The Diffusion of Technological Innovation in the Commercial Banking Industry

Author(s): Johannes M. Pennings and Farid Harianto

Source: Strategic Management Journal, Vol. 13, No. 1, (Jan., 1992), pp. 29-46
Published by: John Wiley & Sons
Stable URL: http://www.jstor.org/stable/2486609
Accessed: 09/04/2008 09:07

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=jwiley.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org
 Strategic Management Journal, Vol. 13, 29-46 (1992)

THE DIFFUSIONOF TECHNOLOGICALINNOVATION

IN THE COMMERCIAL BANKINGINDUSTRY
JOHANNES M. PENNINGS
The WhartonSchool, University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.
FARIDHARIANTO
Institute of Management Education, Jakarta, Indonesia

This paper examines the propensity of organizations to adopt technological innovations.

Technological innovations evolve from the stock of skills which organizations have
accumulated over time. Linkages with extramural sources of technology are presumed to
be important as well. Hypotheses are tested on a sample of commercial banks. Findings
show that prior experience in information technology, in tandem with a variety of interfirm
linkages, will affect the banks' decision to adopt this innovation.

INTRODUCTION Prominent among these are other organizations-

whether competitors or firms in other industries-
This paper examines the propensity of organiza- whose know-how is germane to innovative efforts.
tions to adopt technological innovations. We start Technological spillover from competitors and
from the assumption that such innovations are a from extra-industry sources becomes accessible
direct outgrowth of skills and abilities which have through the presence of interfirm linkages. During
been accumulated in the past. Following Nelson a firm's life, networks of interorganizational
and Winter (1982), we can construe firms as a relationships are developed. Such relationships
repository of skills and abilities. This stock of broaden environmental exposure and reduce the
skills gives a firm strategic leverage to pursue distance which a firm has to bridge when it
certain innovations but also limits discretion as reaches out to external know-how to adopt
to the direction or content of its innovative technological innovations.
efforts. Technological innovations evolve from a These conditions-technological experiences
firm's past accomplishments and in turn furnish and linkages with other firms-are crucial for
a new assortment of skills. explaining innovation and form the cornerstone
Innovation represents the adoption of a new of the present study. We seek to explain why
idea, process, product or service, developed some firms innovate, and, if they do so, why
internally or acquired from the external environ- they innovate in a particular direction. In this
ment. The adoption of innovation follows from regard we will consider a firm's past and its
and is contingent upon an organization's reper- external environment. Elements of a dynamic
toire of technical, strategic and administrative theory of organizational innovation, involving
skills (Nelson and Winter, 1982). These skills past experiences and interfirm linkages, are
combine with extramural sources of know-how. formulated below. We subsequently set out to
study a sample of organizations over time to
examine how their innovative behavior was
affected by the past. These conditions include
Key words: technological innovations, firm skills, various technological skills which are germane to
interfirm linkages, learning, service the innovation under study, along with interfirm

0143-2095/92/010029-18$09.00 Received 14 February 1990

(? 1992 by John Wiley & Sons, Ltd. final revision received 10 July 1991
30 J. M. Pennings and F. Harianto

linkages that funnel extramuralknow-how into qualitative recombination of know-how residing

the firm. in their human and capital assets. Many of a
The empiricalstudy is based on the adoption firm's skills cannot be transferred because the
of video banking by U.S. commercial banks current stock of skills focuses the efforts of
duringthe period 1977-87. Video bankingis an people in organizations in specific directions. The
innovation. It is also a particular kind of improvement of existing deficiencies is inspired
innovation, permittingus to answer the earlier by or predicated on currently available know-
question of direction or content of innovation. how and is therefore consistent with the old
We believe that the banking industry provides notion that organizational search is 'problemistic'
an ideal setting for testing a history-basedtheory (Cyert and March, 1963). New solutions are
about organizationalinnovation.While its testing similar to or in the neighborhood of current
is sector specific, the results could, however, be solutions. On the voyage to new territories we
generalizedto cover other types of organizations are attracted to what looks familiar to or
and different classes of innovations. This study, consistent with what we already know.
we believe, will demonstrate the virtues of These comments are in accord with the
longitudinal research and provide important literature on first mover advantages. Early
advances in innovation research that go beyond adopters of new technology acquire non-trivial
the financialservices sector. advantages over laggards (Lieberman and
Montgomery, 1988). When prior technology is
an integral part of new technology, mimicking is
BACKGROUND difficult and first mover advantages of innovating
firms occur.
Elementsof a theory
We argue that where firms innovate, they
Two sets of innovation factors-external net- should consider extramural knowledge, particu-
working and accumulated know-how-are the larly when the firm is specialized and encounters
core ingredientsof our theoreticalframeworkon an intrusion of new technologies. Aggressive
organizationinnovation.This position represents networking can enhance the benefits of first
a continuation of the theme promulgated in movers and reduce the uncertainty which the
'classic'books, such as Schumpeter's,The Theory encroaching technologies create.
of Economic Development(1934), which argues Since firms usually face a distance between
that innovation is merely a recombination of proprietary and extraneous technologies, inno-
existing skills, resources and other assets. How- vation helps bridge this distance. The smaller
ever, innovation need not be restricted to the this distance, the greater the likelihood that firms
single firm but also occurs frequently under will adopt new technologies. Prior experience
interfirm umbrellas. For example, data trans- with new technologies shortens the distance, and
missionand storagehave become heavilyamalga- so does networking with extramural sources.
mated,renderingthe distinctionbetweentelecom- Networking attenuates the limits of home grown
munication(transmission)andcomputer(storage) skills which could thwart innovation. When a
technology fuzzy. firm has no, or only a few, relevant partners,
We believe that a firm will enjoy a major and if its proprietary skills are remote from those
innovation advantage if it can manage the flow required for an innovation, it is less likely to
of ideas that enter across its boundaries. This innovate. If a firm seeks to short-circuit the
advantageis major when the firm belongs to an distance between the present and future state,
industrythat undergoes an infiltrationof 'alien' we would expect a higher likelihood of innovation
technologies-for examplebiotechnologyinto the failure (Kazanjian and Drazin, 1987).
pharmaceutical,andsemiconductorsinto machine
tool industries. The merging of disparateskills,
Service vs. manufacturing
spanningdifferentindustries,is facilitatedby the
networking resulting from interorganizational Most studies of innovation have been based on
relationships. manufacturing organizations. Thus, a major
Consistentwith Schumpeter(1934) and Nelson research preoccupation has been with process vs.
and Winter (1982), we define innovation as a product innovation (compare Abernathy and
 Financial Services Innovation 31

Utterback, 1978). Much of our knowledge of Economist, 1989). Information technology is

organizational innovation is really manufacturing profoundlyalteringinternaloperationsand deliv-
and product innovation. ery of financialservices. Although the financial
This bias toward the manufacturing sector is services sector resemblesother service sectors in
remarkable in view of its declining role in Western being outside the mainstream of innovation
societies. Less than one in five individuals is research, one segment-banking-provides a
currently emnployed in marnufacturing, and this fruitfulsettingin whichto test hypothesesbecause
proportion continues to shrink. On the other of the technologicaltransformationswhich have
hand, the service sector is growing phenomenally. occurred.
Its growth is partly dependent on the infusion of Banking parallels other service providingsec-
information technology, which has enabled many tors such as telecommunicationsand transpor-
of these firms to deal with explosive information tation. In all these industries,firmshave become
processing requirements. Many service sectors more proximatebecause service deliverynecessi-
have given us laboratories for examining the tates interorganizationalcoordination.Proximity,
emergence of organizational networks spanning whether due to service complementarity or
multiple industries. Innovation rates in the service geographicvicinity, is an importantinducement
sector are astounding (The Economist, 1989). for innovation(e.g. Enright, 1989, Abrahamson
It is beginning to appear that the infusion of and Rosenkopf, 1990). Proximityenhances com-
information technology takes a different form munication, diminishes thresholds for resource
and follows a different path from innovation in sharing and accelerates the diffusion process.
the manufacturing sector. However, until 1980, Compared with manufacturing, these sectors
information technology was traditionally alien to exhibit diminished technological separability
such service sectors as transportation, travel, (Williamson,1975).
banking and insurance, health care, and telecom- Commercial banks have established check
munications. In many of these sectors, there is clearinghouses, have engaged in correspondence
a strong need for interfirm coordination to banking, and have also created alliances to
manage interdependencies which arise during the develop and share new technologies. These
production and delivery of services. It should linkages have been supplemented by other
therefore not surprise us that we are witnessing connections with other service providers; most
a proliferation of interorganizational structures significantly computer and telecommunications
such as airline reservation systems, credit card firms. Financial transactions require banks to
consortia and health maintenance organizations. pool their operations, and thus high levels of
While interorganizational linkages also occur in networkingwithinandacrosssectorshave become
the manufacturing sector they tend to have a standard.
short longevity and appear to be largely faddish This networkingis most visibly illustratedby
(Kogut, 1988). automatedtellermachines(ATM). ATM involves
both telecommunications and computer tech-
niology,and could not have been adoptedwithout
TRENDS IN BANKING the benefit of commensurate learning curves.
For example, ATM requires data storage and
The financial services sector is an interesting case processingand is tied to other components in a
for innovation researchers. Commercial banks bank's information system. Even more pro-
have been very prolific in the establishment of nounced are its interbank linkages; interfirm
joint ventures, licensing arrangements, and other networking is particularly noticeable around
types of interfirm partnerships. They have been ATM.
major players in the merger and acquisition wave In addition to their proclivity to form webs
of the eighties (Harianto and Pennings, 1990) of interfirmnetworks, commercial banks have
and have also been important users of R&D become major users of informationtechnology,
generated in other sectors. especiallycomputersand telecommunications.In
Despite its popular image as a staid, heavily termsof dollaroutput, financialservicesabsorbed
regulated industry, commercial banks have under- almost 66 percent of the total revenue of the
gone significant transformations since 1975 (The information industry in the period 1970-80
32 J. M. Pennings and F. Harianto

(Compaine, 1984). Indeed, this sector is under communication networks for the delivery of such
going a microelectronic influx and is by far information and services). Video banking requires
the most important user of U.S. information a great deal of interfirm coordination. Any
technology. Harianto and Pennings (1990), who organization considering the adoption of video
rely on Scherer's (1982) estimates, observe that banking needs significant skills in information
as far back as 1974, the financial service sector technology and stands to benefit from connections
absorbed most of the R&D outlays of the with firms that furnish pertinent hardware and
computer industry, rendering them highly sym- software.
biotic. They also reveal that banks have an By focusing upon a single innovation, we can
unabating appetite for new technology, which in compare firms which embarked on the innovation
fact has accelerated after the publication date of with those that did not. Longitudinally, we relate
the Scherer study (1982). video banking to interorganizational activities
The capacity of commercial banks to innovate and specific technology skills. Innovation is
around information technology depends very still a poorly understood phenomenon, and its
much on whether they themselves have partici- antecedents appear only accessible if the research
pated in the development of cumulated computer design allows a high degree of control of the
skills, endowing them with what Cohen and history leading up to the innovation event.
Levinthal (1989) call 'absorptive capacity.' This
capacity diminishes the threshold for extracting
know-how from others. Through interfirm chan- HYPOTHESES
nels, banks will have easier access to different
but complementary technology. Such channels We agree with Mohr (1982) that it is more
point to the importance of external conditions productive to create middle-range theories that
that are conducive to innovation. Some types of fit specific innovations, whose specific antecedents
networks and the skills employed in forming can be identified and measured. Therefore, it is
them indicate a proclivity towards other types more crucial to identify the innovation than to
of interfirm relationships-for example, with develop general theories which account for
computer firms. And such exposure to extramural any kind of innovation. Such a perspective is
technology, together with spillovers from compet- analogous to epidemiology and biology in which
ing banks, enhances their current capacity to the diffusion of distinct diseases or organisms
adopt new technology. can be traced to particular antecedents which
occur in a specific sequence. For any innovation,
we need to be aware of the cumulative experi-
Combination of new technology and networking:
ences, the ingredients of which become re-
Video banking
fashioned into the new product or service. The
The empirical research in this study addresses implication is that we should measure those
the introduction of a specific new form of experiences which are germane to video banking.
information technology: video banking services. For banks, the relevant technological skills
This innovation represents another illustration of required to introduce video banking/videotex
the spreading of new technology into the service services include back office automation and
sector. transaction oriented technology. We hypothesize,
In general, videotex refers to computer-based therefore, as follows:
interactive systems that electronically deliver
screen text, numbers and graphics. It also permits Hypothesis 1: that the more firms have
interactions: for example purchasing an airline accumulated experiences in computer and tele-
ticket, transferring funds between bank accounts, communication areas, the higher the likelihood
or trading for one's own securities account. that they will embark on the video banking
Videotex brings together firms from the infor- services.
mation delivery, financial and merchandizing
sectors on one side (as information and service Amassed experiences can also be represented
providers) and computer and telecommunications by the concept of 'learning by doing' (Arrow,
firms on the other (as providers of systems and 1962), which results in improved efficiency
 Financial Services Innovation 33

because the firm has increasingly fine-tuned This interdependence is manifest in the volume
its routines. Conventionally, this approach is of technology transactions and the proliferation
distinguished from 'learning by using,' where of interfirm networking. Other sectors with
learning refers to enhanced capacity to do other noticeable (albeit less documented) networking
things than what a firm has already accomplished vis-a-vis the financial sector are retailing and
(compare Rosenberg, 1982). Increased invest- publishing. The proliferation of point-of-sales
ment in capital goods reflects learning by doing networks requires some sort of arrangement with
and permits a firm to produce its output more retailers at the local, regional or national level.
efficiently. High levels of capital investment Similarly, providing basic financial services (e.g.
should also signal a readiness to expand invest- stock brokerage) may involve the use of data bases
ment programs into new generations of equipment offered by publishing houses. These linkages
to further secure efficiency benefits. Such invest- are especially relevant for the creation of an
ment also promotes a firm's 'absorptive capacity' integrated video banking service, in which cus-
(Cohen and Levinthal, 1989). tomers are able to execute financial transactions
Lieberman (1984) has likewise argued that (e.g. transfers between accounts, bill payments,
capital investments can be construed as an securities trading) as well as teleshopping and
absorption of new technology, with the caveat other forms of on-line transactions. In short,
that the more specific the type of capital linkages with new technology suppliers, competi-
investments, the more valid is the inference about tors and transactional service firms are conducive
a firm's application of information technology. to innovation. We hypothesize that:
A firm's capital investment history and its accrued
performance reflect accumulated experience, Hypothesis 3: the more that banks have
which in turn reflects a propensity to allocate developed interfirm linkages with firms from
capital expenditures to new technologies such as computer, telecommunications, stock broker-
video banking. This argument is analogous to age, insurance and other transactional pro-
that of Amit, Livnat and Zarowin (1989), who viders, the more likely they will venture into
claim that a firm's capital expenditures pattern the video banking services.
will induce it to finance internal diversification
through similar capital expenditures programs. In the research reported, we tested these
The conduct of these firms is quite different from hypotheses while holding certain company and
firms that resort to acquisitive diversification. industry attributes constant.
We contend that firms with comparatively high
levels of capital expenditures will display a higher
propensity to invest in videotex. Therefore: RESEARCH DESIGN
Sample and data collection
Hypothesis 2: the higher capital investments
in systems and equipment, or their derived Our research is based on a sample of 152 of the
productivity index, the more likely that firm largest 300 banks in the United States, covering
will engage in video banking services. an 11-year period, 1977-87. The original listing
was obtained from the American Banker (1987).
We mentioned earlier that firms operating in The 152 banks were 'at risk' in adopting video
a web of interindustry linkages enjoy access to banking. The other banks had to be deleted
extramural technologies. We distinguish two because of insufficient data, because they were
types of linkages. The first class contains various private or because they had been taken over by
information technology firms. Their technologies foreign banks. Of the total American Banker
represent a powerful motive for joint ventures listing, 53 actually did adopt the new technology.
and affect the choice of strategic partners. Other Forty-nine of the subset of 152 are known to
linkages are confined to the banking sector, and have introduced video banking services in the
similarly enhance spillover of ideas. period 1980-87. They are slightly bigger than the
We have indicated that there is a mutual excluded banks. Typically, the included banks
dependence between the financial sector and the are also firms whose records are better accessible
computer and telecommunications industries. and which still existed in 1988.
34 J. M. Pennings and F. Harianto

Financial data of the firms were collected from in the cohort design, the research proceeds from
their annual reports and from Moody's Bank & cause to effect. A simple random sample, or a
Finance Manual (published annually), while the stratified sample of units, is selected and classified
remaining missing data were furnished by banks. according to its exposure to the hypothesized
Banks that had introduced video banking are independent variables. The dependent variables
identified from the videotex directory (Arlen are measured and the exposed vs. unexposed
Communication, Inc., 1985), the 1985/1986 Direc- units are compared on the basis of this variable
tory of Electronic Funds Transfer, and the (Schlesselman, 1982).
American Banker's surveys (1985-87). The case-control design is highly appropriate
We collected data about innovation activities for innovation adoption studies, since adoptions
and interfirm linkages from secondary, public in a particular year are comparatively rare. In a
sources, as they are readily available. The criteria cohort study, a stratified sample of banks
for selecting the information were: (1) the would be selected with the strata based on the
information must contain multiple years of hypothesized independent variables. Afterwards,
observation and (2) it must contain data from data would be collected on whether or not
sources relevant for banks (e.g. American Bankers, videotex was adopted. Obviously, such a design
Wall StreetJournal), computers and telecommuni- would require a large data base to possess enough
cations (e.g. Computer World). The only data adoption events. The case-control design requires
base that met these criteria was the Predicast fewer firms. There is no sampling procedure
Index on U.S. Corporations. We found, however, with the intent to ascertain precise population
that the Index was biased toward big firms. The estimates. In the present study, banks are selected
correlation between company size (log asset) and on the basis of whether or not they adopted
the number of news entries reported in the index video banking, together with the collecting of
is 0.74. We were, therefore, careful to control relevant independent variables. Adopters are
the possible confounding effect of size. For a then compared with non-adopters.
review on the benefits and disadvantages of using Most of our explanatory variables are time-
the Predicast Index in general, see Hladik (1985). dependent. The use of discrete-time event history
analysis is therefore preferable (see Allison, 1984,
for a fuller discussion on the conditions required
Model specification
to employ the method). The model, which uses
The study seeks to identify the role of firms' the logistic regression method, is specified as:
past experience in technological endeavors and
interfirm linkages in predicting their propensity
Log P/(1-P) = B() + Bi(t),
to implement a certain innovation, namely,
video-banking services. This study centers on a
dichotomous dependent variable: whether or not where P is the probability that a firm has the
a firm employs the video-banking services. To event, B( is the intercept, and Bi(t)'s are the
test the hypotheses, we can construe the problem parameter estimates of the covariates (i.e. the
as predicting a single-event history. time-varying independent variables). Estimates
In the present design, adopting banks are of parameter B are obtained from the maximum
compared with non-adopting banks on the basis likelihood method. The approach of discrete-
of a number of attributes; i.e. the study has time method is to pool and treat the longitudinal
many of the features of a case-control design. data as cross sectional (Allison, 1984). A bank
Such designs are common in research on lung will contribute firm years in proportion to the
cancer, toxic shock syndrome and epidemiological timing of adoption; if the adoption takes place
investigations on rare outcomes. in year one, one firm year is included in the
Unlike the cohort design, the case control analysis, and at year two, two firm years, and
study traces effects to antecedent conditions so on. Non-adopting firms contribute 7 years-
(Anderson, et al., 1980; Schlesselman, 1982). the length of the window. To identify the effect
Units with the variable of interest and those of calendar time (year), a set of (n-i) dummy
without are compared with respect to the variables are entered into the model, where n
hypothesized explanatory variables. In contrast, reflects the number of years of observations.
 Financial Services Innovation 35

banking system), marketing agreements (e.g.

Measurement
BancOne agreed to market Manufacturer Han-
The main independent variables are the accumu- over's cash management), R&D agreements (e.g.
lation of experience in information technology Citicorp and its partners contracted David Sarnoff
and in interfirm linkages. Each experience can Laboratories of RCA to develop a video banking
be counted, and cumulatively they amount to the software), and supply agreements (e.g. a bank
magnitude of innovation-relevant know-how. contracts Dow Jones to supply financial data).
They can also be decomposed into distinctive These interfirm linkages are coded further by the
sets of experiences; for example, hardware, nature of the partner's business (e.g. computer,
software, telecommunications, joint venture, or telecommunications, banks, other fi-
M&A. Information technology in the banking nancial services, publishing, retailers, or other
sector consists of back-office automation and transactional providers), and whether the arrange-
transactional-oriented technology. The first ment involves various kinds of 'transfers' of
includes computer technology (e.g. installation technology.
of on-line terminals for tellers and platform Parenthetically, it should be noted that the
personnel, departmental computing, branch sys- inclusion of a variable measuring interfirm
tem integration and software upgrades) and linkages among the banks themselves serves to
telecommunication technology (e.g. the use of highlight a unique feature of the banking industry
satellite and fibre optic networks). Transactional itself. Some firms are active in financial net-
technology is categorized further into ATM working, others with firms providing dissimilar
networks (e.g. introduction of proprietary ATM financial services such as insurance and broker-
networks, ATM for the blind), point of sales age. Inclusion of such variables is highly appropri-
systems (e.g. purchase through debit card in ate in industries whose service delivery entails a
supermarkets, gas stations and national retailer great deal of interfirm coordination.
networks), home banking services (e.g. bank by In our coding, a bank was assigned one point
phone for bill payments and transfer between for each of the above categories whenever it
accounts) and corporate electronic banking (e.g. generated an event reported in the Predicast
electronic cash management and clearing houses). Index. The score is the cumulative number of
Additionally, the firm's experience in infor- events for each category recorded since 1977,
mation technology is measured by two indicators: our starting year of observation. The final score
(1) its investment expenditure in systems and for each bank is the cumulative number of events
equipment (measured as the percentage of capital up to the point of adoption. Events were classified
spending in equipment to asset) and (2) by a as technological or interfirm, where these catego-
productivity efficiency index (measured as the ries were broken down further in certain subcate-
ratio of interest revenue to non-interest expenses, gories. For example, technological events could
consisting of equipment and personnel expenses). further be classified as ATMs, point of sales,
Naturally, this index is also sensitive to other credit card, debit card, or bank by phone.
factors, particularly those residing in the environ- Networking events were further partitioned into
ment, but is often used in the banking community internal development, licensing, joint ventures
to gauge the productivity of equipment outlays. or mergers and acquisitions. All events, which
While the Predicast based measure can be are pertinent to a bank's history were obtained
interpreted as 'learning by doing,' these two from the Predicast Index (1977-88).
financially derived measures of capital invest- The Predicast volumes provide the titles of
ments approximate what might be called 'learning articles which have appeared in a variety of
by using.' Both types of learning are assumed to media. The items therein were treated as
foster strategic readiness to innovate in certain information on a particular bank's relevant
technological directions. experiences or activities, and could be broken
Interfirmlinkages include joint ventures, licens- down in terms of different types of technology
ing or contracting the rights to use, produce or (e.g. credit card, point of sales, ATM, corporate
market certain products, systems or services and consumer video banking), type of exchange
without transferring the ownership rights (e.g. relationship (e.g. joint venture, licensing agree-
several banks licensed Chemical's Pronto video- ment or acquisition vs. internal development),
36 J. M. Pennings and F. Harianto

and type of partners (hardware, software, tele- firms-as stipulated in Hypothesis 3-can thus be
communications, banks, and other financial insti- examined. Particularly important is the distinction
tutions such as brokerage houses and insurance between linkages that foster acquisition of extra-
firms, merchandizing, and newspaper organi- mural skills vs. spillovers that originate from
zations). The source was also used to count the competing financial institutions.'
frequency of administrative innovations. These The dependent variable is the adoption event.
included reorganizations of organization design, The firms received a score of 1 if they adopted
compensation systems, sales networks, coordi- video banking during the period 1981-87, and 0
nation of services, and personnel training pro- otherwise. The very first adoption took place in
grams. These represent elements of the firm's 1981, rendering the data set left censored at this
history, which encompasses a string of experiences year. The data are right censored up to the point
with projects, ventures, interfirm networks, and of the study, 1988.
so on, leading up to other related innovations. The list of video banking adoptions was
Examples of Predicastentries are: obtained from the Arlen Communication Inc.
(1985) Videotex Directory, the 1985/1986 Retail
'Reorganization to form community banking Electronic Fund Transfer Directory, and the
and national consumer sectors' (administrative American Banker surveys (1985-87). This collec-
innovation); 'To use artificial intelligence for tion was a trial and error process, working
foreign exchange trading' (technological inno-
vation-software);'To jointly form Fleet/Norstar backwards and forwards, checking entries in
Financial with Fleet Financial for $1.3bln' different publications against one another. Several
(interorganizational-banks);'BancOne jointly listings of commercial banks, together with listings
develops with E.D.S. video software' (inter- of video banking adopters, led to the creation of
organizational-telecommunications). a file of banks, including most (49 out of 53)
that adopted the innovation.
These entries represented a direct 'translation' While testing the hypotheses on innovation
of experiences, and were counted to arrive at diffusion, we control for a number of firm and
cumulative scores. No attempt was made to environment attributes. Firm variables include
weigh them, nor did we attempt to break them size, return on assets. Industry attributes include
down further. As reported elsewhere (Pennings demand, number of competitors and degree
and Harianto, 1991), triangulation procedures of 'shake-out.' Finally, the study controls for
showed that the Predicast-basedmeasures have 'administrative' or non-technical innovations.
reasonable degrees of validity and reliability. These innovations are construed as mirroring a
A coding scheme was developed to classify the firm's general proclivity to innovate and should
experiences in the pertinent categories; it reflects be held constant when assessing the effects of a
the previously mentioned requirement of deduc- specific set of experiences upon subsequent
ing the content of the blended technologies innovations (Nord and Tucker, 1987). Table I
and interfirm networking. Within the class of provides a condensed description of all the
technological experiences, the coding is mutually independent variables in the study.
exclusive: entry into one category precludes entry Equipped with 11 years of data on the banks'
into others. In the case of interorganizational technological experiences and successive webs of
linkages, however, an occurrence of technical interfirm relationships, we are now in a position
projects employing interfirm arrangements (e.g. to present some findings on the incidence of
'BancOne jointly develops with E.D.S. video video banking. Since we can decompose these
software') should be coded as both a technological covariates into more specific categories, we can
experience and as an interorganizational linkage. also indicate what types of experiences and which
All interorganizational linkages up to the year classes of interfirm linkages are most conducive
of adoption or the year of right censoring (i.e. for its adoption.
1988) yielded a networking score, which can be
decomposed into sepcific types of linkages. These
cumulative linkages can be examined as a time
dependent covariate of innovation. The predictive I A copy of the coding scheme is available on request from
power of interfirm linkages with various types of the authors.
 Financial Services Innovation 37

Table 1. Summary independent variables, their measurement and sources

Variable Measurement

Technological experience:
1. Cumulative experience, Cumulative number of events involving all types of
Information technology hardware, software or telecommunications, 1977-to-
date, (Predicast)
2. Cumulative experience, back-office technology Cumulative number of automation projects in data
processing, 1977-to-date (Predicast)
3. Cumulative experience, systems and equipments Cumulative number of events involving ATM, credit,
debit cards, POS, and electronic fund transfer, 1977-
to-date (Predicast)
4. Productivity index of systems and equipment Net income/(equipment, systems & personnel
expenses); 3-year average (annual reports)
5. Investments systems and equipment Total $ investment in systems and equipment/company
asset; 3-year average (annual reports)

Networking:
6. Cumulative linkages, interorganizational Cumulative number of interfirm arrangements includ-
ing joint ventures, licensing agreements, turn-key
contracts and mergers and acquisition, 1977-to-date
(Predicast)
7. Proportion of linkages for technological purposes Cumulative number of links for technological inno-
vations/total number of cumulative links, 1977-to-date
(Predicast)
8. Cumulative linkages with computer and telecom- Cumulative number of interfirm arrangements with
munictions firms information technology firms, 1977-to-date (Predicast)
9. Cumulative linkages with other financial insti- Cumulative number of interfirm arrangements involv-
tutions ing banks and other financial services firms, 1977-to-
date (Predicast)
10. Cumulative linkages with complementary service Cumulative number interfirm arrangements, with
providers retailers, publishers and data base providers, 1977-to-
date (Predicast)

Control variables-company attributes

11. Cumulative experiences, administrative innovation Cumulative number of sales, personnel, and general
administrative innovations, 1977-to-date (Predicast)
12. Company size Log $ asset (annual reports, Moody's)
13. Return on equity Net income/equity (3-year average) (annual reports,
or Moody's)
14. Cash flow (Net income + depreciation + provision for loan
losses)/Asset (annual reports, or Moody's)
15. Consumer business $ Consumer loans/$ total loans (annual reports, or
Moody's)

Control variables-industry attributes

16. Size of consumer demand Log ($ consumer loans in the state) (FDIC)
17. Growth of consumer demand ($ Consumer loans in the state, this year)/
($ Consumer loans, last year) (FDIC)
18. Number of banks in the state Log (number of banks) (FDIC)
19. Mortality index (Number of banks, this year-number of banks, last
year)/(number of banks, this year) (EDIC)
38 J. M. Pennings and F. Harianto
Table 2. Distribution of banks venturing into video banking services

Total 81 82 83 84 85 86 87

Internal developments 2 3 1 4 4 3 2 19
Licensing/other agreements 1 2 2 11 1 6 1 24
Joint ventures 0 0 1 1 4 0 0 6
Total 3 5 4 16 9 9 3 49

RESULTS dependent. The chi-square statistics which are

associated with models 5.1 and 5.2 permit the test
Tables 2-6 report the results of this research. Table of the null hypothesis, which states that the hazard
2 shows the annual distribution of the numbers of rate does not vary over time, net of other variables.
banks introducing video banking services. Out of The incremental chi-square is 13.54; since its
49 ventures, 19 were internal development ventures, degrees of freedom are six (equivalent to the
24 involved licensing or other contractual agree- number of dummy variables entered into model
ments, and 6 were accomplished through joint 5.2), it is significant at the 5 percent level. We
venturing. One bank got into the service because need to examine the emergence of each and every
it acquired a bank that had already introduced the innovation, recognizing that they are context
service. There were 22 ventures with limited specific. If and when such research accumulates,
video banking services, compared to 27 involving we may then broaden technological innovation
integrated videotex services. The distribution of diffusion models. Since interfirm linkages and
events over time has an inverted-U shape. In the technological experiences are highly intercorrelated
time window 1980-88, there was a rapid growth (r= 0.76), they are entered alternately in the
in the hazard rate, which can be defined as the model, to become Models 5.3 and 5.4.
ratio of the number of 'events'-banks adopting The results confirm the first and third hypoth-
videotex-to the number of uncensored obser- eses that cumulative experiences in technology
vations. This rate was reaching its peak in 1985, and in interfirm linkages are significantly con-
after which it declined toward zero in 1988. ducive to the introcluction of video banking
The years 1984-85 evoke the impression of a services. In contrast, two variables, the pro-
bandwagon, in which numerous banks are mimick- ductivity index and investment in systems and
ing their trend setters; video banking is another equipment, have no significant effect (Hypothesis
expression of 'institutionalisomorphism'(DiMaggio 2). Technological experiences and interfirm link-
and Powell, 1983). ages 'compete' in explaining the probability of
Descriptive statistics about the video banking the event. Exclusion of either variable renders
ventures and the independent variables are shown the other highly significant, as Models 5.3
in Table 3. From Table 4, it appears that five and 5.4 indicate. Of course, this effect of
independent variables are highly intercorrelated. multicollinearity would not have been discernable
Since a large number of observations (pooled if we had only supplied Model 5.2. If the variables
data) are involved, a high level of significance are entered jointly, the interfirm linkages variable
obtains at a relatively low correlation coefficient. is the most prominent and highly significant
Tables 5 and 6 report the results of the logistic (Model 5.2). In view of the fact that many
regressions, taking care of the multicollinearity technological activities entail some form of
problems. The two tables are very significant in networking, it should not be surprising that
terms of their goodness of fit to the overall technologically active firms also figure promi-
models, as indicated by their likelihood ratio or nently in interorganizational strategies.2
chi-square statistic.
Table 5 provides four different models. Models 2
A collinearity diagnosis, yielding 'variance-decomposition
5.1 and 5.2 are equivalent, except that in 5.1 the proportions' (Beisley, Kuh and Welsch, 1980), revealed
dummy variables corresponding to the 7-year time that technological experience was the only variable whose
'tolerance' remained below the 0.2 level (where this level is
frame were excluded. A comparison of the two deemed to be minimally acceptable). The variable had a
columns suggests that the innovation rate is time tolerance level of 0.17, and a variance inflation of 5.71.
 Financial Services Innovation 39

Table 3. Means and standarddeviation of independentvariables

Variable Mean S.D.

Cumulativeexperience, informationtechnology 0.688 1.643

Cumulativeexp., back-officetechnology 0.222 0.792
Cumulativeexp., transactiontechnology 0.354 0.885
Productivityindex, systems equipments 3.446 1.040
Investmentsystems & equipments 0.344 0.192
Cumulativelinks, interorganizational 0.447 1.017
Cumulativelinkages computer& telecomm. 0.037 0.223
Cumulativelinkages, retailers& publishers 0.028 0.185
Cumulativelinkages, fin. institutions 0.359 0.819
Proportionof interorganizationallinkagesfor technologicalinnovations 0.013 0.095
Cumulativeexperience, administrativeinnovations 0.968 1.807
Log assets in ($ millions) 3.648 0.438
Return on equity, 3-year average (%) 12.645 4.842
Cash flow (%) 1.405 0.732
Consumerbusiness (%) 0.218 0.106
Demand of size, consumerbanking($ Billions) 3.691 0.443
Demand growth, 3-year average (%) 9.797 4.682
Numberof banks in the state (Log) 2.333 0.467
Mortalityindex (%) -1.035 5.610

Table 6 presents the results for a more The significance of company size is consistent
disaggregate level of interorganizational experi- with Bantel and Jackson's (1989) study. It is the
ences. These results show which type of interfirm only company attribute which predicts the odds
experiences are most conducive to innovation. of entering the videotex industry. The industry
Various types of technological experiences are attributes have no significant effect, although one
also included in the model. Experiences have variable, size of consumer demand, comes close
been partitioned into specific categories. It is not to being significant.
merely the amount, but the nature of the In Tables 5 and 6, the dummy variables
experiences, which is crucial. The component corresponding to the years 1984 and 1985
technological experiences are comparatively insig- (particularly 1984) appear significant. The effect
nificant. In contrast, the results provide detailed of the year 1984 can be attributed to the Federal
information on the third hypothesis, particularly Reserve Board's decision allowing commercial
that experiences with computer firms (hardware banks (in this case, Citicorp) to engage in full-
and software), are most conducive to the introduc- fledged data processing services. During the
tion of video banking services. Linkages with period 1984-85, there were 15 banks that licensed
telecommunication firms, or with retailers, pub- the Automatic Data Processing's 'Home Banking
lishing houses and the like, have no significant Interchange' systems.
effect. Linkages with other commercial banks are
significant-a striking result. It should be noted
that some of these strategic alliances were CONCLUSION AND DISCUSSION
founded for technological reasons (e.g. licensing
Chemical's Pronto systems, the Video Financial This paper has shown that accumulation of
Service joint venture). In short, among these information technology experiences and interfirm
disaggregate effects, those involving interfirm networking is conducive to innovation. While
conduct appear to be most conducive to techno- these two sets of antecedents are correlated, each
logical innovation, while interorganizational is significantly related to the propensity to
relationships contribute to the diffusion of inno- innovate. There is some overlap between infor-
vation. mation technology experiences and interfirm
40 J. M. Pennings and F. Harianto

*
18.17.16.15.14.13.12.11.10.9. 8. 7. 6. 5. 4. 3. 2. 1.
% Table
Decimal CashSize Link Link
Link Tech.
4.
Variable
Growth Return
Number Interorg. Admin.
Investm.
Mortality flow
points of Consumer on retail
comp. Back-office
TransactionProductivity
Consumerfinancial
&& syst
indexdemand &
banks equity tech.
tech.
linkages experience Correlation
omitted; demand
loans publ. eq index
firms telec. innovations
a

01120414-42-13033102-01090603130906-29100 1

coefficient
03-0616-01200614-14-0903-07-09-05-09-07-09100 2 coefficients*
of
0509-0631-17-01-125247204755555664100
beyond
3 the
0.09 4
is 0901-0422-14-01-0553622566768887100

0900-0523-14-04-02454219675755100 5
independent
significant 6
at 0706-0319-1203-084766235076100
the 7 variables
0209-0425-1302-1048933550100
0.01
0403-0817-1100-01322814100 8
level.
-0402-02040302011517100 9

0209-0325-1300-0844100 10

0411-0136-36-1001100 11

06-1124-031125100 12

0406070310100 13

-08-3515-25100 14

0253-07100 15

19-27100 16

-00100 17

100 18
 Financial Services Innovation 41

Table 5. Logistic regressionresults dependentvariable:Enteringvideo banking

Variable Model 5.1 Model 5.2 Model 5.3 Model 5.4

Intercept -11.392*** - 12.316*** -12.752*** -12.553***

(2.840) (3.068) (3.041) (3.046)
Cum. experience, inf. technology 0.200 0.206* 0.288***
(0.128) (0.128) (0.114)
Cum. interorganizationallinkages 0.279** 0.222 0.354***
(0.140) (0.158) (0.132)
Proportionof interfirmlinkages for 1.5604** 1.897** 2.153*** 1.821***
technological innovation pro- (0.888) (0.944) (0.904) (0.949)
ductivityindex, system & equip. -0.214 -0.226 -0.232 0.221
(0.200) (0.254) (0.239) (0.244)
Investments,system & equipment -0.233 -0.720 -0.486 -0.959
(1.093) (1.068) (1.092)
Cum. experience, administrative -0.185* -0.166* -0.085 -0.158
innov. (0.097) (0.101) (0.086) (0.100)
Companyattributes
Size 1.269** 1.230* 1.403*** 1.357***
(0.560) (0.584) (0.572) (0.578)
Retum on equity -0.003 -0.002 -0.002 -0.006
(0.040) (0.040) (0.039) (0.039)
Cash flow 0.045 0.048 0.049 0.077
(0.289) (0.378) (0.388) (0.340)
Consumerbusiness 1.742 1.484 1.508 1.579
(1.867) (1.928) (1.901) (1.932)
Market attributes
Size of consumerdemand (Log) 0.823* 0.872* 0.923* 0.813
(0.509) (0.537) (0.528) (0.534)
Growth of consumerdemand 0.021 0.034 0.031 0.036
(0.037) (0.032) (0.031) (0.033)
Number of banks in state 0.196 0.165 0.005 0.222
(0.474) (0.481) (0.466) (0.480)
Mortalityindex 0.016 -0.010 -0.007 -0.008
(0.030) (0.030) (0.010) (0.030)
Year 81 (dummy) 0.218 0.454 0.076
(1.049) (1.002) (1.038)
Year 82 (dummy) 0.743 0.868 0.681
(0.893) (0.877) (0.884)
Year 83 (dummy) 0.856 0.906 0.802
(0.821) (0.816) (0.817)
Year 84 (dummy) 1.913** 1.965*** 1.920***
(0.732) (0.727) (0.728)
Year 85 (dummy) 1.340* 1.419* 1.335*
(0.731) (0.725) (0.732)
Year 86 (dummy) 1.148 1.165 1.157
(0.729) (0.727) (0.729)
-2 Log likelihood 318.21 304.68 307.56 306.63
Chi-squared(D.F. = 19) 58.54*** 72.07*** 69.19*** 70.12***
D.F. 14.00 20.00 19.00 19.00
Number of firm years 867.00 867.00 867.00 867.00
Number of adoptingbanks 49.00 49.00 49.00 49.00
* p < 0.10, **p < 0.05, ***p < 0.01. Standarderrorsare in parentheses.
42 J. M. Pennings and F. Harianto

Table 6. Logistic regression of videotex adoption with technological interorganizational experiences broken
down into specific categories

Variable

Intercept -12.030***
(3.135)
Productivity index -0.284
(0.255)
Prior investment, systems & equipment (3-yr avg.) -0.342
(1.076)
Cum. experience, administrative innovation -0.152
(0.111)
Cum. experience, office technology 0.362
(0.236)
Cum. experience, transaction technology -0.409*
(0.225)
Cum. linkages with telecomm. firms 0.773
(1.045)
Cum. linkages with computer hardware firms 2.755***
(1.092)
Cum. linkages with software houses 1.526**
(0.694)
Cum. linkages with transact. providers 0.791
(0.547)
Cum. linkages with other banks and financial services firms 0.420**
(0.198)
Year 81 (dummy) 0.026
(1.087)
Year 82 (dummy) 0.579
(0.908)
Year 83 (dummy) 0.691
(0.829)
Year 84 (dummy) 1.773**
(0.740)
Year 85 (dummy) 1.256*
(0.740)
Year 86 (dummy) 1.273*
(0.737)
Control variables:
Log asset 1.475***
(0.607)
Return on equity -0.007
(0.042)
Cash flow 0.124
(0.284)
Consumer loan 1.332
(1.919)
Industry attributes:
Size of demand, consumer loan (log) 0.658
(0.553)
Annual growth of demand 0.031
(0.031)
Number of banks in the state 0.124
(0.447)
Mortality index 0.000
(0.031)
-2 log likelihood 297.97
Chi-squared (D.F. = 19) 78.79***
D.F. 24.00
Number of firm years 867.00
Number of adopting banks 49.00
* p < 0.10, **p < 0.05, ***p < 0.01. Standarderrorsare in parentheses.
 Financial Services Innovation 43

linkages because some linkages are formed our position that interfirmlinkingis an important
for the purpose of implementing technological precursorto innovation.
investments. The greater a bank's inclination to It is not clear why Hypothesis 2 failed to
network with information technology firms, the receive support. As was indicated, the two
greater the probability it will adopt video pertinent variables convey a sense of 'learning
banking. When we partition linkages into specific by doing,' consistent with the argument of
categories, as was done in Table 6, the results Lieberman (1984). The productivityindex also
suggest spillovers from competing organizations, incorporatesinformationnot germane to infor-
as well as procurement of computer hardware mation technology. Similarly,investment in sys-
and software technology. These are among the tems and equipmentincluded not only computer
most important linkages in accounting for the or telecommunicationstechnology,but also other
incidence of video banking. investments-for example, typewritersand office
Therefore, both internal and extramural skills furniture. Currently available data sources do
are crucial for innovation. A bank's distinctive not allow the partitioning that better fits the
competence derives not only from internal needs of this research. We stress, however, that
experiences; home grown skills improve its in the current study, a learning indicator has
capacity to absorb external skills, which are been used which was far more specific than the
presumably being tapped when the bank is one used by Lieberman(1984).
heavily involved in networking with information The presentstudyis restrictedin that consumer
technology firms. Accumulating technological attributes were not included in the model.
skills and external networks are crucial events in The size of demand approximatedconventional
a bank's history, if we make an attempt to significance levels, suggesting the presence of
explain their current innovative conduct. Viewing 'marketpull.' Several other, more specific attri-
innovation as following from a chain of events butes were included, but they might not fully
enriches our understanding about the process of capture consumers' attitudes toward electronic
innovation itself and hence reveals the merit of paymentsystemsin general. Other studiesshould
event-history analysis. develop measures which are better attuned to
The results support the core elements of the consumerattitudes and opinions.
theory as specified in Hypotheses 1 and 3. The
cumulative experience in video banking-relevant
Issues of sector and innovationspecificfindings
technologies diminishes the distance between the
skills already residing in the firm and extramural The role of conventional market structurevari-
skills which are combined so as to enable the ables did not surface as prominentlyas it might
implementation of this innovation. Of course, have. Obviously, this study should be replicated
this thinking is highly consistent with the Schum- on other innovations,involvingfirmsin different
peterean notion that innovation essentially rep- industries. Ideally, such industriesshould differ
resents the recombination of existing know-how. on boundarypermeabilityand degree of interfirm
In the present case, this know-how involves partnering,or with respect to some modal forms
multiple industries, necessitating the extension of of strategic alliances such as joint ventures,
a firm's boundaries into other industries. licensing agreements, turnkey contracts and
The single, most striking finding of this study minority interests, and the types of innovations
was the strong effect of the proportion of around which they evolve. Such a comparison
technological linkages. The more a firm commits would permit a detailed multilevelstudy, tracing
itself to technological networking, the greater the diffusion of innovation to technological
its propensity to innovate. Networking and and interfirmexperiences and market structure
technological experiences are correlated and are variables,as well as varyingdegreesof interindus-
therefore somewhat statistically redundant. Both try interdependencies(Pennings, 1981). Some of
variables are predictive of the innovation event. the control variablessuch as changes in demand,
In their redundancy, these variables compete in concentration ratio and amount of shake-outs
accounting for the bank's probability to innovate. could figure more prominentlyin the hypothesis
The prominence of technological networking in testing-particularly if the research is extended
all regression models, however, is a testimony to to other settings. In some ways, the findingsof
44 J. M. Pennings and F. Harianto

this study inevitably reflect the uniqueness of the interfirm structures stabilize knowledge flows
financial services industry. among industries.
The sudden emergence of converging techno- Significantly, it should be pointed out that this
logical trends in the banking industry entails a research did not attempt to account for the
quasi-experiment providing unique opportunities incidence of video banking. We relied on a
for hypothesis testing. Industry studies conducted variant of the case control method (Schlesselman,
in other settings should reveal how far the present 1982) that compares adopters vs. non-adopters.
results are idiosyncratic to the financial services The intercept in this model would not accurately
industry. Likewise, studies in Europe and Pacific reflect the incidence of video banking in the U.S.
Basin countries might disclose to what extent this The estimate of the incidence, the intercept, is
research is U.S. biased (compare Grant, Jammine biased in relation to the proportion of adopters
and Thomas, 1989, who raise a similar issue with vs. non-adopters. The covariates in the study,
respect to strategic diversification studies). It however, are not affected by the retrospective
should also be borne in mind that the banking sampling procedure. These estimates were central
industry is not the only one being inundated to the present study.
with information technology. Transportation,
entertainment and watch and toy industries, to
A 'failed' innovation
name a few, are likewise witnessing an increased
dependence on technology which traditionally is Much can be said about successful vs. failed
not theirs. The drastic transformation of those innovations. It remains difficult to specify what
industries creates opportunities for corporate is meant by failure. Recent experimentation, for
entrepreneurship. We are witnessing the creation example with CompuServe and Prodigy, points
of various interfirm structures to harbor such to wider consumer acceptance of video banking
entrepreneurial activities. They should be studied in the U.S. Prodigy has over one million
in order to understand how firms in a given subscribers now. Its French version, Minitel, has
industry succeed in absorbing the know-how been an astounding success. Compared with
which originally was largely alien. This study of Europe, the U.S. banking and telecommuni-
financial services firms is illustrative of the sort cations industry is still heavily regulated. Regu-
of behaviors that any firm may display under lation has precluded local telephone companies
conditions of increased interindustry interdepen- or long distance carriers from joining the videotex
dence. What sets banks apart, however, is that industry. Even an obvious service as on-line
their service delivery often requires interfirm directory assistance is prohibited. Deregulation
coordination. They resemble therefore such could eliminate some of the barriers for diffusion,
sectors as airlines, telephone companies and however (Aumente, 1987). Also, as shown in
health care providers. As we have seen, competi- Table 1, there are different modes of entry into
tor linkages are also conducive to innovation. It the service (i.e. internal development, licensing,
is therefore obvious that replications should joint venturing and acquisition). They dictate
consider these sectors in the first place. entry strategies with different technological and
Studies on the present level of analysis uncover marketing risks.
how firm-specific traits provide ingredients for Any innovative effort entails strategic risks.
industry and interindustry dynamics which popu- Ideally, a firm wants to enter a market with the
lation level or input-output tables cannot provide. right (or 'dominant') design. However, without
As Astley (1985) has suggested, research should careful planning, two types of strategic errors
move from the level of industry to what they call can be committed. A firm may enter too
the population level. However, the strategic early when design requirements are still in an
implications of interindustry interdependencies experimental stage and the choice results in an
should not only be inferred from input-output unsuccessful design. This 'first mover' dilemma
R&D, as Scherer (1982) suggested, or value represents a development risk; although, since
added flows, but also from the trends in interfirm R&D expenditures and capital intensity in the
exchanges which firms maintain in order to banking sector are comparatively minor, this risk
expand the so-called 'visible hand' (Chandler, is rather small. It was, however, a significant
1976). However, it is not yet evident how various issue in the early eighties when several expensive,
 Financial Services Innovation 45

but ill-conceived, video banking adoptions such mately, the health of any firm is optimal when
as Chemical's Pronto, 'crashed.' A firm might it finds a balance between capitalization of
also enter too late, when the product or service existing resources and the formation of new ones
requirements have become known and have (Penrose, 1959). It may be difficult to delineate
found their way into a dominant design. This an optimal trade-off. Yet, this study suggests that
dilemma amounts more to a marketing risk. This firms continuously add know-how to their existing
error is serious because the laggard will face an pool of skills and promote their readiness to
uphill battle against the adopter who enters with venture into future projects. Furthermore, this
the right design at the right time, i.e. who accumulation of skills is not restricted to process
committed neither of the two errors. Minitel, the and product skills. This study shows that by
French videotex service has become a dominant 'investing' in interfirm contacts, companies
design, while Prodigy and CompuServe are still broaden their exposure to other industries. This
exploring whether any one of them, or a laughing exposure furnishes important strategic advan-
third might become the U.S. dominant design. tages, and endows them with the adaptiveness
In the service sector where strategic mimicking that is particularly needed at the present time.
of services seems easy and fast, technology-based While this has been a study on banking, it
innovations appear to be an exception in that presents also important implications for all those
implementation takes a long time. A firm needs industries, whose core skills become blended with
to commit resources that permit it to respond to know-how that is not traditionally theirs.
new technological developments. At the same
time, and even more so, the firm requires the
flexibility and market intelligence to market new ACKNOWLEDGEMENTS
services at a strategically opportune time. When
new technology continues to diffuse into the This research was supported by National Science
banking sector, the banks should be prepared to Foundation grant SES-8709674 (Johannes M.
respond technologically and commercially. Pennings Principal Investigator) and by the Jones
It is important, therefore, not to dismiss Center of the Wharton School. We appreciate
videotex in the U.S. as an outright failure because the comments of Eric Abrahamson, Paul Allison,
its learning and networking benefits cannot be Graham Astley, Bob Drazin and Larry Robbins.
gauged at the present time. Some of these
benefits may be quite intangible. Videotex
enhances a bank's absorptive capacity in infor-
REFERENCES
mation technology-a technology which is so
critical for this sector. For managers, it remains
Abernathy, W. J. and J. M. Utterback. 'Patterns of
crucial not to dismiss failed innovations as a lost industrial innovation', Technology Review, 80, 1978,
cause but rather to justify adoption as preserving pp. 40-47.
their competitive edge in building and acquiring Abrahamson, E. and L. Rosenkopf, 'When do
new skills. Regulations may change, thus altering bandwagon diffusions roll? How far do they go?
a firm's strategic contingencies. New technologies and When do they roll backwards: A computer
simulation', Academy of Management Best Paper
will further enter the banking sector, creating Proceedings, 1990.
unforeseen strategic contingencies. Finally, as we Allison, P. D. Event History Analysis, Sage, Beverly
have shown, the market of banking services has Hills, CA, 1984.
become volatile. Responding to volatility requires American Banker. Annual Survey, The American
that banks do not take their environment for Banker, Washington, DC, 1977-1988.
Amit, R., J. Livnat and P. Zarowin, 'The mode of
granted. corporate diversification: Internal modes versus
Indeed, many firms are continuously engaged acquisition', Managerial and Decision Economics,
in some form of corporate entrepreneurship to 10, 1989, pp. 89-100.
develop and promote their capacity for future Anderson, A., A. Auquir, W. W. Hauck and D.
activity. Since such activities entail risk of failure, Oakes, Statistical Methods for Comparative Studies,
Wiley, New York, 1980.
the crucial element of managerial responsibility Arlen Communication, Videotex Directory, Arlen
is to gauge the trade-offs between the acceptable Communication, Inc., Washington, DC, 1985.
risk and the growth of future resources. Ulti- Arrow, K. J. 'The economic implications of learning
46 J. M. Pennings and F. Harianto
by doing', Review of Economic Studies, 29, 1962, Lieberman, M. B. 'The learning curve and pricing in
pp. 155-173. the chemical processing industries', Rand Journal
Astley, G. T. 'The two ecologies: Population and of Economics, 15, 1984, pp. 213-228.
community perspectives on organizational evolu- Lieberman, M. B. and D. B. Montgomery, 'First
tion', Administrative Science Quarterly, 1985, pp. mover advantage', Strategic Management Journal,
224-262. 9, 1988, pp. 41-58.
Aumente, J. New Electronic Pathways, Sage, Beverly Mansfield, E. 'Size of firm, market structure, and
Hills, CA, 1987. innovation', Journal of Political Economy, 71, 1963,
Bantel, K. and S. J. Jackson. 'Top management and pp. 477-485.
innovation in banking: Does the composition of Mohr, L. B. Explaining Organizational Behavior,
top management teams make a difference?' Strategic Jossey Bass, San Francisco, CA, 1982.
Management Journal, Special Issue, 1989, pp. Nord, W. R. and S. Tucker. Implementing Routine and
107-124. Radical Innovation, Lexington Books, Lexington,
Belsley, D. A., E. Kuh and R. E. Welch, Regression MA, 1987.
Diagnostics, Wiley, New York, 1980. Normann, R. 'Organizational innovativeness: Product
Braun, E. and S. MacDonald, Revolution in Miniature: variation and reorientation', Administrative Science
The History and Impact of Semiconductor Elec- Quarterly, 16, 1971, pp. 203-215.
tronics, 2nd Edn. Cambridge University Press, New Nelson, R. R. and S. G. Winter. An Evolutionary
York, 1982. Theory of Economic Change, Belknap Press and
Chandler, A. The Visible Hand, Belknap, Cambridge, Harvard University Press, Cambridge, MA, 1982.
MA, 1976. Office of Technology Assessment, Congress of the
Cohen, W. M. and D. A. Levinthal. 'Innovation and United States. Effects of Information Technology
learning: The two faces of R&D', Economic on Financial Service Systems, U.S. Government
Journal, 99, 1989, pp. 569-596. Printing Office, Washington, DC, 1984.
Compaine, B. M. Understanding New Media, Ballinger, Pennings, J. M. 'Strategically interdependent organiza-
Cambridge, MA, 1984. tions'. In P. C. Nystrom and W. H. Starbuck (eds.)
Cyert, R. M. and J. G. March. A Behavioral Theory Handbook of Organization Design, Vol I, Oxford
of the Firm, Prentice Hall, Englewood Cliffs, NJ, University Press, New York, 1981, pp. 433-455.
1963. Pennings, J. M. 'On the nature of new technology as
DiMaggio, P. and W. W. Powell. 'The iron-cage organizational innovation'. In J. M. Pennings
revisited: Institutional isomorphism and collective and A. Buitendam (eds.) New Technology as
rationality in organizational fields', American Socio- Organizational Innovation, Ballinger, Cambridge,
logical Review, 43, 1983, pp. 147-160. MA, 1987, pp. 3-12.
Directory of Electronic Fund Transfer, Directory of Pennings, J. M. and F. Harianto. 'Technological
Electronic Fund Transfer, EFT, New York, 1986. innovation and organizational form'. Working
Economist, The, 'Survey on banking', March 17, 1989. Paper, The Jones Center of the Wharton School
Enright, M. J. 'Geographic concentration, vertical at the University of Pennsylvania, 1991.
structure and innovation'. Working Paper, Harvard Penrose, E. The Theory of the Growth of the Firm,
University, Cambridge, MA, 1989. Wiley, New York, 1959.
Federal Deposit and Insurance Commission. Bank Predicast. Index of American Business, Predicast, New
Operating Statistics, FDIC, Washington, DC, York, 1978-1988.
1976-1983. Revell, J. R. S. Banking and Electronic Fund Transfers.
Federal Deposit and Insurance Commission. Statistics Organization for Economic Cooperation and Devel-
on Banking, FDIC, Washington, DC, 1985-1987. opment, Paris, 1983.
Grant, R. M., A. P. Jammine and H. Thomas. Rosenberg, N. Inside the Black Box: Technology and
'Diversity, diversification, and profitability among Economics. Cambridge University Press, New York,
British manufacturing companies, 1972-1984', 1982.
Academy of Management Journal, 31, 1988, pp. Rumelt, R. Strategy, Structure and Economic Perform-
771-801. ance, Division of Research, Graduate School
Harianto, F. and J. M. Pennings. 'Technological of Business Administration, Harvard University,
innovation through interfirm linkage'. In L. Gomez- Boston, MA, 1974.
Mejia and M. W. Lawless (eds.) Managing The Scherer, F. M. 'Inter-industry technology flows in the
High-Technology Firm, JAI Press, Greenwich, CT, United States', Research Policy, 11, 1982, pp.
1990, pp. 15-42. 227-245.
Hladik, K. J. International Joint Venture:An Economic Schlesselman, J. J. Case-Control Studies: Design,
Analysis of U. S. -Foreign Business Partnerships. Conduct, Analysis, Oxford University Press, New
Lexington Books, Lexington, MA, 1985. York, 1982.
Kazanjian, R. and R. D. Drazin. 'Implementing Schumpeter, J. A. The Theory of Economic Develop-
internal diversification', Academy of Management ment, Harvard University Press, Cambridge, MA,
Review, 12, 1987, pp. 342-354. 1934.
Kogut, B. 'Joint ventures: Theoretical and empirical Williamson, 0. E. Markets and Hierarchies, Free
perspectives', Strategic Management Journal, 9, Press, New York, 1975.
1988, pp. 319-332.