The
Liability of Growth; A Second Critical Period of
Liability
Glenn S. Omura and Eric Craymer
presented at the American Marketing Associationís
Marketing and Entrepreneurship Conference,
Introduction
ÝÝÝÝÝÝÝÝÝÝÝ Since its introduction
into the literature on firm performance by Stinchcombe
(1965), the concept of a ìliability of newnessî, a period of high mortality
shortly after Start-up ( often proposed as one
to three years after) has become widely accepted.Ý
A majority of growth theorists, while differing in many other aspects,
seem to agree on this point (Stinchcombe, 1965;
Starbuck, 1965; Kroeger, 1974; Vesper, 1980; Hannan & Freeman, 1985; Timmons, 1990).Ý One thing they all hold in common is the premise,
albeit sometimes implicit, that small firms are especially vulnerable in
their first several years of existence.
ÝÝÝÝÝÝÝÝÝÝÝ More recently,
though , an increasing number of questions have
arisen in the literature concerning ìthe liability of newnessî.Ý Repeated mentions of survival crises later in
life suggest that, at a minimum, a second wave of failure, which is perhaps
equivalent in severity to start-up failure, occurs during the growth period
of the entrepreneurial firmís development.Ý
There are studies which show empirical evidence of greater severity
in a later ìadolescentî period ( Fichman
and Levinthal, 1991; Bruderl
and Schussler, 1990) and others which have found
empirical proof of an equivalent rise in hazard rate of failure as time
proceeds beyond the first several years (Reynolds and Miller, 1987; Reynolds
and Miller, 1989) but with older firms failing less due to a lower baseline
hazard rate (Amberguey, et al. , 1993).Ý In these and many other publications, authors
speak of firms which have passed through the challenges of new venture start-up
to become successful and productive only to face an entirely new, but equally
serious, test of the ability of the entrepreneur and the firm.Ý
ÝÝÝÝÝÝÝÝÝÝÝ These
studies raise serious questions raised as to the existence and comparative
severity of ìthe liability of newnessî. Despite many instances of its mention,
few studies have shown any proof of itís existence
or suggested a plausible mechanism by which it might operate.Ý This paper discusses a study whose purpose is
to rectify both of those short-comings.Ý
Growth Models
ÝÝÝÝÝÝÝÝÝÝÝ In this
study, we seek an explanation of a second critical period of potential failure
that occurs to firms which have already passed the age associated with a
"liability of newness" and moved into a period of successful growth.Ý
Looking more broadly at the mentions of later in life crisis cited
in the Introduction, one could ask what are the common underlying elements
that join them into a unique class of events?Ý Is there a larger
model at work of which the later in life crisis is a smaller part?
ÝÝÝÝÝÝÝÝÝÝÝ In those
instances cited in the Introduction, one element is that of change; a negative
change in performance and a numerically positive change in the probability
of discontinuation.Ý As the firm changes,
the very characteristics that supported earlier success may "start
turning into barriers" which might lead to impairment and even failure
(Hofer and Charan, 1984).Ý Another element is the passing of time.Ý Change does not occur as a function of time
but rather as a function of developmental maturity which occurs as time
passes.Ý As put by Quinn and Cameron
(1983), "a consistent pattern of development seems to occur in organizations
over time" with both behaviors and structures of one period distinguishably
different from those of another.Ý Lastly, there is an element of size.Ý Not all small companies become big companies,
but most companies that are very large today started out small (
ÝÝÝÝÝÝÝÝÝÝÝ Unfortunately,
as so often seems the case in entrepreneurial research, while there is much
agreement as to the appropriateness of a growth model in new venture performance,
there is very little agreement as to which model, of the many available,
to make use of (Aldrich, 1992).Ý As
described by Hofer and Charan (1984) there are
four major types of growth models; the life cycle growth models, the stages
of development models, the evolutionary models, and the transition models.
ÝÝÝÝÝÝÝÝÝÝÝ The
life cycle models propose that firms progress through different stages of
maturity in a fashion similar to biological organisms with stages of birth,
growth, maturity, and death.Ý Change
occurs in a relatively set sequence as a result of the passing of time and
is always in the form of a refinement of the earlier stage until the eventual
decline of death.ÝÝÝ
ÝÝÝÝÝÝÝÝÝÝÝ Stages
models hold in common with the life cycle models the concept of passing
through a series of phases.Ý While
the firm and the founder remain basically the same, their focus and activities
change.Ý Unlike the life cycle models, however, it is
not necessary to pass through all stages nor to
do it in any set sequence.Ý In stages
models triggering change is not limited to the aging process but can also
be due to events such as increased size and organizational complexity (Galbraith
and Nathanson, 1978; Timmons, 1990; Terpstra
and Olson, 1993).
ÝÝÝÝÝÝÝÝÝÝÝ In evolutionary
models phases are also seen as developmental improvements but rather than
the slow and steady progression seen in life cycle models, they suggest
cataclysmic break-through change which alters the form of the organization
in a drastic manner.Ý Change follows
a predictably hierarchical sequence driven by the growth of the organization
but in order to move to the next phase the organization must alter its form
so as to be entirely different from the form it had prior to the change.
ÝÝÝÝÝÝÝÝÝÝÝ Ý Transitional models propose change that is both
drastic and unpredictable.Ý Driven
by changes from within or without, the firm finds it necessary to abruptly
and disruptively change their method of operation in order to survive.Ý In the transitional period the firm leaves behind
the safetyÝ of
routinized decisions and attempts to develop a
new organizational strategy and structure that better match its new circumstances.Ý
During these transitional periods the firm is extremely vulnerable.
ÝÝÝÝÝÝÝÝÝÝÝ Each
of the four model types possesses some theoretical strengths, but each also
displays at least one flaw which prevents it from explaining a later in
life crisis arising after initial success.Ý
Life cycle models insist on an invariant sequence of stages that
are a result of the "aging" of the organization and, thus, do
not allow for any decrease in performance except in the final decline unto
death.Ý This leaves us with no way to explain the pattern
of growth, decline, and new growth often mentioned in the literature. While stages models allow for forward
and backward movements and a multitude of sequences, they do not
explain how the actual transition from stage to stage takes
place, nor do they consider the particular vulnerability of firms
during it, as cited above.Ý Transition
models speak directly to this problem but fail to provide guidance as to
what might occur next or to provide a general model that could explain how
the stages or the transitions that join them operate.Ý
Evolutionary models, while in some cases allowing for de-evolution,
still tend to suggest an incremental step up or down a fixed ladder of structures
and strategies.Ý As put by Galbraith and Nathanson
(1978:103), "The problem with all of these models is not that they
are wrong, but that they are only partially correct."
ÝÝÝÝÝÝÝÝÝÝÝ Perhaps
the real problem lies in the fact that each of the four model types are looked at as being mutually exclusive. It is possible that
each model type accurately mirrors one part of a larger growth process.Ý
This could explain the mixture of anecdotal and empirical evidence
each model has found and could also explain why researchers have, as yet,
been unable to determine and agree upon a single prevalent model.
ÝÝÝÝÝÝÝÝÝÝÝ If each
of the model types holds some element which accurately describes a part
of the growth process experienced by small firms, then perhaps an amalgam
model is called for.Ý By pulling the
strengths of each model together, a new and synergistically superior description
may be had.Ý One model which moved
in this direction was the Churchill & Lewis (1983) model.Ý While primarily a stages model it also incorporates
many of the elements ofÝ the life cycle models and thus overcomes
many of the weaknesses of each model type considered alone.
ÝÝÝÝÝÝÝÝÝÝÝ Unlike
the majority of growth models which depict a firm's development following
a curve, Churchill & Lewis utilize a decision tree structure (see Exhibit
I ).Ý They eschew the
popular but narrow indications of a firm's development, sales and profits,
and instead characterize the stages by an index which simultaneously considers
size, diversity, and complexity.Ý Each
stage is described by its attributes along five management factors dealing
with strategy, structure, and the owner's relationship to the business.Ý This also allows them to differentiate the changing
importance of various key management factors as the business progresses
through the stages.
ÝÝÝÝÝÝÝÝÝÝÝ In the
model the firm is propelled forward by entrepreneurial and strategic drive
and is limited by environmental selection.Ý
In the earlier stagesÝ
(one and two)Ý the firmís
movement is almost entirely due to environmental selection and adaptation.Ý Then in stage three, with a place for it within
the competitive environment secured, strategic choice enters and the firmís
progress is the result of the counterbalance and interaction of the strategy
and the environment.Ý Finally, if
all goes well, the firm has achieved the requisite resources to allow outcome
to be largely predicated upon strategy alone.
ÝÝÝÝÝÝÝÝÝÝÝ By entering
volitional strategic choice as a variable, the Churchill and Lewis model
allows for the possibility of the non-monotonic movement that a life cycle
model precludes.Ý In addition, it
appears that they also include elements of the transitional models.Ý Each stage may meet with success or the failure
of ìfoldingî and, if successful, they may hold indefinitely at that point.Ý
The only way that risk is re-introduced is when they move, or attempt
a transition, to the next stage.Ý It
is only at this point that the model holds the possibility of ìfoldingî,
exactly the type of risk of failure the transition models imply.
ÝÝÝÝÝÝÝÝÝÝÝ Two
weaknesses remain.Ý The model still
cannot explain how a firm might come to be at risk for the later in life
crisis nor does it explain the reports of firms moving more than one stage
or phase at a time.Ý It is possible
by incorporating further elements of the other model types to strengthen
the Churchill and Lewis model.Ý The
first weakness is overcome if the transitional periods which threaten firm
survival in the transitional models are added between the phases of Churchill
and Lewis.Ý There is much in the literature to suggest that
the changes in strategy and structure suggested by Churchill and Lewis to
occur as a firm passes from stage to stage could trigger the increase in
vulnerability ( Chandler, 1962;Ý Galbraith and Nathanson,
1978;Ý Hofer and Charan,
1984;Ý Hrebniak
and Joyce, 1985;Ý Cowen, et al. ,
1984;Ý Sandberg, 1987).
ÝÝÝÝÝÝÝÝÝÝÝ The
second weakness is overcome if a firms trajectory
is considered.Ý It has been hypothesized
by Greiner (1972) that faster growing firms might pass through evolutionary
phases more quickly due to a higher trajectory.Ý If faster growth shortens the time a firm spends
in a phase it is likely that it could also shorten the time spent in transition.Ý
If it shortened that time enough itÝ
is possible that the movement through a phase might be short
enough so as to appear to be a ìjumpî of more than one phase.ÝÝ
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ
With these additions, the model successfully overcomes several of
the discrepancies between previous models and the behavior suggested in
the studies cited in the introduction.Ý It
allows for variation in sequential upward movement such as that found by
Drazin and Kazanjian
(1990) who allow for three alternatives:Ý
firms may prematurely move into decline (jumping forward several
stages), they may regenerate and "regress" out of decline back
to an earlier and more successful phase (moving backward over the curve),
or they may stay in the same phase for a period of time (the persistence
referred to by Aldrich, 1979).Ý Churchill and Lewis allow, as possible outcomes,
retrenchment (falling back one or two stages), remaining stuck in a holding
option for as long as its entire life, and, with the extensions proposed,
the appearance of a forward leap beyond the next sequential stage.Ý
If these three variations do occur [(1) forward or backward
movement through stages, (2) non-sequential movement while moving in either
direction, and (3) cessation of movement while at one stage for an
indefinite period of time] it would suggest the following proposition:
ÝÝÝÝÝÝÝÝÝÝÝ PROPOSITION I:Ý Due to the existence of movement on the curve
which is not necessarily monotonic, the relationship between development
and age will tend to be non-linear.
Venture
Types
ÝÝÝÝÝÝÝÝÝÝÝ By entering
human volition ( in the form of strategic choice)
as a variable that interacts with determinism to alter outcome, the possibility
of differing patterns of growth and performance also enters (Hrebniak and Joyce, 1985).Ý
If it is assumed that all firms follow a common and pre-determined
pattern of development then a similar level of performance for any given
level of development would be expected.Ý
The literature suggests this is not the case.Ý
Strategic choice opens the possibility for differentiation of performance.
ÝÝÝÝÝÝÝÝÝÝÝ First,
the firm can choose to grow or not.Ý Those
who do not may become the mom-n-pop or lifestyle businesses discussed in
the literature ( Cooper, 1982; Sandberg, 1986;
Timmons, 1990).Ý Those who do will
be differentiated on the basis of their success in pursuing that growth. As a result of how well they perform within it, they can be grouped into
categories.Ý By integrating and collapsing
the various theories "types" into categories, four groups are
created:Ý those who fail, those whose
sales continue to increase moderately over time as would be expected by
most life cycle models and thus are "on track," those whose performance
is higher and occurs faster than "on track" and makes them "over
achievers," and those who attempted to beÝ
(or actually were) either "on track" or "over achievers"
but have since declined and are now struggling with slow and/or negative
growth and poor performance, the "under achievers."
Ý ÝÝÝÝÝÝÝÝÝ All of the above would lead us to believe
the following:
ÝÝÝÝÝÝÝÝÝÝÝ PROPOSITION II:Ý Amongst firms pursuing growth (all firms, minus
the no-growth firms) we would expect to find three groups distinguishable
on the basis of their performance; under
achiever with poor performance for their developmental maturity, "on
track" with average performance for their developmental maturity, and
over achiever with high performance given
their developmental maturity.Ý (Due
to its demise, the "extinguished" group would not be present except
in historical record.)
ÝÝÝÝÝÝÝÝÝÝÝ It would
be within the ìunderachieversî group that we would expect to find the subject
ofÝ our search,
those firms going through a transitional phase and showing signs of decline.
who might again climb to ìon trackí or ìover achieverî
status in a renewed stage of growth.Ý Given
that the model suggests that firms renew their risk each time they decide
to move to another stage of development and thusÝ reset the liability of newness clock (Stinchcombe, 1965), and that it also suggests that the firm
gains the ability to make this choice in its later stages, it would lead
us to believe the following:
ÝÝÝÝÝÝÝÝÝÝÝ PROPOSITION
III:Ý Due to the continued liability
a firm faces each time it moves from stage to ÝÝÝÝÝÝÝÝÝÝ
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ
stage and its inability to have an active choice in making
the move which triggers the
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ
liability until its later developmental stages, there will
be a group of declining firms
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ
much older than the one to three years associated with the
liability of ìnewnessî.
Hypotheses
ÝÝÝÝÝÝÝÝÝÝÝ The
Churchill and Lewis model, with the preceding adaptations, could explain
a later in life crisis.Ý If we believe
that this model may be in operation, it would lead us to make certain assumptions
concerning what patterns we would find when looking at a large group of
firms.Ý To test these assumptions,
the following hypotheses (based on the propositions specified in the preceding
two sections of this paper) were developed.
ÝÝÝÝÝÝÝÝÝÝÝ Due
to the possible forward and backward movement by a firm on the curve and
the varied performance of individual firms at a given point on the curve,
the following hypotheses are suggested:
ÝÝÝÝÝÝÝÝÝÝÝ
H1a:Ý The relationships between age and sales and
ÝÝÝÝÝÝÝÝÝÝÝ H1b:Ý Age and performance will be non-linear.
By adding the ability of a firm to make a volitional choiceÝ to not evenÝ participate in the life cycle growth
curve, it
becomes possible to expect a group of firms which remain at low sales for
extendedÝ and even
eternal
periods of time.Ý This would explain
the lifestyle mom-n-pops and the start small/stay small
firms.Ý We are not interested in such firms but rather
in the firms who choose growth and thus risk a
later in
life crisis.
ÝÝÝÝÝÝÝÝÝÝÝ When
a firm chooses to move onto the growth curve, it may meet with variable
success.Ý Some firms would be expected
to meet with average success, moving steadily up in age and performance.Ý Other more successful firms would be on an accelerated
path, performing better at an earlier age.Ý The remaining firms would be those less successful,
possibly starting successfully on a growth curve only to fall prey to transitional
vulnerability.Ý The varied success
of these three groups as they follow the growth curve would lead us to form
the following hypothesis:
ÝÝÝÝÝÝÝÝÝÝÝ H2:Ý Of those firms on the growth curve, there will
be three sub-groups with significantly
ÝÝÝÝÝÝÝÝÝÝÝ different
levels of performance given their developmental maturity.
Since the model suggests that the firms on the growth
curve have met with at least some initial success and that decline may occur
at any stage, even later ones, the following could be true:
ÝÝÝÝÝÝÝÝÝÝÝ ÝH3: The less successful, declining sub-group
will be much older than the one to three years of
ÝÝÝÝÝÝÝÝÝÝÝ age generally considered to be the period of "liability
of newness."
Methods
ÝÝÝÝÝÝÝÝÝÝÝ In order
to search for that evidence a database of
ÝÝÝÝÝÝÝÝÝÝÝ To look
for the existence of the hypothesized venture types a 3x3 matrix was developed
with sales on the horizontal edge and age on the vertical edge.Ý Ranges of values for each of the variables were
developed by loosely integrating the ranges for stage transitions suggested
in the literature ( Vesper, 1980; Timmons, 1990;
Osborne, 1992; Reynolds, 1993) resulting in the following gradients:
ÝÝÝÝÝÝÝÝÝÝÝ Age (from top to bottom)
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ Young: Firms with 1-9 years
of age
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ Middle: Firms with 10-19
years ofÝ age
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ Mature: Firms with 20+
years of age
ÝÝÝÝÝÝÝÝÝÝÝ Sales (from left to right)
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ Low: Firms with sales of
$1 to $500,000
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ Moderate: Firms with sales
of $501,000 to $5,000,000
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ High: Firms with sales
over $5,000,000
Within the matrix and according to the model, the ìon
trackî firms would occupy the three cells on the top left to bottom right
diagonal (age young /sales low, age middle /sales moderate, and age mature
/sales high) .Ý
The ìover achieversî would be in the cells above that diagonal (age
young /sales moderate, age young /sales high, and age middle /sales high)
and the ìunder achieversî would occupy the cells below the diagonal (age
middle / sales low, age mature /sales low, and age mature /sales moderate).
The Variables
ÝÝÝÝÝÝÝÝÝÝÝ The
hypotheses are all based on performance over time.Ý In order to test them it was necessary to develop
substitute measures for both developmental growth and performance.Ý Three measures that are typically used in the
literature to approximate growth are sales (Cragg
and King, 1988), age (Reynolds and Miller, 1987), and number of employees
(Van de Ven, et al. , 1984).Ý To these we added a variable for asset size,
square feet of plant.Ý
ÝÝÝÝÝÝÝÝÝÝÝ A perhaps
more difficult task is to identify similarly good substitutes for the dependent
variable(s).Ý If the above measures
are used to infer development, there must also be a way to measure performance
as it varies over time and from firm to firm.Ý In many venture performance studies, profit
has been used as a measure of success, but Cooper (1982) warns us that this
may not be a good measure as many small ventures have goals other than profit.Ý Van de Ven, et al.,
(1984:90) states the case that "performance measures for company start-ups
must differ from traditional measures of performance" as they are usually
cash hungry and growth oriented, foregoing today's profits to build the
future.
ÝÝÝÝÝÝÝÝÝÝÝ In this
research, we seek to find a situation in which a firm experiences a decrease
in performance.Ý It is not a loss
of the ability to perform or there would be no evidence of rejuvenation.Ý
The fact that the ability to be successful exists
but is not utilized to do so suggests slack resources or, more accurately,
below par productivity.Ý In the current study, indications of this might
be sales per employee (a measure of labor productivity) or sales per square
foot (a measure of asset productivity).Ý
A similar performance measure was used by Anikeef,
et al. (1993) in a 1993 study of the real estate industry.Ý While neither of these were original variables
in the database, they were easily formulated by using original variables.
ÝÝÝÝÝÝÝÝÝÝÝ Sales per Employee:
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ This was
measured, by firm, as sales divided by number of employees and could be
considered an indication of how efficiently labor is utilized.
ÝÝÝÝÝÝÝÝÝÝÝ Sales per Square Foot:
ÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝÝ This is measured,
by firm, as sales divided by square feet of physical plant and could be
considered an indication of how efficiently assets are utilized.
Results and Discussion
ÝÝÝÝÝÝÝÝÝÝÝ The hypothesis were tested using procedures available in the SPSS-X
software package.Ý Specific tests
used included Descriptive Means, Linear Regression, and Analysis of Variance.Ý
ÝÝÝÝÝÝÝÝÝÝÝ To test Hypothesis
1a, which claims that the relationship between age and sales are non-linear
due to non-monotonic movement, linear regression was used.Ý As can be seen in Exhibit IV, there appears
to be interaction between the variables, significant at the .01 level.Ý What can also be seen, though, is that the percentage
of variation it explainsÝ is relatively low (Adjusted R
[SoM1]
2
value of .02618).
ÝÝÝÝÝÝÝÝÝÝÝ The evidence,
then, did not support Hypothesis 1a.Ý The
results show that the relationship does appear to be linear.Ý While not providing proof for our hypothesis,
we do not believe that this evidence rules out the non-linear movement proposed.Ý
It is possible that, for the most part, firms tend to follow a normal
life cycle pattern of sales increasing over age.Ý
Even with this general tendency, it is possible that certain firms
could indeed leap forward, fall backwards or stall.Ý
Indeed, the mere existence of the significant number of members in
both the "over achiever" and "under achiever venture types
speaks strongly to its truth.
ÝÝÝÝÝÝÝÝÝÝÝ These findings
are consistent with the findings of Reynolds & Miller (1987b) that,
while age/sales correlations are statistically significant, they account
for such a small proportion of the variance as to be of little help in predicting
sales.Ý They are also consistent with
our expectation that sales would be correlated to age but not primarily
a function of it, and thus may not directly disprove Hypothesis 1a.
ÝÝÝÝÝÝÝÝÝÝÝ To test Hypothesis
1b, which claims that the relationship between age and performance are non-linear
due to variation in performance, linear regression was once again used.
Linear regressions were run on the variables sales per employee, sales per
square foot, and age.Ý In each case,
one of the sales based performance measures was assigned as the dependent
variable, and age was assigned as the independent variable.Ý Some linear relationship was found for both
performance measures regressed against age (both were significant at the
.01 level) but, again, the percentage of variance explained was small (both
had adjusted R2 <.002), even smaller than that in the age/sales
analysis (Exhibit II).
ÝÝÝÝÝÝÝÝÝÝÝ To test Hypotheses
2 and 3 the firms were arrayed in the 3x3 matrix and divided into the three
hypothesized venture types as described in the methods section. Looking
at the means of each growth and performance measure for each of the three
subgroups it can be seen that on all measures the "under achievers" have the lowest means
(See Exhibit III). Interestingly,
though, while the "under achievers"
are always the lowest means of the three, the highest values are not always
found in the "over achievers."Ý
As would be expected, the high performers had higher sales per employee
than the "on tracks", but had lower overall sales and sales per
square foot.Ý This may be an anomaly or it may be explained
by one of two other possibilities.Ý It
could be due to the design of the study.Ý
In setting up the 3 x 3 matrix, all of the firms in
, the bottom right cell (with age mature /sales high) are considered
to be "on track".Ý In actuality,
it is likely that some of the over
achievers would continue to superior performance into their mature years.Ý If this was the case, that cell would contain
a mixture of "on track" and "over achieversî.Ý By including "over achievers" in the
on track averages, the on track means might be artificially inflated.Ý It may also be evidence of a build up of capacity
in preparation for accelerated growth as suggested by several researchers
(for example, see:Ý Cooper, et al.,
1992b).Ý Thus, as high performers
built capacity to handle high growth, their plant size would increase while
their sales remained constant causing the current productivity of their
assets (sales per square foot) to fall even as their labor productivity
(sales per employee) remained high.
ÝÝÝÝÝÝÝÝÝÝÝ To test the
statistical significance of these differences, ANOVA's were run on the three
subgroups means for the following variables:Ý
Sales per Employee, Sales per Square Foot, Number of Employees, Number
of Square Feet of Plant, and Age.
ÝÝÝÝÝÝÝÝÝÝÝ The ANOVA's
showed a significant difference on all variables when grouped by type and
thus support Hypothesis 2Ý
(See Exhibit IV).Ý The
under achievers were the oldest and had the lowest values for performance
and size.Ý Over achievers were exceptionally young, smaller
than the on-tracks in size of plant and employment, relatively close to
"on-tracks" in sales and performed better (as measured by sales
per employee but not sales per square foot as discussed above).Ý The fact that the under achievers are significantly
older than the other two groups and, by looking at their productivity measures,
performing significantly less well, suggests that their problems have arisen
later in life.Ý Since we have removed
those firms which choose "no growth", this group must represent
a combination of older firms which were never successful and firms which
were once successful and now are not.Ý It is this later group, those who pursued growth
and have fallen into decline only as a result of their choice to pursue
it, that we believe exhibit the liability of growth.Ý It is only by the choice of pursuing growth
that the firm has been put at risk, and those under achievers are paying
the price for it.Ý In this manner
they see decreased performance and renewed risk of failure much later in
life than that of the one to three years postulated by ìthe liability of
newnessî.Ý This coincides with the recent research finding
a later in life "liability of adolescence" (Bruderl & Schussler, 1990; Fichman & Levinthal, 1991) and
supports Hypothesis 3.
Summary
and Implications
ÝÝÝÝÝÝÝÝÝÝÝ When
the sample was divided into three groups based on the hypothesized distribution
by age and sales, evidence was found of significantly different performance
levels by venture type as suggested by Hypothesis 2.Ý There was a group ofÝ ìon tracksî with average performance
for their given developmental level, a group ofÝ
ìover achieversî with high performance for their given level of development,
and a group ofÝ ìunder achieversî
with low performance for all levels of development.Ý The ìunder achieverî group exhibited the lower
performance and age older than the one to three years suggested by the ìliability
of newnessî supporting Hypothesis 3.Ý While
the non-linearity of age and sales and age and performance of Hypothesis
1a and 1b were not supported, it is possible that this was due to a central
tendency of the firms to on average follow a pattern of increased sales
and performance with increased development.Ý
This would not preclude the possibility that some firms do not stay
on this developmental track, jumping ahead and falling back.Ý Even though the relationship showed linearity,
the percentage of variation explained by age was so small as to be of practically
no aid in predicting either sales or performance.
ÝÝÝÝÝÝÝÝÝÝÝ Due
to the cross sectional nature of the study, we were unable to prove that
the extended model was in operation.Ý That
is, we could not prove a pattern of success followed by the struggle of
transition, resulting in renewed success, continued struggle, or failure.Ý The presence of three groups of firms with significantly
different performance levels given their development does, however allow
for the possibility that the model is in operation.Ý It could explain the discrepancies between the
behavior predicted by other models and the actual observed firm behavior
discussed in the literature.
ÝÝÝÝÝÝÝÝÝÝÝ Unlike
the biological growth models which mimic an organismic
necessity to progress along its natural development, the inclusion of choice
allows for the firm's ability to actively choose to grow or not. Inclusion
of the role of active choice allows for the possibility that firms may drop
off the curve or accelerate their trajectory along it, resulting in the
non-linear movement.
ÝÝÝÝÝÝÝÝÝÝÝ Once
a firm pursues growth, it must then succeed at it.Ý The extent to which it does so is the underlying
basis for many of the performance typologies in the literature (for example:Ý
Vesper, 1980; Reynolds & Miller, 1987; Cooper, et al., 1992).Ý
In this study, support was found for Hypothesis 2, showing evidence
of three distinct sub-groups of growth firms:Ý
over achievers, those on track, and under achievers.Ý Support was found for Hypothesis 3, showing
evidence of an older and lower performing group of firms which might be
the result of the later in life crisis.
ÝÝÝÝÝÝÝÝÝÝÝ Additionally,
the model theorizes that the danger experienced during transitions may be
due to the shifting needs, activities, and market focus the firm must make
to move from phase to phase.Ý If true,
this would have distinct implications for the firm, marketing managers,
consultants, and researchers.Ý Firms
should organize and manage with these changes in mind, planning ahead
shift strategy and structure to reduce or remove the transitional decline.Ý Firms most at risk would be those pursuing a
high growth strategy as they would move through more transitions in a given
time and have less time to make the passage through the transition.Ý Marketing managers should design their efforts
to support the firms changing needs and focus, altering strategy and tactics
to match the and meet the goals of each phase it
passes through.Ý Consultants should
be aware of the phenomenon of this later in life threat, helping firms to
predict its onset (possibly with the help of productivity measures) and
aid them in making the transition more safely.Ý
Researchers need to be aware of this later in life crisis and take
it into account in modeling and measuring.Ý
It cannot be assumed that performance is determined entirely by environmental
interaction or a predetermined growth curve.Ý
With volition comes the possibility for variation.Ý Within each group and within each phase, the
performance measures must change in order to account for the variation both
due to differing performance within the stage and the difference of performance
goals between the stages.Ý Hopefully
the extended Churchill and Lewis model will aid in these endeavors.
ÝÝÝÝÝÝÝÝÝÝÝ There
is increasing evidence in the literature that a later in life threat that
is at least as likely if not more likely than the earlier threats theorized
by the liability of newness.Ý This
later life threat could be considered a "liability of growth" as it can only become a threat after a firm has actively
chosen to continue historically successful growth.Ý It is only by pursuing this secondary growth
that a firm can incur the liability of not achieving it.Ý This study has provided some support for an
integrative growth model based on Churchill & Lewis' (1983) stage model.Ý
With this model there is the possibility of a later in life crisis,
the "liability of growth," which is only possible through
growth and which might be predicted by asset and labor productivity measures.Ý
It is hoped that future research will prove these possibilities to
be probabilities.
BIBLIOGRAPHY
Aldrich, Howard E.
(1979).Ý Organizations
& environments.Ý
Amburguey, Terry L., Kelly, Dawn, & Barnett, William P. (1993).Ý
Resetting the clock:Ý
The dynamics of organizational change and failure.Ý
Administrative Science Quarterly, 38, 51-73.
Anikeef, Michael A., Ven Sriram & Sapienza, Harry J.
(1993).Ý Age, size, and product market strategy in a
fragmented industry.Ý An
analysis of real estate developers.Ý
Unpublished paper presented at the AMA Conference on Research at
the Marketing/Entrepreneurship Interface,
Bruderl, Josef, & Schussler, Rudolf
(1990).Ý Organizational mortality:Ý The liabilities of newness and adolescence.Ý
Administrative Science Quarterly, 35, 530-547.
Bygrave, William D., & Hofer, Charles W. (1991).Ý Theorizing about entrepreneurship.Ý Entrepreneurship
Theory & Practice, 15(4), 13-22.
Buchele, Robert (1967).Ý Business
policy in growing firms.Ý
Churchill,
Neil C., &
Cooper,
Arnold C. (1982).Ý The entrepreneurship -- small
business interface.Ý The Encyclopedia of Entrepreneurship, 193-208.Ý
ed.
Cooper, Arnold C.,
Woo, Carolyn Y., & Dunkelberg, William C. (1992).Ý Entrepreneurship and the initial
size of the firm.Ý Journal of Business Venturing, 4(5), 317-332.
Cowen,
Scott S., Middaugh, J. Kendall II, & McCarthy,
Kevin (1984).Ý Corporate life cycles and the evolution of management:Ý
Part I.Ý Management
Decision, 22(2), 3-11.
Cragg, Paul B., & King, Malcolm (1988).Ý Organizational
characteristics of firm's performance revisited.Ý Entrepreneurship:ÝÝ Theory & Practice, 12(2), 49-64.
Drazin, Robert, & Kazanjian,
Robert K. (1990).Ý A re-analysis of Miller and
Friesen's life cycle data.Ý Strategic Management Journal, 11, 319-325.
Fichman, Mark, & Levinthal, Daniel
A. (1991).Ý Honeymoons and the liability of adolescence:Ý
A new perspective on duration dependence in social and organizational
relationships.Ý
Galbraith,
Jay R., & Nathanson, Daniel A. (1978).Ý Strategy implementation:Ý
The role of structure and process.Ý
Greiner, Larry E.
(1972).Ý Evolution and revolution
as organizations grow.Ý Harvard Business Review, 50(4), 37-46.
Hannan, Michael T., & Freeman, John (1984).Ý Structural inertia and organizational change.Ý American
Sociological Review, 49:149-164.
Hofer, Charles W.,
& Charan, Ram (1984).Ý The transition to professional management:Ý
Hrebniak, Lawrence G., & Joyce, William F. (1985).Ý Organizational adaptation:Ý Strategic choice and environmental determinism.Ý
Administrative Science Quarterly,
Kroeger, Carroll V. (1974).Ý
Managerial development in the small firm.Ý
Meyer,
Bill T., & Merrill, David W. (1984).Ý The life cycle
of the organization - Part I.Ý Journal
of Commercial Lending, 66 (Spring), 19-37.
Osborne, Richard
L. (1992).Ý Entrepreneurial
renewal.Ý Business Horizons, November/December, 58-62.
Quinn,
Robert E., & Cameron, Kim (1983).Ý Organizational lifecycles and shifting criteria
of success:Ý Some preliminary evidence.Ý
Management Science, 29(1), 33-51.
Reynolds, Paul D.
(1993).Ý High performance entrepreneurship:Ý What makes it different.Ý Unpublished paper, February
21.
Reynolds,
Paul D., & Miller, Brenda (1987).Ý 1987
Reynolds,
Paul D., & Miller, Brenda (1989).Ý New firm survival:Ý Analysis of a panel's fourth year.Ý Paper presented at Ninth Annual Babson College Entrepreneurship Research Conference,
Sandberg,
William R., & Hofer, Charles W. (1987).Ý Improving new
venture performance:Ý The role of
strategy, industry structure, and the entrepreneur.Ý Journal
of Business Venturing, 2(1), 5-28.
Starbuck, William
H. (1965).Ý Organizational
growth and development.Ý In The Handbook of Organizations, ed. March, James G.Ý
Stinchcombe, Arthur L.. (1965).Ý Social structure and
organizations, in The Handbook of Organizations, ed. March, James
G., Chicago:Ý Rand McNally.
Terpstra, David E., & Olson, Philip D. (1993).Ý Entrepreneurial
start-up and growth:Ý A classification
of problems.Ý Entrepreneurship Theory & Practice,
17(3), 5-19.
Timmons, Jeffry A. (1990).Ý New Venture Creation.Ý
Van de Ven, Andrew H., Hudson, Roger, & Schroeder, Dean M. (1984).Ý
Designing new business start-ups:Ý
Entrepreneurial, organizational, and ecological considerations.Ý Journal
of Management, 10(1), 87-107.
Vesper,
Karl (1980).Ý New Venture Strategies.Ý