work. The terms of the gift from Leland and Jane Stanford
prevented the sale of any of the university’s 8,800 acres. Instead, they were leased: some
for residential use, some for a shopping center, and most notably, some for industrial
use.31
1.3 Stanford Industrial Park
The Stanford Industrial Park, established in 1951, was a harbinger of the suburban
industrial cluster, accessible by automobile. Ironically the Park’s location on the
southeast corner of the campus was chosen partly because of its railroad access (a feature
of early 20th century industrial centers), but by 1960 the tracks running diagonally
through the Park were gone. This industrial park mainly trafficked in brainpower. The
Park became a prized location for companies seeking easy access to recent graduates and
the consulting capabilities of faculty, as well as the ability to offer their employees the
opportunity to earn graduate degrees. As Rice University Chancellor Cary Croneis said in
1965, “industry now goes where the scientific talent is—or where it can be trained.”32
Principal magnets for such talent were universities such as Stanford. In order to achieve
access to brains, O’Mara notes, tech employers sought “land that was plentiful, but not
necessarily cheap.”33 In that respect, Stanford Industrial Park was the opposite of what
Storper and Walker call “false cheapness.” They argue, “. . . the same conditions of
underdevelopment that make labor or land inexpensive generally make them less
productive to use.”34
The Park’s first tenant was Varian Associates, whose initial product had been first
developed at Stanford in the 1930s. The most significant tenant of Stanford Industrial
9
Park would be Hewlett-Packard, whose initial product had also been developed at
Stanford in the 1930s.35 In 1956, H-P moved its corporate headquarters to the Park.
Approximately thirty years later, the park housed 55 companies with 26,000 workers.36
The Stanford Industrial Park became an important factor in the geographic shift in tech
from San Mateo County southward.
There were other factors. In 1952, with an eye to proximity to Stanford
University, IBM established a laboratory in San Jose, seeding the disk drive industry. In
1956, Lockheed established the region’s foremost aerospace center by moving its
Missiles Systems Division from southern California to Sunnyvale, and establishing a
research laboratory at Stanford Industrial Park. The greatest regional catalyst for start-ups
came when William Shockley, head of the team that invented the transistor at AT&T’s
Bell Telephone Laboratories, brought semiconductors to the area in 1955 (with
encouragement from Frederick Terman, Stanford’s Dean of Engineering) by establishing
a division of Beckman Instruments (a tenant of Stanford Industrial Park) in Mountain
View. In 1957, eight disgruntled employees of Shockley Semiconductor Laboratories
spun off Fairchild Semiconductor which, like its parent, was located south of Palo Alto.37
By 1986, approximately one hundred semiconductor firms (including Intel and Advanced
Micro Devices) had spun off from Shockley, Fairchild, and the ‘Fairchildren,’ moving
the Valley’s tech center southward as their buildings replaced orchards in Mountain
View, Sunnyvale, and Santa Clara.38
One reason the region’s tech center did not stop its southeastern path in the Palo
Alto area was that Stanford Industrial Park proved to be not just a destination, but also a
model. Urban industrial sites had often been multi-stories on a relatively small footprint,
10
and near rail facilities. Stanford’s suburban Industrial Park was the opposite, with “extra
acreage for the ‘land-consuming, one-story’ structures required for ‘horizontal-line
production methods,’ for parking and loading space, and for plant expansion. . .”39 The
success of Stanford Industrial Park encouraged the development of alternatives south of
Palo Alto. Such developments (albeit without the university landlord) became the norm
in Santa Clara County. By 1967, there were 38 industrial parks between Palo Alto and
San Jose.40
Some of the occupants of the new industrial parks in southern Santa Clara County
had been early lessees at Stanford Industrial Park. In the 1960s, Varian Associates bought
640 acres (nearly the entire ultimate size of the Stanford Industrial Park) in Cupertino and
Santa Clara.41 In 1977, Hewlett-Packard and Watkins-Johnson (both also headquartered
at Stanford Industrial Park) built plants in San Jose, haven chosen the location for its
proximity to the homes of their employees.42 Although overall, Silicon Valley appears to
support an argument that people follow jobs, here was a chapter of the Valley’s story
where jobs followed people.43
By 1981, Silicon Valley had more than one hundred industrial parks totaling
nearly 10,000 acres. Findlay suggests that “because most of the land available for large
projects was in the southern part of the valley, the majority of industrial parks were now
situated within the city limits of San Jose and Santa Clara.”44 One such park, Sunnyvale’s
International Science Center, lacked immediate proximity to Stanford, but compensated
with features of an incubator: shared computer and data centers, an Armed Services
Technical Information Agency Office, and a branch of the U.S. Patent Office. The
11
Center, which included Fairchild Semiconductor, Signetics, and Lockheed (among
others), also featured a facility of the University of California extension program.45
1.4 Academic Programs
Storper and Scott emphasize the creation of human capital associated with agglomeration,
including “by educational and training programs that themselves evolve in response to
the demands of local production systems.”46 Such was the case in Silicon Valley.
Although activity involving Stanford University has received the lion’s share of attention,
by the 1960s, a division of labor developed with respect to engineering education in the
Valley. New programs stretched along the Peninsula’s southeast corridor, the very path
Silicon Valley’s geographic center would follow during the next two decades: in
Sunnyvale (UC Berkeley), Santa Clara (Santa Clara University), and San Jose (San Jose
State University).
This was a far cry from the situation in 1954 when Stanford established its Honors
Cooperative Program (HCP). The HCP began as an effort to provide graduate education
in engineering to employees of the Valley’s firms. The company paid tuition and the
employee completed the degree in two years, while receiving full salary. After five years,
the program enrolled 324 students—more than 40 percent of Stanford’s graduate
enrollment in engineering.47 The problem for some employers and would-be students was
HCP’s selectivity. Engineering Dean Frederick Terman had designed the program with a
high demand/supply ratio in mind, so Stanford’s program was not large enough to keep
up with demand. From 1955 to 1957, complaints arose from the region’s employers about
insufficient local availability of advanced engineering training. Particularly vocal were
12
firms in cities south of Palo Alto: Sunnyvale (Westinghouse) and San Jose (FMC,
General Electric, and IBM). The primary target for such grievances was across the Bay:
Why couldn’t UC Berkeley establish an engineering center in the Valley to offer graduate
courses in the 1950s? Cal’s professors were reluctant to make a more than two-hour
round trip commute to teach in the Valley, so Engineering Dean M. P. O’Brien failed to
follow through on his earlier commitment to such a program. The results were the firing
of O’Brien by UC President Clark Kerr, and an opportunity for universities based south
of Palo Alto.48
Beginning in 1959, Santa Clara University (one of the first four accredited
engineering schools in California) offered engineering graduate courses between 7:00 and
9:00 in the morning. By 1963, enrollments in the “early bird” program exceeded 600—
nearly the same level as Stanford’s HCP.49 The division of labor between Stanford and
Santa Clara was two-fold. Because of its superior national reputation, Stanford could
(other things being equal) cherry pick the top students. Yet there was a geographic
division of labor as well. Employees of companies in the Stanford Industrial Park favored
the Stanford program. Hewlett-Packard had 45 students in the Stanford’s HCP, but only
five in Santa Clara’s early bird program. Lockheed, half way between Palo Alto and
Santa Clara, was the leading supplier of students in both programs. Meanwhile, San Jose-
based FMC and IBM were sparsely represented in HCP, but had nearly 100 students
between them in the early bird program.50
Also in 1959, San Jose State received accreditation in civil and electrical
engineering, as well as permission from the state to offer a master’s degree in
engineering. San Jose State has since been a leading provider of engineers to firms in the
13
Valley.51 So the migration of firms in a southeasterly direction followed not only
available land and a growing number of industrial parks, but also followed educational
opportunity—from Palo Alto to Santa Clara and San Jose.
1.5 Economy Two
The Valley’s growing number of indigenous firms attracted financial resources
southward. Prior to 1970, most of the investment in Peninsula firms not provided by their
founders came from the north. The Crocker National Bank of San Francisco (William
Crocker’s father was one of the Central Pacific railroad’s ‘Big Four’) was an early
financier in high tech firms such as Federal Telegraph and Ampex. Joseph and Henry
McMicking and, later, Reid Dennis, were other early San Francisco investors in Ampex
in the 1940s and 1950s.52 Another example of San Francisco’s influence came in the
semiconductor industry. When Arthur Rock came to the Bay Area after a 1957
investment in Fairchild Semiconductor, he set up shop in San Francisco—and also
became a key financier for Intel and then Apple. Also based in San Francisco, Thomas J.
Davis, Jr. would team up with Rock in 1961, having made the first significant investment
in the microwave firm Watkins-Johnson Company in 1957 as the investment brains
behind the Kern County Land Company.53
As the region became increasingly dominated by start-ups, the Valley developed
its own sources of financing closer to the entrepreneurial activity. In the late 1960s, Davis
established the Mayfield Fund, one of the early local venture capital firms prior to the
1970s clustering of VC’s on Sand Hill Road.54 The phenomenon of venture capitalists’
proximity to their investments has been well chronicled, especially in the Valley.55 The
14
departure of venture capital firms from San Francisco marked a shift from an era when
tech followed money (prior to 1960, when the Valley’s center was north of Stanford
University) to an era when money followed tech. The latter phenomenon enabled much
of the industry development that moved the region’s tech center southward.
Legal expertise followed a different path to a similar outcome. The leading patent
attorneys serving the region from 1920s until the 1960s were based in San Francisco,
including the firms of Donald Lippincott and Paul D. Flehr.56 As tech enterprise
multiplied in Santa Clara County and increasingly congested traffic between the Valley
and San Francisco caused travel time to increase, Valley-based attorneys enjoyed a
competitive advantage in high-tech law. No firm would exploit that advantage more
successfully than the Palo Alto firm Wilson, Sonsini, Goodrich, and Rosati—which
expanded from a dozen of attorneys in 1975 to more than 700 by 2000.57
Not until after 1970, was ‘Economy Two,’ as Kenney and von Burg call it, in
place: organizations in or adjacent to Santa Clara County (such as venture capital and law
firms) whose primary mission is to enable start-ups. The development of Economy Two,
combined with elements of Economy One (universities and research institutions such as
Stanford University, Xerox PARC, and Stanford Research Institute, for which helping
start-ups was ancillary to their mission) reinforced the appearance of a self-contained
region that observers emphasized in the 1980s.58
15
2. Bay Area Firm Location Data
In “Cluster Evolution and a Roadmap for Future research,” Boschma and Fornahl suggest
the need to “perform longitudinal studies on clusters.” Few such studies had been done
“because data availability on clusters over a long period of time is a real problem
[especially for] the formative stage of cluster development.”59 We have gathered such
data for Silicon Valley’s spatial change, dating to when the region had dozens, rather than
thousands of tech firms.
We provide data and visual support showing that the geographic center of high-
tech activity shifted approximately 24 kilometers (km) to the southeast over this 37-year
period, moving 16.47 km to the east and 17.33 km to the south. Meanwhile, the
migration pattern of existing firms was very similar over the same period, with relocating
firms moving on average approximately 3.16 km east and 3.08 km south every decade.
This evidence strongly supports the theory that early high tech development occurred in
proximity to San Francisco’s legal and financial expertise, but as these capabilities
became more widely available near Stanford University, new companies were founded in
(and existing ones migrated to) Silicon Valley—and continued to move south toward
available land and proximity to other academic programs.
To determine the extent to which the geographic center of high tech industry
shifted from the northern part of the San Francisco Peninsula toward the South Bay, we
collected and geocoded an exhaustive list of high tech firms operating in the Bay Area
between 1953 and 1990. We acquired five cross-sections of firm-specific data, each
spaced approximately 10 years apart, beginning in 1953 and ending in 1990. Our initial
firm data for 1953 comes from an archived copy of the membership roster of the West
16
Coast Electrical Manufacturers Association (WCEMA, a trade group co-founded by
David Packard in 1943).60 Each entry in the membership index consists of 1) firm name,
2) location mailing address, 3) names of corporate executives, 4) date of establishment
and facility size, and 5) a brief description of the firm’s products.
For the next three time periods (1960, 1970, and 1980), we obtained data from the
Dun & Bradstreet Metalworking Directories at the Library of Congress. We use the Dun
& Bradstreet Metalworking Directories because they are the earliest source (to our
knowledge) of SIC-classified locational firm data. Moreover, because early tech
companies were heavy users of metal, the industries relevant to Silicon Valley’s
development are represented in the guides. For the 1990 cross-section, we purchased the
data directly from Dun & Bradstreet.61 Each observation consists of 1) firm name, 2)
location mailing address, 3) number of employees, and 4) Standard Industrial
Classification (SIC) code. All firms from the WCEMA roster and all Dun & Bradstreet
sourced firms with high-tech SIC codes located within the California counties (from
north to south) of San Francisco, San Mateo, Santa Clara, and Santa Cruz.i
Using the street addresses, each observation was geo-spatially coded (i.e. assigned
an ‘X’ and ‘Y’ state-plane coordinate measured in meters). The dataset contains a total
of 3,534 spatially matched observations, which are distributed by year as follows: 29
observations from 1953, 47 observations from 1960, 148 observations from 1970, 258
observations from 1980, and 3,052 observations from 1990. Data plots are provided in
Figures 1 to 5.
i The relevant high-tech SIC codes are: 357 (Computer and Office Equipment), 366 (Communications
Equipment), 367 (Electronic Components and Accessories), 376 (Guided Missiles and Space Vehicles), 38
(Instruments), and 737 (Computer Programming and Data Processing).
17
3. The Location and Movement of Bay Area Firms
3.1 Evidence of High Tech Clusters
Consistent with the findings of other researchers, we find that Bay Area tech firms form
geo-spatial clusters in each of our decadal snapshots.62 Looking at Table 1, we see that
the mean nearest neighbor, which is a measure of the average distance between any given
firm and its closest neighbor, steadily declined from 1.76 km in 1953, down to a mere
221 meters in 1990, confirming that firms were located in ever-closer proximity. To
formally test this hypothesis, mean nearest neighbor tests were performed for each of the
five time periods. The test is based on Monte Carlo simulations, whereby a sample of
points (equal to the number of observations from the original dataset) is randomly chosen
from a bivariate uniform distribution over the study area. The mean nearest neighbor is
calculated from each round of the simulation and an empirical distribution of mean
nearest distances is constructed. Because this distribution is asymptotically normally
distributed, it implies a simple asymptotically valid test (under the null hypothesis):
!"#$%&'
()%"'
*+%,-.' (1)
where /& is the actual sample mean nearest neighbor for period t (e.g. the sample mean
nearest neighbor for 1960), 0%1' is the average mean nearest neighbor from the period-t
Monte Carlo simulations, and 2$%&' is the standard deviation of the mean nearest
neighbors from the period-t Monte Carlo simulations. The test results are provided in
Table 2, and clearly reject the null hypothesis that firms are spatially randomly
distributed at any standard level of significance. We therefore accept the alternative
hypothesis of high tech firm clustering.
18
An additional spatial clustering test called the quadrat-count chi-square test was
conducted with similar results. This test divides the study region into a grid of q, equally
sized squares and counts the number of firms (k) lying in each square, and the following
test statistic is calculated:
34#5 6
7
489
5*:;#<
= (2)
where >? is the number of firms in area i, and @ is the average number of firms per square
(i.e. @ A BCD, where n is the sample size). Table 2 provides the test results for the
quadrat-count chi-square test, which also clearly reject the null hypothesis that firms are
spatially randomly distributed at any standard level of significance.ii
3.2 Description of Cluster Patterns
Beginning in 1953, there appear to be several high-tech clusters throughout the San
Francisco Peninsula, in San Francisco, San Bruno, San Carlos, San Mateo, Redwood
City, Palo Alto, Sunnyvale, and San Jose (see Figure 6). In that year, the mean center of
high tech firms was north of the U.S. Highway 101 between San Mateo and Redwood
City (see Figure 1). This clustering pattern continues into 1960, with additional ‘hot
spots’ (i.e. areas with a high concentration of high tech firms) appearing in Santa Cruz
and near the intersection of California Highway 85 and U.S. 101 (see Figure 7). In that
year, the mean center of high tech firms was northeast of the U.S. 101 near Redwood
City (see Figure 2). By 1970, the hotspot patterns noticeably changed, with San Francisco
diminishing in intensity, and the only hotspot north of Redwood City lying between San
ii See O’Sullivan and Unwin, 98-105 for a discussion of both the mean nearest neighbor and quadrat-count
tests.
19
Bruno and San Mateo (see Figure 8). There appears to be a merger of the high tech
clusters in what one would traditionally identify as Silicon Valley, encompassing
Stanford University, Mountain View, Sunnyvale, and Santa Clara. Additional hotspots
emerge south of San Jose, at the intersection of California Highway 85 and California
Highway 17 (near the city of Los Gatos), and near Watsonville. In addition, the mean
center (see Figure 3) migrated sharply to the southeast, lying due east of Stanford
University and adjacent to U.S. 101. In 1980, the intensity of hotspots north of Redwood
City continued to decline, with the Redwood City hotspot splitting in two, with the
respective halves centered over Redwood City and San Carlos (see Figure 9). In
addition, the continuous cluster of high tech firms stretching from Stanford University to
Santa Clara also fragmented, with a distinct cluster around Stanford breaking off from the
rest of the Valley, as well as another distinct cluster forming around Cupertino. The mean
center continued to migrate in a southeasterly direction, lying between Moffett Airfield
and Sunnyvale (see Figure 4). By 1990, there is a proliferation of high-tech firms
scattered over the entire region, including far more remote locations like Morgan Hill and
Gilroy (see Figure 10). The greatest intensity of firms occurs along U.S. 101 corridor,
stretching continuously from San Mateo in the north to San Jose in the south, with finger-
like protrusions extending out to Cupertino and Los Gatos. Interestingly, the mean center
shifted very little between 1980 and 1990, moving just 650 meters to the east and 1.99
km south, lying directly over Sunnyvale (see Figure 5).
20
3.3 Evidence of a Shifting Geospatial Center
To more formally test the hypothesis that the mean center of the spatial firm distribution
shifted between 1953 and 1990, a bootstrap technique is used whereby 70% of each
period’s sample data were re-sampled (with replacement) each round and used to
calculate the mean center. This is repeated 1,000 times and 10%, 5%, and 1% confidence
intervals are constructed for each time period. The formal hypothesis is as follows:
EFG H&I< A H& (3)
E<G H&I< J H&
where H& is the mean center, x-y coordinate vector in period t (e.g. 1960), and H&I< is the
mean center in the following period (e.g. 1970). Taking the information available at time
t as given, this implies a simple decision rule for a 1%-level test: reject the null
hypothesis (EF) if the straight-line distance between the mean centers exceeds the
bootstrapped threshold (KL ML H&I< N H&O 0<P%1', where 0<P%1' is the bootstrapped 1%
confidence distance (circle radius) in period t).
The results of this bootstrap procedure are provided in Table 3. With the
exception of the 1953 and 1960 mean centers, all of the remaining mean centers (i.e.
1970, 1980, and 1990) move a statistically significant distance relative to the previous
decade. The 1970 mean center is 11.58 km from the 1960 mean center, which far exceeds
the 1% confidence circle radius of 9.54 km. Likewise, the 1980 mean center is 6.33 km
from the 1970 mean center, which exceeds the 1% confidence circle radius of 5.27 km.
Finally, the 1990 mean center is 2.17 km from the 1980 mean center, which exceeds the
5% confidence circle radius of 2.01 km. In other words, there is a statistically significant
break/change in the mean center of the firm distributions between 1960 and 1970, which
21
supports our earlier finding that the high tech firms began breaking free of San
Francisco’s influence at around this time.
3.4 Intraregional Firm Relocation Patterns
Our data allows us to investigate the movement of firms within the Silicon Valley study
area. A relocation of firms away from San Francisco and toward the South Bay would
provide strong evidence that important support services like venture funding, legal
counsel, and academic programs were now readily available to accompany available land
in Silicon Valley.
Firms were matched between 1953 and 1960, 1960 and 1970, 1970 and 1980, and
1980 and 1990, to determine what proportion of firms moved in any given 10-year
period, and the net migration patterns of these firms. The results are provided in Table 4.
Overall, roughly half of all high-tech firms relocated in any given 10-year period. Among
those firms that did relocate, the average distance traveled was 7.95 km. The mean center
of the relocating firms shifted approximately 3.16 km east and 3.08 km south each
decade. These intra-regional migration patterns match the overall evolution of the
broader sample, in which the mean center of all firms drifted on average 4.12 km east and
4.33 km south each decade.
22
4. Conclusion: Moving Targets and Evolving Infrastructure
American suburbanization has been well chronicled, including the movement of
manufacturing, research and development, and administration away from inner cities.
The number of industrial parks in the United States increased by nearly a factor of ten
from 1940 to 1957.63 A study of American tech firms in the early 1980s found that more
than two-thirds of founders who moved at the time of founding departed from large
cities.64 In the San Francisco area, this phenomenon was, if anything, exaggerated. By
1970 almost three fourths of all work trips in the SF metropolitan area, noted a scholar of
suburbanization, ‘were by people who neither lived nor worked in the core city.’65 This
article provides a more detailed glimpse than has been previously available of how the
Silicon Valley story fits that larger narrative of industry dispersion.
Silicon Valley did not spring up as a collection of garages at the doorstep of
Stanford University and then expand outward. As we have shown, the geographic critical
mass of Silicon Valley has been a moving target for decades. The infrastructure
supporting the region when its geographic center was in San Mateo County was different
in many ways from the one supporting the region when it was centered in Santa Clara and
San Jose. At the beginning, much of its infrastructure was in San Francisco, while
Stanford University and available land acted as magnets to gradually pull the high tech
region’s epicenter southward. With a steady movement of the geographic center of high
tech activity toward the South Bay Area, the ecosystem upon which the Valley depends
shifted in key areas: capital (from San Francisco banks to Menlo Park-based venture
capitalists), law (from San Francisco patent attorneys to tech-centered law firms in Palo
Alto), and higher education--not only with developments at Stanford, creation of Santa
23
Clara University’s early bird program and with the 1959 accreditation and subsequent
growth of San Jose State’s engineering school.66
Since Michael Porter popularized the subject of clusters, one of the calls in the
literature has involved the question of change—and the evolution of clusters in places
from Sweden to Greece to Akron, Ohio have been become the subject of scholarly study,
providing a variety of lessons for national policy makers and local officials.67 The results
of this study also suggest broader lessons regarding the spatial evolution of industry. Our
central finding relates to a point made by Bresnahan, Gambardella, and Saxenian, “. . .
that the economic factors that give rise to the start of a cluster can be very different from
those that keep it going.”68 Our research suggests that such is also the case for
institutional support associated with economic factors as a cluster’s center moves over
time.
24
References
Adams, S. B. “Arc of Empire: Federal Telegraph Company, the U. S. Navy, and the
Beginnings of Silicon Valley,” Business History Review (forthcoming).
Adams, S. B. “A Garage, an Idea, and an Ecosystem: Reinterpreting the H-P Story,”
paper presented at Business History Conference, Portland, OR, April 2, 2016.
Adams, S. B. “Born on Third Base: Silicon Valley’s Rich Inheritance,” paper presented
at Business History Conference, Miami, FL, June 26, 2015.
Adams, S. B. “Growing Where You are Planted: Exogenous Firms and the Seeding of
Silicon Valley,” Research Policy 40, no. 3 (2011): 368-379.
Adams, S. B. “Their Minds Will Follow: Big Business and California Higher Education,
1954-1960,” paper presented at Business History Conference, St. Louis, MO, April 2011.
Adams, S.B. “Stanford and Silicon Valley: Lessons on Becoming a High-Tech Region,”
California Management Review 48, no. 1 (2005): 29-51.
Alpers, M. L. “Valley of Heart’s Delight: Orchards to Hard Drives in the San Jose
Mercury News,” Masters Theses. Paper 3743, 2010.
Boschma, R., and D. Fornahl. “Cluster Evolution and a Roadmap for Future Research,”
Regional Studies 45, no. 10 (2011): 1295-1298.
Boschma, R., F. Neffke, and M. Henning. “How do Regions Diversify over time?
Industry Relatedness and the development of new growth paths in regions, paper
presented at conference on “Opening up Innovation: Strategy, Organization and
Technology.” Imperial College, London, June 16-18, 2010.
Bresnahan, T., A. Gambardella, and A. Saxenian. “’Old Economy’ Inputs for ‘New
Economy’ Outcomes: Cluster Formation in the New Silicon Valleys,” Industrial and
Corporate Change 10, no. 4 (2001): 835-860.
Carlton, D.L., and P.A. Coclanis. “The Uninventive South? A Quantitative Look at
Region and American Inventiveness,” Technology and Culture 36, no. 2 (1995): 302-326.
Cooke, P. Knowledge Economies: Clusters, Learning and Cooperative Advantage.
London: Routledge, 2002.
Cooke, P. “Regional Innovation Systems, Clusters, and the Knowledge Economy,”
Industrial and Corporate Change 10, no. 4 (2001): 945-974.
Cooper, A.C. “The Role of Incubator Organizations in the Founding of Growth-Oriented
Firms,” Journal of Business Venturing 1 (1985): 75-86.
25
David, P.A., and G. Wright. “Increasing Returns and the Genesis of American Resource
Abundance,” Industrial and Corporate Change 6, no. 2 (1997): 203-245.
De Socio, M. “Regime Network Restructuring in Akron, Ohio, 1975-2009: A
Longitudinal Social Network Analysis,” Growth and Change 43, no. 1 (March 2012): 27-
55.
Dun & Bradstreet. Metalworking Directory. New York: Dun & Bradstreet, 1960.
Dun & Bradstreet. Metalworking Directory. New York: Dun & Bradstreet, 1970.
Dun & Bradstreet. Metalworking Directory. New York: Dun & Bradstreet, 1980.
Dun & Bradstreet. Metalworking Directory. New York: Dun & Bradstreet, 1990.
Findlay, J. Magic Lands: Western Cityscapes and American Culture after 1940.
Berkeley, CA: University of California Press, 1993.
Flehr, P. D. Inventors and their Inventions. Palo Alto, CA: Pacific Books, 1990.
Gillmor, C.S. Fred Terman at Stanford: Building a Discipline, a University, and Silicon
Valley. Stanford, CA: Stanford University Press, 2004.
Hanson, D. The New Alchemists: Silicon Valley and the Microelectronics Revolution.
Boston: Little, Brown and Company, 1982.
Hasegawa, S. Engineering the Future: A History of the San Jose State University College
of Engineering. Sausalito, CA: Oral History Associates, 1992.
Jackson, K. T. Crabgrass Frontier: The Suburbanization of the United States. New York:
Oxford University Press, 1985.
Kenney, M., and R. Florida. “Venture Capital in Silicon Valley: Fueling New Firm
Formation.” In Understanding Silicon Valley: The Anatomy of an Entrepreneurial
Region, edited by M. Kenney, 98-123. Stanford, CA: Stanford University Press, 2000.
Kenney, M., and U. von Burg. “Institutions and Economies: Creating Silicon Valley,” in
Understanding Silicon Valley: The Anatomy of an Entrepreneurial Region, edited by M.
Kenney, 218-240. Stanford, CA: Stanford University Press, 2000.
Klepper, S. “Silicon Valley—A Chip Off the Old Detroit Bloc,” in Entrepreneurship,
Growth, and Public Policy, edited by Z. J. Acs, D. B. Audretsch, and R. J. Strom, 79-115.
New York: Cambridge University Press, 2009.
Lécuyer, C. Making Silicon Valley: Innovation and the Growth of High Tech, 1930-1970.
Cambridge, MA: The MIT Press, 2007.
26
Longhi, C. “Networks, Collective Learning and Technology Development in Innovative
High technology Regions: The Case of Sophia-Antipolis,” Regional Studies 33, no. 4:
333-342.
Madrigal, A. C. “Not Even Silicon Valley Escapes History,” July 23, 2013.
http://www.theatlantic.com/technology/archive/2013/07/not-even-silicon-valley-escapes-
history/277824/
Malone, M. The Big Score: The Billion Dollar Story of Silicon Valley. Garden City, New
York: Doubleday, 1985.
Matthews, G. Silicon Valley, Women, and the California Dream: Gender, Class, and
Opportunity in the Twentieth Century. Stanford, CA: Stanford University Press, 2003.
McWilliams, C. California: The Great Exception. Santa Barbara, CA: Peregrine Smith,
Inc., 1976.
Miller, R., and M. Cote. “Growing the Next Silicon Valley,” Harvard Business Review
63, no. 4 (1985): 114-124.
Norberg, A. L. “The Origins of the Electronics Industry on the Pacific Coast,”
Proceedings of the IEEE, 64, no. 9 (1976): 1314- 1322.
Olmstead, A.L., and Rhode, P. “An Overview of California Agricultural Mechanization,”
Agricultural History 62, no. 3 (1988): 86-112.
O’Mara, M. P. Cities of Knowledge: Cold War Science and the Search for the Next
Silicon Valley. Princeton, N. J.: Princeton University Press, 2005.
Porter, M. The Competitive Advantage of Nations. New York: Free Press, 1998.
Rao, A. “The Helpers: Lawyers, Accountants & the Regional Investment Banks” in A.
Rao and P. Scaruffi, A History of Silicon Valley, 2010.
http://www.scaruffi.com/politics/sv.html
Rogers, E., and J. Larsen. Silicon Valley Fever: Growth of High-Technology Culture.
New York: Allen and Unwin, 1984.
Saxenian, A. Regional Advantage. Cambridge, MA: Harvard University Press, 1996.
Saxenian, A. “In Search of Power: The Organization of Business Interests in Silicon
Valley and Route 128,” Economy and Society 18, no. 1 (1989): 25-70.
Scott, A.J. “Economic Geography: the Great Half Century,” Cambridge Journal of
Economics 24 (2000): 483-504.
27
Scott, A.J. Regions and the World Economy: The Coming Shape of Global Production,
Competition, and Political Order. New York: Oxford University Press, 1998.
Storper, M., T. Kemeny, N.P. Makarem, and T. Osman, The Rise and Fall of Urban
Economies. Stanford, CA: Stanford University Press, 2015.
Storper, M., and A.J. Scott. “Rethinking human capital, creativity, and urban growth,”
Journal of Economic Geography 9, no. 2 (2009): 147-167.
Storper, M., and R. Walker. The Capitalist Imperative. Cambridge, MA: Blackwell,
1989.
Sturgeon, T. J. “How Silicon Valley Came to Be.” In Understanding Silicon Valley: The
Anatomy of an Entrepreneurial Region, edited by M. Kenney, 15-47. Stanford, CA:
Stanford University Press, 2000.
Suchman, M. C. “Dealmakers and Counselors: Law Firms as Intermediaries in the
Development of Silicon Valley.” In Understanding Silicon Valley: The Anatomy of an
Entrepreneurial Region, edited by M. Kenney, 71-97. Stanford, CA: Stanford University
Press, 2000.
Tajnai, C. From the Valley of Heart’s Delight to Silicon Valley: A Study of Stanford
University’s Role in the Transition, Stanford, CA: Stanford Department of Computer
Science, 1996.
University of Santa Clara School of Engineering, “The Second Decade of Achievement,”
1968. School of Engineering Papers, classification 23DB2, Box 2 of 2, Accession 984-
004. Office Files, Dean School of Engineering, folder “The Second Decade of
Achievement.”
Wal, A.L.J. “Cluster Emergence and Network Evolution: A Longitudinal Analysis of the
Inventor Network in Sophia-Antipolis. Regional Studies 47, no. 5 (2013): 651-668.
Wal, A.L.J., and R. Boschma. “Co-evolution of firms, industries and networks in space,”
Regional Studies 45, no. 7 (2011): 919-933.
WCEMA. Product List—Membership Roster. Los Angeles, CA: West Coast Electronics
Manufacturers Association, 1953. Varian Associates Papers (73/65c), box 4, folder 44.
Bancroft Library, University of California, Berkeley.
28
Notes
1 Scott, “Economic Geography,” p. 486.
2 Porter, The Competitive Advantage of Nations; Saxenian, Regional Advantage; Scott, Regions and the
World Economy.
3 Scott, “Economic Geography,” pp. 492-493.
4 Wal and Boschma, “Co-evolution of firms, industries and networks in space,” p. 4.
5 Cooke, Knowledge Economies, p. 194.
6 Miller and Cote, “Growing the Next”; O’Mara, Cities of Knowledge, p. 1
7 Sturgeon, “How Silicon Valley Came”; Lécuyer, Making Silicon Valley; Norberg, “Origins”; Adams,
“Arc of Empire.”
8 Hanson, The New Alchemists; Rogers and Larsen, Silicon Valley Fever; Malone, The Big Score.
9 Malone, Big Score, 8.
10 Rogers and Larsen, 28.
11 Madrigal, “Not Even Silicon Valley,” 1, 2, 6.
12 Saxenian, Regional Advantage; O’Mara, Cities of Knowledge; Lécuyer, Making Silicon Valley.
13 Storper, et al, The Rise and Fall of Urban Economies.
14 Adams, “Growing Where.”
15 Tajnai, “From the Valley of Heart’s Delight”; Alpers, “Valley of Heart’s Delight.”
16 Malone, The Big Score, 428.
17 Rogers and Larsen, Silicon Valley Fever, 28.
18 Saxenian, Regional Advantage, 11.
19 O’Mara, Cities of Knowledge, 97.
20 Matthews, Silicon Valley, Women, 137-140.
21 Saxenian, Regional Advantage, 12.
22 McWilliams, California, 31-33, 216-218, 229-232.
23 Olmstead and Rhode, “An Overview”; David and Wright, “Increasing Returns.”
24 Carlton and Coclanis, “The Uninventive South,” 322-323.
25 Adams, “Born on Third,” 17-18.
26 Sturgeon, “How Silicon Valley,” 34-38.
27 O’Mara, Cities of Knowledge, 18-19.
28 Lécuyer, Making Silicon Valley, 21-30, 53-55.
29 Gillmor, Fred Terman, 161-167; Lécuyer, Making Silicon Valley, 95-100.
30 Lécuyer, Making Silicon Valley, 30, 40, 100; Sturgeon, “How Silicon Valley,” 41-44.
31 Findlay, Magic Lands, pp. 125-126.
32 O’Mara, Cities of Knowledge, p. 70.
33 Ibid, 68.
34 Storper and Walker, Capitalist Imperative, p. 73.
35 Adams, “A Garage, an Idea, and an Ecosystem.”
36 Findlay, Magic Lands, p. 140.
37 Adams, “Growing Where.”
38 Klepper, “Silicon Valley,” 85.
39 Findlay, Magic Lands, p. 119.
40 Ibid, 151.
41 Ibid 152.
42 Ibid 157.
43 Storper and Scott, “Rethinking human capital,” p. 147.
44 Findlay, Magic Lands, p. 152.
45 Ibid 152.
46 Storper and Scott, “Rethinking human capital,” p. 162.
47 Adams, “Stanford and Silicon Valley.”
48 Adams, “Growing where you are planted.”
49 “The Second Decade of Achievement,” pp. 3 and 22; Adams, “Stanford and Silicon Valley.”
50 “The Second Decade of Achievement.”
29
51 Adams, “Their Minds Will Follow.”
52 Sturgeon, “How Silicon Valley,” 20, 45-46.
53 Kenney and Florida, “Venture Capital,” 106-109; Lécuyer, “Making Silicon Valley,” 166.
54 Ibid, 115.
55 Cooke, “Regional Innovation Systems,” 967.
56 Sturgeon, “How Silicon Valley,” 35, 40; Flehr, Inventors; Adams, “A Garage,” 6-9.
57 Suchman, “Dealmakers,” 74; Rao, “The Helpers.”
58 Kenney and von Burg, “Institutions.”
59 Boschma and Fornahl, “Cluster evolution,” p. 3.
60 WCEMA, Product List—Membership Roster; Saxenian, “In Search Of,” 33.
61 Dun & Bradstreet, Metalworking Directory, 1960, 1970, 1980, 1990.
62 Lécuyer, Making Silicon Valley; O’Mara, Cities of Knowledge; Rogers and Larsen, Silicon Valley Fever;
Saxenian, Regional Advantage.
63 O’Mara, Cities of Knowledge, 63-64.
64 Cooper, “The Role Of,” 82.
65 Jackson, Crabgrass Frontier, 267.
66 Hasegawa, Engineering the Future, 81.
67 Longhi (1999); Wal (2013); Boschma, et al (2010); de Socia (2012).
68 Bresnahan, Gambardella, and Saxenian, “’Old Economy’ Inputs for ‘New Economy’ Outcomes,” p. 835.
30
Table 1 – Spatial Summary Statistics (1953 to 1990)
Descriptive Point-Pattern Statistics
!!
1953!
1960!
1970!
1980!
1990!
Geomean1,2!
!!
!!
!!
!!
!!
!!!X!
-4.31!
-1.66!
5.89!
11.51!
12.16!
!!!Y!
11.42!
8.04!
-0.69!
-3.92!
-5.91!
Distance!to!Stanford!
12.21!
8.21!
5.93!
12.16!
13.52!
!!
!!
!!
!!
!!
!!
Standard!Distance!
19.86!
22.48!
19.90!
14.72!
19.22!
Mean!Nearest!Neighbor!
1.76!
2.22!
0.67!
0.46!
0.21!
!!
!!
!!
!!
!!
!!
Observations!
29!
47!
148!
258!
3,106!
Notes:!
!!
!!
!!
!!
!!
1)!All!distances!measured!in!kilometers!(km)!
!!
!!
!!
2)!Geomeans!are!measured!relative!to!Stanford!University!using!state-plane!
!!coordinates!(i.e.!Stanford:!X:!0,!Y:!0).!!Positive!values!of!X!correspond!to!points!east!
!!of!Stanford,!while!positive!values!of!Y!lie!north!of!Stanford.!
!!
!!
!!
!!
!!
!!
!!
!!
Table 2 – Tests of Random Spatial Distribution (1953 to 1990)
Cluster Test Statistics
!!
1953!
1960!
1970!
1980!
1990!
Mean!Nearest!Neighbor!Test!
6.42!
7.90!
17.90!
23.71!
81.70!
!!!p-value!
0.00!
0.00!
0.00!
0.00!
0.00!
!!
!!
!!
!!
!!
!!
Quadrat-Count!Chi-square!Test!
4,628!
9,467!
21,204!
26,060!
108,398!
!!!p-value!
0.00!
0.00!
0.00!
0.00!
0.00!
Notes:!
!!
!!
!!
!!
!!
The!null!hypothesis!under!both!tests!is!that!there!is!no!spatial!clustering.!
!!
This!null!hypothesis!is!strongly!rejected!for!all!tests!and!time!periods.!
!!
!!
!!
!!
!!
!!
!!
31
Table 3 – Bootstrapped Mean Centers and Confidence Circles
Bootstrapped Mean Center1
!!
1953!
1960!
1970!
1980!
1990!
Bootstrap!mean!center2,3!
!!
!!
!!
!!
!!
!!X-coordinate!
-4.33!
-1.63!
5.95!
11.47!
12.16!
!!Y-coordinate!
11.46!
8.01!
-0.75!
-3.85!
-5.91!
10%!confidence!circle!radius4!
7.38!
6.15!
3.10!
1.69!
0.64!
5%!confidence!circle!radius!
8.68!
7.45!
3.80!
2.01!
0.76!
1%!confidence!circle!radius!
11.10!
9.54!
5.27!
2.63!
1.02!
!!
!!
!!
!!
!!
!!
Change!in!bootstrap!mean!center!
!!
!!
!!
!!
!!
(from!previous!mean!center)!
!!
!!
!!
!!
!!
!!X-coordinate!change!
---!
2.70!
7.58!
5.52!
0.69!
!!Y-coordinate!change!
---!
-3.45!
-8.76!
-3.10!
-2.06!
!!
!!
!!
!!
!!
!!
Distance!from!previous!mean!center!
---!
4.38!
11.58***!
6.33***!
2.17***!
!!
!!
!!
!!
!!
!!
Notes:!
!!
!!
!!
!!
!!
1)!Seventy!percent!of!original!dataset!for!each!period!was!randomly!selected!with!
replacement;!process!repeated!1,000!times!
2)!All!distances!measured!in!kilometers!(km)!
!!
!!
!!
!!
3)!Geomeans!are!measured!relative!to!Stanford!University!using!state!plane!
coordinates!(X:!0,!Y:0)!!
4)!A!confidence!circle!with!a!given!p-value!refers!to!the!probability!of!a!firm!
lying!outside!the!confidence!circle!
!!
!!
!!
!!
!!
!!
!!
32
Table 4 – 10-Year Intraregional Firm Relocation Patterns
Intraregional Firm Movements
!!
1953-1960!
!!
1960-1970!
!!
1970-1980!
!!
1980-1990!
Average!Firm!Movement1,2!
!!
!!
!!
!!
!!
!!
!!
!!Dx-coordinate!
1.78!
!!
2.87!
!!
4.91!
!!
3.07!
!!Dy-coordinate!
-4.73!
!!
-4.17!
!!
-2.18!
!!
-1.22!
!!Average!Firm!Move!Distance!
5.52!
!!
7.04!
!!
8.26!
!!
10.97!
!!Direction!of!Change!
South!East!
!!
South!East!
!!
South!East!
!!
South!East!
!!
!!
!!
!!
!!
!!
!!
!!
Period-over-period!matches!
10!
!!
21!
!!
50!
!!
106!
Number!of!firms!that!moved!
5!
!!
10!
!!
28!
!!
61!
Relocation!rate!
50%!
!!
48%!
!!
56%!
!!
58%!
!!
!!
!!
!!
!!
!!
!!
!!
Notes:!
!!
!!
!!
!!
!!
!!
!!
1)!Average!firm!movements!are!based!only!on!those!firms!that!relocated!during!a!
stated!time!period.!
!!
!!
!!
!!
!!
!!
!!
2)!All!distances!measured!in!kilometers!(km)!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
33
Figure 1
34
Figure 2
35
Figure 3
36
Figure 4
37
Figure 5
38
Figure 6
39
Figure 7
40
Figure 8
41
Figure 9
42
Figure 10
