Journal of High Technology Management Research 26 (2015) 1–13

Contents lists available at ScienceDirect

Journal of High Technology Management Research

Growth options and acquisition likelihood in high tech

Sean M. Davis a,⁎, Jeff Madura b
a
University of North Florida, Coggin College of Business, 1 UNF Drive, Jacksonville, FL 32224, USA
b
Florida Atlantic University, College of Business, 777 Glades Road, Boca Raton, FL 33431, USA

a r t i c l e i n f o a b s t r a c t

Available online 19 May 2015 With high tech firms now representing the majority of acquisitions among all pubic, non-
regulated firms, we attempt to determine how a tech firm’s growth options influence its
Keywords: likelihood of being acquired. In particular, we develop a new growth options proxy called
Growth options Gamma (Γ) to represent the return relative to investment in research and development. We
Real options find that Γ is inversely related to the likelihood of being acquired. Robustness tests show that
Mergers this relationship holds regardless of the subperiod assessed, the size category assessed, whether
M&A tech firms are engaged in friendly or hostile acquisitions, the method used to identify tech
High tech firms, and whether the R&D definition includes capital expenditures. The relationship is even
more pronounced when tech targets have a relatively low valuation (based on the market-book
ratio). Furthermore, we find that tech firms with a high Γ are less likely to acquire targets. In
general, tech firms with a high Γ appear to prefer organic growth rather than expansion by
combinations with other tech firms.
© 2015 Elsevier Inc. All rights reserved.

1. Introduction

Tech firms represent about one third of all non-regulated, publicly traded firms, but increasingly dominate the market for
corporate control, with tech acquisitions now representing over 60 percent of all deals. Yet, research is limited in identifying the
profile of a tech firm that makes it an appealing target. Our objective is to identify the characteristics that increase a tech company’s
likelihood of being acquired.
Tech firms are characterized by their high level of research and development and their continual need to grow and expand in the
face of rapid product obsolescence. They rely heavily on mergers as a means of expediting shifts in their strategies and operations, and
exercising their growth options. Research and development expenditures are those that management undertakes to create growth
opportunities within the firm. R&D can help extend the life of existing projects or technologies in use by the firm or create new
projects or technologies. They can be patentable or not. They can be put in use right away, shelved for future use, or abandoned.
They help create options for management to expand, contract, or switch the use of assets in the firm – part of the classic “financial
flexibility” outlined by Trigeorgis (1993). Because of the unusually high levels of research and development in tech firms, and the
unique operating characteristics surrounding those high R&D expenditures, the results of general studies on mergers cannot be
directly applied to provide inferences about the likelihood of acquisition for high-technology firms. The unique characteristics of
tech firms, high R&D and how R&D interacts with other firm level characteristics may significantly affect the desirability of a target,
therefore affect the likelihood of acquisition.
While much research has attempted to identify characteristics of firms that attract takeovers, the inferences from these studies are
not directly applicable to the technology industry. Previous literature suggests that undervalued firms (based on market to book
ratios) are more likely to be acquired. However, over 40 percent of tech mergers represent glamour firms with a high market to

⁎ Corresponding author. Tel.: +1 904 620 3928; fax: +1 904 620 3861.
E-mail address: sean.m.davis@unf.edu (S.M. Davis).

http://dx.doi.org/10.1016/j.hitech.2015.04.001
1047-8310/© 2015 Elsevier Inc. All rights reserved.
2 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13

book ratio. This suggests that while undervalued firms may represent the typical merger target, a large proportion of tech firms may
attract takeover bids based on significantly different characteristics. Since tech firms are often viewed as growth firms with unique
growth opportunities and characteristics, we apply a growth options analysis to facilitate our investigation. High research and devel-
opment expenditures differentiate tech firms from non-tech firms. Therefore, we focus our analysis within the tech sectors, with the
goal of examining whether and how growth options characteristics influence the appeal of firms as takeover candidates.
We develop a new proxy for managerial effectiveness that estimates its returns relative to its investment in R&D. We hypothesize
that since tech firms with a higher Gamma ( Γ ) are more efficient and priced higher, they are less desirable takeover targets. Converse-
ly, tech firms that are inefficient in utilizing their R&D (low Γ firms) should be cheaper, and have more potential for improvement if
they are acquired. Thus, low Γ firms should be more appealing within the market for corporate control. To our knowledge, this growth
options construct has not been given attention in the literature on mergers. We find that Γ is inversely related to the likelihood of being
acquired. Robustness tests show that this relationship holds regardless of the subperiod assessed, the size category assessed, whether
tech firms are engaged in friendly or competitive acquisitions, the method used to identify tech firms, and whether the R&D definition
includes capital expenditures. The relationship is even more pronounced when tech targets have a relatively low valuation (based on
the market-book ratio). Furthermore, we find that Γ significantly affects the likelihood that tech firms will acquire other firms. In
general, tech firms with a high Γ appear to prefer organic growth rather than expansion by combinations with other tech firms.
Our study offers relevant implications for tech firm managers and board members, who may position themselves to attract bids, or
may attempt to discourage bidders in order to remain independent. Our study also offers relevant implications for investors in tech
companies who benefit from the large premium that is typically received by tech targets.

2. Review of the literature

We rely on academic literature concerning growth options and the probability of being acquired to inform the research questions
of our analysis. We view the terms growth options and real options as synonymous, but we use the term “growth options” since it is
more commonly used in empirical studies.

2.1. Review of literature on growth options and research and development

Berk, Green, and Naik (1999) develop a general equilibrium model for growth options and assets in place, and while they focus
more on the evolution of risk, they find that “firms that perform well tend to be those that have discovered particularly valuable
investment opportunities.” The portion of firm value that is derived from growth options can be considered a “portfolio of options.”
They emphasize that the exercise of a firm’s investment policy involves the exercise of growth options by management, and this
helps motivate a growth options analysis of technology mergers. The Berk et al. (1999) model (hereafter “BGN model”) serves as
the foundation from which our growth options analysis is based. The BGN model states that the value of a firm is made up of the
firm’s present value of its assets in place (AIP) and the present value of its growth options. The AIP are the cumulative value of the
projects currently "alive," and are represented by the book value of capital invested in these projects. The value of a firm’s growth
options are therefore represented in the present value of growth opportunities (PVGO) in the following formulas:

V i ¼ AIP i þ PVGOi ð1Þ

PVGOi ¼ V i −AIP i ð2Þ

Many studies have used this model to examine growth options, but the methods used to create a proxy for PVGO varies. For
example, Andres-Alonso, Azofra-Palenzuela, and Fuente-Herrero (2006) measure AIP as a perpetuity of free cash flow discounted
by cost of capital. Bernardo, Chowdhry, and Goyal (2007) use a book value measure of AIP based on debt and equity. Cao, Simin,
and Zhao (2008) examine five proxies for growth options, including a proxy for PVGO based on trailing ROE, and find that the market
to book ratio has the highest significance and explanatory power.
Using a market-based measure of growth options value captures information (including market sentiment, perceived industry
opportunities, and exposure to economic conditions) about the firm and its future prospects that is not observable on financial
statements or in other non-market measures. Those tech industries with the highest R&D intensities have the most technological
products, or utilize the most technology in the creation of their products. How R&D is used within the firm, where and when resources
are allocated, which projects are deferred, delayed or expanded are all real options for management. The range of choices available to
managers is described by Trigeorgis (1993) as part of the “financial flexibility” available to management in the exercise of their
options. Thus, we suggest R&D expenditures are a proxy for the unobservable assets within the firm, specifically related to its growth
opportunities.
Ho, Tjahjapranata, and Yap (2006), Andres-Alonso et al. (2006) found that R&D investment had a positive impact on the growth
opportunities of a firm. Oriani and Sobrero (2008) found that R&D was positively related to the market value of the firm. Coad and Rao
(2008) found that R&D is positively related to sales growth. These studies support the theory that R&D leads to the creation of growth
opportunities in a firm. However, the impact of R&D on growth options or firm value may vary among technology firms. Ho et al.
(2006) found that R&D interacts with the size and leverage of the firm to change the direction and intensity of the effects. Coad
and Rao (2008) found a similar directional shift in the coefficient for R&D as a predictor of sales growth for tech firms. These findings
motivate an examination of the unique characteristics of tech firms that cause some tech firms to be more attractive takeover
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 3

candidates. We build on this foundation by assessing how a firm’s appeal as a takeover candidate changes with its efficiency in
converting R&D into growth options.

2.2. Review of the studies on the likelihood of being acquired

Palepu (1986) examined how firm and industry characteristics lead to the likelihood of being acquired. His main results relate to
growth, liquidity, leverage, undervaluation, and management effectiveness. Powell (1997) expanded the model developed by Palepu
(1986) by also examining how free cash flow and tangible fixed assets affect takeover likelihood. Powell finds that targets in friendly
mergers were smaller, less liquid, and more levered. In contrast, targets in hostile acquisitions are larger and have significantly higher
valuations as measured by the market to book ratio, a traditional growth options proxy.
In general, studies suggest that most targets of acquisition have poor growth and are resource constrained. The targets are gener-
ally smaller, have less efficient management teams in place, lower liquidity, higher leverage, and lower market valuations relative to
their peers. Many of these characteristics fit the classic profile of a distressed (or value) stock.
Heeley, King, and Covin (2006) found that higher R&D spending increases the likelihood of being acquired for all firms that report
R&D expenditures, but they do not control for the effects of industry variation in R&D spending. Desyllas and Hughes (2009) found
that R&D predicts the likelihood of being acquired, but they limited their analysis to R&D productivity as measured by patent intensity.
They found that targets had low or no recent patent activity but higher patent inventories, and had lower profitability and liquidity.
While there are many industries that generate and utilize their own patents, there are also many tech industries that do not
(e.g. software). The production of patents is not uniform across tech industries and cannot be reliably observed. A patent could
have been an extension of an existing patent that adds no new meaningful value, or its use could be delayed for a later date. Patent
numbers alone are not indicative of the growth opportunities that patents may represent. Therefore, by examining R&D productivity
in a growth options construct, we can examine all tech industries regardless of whether they rely on patents.
In summary, the broader literature suggests that low growth, undervalued firms with low performance and small size are more
likely to be takeover targets. Yet this literature also suggests that high growth firms may be targeted in hostile acquisitions, suggesting
that parsing the sample into subsamples and an investigation of interaction terms may reveal more insight as to how growth options
affect the appeal of tech takeover candidates. R&D leads to the formation of growth opportunities to varying degrees across firms and
industries, and growth options due to R&D are conditioned by size, leverage, market uncertainty and technological uncertainty. The
very limited research regarding the appeal of tech firms as takeover candidates is focused on a few characteristics. We attempt to
build on the existing foundation of related research by developing and testing a more complete model specifically applied to tech
firms over the 1986–2011 period. We also consider how the characteristics that make tech targets appealing targets may change in
response to changes in economic growth.

3. Hypothesis development

We hypothesize that the following characteristics may significantly influence the probability that a tech firm will be acquired.

3.1. Market to book ratio

The market to book (MTB) ratio is frequently used in the literature on growth options and as a proxy for valuation in predicting
acquisition likelihood. Market value captures sentiment for a firm’s past performance as well as its current and future prospects.
Berk et al. (1999) model the present value of a firm’s growth opportunities (PVGO) as the market value of a firm that is greater
than the firm’s book value. Furthermore, Cao et al. (2008), Bernardo et al. (2007), and Lantz and Sahut (2005) find that there is a pos-
itive relation between risk and a firm’s growth opportunities. To the extent that growth options make firms more appealing as targets,
a higher market to book ratio (MTB) should increase the likelihood of being acquired.
However, Cao et al. (2008)) find that idiosyncratic risk is highly related to growth options and this suggests that low MTB targets
would make lower risk acquisitions. Further, because firms with a high MTB are priced high, they are less appealing as targets. Their
prevailing high share prices and the necessary control premium might discourage potential acquirers from pursuing them. According
to Powell (1997), a low MTB is more likely to receive takeover bids than its more richly priced peers. The firm may even be in distress.
A high MTB discourages acquirers because the higher priced firms are growth firms.
Massa and Zhang (2009) find evidence that style investing influences the market for corporate control, and acquiring firms pursue
“popular” targets based on investor sentiment. This sentiment changes over time as growth and value mergers shift in popularity.
They find that acquirers and targets may enter “cosmetic” mergers for a perceived halo effect to the combined firm.
We hypothesize that whether through high growth options or high firm value, the market to book ratio will be inversely related
with acquisition likelihood. This proxy is the market value of assets to book value of assets ratio (MTB) and is measured as the sum of
total assets minus common equity plus the market value of common shares outstanding divided by total assets [see Cao et al. (2008)].

3.2. Research & development (R&D) expenditures

Research and development (R&D) is the primary way that many tech firms create, maintain or extend the life of new growth
opportunities, and the amount of R&D to sales or assets, R&D intensity, is often a defining characteristic of a tech firm. Studies by
Oriani and Sobrero (2008), Tsai and Wang (2005), Andres-Alonso et al. (2006), and many others find a positive relation between
4 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13

R&D expenses and growth of the firm overall. Since R&D intensity contributes to growth, it may make firms more appealing targets.
Heeley et al. (2006) find a positive relation between a firm’s R&D “stock” (a proxy for the current value of trailing R&D expenditures)
and the likelihood of being acquired. Desyllas and Hughes (2009) found that R&D is positively related to the likelihood of being
acquired, but they focus on patents and patent productivity as the key factors in acquisition likelihood. Further, Higgins and
Rodriguez (2006) find that a firm’s declining R&D productivity may be the underlying motivation for its acquisition of R&D intensive
firms.
We also consider a counter hypothesis that a firm’s appeal as a takeover candidate is inversely related to its R&D. The
impact from R&D alone may not be easily perceived by the marketplace. Eberhart, Maxwell, and Siddique (2004) found evidence of
R&D investments being mispriced and found that investors were slow to recognize the long term benefits from increased R&D.
Further, if firms with high R&D intensity have higher growth, they would have higher market values and would be less appealing
targets.
We hypothesize that R&D, as measured by the ratio of R&D annual expenditures in proportion to annual sales will be inversely
related with acquisition likelihood. 1

3.3. Growth options efficiency (gamma or Γ)

Since Eberhart et al. (2004) find that R&D investments may be mispriced, the management of acquiring firms may attempt to
develop their own valuation of R&D and other intangible assets when assessing potential growth options of prospective targets.
While the R&D intensities of firms measure the investment in firm growth, they do not account for how investments convert into
new growth opportunities. Firms with an exceptional ability to efficiently convert R&D into growth opportunities may be viewed
as lower risk targets. These firms may prefer to grow organically, and may prefer to remain independent rather than be absorbed
by a firm that does not extract efficiencies from its investment in growth options. In addition, it might be priced sufficiently high
because of its efficiency in exercising growth options, such that it discourages takeover attempts. Conversely, a firm that is less
efficient in exercising its growth options (as measured by a lower relative return on its R&D investment) may be more appealing
within the market for corporate control, because it can be purchased at a lower price and has much more potential for improvement.
In addition, its managers or board of directors may have difficulty avoiding takeover attempts if it has not been able to efficiently
convert its R&D into value.
Oriani and Sobrero (2008) use a measure of firm value derived from Griliches (1981) to show how R&D is positively related to firm
value. They calculate a variable, γ, as a proxy for the market valuation of R&D. We extend their approach to an analysis of growth
options where γ represents the return relative to investment from R&D. The gross return on investment from R&D expenditures,
γi, indicates the contribution of $1 of R&D investment toward the replacement of or increase in growth options value. Therefore, γi
is calculated as the present value of a firm’s growth options value, PVGO, divided by the accumulated inventory or stock of R&D
expenses (R&D Stock).2

γi ¼ PVGOi =R&D Stocki ð3Þ

The measure of PVGO used is Total Firm Value minus Assets in Place (AIP) where AIP = Book Value of Long Term debt + Book
Value of Common Equity, based on Bernardo et al. (2007).
While the degree to which a tech firm converts its R&D into growth options is not directly observable, managers of acquiring firms
know how well prospective targets perform relative to one another in capitalizing on their R&D investments. Firms in the bottom
decile in terms of converting their R&D into growth options have a γ of about 0.63 on average. Growth options for these firms are
being created at a rate below their cost of replacement, suggesting financial or at least operational distress. Conversely, tech firms
in the top decile in converting their R&D into growth options have a γ greater than 38.95. We find that a median tech firm realizes
$5.37 in PVGO for $1.00 in R&D Stock. To distinguish our proxy from the one used by Oriani and Sobrero (2008), and because γ is
significantly skewed, we use and refer to Γ, the natural logarithm of γ, going forward.

Γ i ¼ ln ðγi Þ ð4Þ

We hypothesize that high Γ firms are less likely to be acquired. A firm with high growth options efficiency would have high Γ and
would have less need to be acquired. They can remain independent, and their high Γ suggests these firms may succeed at growing
organically. In addition, they may be priced too high to attract acquirers, and may have less potential for improvement if acquired
by another firm.

1
Generally Accepted Accounting Principles (GAAP) have required R&D expenditures to be expensed in the period the costs are incurred since 1975 with few excep-
tions. However, in 1985, Statement of Financial Accounting Standards (SFAS) 86 created a standard where software R&D could be capitalized based on where the ex-
pense fell in the life-cycle of the software product (Oliver, 2003). Because some tech firms may have significant capital expenditures (CAPX), we incorporate CAPX in a
version of Γ for robustness.
2
To calculate R&D Stock, we simplify the method used by Griliches and Mairesse (1984) and Oriani and Sobrero (2008) to depreciate each year’s R&D expense by 20%
starting with the prior year. R&D Stock is therefore the sum of 100% of the trailing year’s expenses, plus 80% of the prior year’s expenses, 60% of the next prior year, etc.
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 5

4. Firm characteristics and control variables

Some firm specific characteristics have been identified in previous studies as being associated with the likelihood of being
acquired. We include control variables to account for these characteristics.

4.1. Return on equity

Firms with poor earnings performance are considered likely targets because they have suboptimal performance as measured by
return on equity. Changes would be expected eventually through new management (internal change) or an acquisition (external
change). Palepu (1986) found that low average excess return and low return on equity increased the likelihood of being acquired.
Powell (1997) and Brar, Giamouridis, and Liodakis (2009) examined the impact of return on equity (ROE) on the likelihood of
being acquired, although these studies did not focus on tech firms. We hypothesize that within the technology industry, firms with
inefficient management are more likely to be acquired. Inefficient management is proxied by low return on equity (ROE), which is
income before extraordinary items divided by common equity.

4.2. Firm liquidity, leverage, free cash flow, and size

Palepu (1986), Israel (1991), and Powell (1997) examine the effect of liquidity and leverage on the likelihood of being acquired.
We expect that low liquidity increases the likelihood of being acquired. Liquidity (LIQ) is measured as cash and short term
investments to total assets.
Israel (1991) finds that high levels of debt are defensive and significantly reduce the likelihood that a firm is acquired. However,
Powell (1997) found that leverage increased the likelihood of being acquired. Prospective targets may be less capable of continuing
independently if they have substantial debt, especially in tech industries where future cash flows tend to be unstable. We measure
leverage (LEV) as total debt (debt in current liabilities, total long term debt, and preferred stock) to total common equity.
Powell (1997) examine the role that free cash flow has in motivating acquisitions. For tech firms, access to cash is critical, and free
cash flow could have a major impact on the likelihood of being acquired. To the extent that the free cash flow of targets might help
prospective bidders absorb the cost of an acquisition, it could attract takeovers. Free cash flow (FCF) is proxied as the ratio of
EBITDA (minus interest and dividend payments) divided by sales.
Firm size tends to affect the likelihood of being a target because small firms are easier to acquire than larger firms. The time and
managerial effort needed to integrate a target into a new parent are lower, and smaller targets are more affordable. SIZE is measured
as the natural logarithm of the firm’s market capitalization in 2009 dollars. Some studies have used total assets as the proxy for size but
this may bias results in an examination of tech firms, since some industries are more capital intensive than others (e.g. telecommu-
nications versus software).
Previous studies also examine the impact of sales growth and fixed assets on the likelihood of being acquired. We tested these
variables, but found that they did not add significantly to our analysis, and therefore removed them for parsimony and space.

4.3. Periods of economic expansion

Economic expansion tends to result in rising equity values, and a perceived reduction in uncertainty. These conditions may
significantly affect the profile of prospective targets that would be most appealing to bidders. Harford (2005) finds that economic,
industrial, and technological shocks are the cause of increased merger activity. Powell and Yawson (2005) find that industry shocks
and stock market performance increase takeover activity. Dittmar and Dittmar (2008) find that merger activity is positively related to
the change in gross domestic product. Oriani and Sobrero (2008) find that market uncertainty significantly moderated the effect of
R&D on value. We hypothesize that tech firms will tend to become more appealing takeover candidates during favorable economic
conditions, and we control for this in our analysis.
GDP growth (GDP) is used to represent the level of economic expansion for this analysis. We measure the quarterly growth rate
from the seasonally adjusted annual rate provided by the Bureau of Economic Analysis. GDP is used as a continuous variable in
regressions.

5. Research methodology and design

The sample used for our study is drawn from the Standard & Poors Compustat and Securities Data Corporation Platinum (SDC)
databases. We examine a sample of all US firms that are publicly traded from January 1986 to December 2011. Because we are
predicting acquisition likelihood in the year prior to a merger’s announcement, we analyze data from 1985 to 2010. Our analysis
focuses on technology firms only. However, a sample of descriptive statistics for all firms was collected for comparison with tech firms.
The sample of target firms are compiled from the SDC database for all successful acquisitions of greater than 50% of a tech target
firm’s outstanding stock. A further constraint is that our database of SDC data must successfully merge by CUSIP with the annual
Compustat balance sheet data.3 Financial and utilities firms were removed and this yields a sample with 2,505 non-regulated mergers

3
CUSIPS were also manually matched between SDC and Compustat where possible to increase the sample.
6 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13

announced from 1986 to 2011, of which 1,166 were high tech firms. Deal characteristics are available on this sample from SDC
Platinum, but the sample drops to 927 tech mergers once these are merged with Compustat. Acquisition likelihood is then estimated
for these firms from the total sample of 44,299 firm years, and by subsamples, as noted above. We also estimated acquisition
likelihood using quarterly Compustat data and found the results were not significantly improved. These results are not shown for
space
An inverse Mills ratio test was performed on the sample after identifying two sources of potential sample bias. Sample bias could
have been introduced: 1) due to the accepted method of converting all missing values for R&D expenditures to zero; and 2) where
observations were excluded from the final regression due to missing variables in the specified model. Sample bias was not indicated,
and no Heckman correction was applied.
Logistic regressions using maximum likelihood methods were applied to predict acquisition likelihood. The likelihood ratio, score
and Wald tests are all highly significant at less than the 1% level for all models tested. Only the chi-square for the likelihood ratio is
shown for space. The basic model for the logistic regression is specified as:4

Acquiredi ¼ α þ β1 MTBi þ β2 R&Di þ β3 Γ i þ β4 ROEi þ β5 LIQ i þ β6 LEV i

ð5Þ
þ β7 FC F i þ β8 SIZEi þ β9 GDP þ β10 R&D missing i þ ε i

We also examined interaction variables for Γ with MTB and R&D, and an alternative measure of Γ that includes capital expenditures
was also examined for robustness.
We examined these variables further to predict the likelihood that a firm would make an acquisition, and this basic model is
specified as:

MadeAcquisitioni ¼ α þ β1 Acq MTBi þ β2 Acq R&Di þ β3 Acq Γ i þ β4 Acq ROEi

þ β5 Acq LIQ i þ β6 Acq LEV i þ β7 Acq FC F i þ β8 Acq SIZEi ð6Þ
þ β9 GDP þ β10 Acq R&D missing i þ εi

Multicollinearity is not indicated, but heteroscedasticity is indicated. Coefficients, standard errors and p-values are reported as
corrected for all logit models. The estimates from the logistic regressions are reported in terms of marginal effects where the sample
mean of the effects is multiplied by 100.
Tech firms can be classified into five sectors and twenty industries according to the industry trade groups TechAmerica Foundation
and the Biotechnology Industry Association. We use these classifications because they provide more delineation and industry-level
detail than relying on Fama-French or other groupings, many of which may include a blend of tech and non-tech firms.5
Logistic regression analysis is applied to calculate the likelihood of being acquired for the full sample of tech firm acquisitions and
subsamples. Subsamples are parsed by time period, SIZE, friendly and competitive, and with a high tech grouping defined by Fama
French.
The entire sample is Winsorized with outlier values being replaced with values equivalent to the 1st and 99th percentile respec-
tively for each variable. The Winsorized values are based on industry values rather than the larger sectors’ or overall sample’s values.
Pearson correlations for the variables used in the regressions are examined, and all correlations are low. The average of the absolute
values of correlations is 0.124, and no value is greater than 0.52.

5.1. Isolation of subsamples

We also apply our likelihood analysis to various subsamples, because the impact of growth options on the tech firm’s likelihood of
being acquired might be conditioned on other characteristics. First we parse our sample by time, examining the period from 1986 to
2000 and then 2001 to 2011, to determine if the period up to and including the tech boom is significantly different from the more
recent period. Second, we parse our sample by SIZE at firms less than or greater than $100 million (2009) to see if the inclusion of
very small firms biases the analysis. Third, we apply our analysis to separate subsamples of Friendly versus Competitive transactions.
For robustness, we also examined a sample of high tech firms as defined by Fama-French, as our definition of tech includes some
industries they exclude.

6. Results

Table 1 shows the number and value of mergers by time period, and the sample is further parsed by tech and non-tech. It shows
that tech M&A, as a proportion of all M&A excluding banks and utility firms, increased dramatically over time. The number of tech
M&A transactions in proportion to the number of total M&A transactions rose from 34.9% in the 1986–1993 period to 61.22%in the

4
A firm is “Acquired,” and this variable is set to “1,” when an announcement is made for an acquisition in the next calendar year. Therefore, acquisition likelihood is
predicted using balance sheet and related data in the year prior to an acquisition’s announcement, e.g. 2010 data predicts the likelihood of acquisition in 2011. The same
method is used for firms that “MadeAcquisition.”
5
Fama-French industry groups have non-tech and high-tech industries intermingled. As an example, using their 49 industry portfolios, the group “Business Services”
includes Computer Programming, Information Retrieval, and Computer Maintenance. It also includes Commercial Printing, Credit Reporting Agencies, Industrial Laun-
derers, and Trailer Rental & Leasing.
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 7

Table 1
Tech and non-tech M&A deal value by time period.

All Mergers & Acquisitions 1986–2011 ⁎ n Share of Deals Total Deal Value in $2009 (millions) Share of Value Mean StDev Median

ALL MERGERS 2,505 100.0% 4,691,632 100.0% 1,873 7,919 287

1986–1993 711 739,263 1,040 3,614 190
1994–2001 1,278 2,653,859 2,077 9,505 292
2002–2011 516 1,298,510 2,516 7,834 485
NON-TECH 1,339 53.5% 2,525,512 53.8% 1,886 7,298 369
1986–1993 463 65.1% 516,587 69.9% 1,116 3,123 268
1994–2001 676 52.9% 1,387,333 52.3% 2,052 9,106 366
2002–2011 200 38.8% 621,592 47.9% 3,108 7,155 875
HIGH TECH 1,166 46.5% 2,166,121 46.2% 1,858 8,580 205
1986–1993 248 34.9% 222,676 30.1% 898 4,389 107
1994–2001 602 47.1% 1,266,526 47.7% 2,104 9,942 213
2002–2011 316 61.2% 676,918 52.1% 2,142 8,225 347
⁎ The sample includes all completed mergers/acquisitions of public US firms by public US firms announced in the time periods shown. Firms in the financial and
utilities industries have been removed. The number of tech firms rises from 21% in 1986 to 30% by 2011. Source: Securities Data Corporation (SDC) Platinum and
Compustat.

2002–2011 period, while the value of tech M&A transactions in proportion to the value of total M&A transactions rose from 30.1% in
the 1986–1993 period to 52.1% in the 2002–2011 period. Table 2 summarizes the tech deals by sector and industry. Mergers involving
biotech, high tech manufacturing and software firms represent about 87% of all tech merger transactions.
Descriptive statistics for the variables used in our analysis are provided in Table 3. All firms are examined from 1985 to 2010 using
annual financial data to predict acquisition likelihood in the following year. For example, the financial data in 2010 are used to predict
likelihood of acquisition in 2011. Values are in the form of firm years and are provided for All Firms, Tech Firms, Tech Acquisitions and
then by subsamples. We show descriptive statistics for the subsamples used in our regression analyses: Friendly versus Competitive
transactions, transactions from 1986–2000 versus 2001–2010, and mergers by SIZE less than $100 million versus greater than $100
million market capitalization respectively.
Table 3, Panel A, shows that tech firms have significantly different characteristics from the sample of all firms. Tech Firms have sig-
nificantly higher MTB, R&D and LIQ than the overall sample of firms, and they have lower ROE and LEV. Tech firms that are acquired
have significantly lower ΓRD, MTB, and R&D than tech firms that are not acquired. The mean of R&D/Sales is 1.826 for all tech firms, but
is only 0.428 for acquired tech firms. When using R&D/Assets as a proxy (not shown in the table to conserve space), the mean is 0.18
for non-acquired tech firms and is 0.14 for acquired firms, and the difference is significant. This result is consistent with that when
using R&D/Sales as the proxy, and suggests that tech targets have significantly lower R&D expenditures than their non-acquired
tech peers.

Table 2
M&A deal value by tech sector and industry group.

M&A by High Tech Sectors and Industries – 1986–2011 n Share of Deals Total Deal Value Share of Value Mean StDev Median

ALL HIGH TECH 1,166 100.0% 2,166,121 100.0% 1,858 8,580 205
BIOTECH & PHARMACEUTICALS ⁎ 330 28.3% 830,607 38.3% 2,517 10,843 273
Biotech, Pharmaceutical, Medical Mfr'g 228 648,911 2,846 12,397 259
Biotech more broadly defined 102 181,696 1,781 6,067 304
COMMUNICATIONS SERVICES 115 9.9% 704,353 32.5% 6,125 18,576 636
Internet Services 56 66,070 1,180 1,979 251
Telecommunications Services 59 638,283 10,818 25,071 1,414
ENGINEERING AND TECH SERVICES 29 2.5% 12,043 0.6% 415 713 170
Computer Training 1 211 211 - 211
Engineering Services 16 8,381 524 860 173
R&D & Testing Labs 12 3,451 288 498 47
HIGH-TECH MANUFACTURING 375 32.2% 391,299 18.1% 1,043 3,117 194
Communications Equipment 83 75,238 906 2,576 142
Computers and Peripheral Equipment 109 156,461 1,435 4,251 200
Consumer Electronics 4 387 97 51 97
Defense Electronics 15 19,027 1,268 1,719 617
Electronic Components and Accessories 53 39,910 753 2,373 209
Measuring Instruments 46 34,096 741 3,006 179
Photonics 11 10,421 947 1,664 191
Semiconductors 54 55,757 1,033 2,521 357
SOFTWARE 317 27.2% 227,818 10.5% 719 1,895 161
Software Services 203 149,282 735 1,720 183
Systems Design & Related 114 78,537 689 2,181 140
⁎ Values are millions in $2009. Deal values are from SDC Platinum. Biotechnology is defined "narrowly" or "broadly" by industry. Biotech research and manufacturing
for human health is generally classified as "narrow biotech." A small portion of the sector is broadly defined as biotech. These industries include manufacturing organic
oils and chemicals, fertilizer, pesticides, and medical labortories. Source: TechAmerica Foundation and Biotechnology Industry Association.
 8
Table 3
Descriptive statistics for the independent variables examined.

Panel A All Firm Year Observations Tech Firm Acquisitions in the Year Prior to Announcement

S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13
All Firm Years Tech Firm Years All Tech Acquisitions Friendly Competitive

n 163,892 44,299 927 711 216

med (1) mean st dev med (2) mean st dev t-test (1)-(2) med (3) mean st dev t-test (2)-(3) med (4) mean st dev med (5) mean st dev t-test (4)-(5)

ΓRD n.m. n.m. n.m. 1.756 1.821 1.539 n.m. 1.399 1.529 1.506 *** 1.346 1.487 1.503 1.629 1.668 1.511
MTB 0.392 0.582 0.780 0.669 0.839 0.877 *** 0.520 0.648 0.671 *** 0.528 0.639 0.667 0.505 0.676 0.688
R&D/Sales 0.000 0.624 8.248 0.091 1.826 14.172 *** 0.097 0.428 3.652 *** 0.100 0.498 4.134 0.087 0.185 0.502
ROE 0.069 −0.033 1.520 0.043 −0.121 2.140 *** 0.036 −0.211 1.671 0.024 −0.265 1.762 0.070 −0.021 1.293 **
LIQ 0.086 0.188 0.233 0.227 0.304 0.271 *** 0.229 0.290 0.248 * 0.223 0.293 0.252 0.237 0.282 0.235
LEV 0.187 1.024 3.414 0.053 0.499 2.353 *** 0.061 0.420 1.400 0.064 0.456 1.543 0.056 0.295 0.712
FCF 0.040 0.082 0.118 0.000 0.071 0.101 *** 0.056 0.082 0.094 *** 0.040 0.074 0.090 0.087 0.107 0.105 ***
SIZE 4.660 4.723 2.425 4.532 4.631 2.346 *** 4.964 4.994 1.981 *** 4.845 4.834 1.985 5.249 5.542 1.871 ***

Panel B Acquired Tech Firms by Firm Size Acquired Tech Firms by Time Period

SIZE b $100mil SIZE N =$100mil 1985 to 2000 2001 to 2010

n 365 562 646 281

med (6) mean st dev med (7) mean st dev t-test (6)-(7) med (8) mean st dev med (9) mean st dev t-test (8)-(9)

ΓRD 0.767 0.888 1.438 1.822 1.945 1.400 *** 1.402 1.572 1.513 1.393 1.431 1.487
MTB 0.301 0.401 0.612 0.692 0.829 0.655 *** 0.492 0.645 0.687 0.565 0.653 0.634
R&D/Sales 0.097 0.629 5.404 0.097 0.272 0.979 * 0.084 0.322 1.363 0.130 0.682 6.380
ROE −0.063 −0.403 2.265 0.065 −0.060 0.961 *** 0.043 −0.242 1.752 0.021 −0.137 1.461
LIQ 0.191 0.272 0.254 0.251 0.304 0.243 ** 0.199 0.265 0.243 0.310 0.349 0.251 ***
LEV 0.128 0.705 2.024 0.036 0.211 0.543 *** 0.070 0.379 0.995 0.046 0.516 2.053
FCF 0.000 0.043 0.068 0.098 0.112 0.101 *** 0.062 0.083 0.094 0.034 0.078 0.095
SIZE 3.428 3.188 1.079 5.940 6.318 1.344 *** 4.792 4.832 1.963 5.422 5.374 1.975 ***

The sample includes all firms after financial and utilities firms are removed, tech firms and acquired tech firms where annual Compustat data sucessfully merges with SDC Platinum data. Acquisitions are defined as those announced
in the period from 1986 to 2011 that successfully closed then or later. The Compustat data and ratios shown are from 1985−2010 because acquisition likelihood will be estimated based on available data from the year prior to the
announcement. Values shown are the logarithmic values used in regressions for Γ, MTB, and SIZE. The 927 tech acquisitions above are lower than the 1166 acquisitions shown in Table 2 due to missing balance sheet data for some
observations once SDC data is merged with Compustat. This sub-sample of tech acquisitions represents those that are the focus of the analysis to follow.
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 9

Table 4
Logistic regression analysis of the likelihood of being acquired for tech firms from 1986 to 2011.

DV = 0 41,508 DV = 0 41,508

DV = 1 925 DV = 1 925

Parameter Est. t val. Est. t val.

Intercept −6.282 −34.46*** −6.353 −34.43***

MTB −0.198 −1.21 −0.440 −2.61***
R&D −0.038 −1.45 −0.244 −2.96***
ΓRD −0.418 −5.51*** −0.364 −4.79***
ΓRD*HighMTB 0.107 1.26
ΓRD *R&D −0.086 −2.47**
ROE −0.059 −1.53 −0.060 −1.50
LIQ 0.913 2.65*** 0.559 1.73*
LEV 0.035 0.58 0.023 0.36
FCF −0.728 −0.79 −0.281 −0.33
SIZE 0.203 5.30*** 0.192 5.07***
GDP 44.352 8.42*** 43.682 8.31***
R&D_missing 0.637 1.24 0.581 1.17
Likelihood Ratio χ2 212.3*** χ2 211.3***

The logistic regression predicts the likelihood of a firm making an acquisition in the year to follow. The dependent variable (DV) equals 1 when an acquisition is made,
0 when not, for all tech firms in the sample over the period from 1986 to 2011. The estimates from the regression are reported in terms of marginal effects where
the sample mean of effects is multiplied by 100. Heteroscedasticity in logistic regressions has been corrected where indicated. Significance is reported as ***, **, and
* for p-values b0.01, b0.05, and b0.10 respectively.

Competitively targeted tech firms have significantly higher ROE and FCF than friendly target firms, and they are also significantly
larger. In Table 3, Panel B, acquired firms are parsed into subsamples by size and time period. Tech targets below $100 million in
market capitalization exhibit some dramatically different characteristics than larger firms. They have lower mean values of ΓRD,
MTB, ROE, LIQ and FCF, while having higher LEV. The panel comparing tech targets over time shows that tech firms tend to have higher
LIQ and SIZE in the most recent subperiod.
Table 4 shows the results for logistic regressions on the likelihood of a tech firm being acquired from 1986 to 2011. The logistic
regression analysis assesses the tech firm’s financial data in a given year to predict the likelihood that the firm will be targeted for
acquisition during the following year. Therefore, accounting data from 1985 to 2010 is used in the analysis. Table 4 shows the results
from applying our model to the entire sample of tech firms. The first model uses MTB, R&D, and ΓRD as the key growth options
variables, while the second model also includes interaction terms for ΓRD with MTB and, R&D.
Table 4 shows that the coefficient of MTB is negative in both models. It is significant in the second model, which is consistent with
the univariate results described earlier, and offers evidence that tech firms with high valuations are less likely to be acquired. This
result is also consistent with the notion that firms are more vulnerable to a takeover if they are distressed or at the very least are
out of favor with investors.
The coefficient of R&D is negative in both models, and significant in the second model when interaction terms are included. These
results suggest that a firm’s R&D intensity by itself does not attract bidders, and may even discourage bidders. This is consistent with
prior literature, except for Heeley et al. (2006) and Desyllas and Hughes (2009) who found a positive relationship between R&D and
acquisition probability. We suggest that the differences in our methodology and theirs may account for the different results.6
The coefficient of ΓRD (our measure of growth options efficiency) is negative and significant in both models, which suggests that
firms that generate a high return relative to their R&D are less likely to be acquired. High ΓRD firms may be able to fend off acquisition
attempts and/or may prefer to grow organically. Recall that tech acquisitions were shown in the descriptive statistics in Table 3 to
have significantly lower ΓRD than other tech firms. This result is reinforced by the inverse relationship between ΓRD and likelihood
of acquisition in Table 4 even when controlling for other characteristics.
Further, as mentioned above in our development of Γ, the worst performing firms with regard to ΓRD have a relative return below
the cost of replacement. These firms are so inefficient at converting their R&D into growth options that they have negative ΓRD. These
distressed, low growth options firms represent roughly ten percent of our sample, and our results in Table 4 show these negative ΓRD
firms are significantly more likely to be acquired.
In addition, the second model in Table 4 tests the interaction between ΓRD and R&D, and the coefficient is negative and significant.
That is, firms that have a high level of R&D and generate a high return relative to their R&D have a lower likelihood of being acquired,
even after controlling for the individual effects of both variables. Firms with a large investment in R&D that generate high returns on
that investment may be especially willing to grow organically. Alternatively, they may be priced unusually high because of their
success at utilizing their investment in R&D, and do not appeal to those bidders that are hoping to expand by acquiring a low cost

6
Heeley et al. (2006) sample from all firms, not limiting their analysis to tech, and their matched sample was matched by size and explicitly not by industry. This
suggests their method may have captured the likelihood that a tech firm will be acquired over a non-tech firm, irrespective of R&D. Desyllas and Hughes (2009) focus
on patent stock and patent intensity in their analysis, and they use R&D/Assets which we find to be insignificant. Further, our analysis includes data through 2011 while
the samples of the prior papers end in the years 2000 and 1998 respectively.
10 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13

target. Conversely, tech firms that either have low investment in R&D or have generated weak returns relative to their R&D may have
more potential to be restructured and improved by acquirers. These types of tech firms may also be purchased at a lower cost.
Regarding the firm-specific control variables, the coefficient of ROE is negative in both models as expected, but falls short of being
significant. The coefficient of LIQ is positive and significant, suggesting that more liquid tech firms attract acquirers. The coefficient of
SIZE is positive and significant in both models, which suggests that larger tech firms are more likely to be acquired. The coefficient of
GDP is positive and significant in both models, suggesting that tech firms are more likely to be acquired during periods of relatively
high economic growth. The remaining control variables are not significant.
Table 5 provides results from applying our model containing the interaction terms to subsamples separated by period and firm
size. Results from applying the model to the earlier subperiod are disclosed in columns 1 and 2, while results from applying the
model to the more recent subperiod are disclosed in columns 3 and 4. The coefficient of MTB is negative when the model is applied
to either subperiod, but only significant when the model is applied to the earlier subperiod. The coefficient of R&D is negative and sig-
nificant when applying the model to either subperiod, which is consistent with results in Table 4 that represent the entire sample.
The coefficient of our measure of growth options efficiency ΓRD, is negative and significant when the model is applied to either
subperiod, which is consistent with our results in Table 4 that represent the entire sample. These results suggest that the inverse
relationship between the relative return on R&D and the likelihood of being acquired is maintained in both subperiods. High ΓRD
firms in tech industries appear able to fend off acquisition attempts and/or may prefer to grow organically, regardless of the period
assessed.
Regarding the interaction terms, the coefficient of the interaction term ΓRD*MTB is not significant in models applied to either
subperiod. In addition, the coefficient of the interaction term ΓRD*R&D is negative and significant when we examine the early
subperiod, but it is not significant when applied to the more recent subperiod.
Regarding the control variables, the coefficient of SIZE is positive and significant when the model is applied to either subperiod.
These results are consistent with those disclosed in Table 4 for the entire sample. The coefficient of GDP is positive and significant
in the model applied to either subperiod which is consistent with the results representing the entire sample in Table 4, and suggests
that increased economic activity increases the likelihood of being acquired regardless of the subperiod.
We also parse the sample by firm SIZE, and find largely similar results. The coefficients of R&D and ΓRD remain negative and
significant, implying that tech firms with a high investment in R&D and high returns relative to that investment are less likely to be
acquired. These relationships hold for both categories of tech firms segmented by size, and these results are consistent with those
found for the entire sample in Table 4.
We also find that the coefficient of the interaction term ΓRD*MTB is negative and significant for smaller tech firms, which suggests
that the inverse relationship between ΓRD and the likelihood of being acquired is especially pronounced for tech firms that have a high
valuation (based on MTB). However, this coefficient is positive and significant for larger tech firms. One plausible explanation is that
prospective acquirers that can afford to acquire relatively large tech firms are less focused on targets that exhibit cheap valuations.
Regarding the control variables, LEV and SIZE are positive and significant for tech firms in the small size category, which suggests
that more highly levered and larger firms are more likely to be targeted within this category. However, the coefficient of SIZE is neg-
ative and significant for tech firms in the large size category, suggesting that relatively small firms within this subsample are more
likely to be acquired. This is consistent with the view that very large firms are less likely to be acquired, perhaps because many

Table 5
Likelihood of being acquired – parsed by time period and size.

Period from 1986–2000 Period from 2001–2011 Firm Size b $100 million Firm Size N =$100 million
($2009) ($2009)

DV = 0 23,571 DV = 0 17,937 DV = 0 19,647 DV = 0 21,861

DV = 1 647 DV = 1 278 DV = 1 363 DV = 1 562

Parameter Est. t val. Est. t val. Est. t val. Est. t val.

Intercept −6.057 −21.75*** −6.538 −22.19*** −7.250 −22.64*** −4.280 −12.27***

MTB −0.537 −2.10** −0.256 −1.23 −0.229 −1.00 −0.330 −1.24
R&D −0.259 −2.58*** −0.393 −2.75*** −0.184 −2.08** −0.345 −2.07**
ΓRD −0.435 −3.90*** −0.287 −2.95*** −0.295 −3.30*** −0.473 −3.65***
ΓRD*HighMTB 0.137 1.08 0.023 0.21 −0.285 −2.08** 0.280 2.23**
ΓRD *R&D −0.127 −2.78*** 0.057 0.84 −0.068 −1.75* −0.088 −0.97
ROE −0.109 −1.87* −0.006 −0.11 −0.058 −1.45 −0.073 −0.76
LIQ 0.742 1.54 0.801 1.98** 0.286 0.69 −0.476 −0.93
LEV −0.034 −0.29 0.045 1.01 0.095 2.38** −0.029 −0.11
FCF 0.141 0.11 −1.266 −1.18 −0.369 −0.24 0.030 0.03
SIZE 0.230 4.00*** 0.190 4.03*** 0.492 5.22*** −0.264 −3.28***
GDP 44.084 3.92*** 13.895 2.50** 37.356 5.19*** 44.839 5.92***
R&D_missing 0.476 0.64 0.848 1.40 0.128 0.19 1.403 1.88*
Likelihood Ratio χ2 108.7*** χ2 70.3*** χ2 148.6*** χ2 119.6***

The logistic regression predicts the likelihood of a firm making an acquisition in the year to follow. The dependent variable (DV) equals 1 when an acquisition is made,
0 when not, for all tech firms in the sample over the period from 1986 to 2011. The estimates from the regression are reported in terms of marginal effects where
the sample mean of effects is multiplied by 100. Heteroscedasticity in logistic regressions has been corrected where indicated. Significance is reported as ***, **, and
* for p-values b0.01, b0.05, and b0.10 respectively.
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 11

potential acquirers may not be able to afford the takeover or execute the integration of operations of such large firms with their own
operations. Our findings of a conditional effect of size on the likelihood of tech firms being acquired may explain why results for the
size effect are mixed in prior studies testing acquisition likelihood.
Table 6 discloses results from applying our model to separate subsamples of mergers in which firms are targeted in a friendly
(columns 1 and 2) or competitive (columns 3 and 4) manner. The coefficient of the MTB variable and R&D variable is negative in
both subsamples, but only significant in the subsample of friendly acquisitions. The coefficient of ΓRD remains negative and significant,
regardless of the manner by which firms are targeted, which is consistent with the results for the entire sample and other subsamples.
The coefficient of the interaction term ΓRD*MTB is not significant in either subsample. The coefficient of the interaction term
ΓRD*R&D is negative and significant when the model is applied to the friendly acquisitions, which suggests a more pronounced inverse
relationship between the ability to generate returns from R&D and the likelihood of being acquired for tech firms with a relatively high
R&D intensity.
Regarding the control variables, the coefficient of SIZE and GDP remains positive and significant when applied to subsamples of
friendly or competitive acquisitions, consistent with the findings for most other subsamples shown in the previous tables.
We also apply a robustness test for whether our definition of “tech” affects the outcome for our growth options measure ΓRD. The
Fama-French industry classification system categorizes firms in a different manner than the process we use (as described earlier), as it
excludes biotechnology and includes some other low R&D intensive industries. Results from applying our model to the firms defined
as tech by the Fama-French classification system are shown in columns 5 and 6 of Table 6. In general, results are robust to the
classification system for identifying tech firms.
We also test an alternative measurement for ΓRD, called ΓRD_CAPX, which includes R&D and capital expenditures in the Γ calculation.
Some R&D can be capitalized, and some tech firms (such as telecommunications and pharmaceutical firms) have significant capital
expenditures. Results when applying our model and using this variable instead of ΓRD are disclosed in columns 7 and 8 of Table 6.
The coefficient of ΓRD_CAPX is negative and significant, similar to the results derived when using ΓRD as the proxy for the firm’s ability
to generate returns relative to its R&D. In addition, the coefficient of the interaction term between ΓRD_CAPX and MTB is negative and
significant, which implies that tech firms with a high ΓRD_CAPX are especially insulated from bidder takeover attempts when they
exhibit a high valuation (as measured by MTB).
Up to this point, our results are focused on the probability that a tech firm is acquired. Table 7 tests whether our measure of growth
options efficiency (ΓRD or ΓRD_CAPX) can predict the likelihood of a tech firm making an acquisition. Columns 1 and 2 disclose results
when testing ΓRD. The MTB and R&D variables are not significant. However, just as tech firms with a high ΓRD are less likely to be
acquired (as was shown in Tables 4 – 6), they are also less likely to acquire other firms. Because Γ is a proxy for the firm’s ability to
generate returns from its R&D investments, these results indicate that high Γ firms may prefer to grow organically. Conversely, low
ΓRD firms may be more willing to acquire tech targets because they are unable to internally produce their own growth options. This
inference for the likelihood of acquiring is similar to the inference drawn for prospective targets.

Table 6
Likelihood of being acquired – friendly versus competitive targeting, Fama French definitions of high tech, and Γ based on both R&D and CAPX as inputs to growth
options value.

Friendly Acquisitions Competitive Acquisitions Fama French - High Tech Growth options from R&D
Grouping and CAPX

DV = 0 41,508 DV = 0 42,219 DV = 0 29,289 DV = 0 48,648

DV = 1 711 DV = 1 214 DV = 1 634 DV = 1 1,107

Parameter Est. t val. Est. t val. Est. t val. Est. t val.

Intercept −6.320 −31.66*** −10.13 −21.50*** −6.383 −28.99*** −6.482 −39.36***

MTB −0.369 −2.51** −0.060 −0.68 −0.075 −0.37 0.433 2.15**
R&D −0.154 −2.28** −0.146 −0.93 −0.535 −1.04 −0.050 −1.83*
ΓRD −0.249 −3.71*** −0.120 −3.17*** −0.328 −3.57***
ΓRD*MTB 0.093 1.22 0.006 0.16 −0.122 −1.19
ΓRD *R&D −0.056 −2.22** −0.061 −1.06 −0.376 −2.05**
ΓRD_CAPX −0.435 −5.37***
ΓRD_CAPX*MTB −0.189 −3.00***
ΓRD_CAPX*R&D −0.020 −1.23
ROE −0.058 −1.77* 0.009 0.33 −0.067 −1.38 −0.054 −1.53
LIQ 0.513 1.83* 0.107 0.63 1.066 2.70*** 1.226 3.94***
LEV 0.030 0.63 −0.039 −0.69 0.026 0.44 −0.005 −0.13
FCF −1.289 −1.63 0.707 1.83* −0.987 −0.96 −1.734 −2.18**
SIZE 0.116 3.43*** 0.079 4.39*** 0.187 4.19*** 0.212 6.24***
GDP 24.800 5.68*** 23.144 6.91*** 38.886 6.36*** 42.393 8.74***
R&D_missing −0.289 −0.55 0.591 3.28*** 0.345 0.62 0.455 2.27**
Likelihood Ratio χ2 129.6*** χ2 144.8*** χ2 136.1*** χ2 258.6***

Model 3 and 4 are provided for robustness. High Tech as defined by Fama French excludes many industries we include in our analysis, like Biotechnology.
The logistic regression predicts the likelihood of a firm making an acquisition in the year to follow. The dependent variable (DV) equals 1 when an acquisition is made,
0 when not, for all tech firms in the sample over the period from 1986 to 2011. The estimates from the regression are reported in terms of marginal effects where
the sample mean of effects is multiplied by 100. Heteroscedasticity in logistic regressions has been corrected where indicated. Significance is reported as ***, **, and
* for p-values b0.01, b0.05, and b0.10 respectively.
12 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13

Table 7
Likelihood that a public US tech company will acquire another public US tech firm in the following year.

Likelihood of Acquiring Firms Making An Acquisition

DV = 0 21,730 DV = 0 24,913

DV = 1 686 DV = 1 778

Parameter Est. t val. Est. t val.

Intercept −10.03 −29.54 −10.39 −32.43***

Acq_MTB −0.073 −0.17 −0.721 −1.72*
Acq_R&D −0.255 −1.35 −0.153 −1.07
Acq_ΓRD −0.431 −2.94
Acq_ΓRD*Acq_HighMTB 0.010 0.09
Acq_ΓRD *Acq_R&D −0.029 −0.41
Acq_ΓRD_CAPX 0.561 3.38***
Acq_ΓRD_CAPX*Acq_HighMTB −0.128 −1.17
Acq_ΓRD_CAPX*Acq_R&D −0.039 −0.73
Acq_ROE 0.049 0.30 0.084 0.62
Acq_LIQ −0.656 −0.98 −0.028 −0.05
Acq_LEV −1.287 −2.14 −0.664 −1.68*
Acq_FCF 3.775 3.20 1.400 1.41
Acq_SIZE 1.393 19.19 1.306 20.50***
GDP 57.445 6.79 53.592 7.19***
Acq_R&D_missing 2.377 2.89 −0.580 −1.77*
Likelihood Ratio χ2 625.3 χ2 673.8***

Fewer acquiring firms are in this sample than the previous regressions because 1) acquiring firms are limited to tech firms to examine the potential impact of Γ on
acquisition likelihood and many tech firms are acquired by non-tech firms, 2) acquiring firms are limited to $100 million or larger in size, and 3) some observations
are lost in the merge between SDC and Compustat data.
The logistic regression predicts the likelihood of a firm making an acquisition in the year to follow. The dependent variable (DV) equals 1 when an acquisition is made,
0 when not, for all tech firms in the sample over the period from 1986 to 2011. The estimates from the regression are reported in terms of marginal effects where
the sample mean of effects is multiplied by 100. Heteroscedasticity in logistic regressions has been corrected where indicated. Significance is reported as ***, **, and
* for p-values b0.01, b0.05, and b0.10 respectively.

Columns 3 and 4 of Table 7 show results from testing Acq_ΓRD_CAPX, the measure of growth options efficiency in which capital
expenditures are included. For this model, the coefficient is positive and significant. The effect of Γ in this model is moderated by
MTB, which is negative and significant in predicting acquisition likelihood. A comparison of the results for the measure of growth
options efficiency suggests that the relationship between growth options efficiency in using R&D and propensity to acquire is
conditioned on how R&D is defined. While tech firms that are efficient in using only R&D (without capital expenditures) are less
willing to expand through acquisitions, tech firms that are efficient in using R&D and capital expenditures are more willing to expand
by acquisition. This suggests that tech firms that are more capital intensive may approach the market for corporate control differently
than do more research intensive tech firms.
We examined subsamples (not shown) by SIZE and for when “pooling of interests” was allowed pre-2001 and later when only
purchase accounting rules were in place, and the results for Acq_ΓRD_CAPX do not significantly change. However, an analysis of tech
sectors (also not shown) indicates the result for Acq_ΓRD_CAPX is not uniform across all tech firms. Biotechnology and communications
firms have a positive and significant result for Acq_ΓRD_CAPX whereas software firms are negative and significant. This may be an area
for further research, but it is beyond our primary focus of inquiry.
Among the control variables, the coefficient of Acq_LEV is negative and significant, which implies that more highly levered tech
firms are less likely to pursue acquisition. The coefficient of Acq_FCF is positive in both models, and significant in one of them, offering
some evidence that tech firms with high free cash flow are more likely to pursue acquisitions. The coefficient of GDP is positive and
significant in both models, which suggests that the propensity of tech firms to acquire targets is positively related to economic
conditions (consistent with the results found for likelihood of being acquired).

7. Summary

More than one half of all acquisitions involve tech firms, yet, there is very limited research on the profile of tech firms that attract
takeovers. Our goal is to examine how growth option characteristics influence the appeal of tech firms as targets. We develop and
examine a new proxy, ΓRD that measures the return in growth options relative to investment in R&D. To our knowledge, this proxy
has not been given attention in merger literature. It is especially appropriate for investigating mergers among tech firms. Since the
ability to deliver valuable products and services is captured in a firm’s valuation, a portion of firm value is due to its growth options.
We find that Γ is inversely related to the likelihood of being acquired. Robustness tests show that this relationship holds regardless
of the subperiod assessed, the size category assessed, whether tech firms are engaged friendly or competitive acquisitions, the system
used to identify tech firms, and whether the R&D definition includes capital expenditures. The relationship between Γ and the likeli-
hood of being acquired is even more pronounced when tech targets have a relatively high valuation (based on the market-book ratio).
In general, tech firms with a high Γ appear to prefer organic growth rather than expansion by combinations with other tech firms.
 S.M. Davis, J. Madura / Journal of High Technology Management Research 26 (2015) 1–13 13

8. Implications

Among the hundreds of acquisitions examined, and the thousands of potential targets and potential acquirers, the one overriding
and highly significant factor to determine whether a high technology firm will be acquired or will make an acquisition is its ability to
convert its own R&D into usable and valuable growth options. Tech firms that use their R&D investments efficiently are less likely to
merge with other firms. They are more likely to choose organic growth than either seeking or allowing themselves to be acquired.
Since we use a market-based proxy for growth options value, this is consistent with prior literature finding overvalued firms are
less likely targets. However, our study adds to the literature by considering both how firms are valued (MTB) and how firms have
converted their R&D into growth options (Γ) in assessing acquisition likelihood. When both factors are considered at the same
time, firm valuation becomes less significant or completely insignificant, while the efficiency with which a company converts its
R&D into growth options is highly significant in every model examined. Further, firms that have a high Γ may attempt to avoid
merging their R&D operations with other less efficient R&D operations that cannot be easily corrected. Inefficient R&D operations
might not only lead to the exercise of fewer valuable growth options, but might also result in fixed expenses (existing salaries,
labs, and related expenses) that cannot be easily reduced.
An obvious tradeoff is that organic growth tends to be slower than growth resulting from mergers. However, tech firms that have
efficient investment in R&D may be willing to accept a relatively slower form of growth, rather than pursue faster growth that could
disrupt efficiency, in order to avoid the time and cost of restructuring other R&D divisions that were not developed organically.

9. Possible research extensions

Our study demonstrates that when assessing mergers, results from general studies on all manufacturing firms cannot necessarily
be used to generalize about tech firms. In a similar manner, general results on a subsample of tech firms cannot necessarily be used to
generalize about each type of tech firm. Future research is needed to understand how the influence of Γ on mergers varies among tech
firms. In particular, the desire to achieve more efficient investment in R&D by growing organically might be more crucial in some types
of tech firms. Conversely, the need for economies of scale in R&D might require a firm to grow its R&D division by merger. However,
more research is needed to understand whether and how the tech firm's governance structure could ensure proper assessment of R&D
synergies before a merger is consummated. In addition, more research is needed to understand how to measure potential synergies
(or lack thereof) from the merging of R&D divisions of different tech firms. Further research is also needed to assess how Γ may
determine the success or failure of a high tech merger.

References

Andres-Alonso, P., Azofra-Palenzuela, V., & Fuente-Herrero, G. (2006). The real options component of firm market value: The case of the technological corporation.
Journal of Business Finance & Accounting, 33(1), 203–219.
Berk, Jonathan B., Green, Richard C., & Naik, Vasant (1999). Optimal investment, growth options, and security returns. Journal of Finance, 54(5), 1553–1607.
Bernardo, Antonio E., Chowdhry, Bhagwan, & Goyal, Amit (2007). Growth options, beta, and the cost of capital. Financial Management, 36(2), 5–17.
Brar, Gurvinder, Giamouridis, Daniel, & Liodakis, Manolis (2009). Predicting European takeover targets. European Financial Management, 15(2), 430–450.
Cao, Charles, Simin, Timothy, & Zhao, Jing (2008). Can growth options explain the trend in idiosyncratic risk? Review of Financial Studies, 21(6), 2599–2633.
Coad, Alex, & Rao, Rekha (2008). Innovation and firm growth in high-tech sectors: A quantile regression approach. Research Policy, 37, 633–648.
Desyllas, Panos, & Hughes, Alan (2009). The revealed preferences of high technology acquirers: An analysis of the innovation characteristics of their targets. Cambridge
Journal of Economics, 33, 1089–1111.
Dittmar, Amy K., & Dittmar, Robert F. (2008). The timing of financing decisions: An examination of the correlation of financing waves. Journal of Financial Economics, 90,
59–83.
Eberhart, Allan C., Maxwell, William F., & Siddique, Akhtar R. (2004). An examination of long-term abnormal stock returns and operating performance following R&D
increases. Journal of Finance, 59(2), 623–650.
Griliches, Zvi (1981). Market value, R&D and patents. Economic Letters, 7(2), 183–187.
Griliches, Zvi, & Mairesse, Jacques (1984). Productivity and R&D at the firm level. R&D, patents, and productivity. Chicago, IL: University of Chicago Press, 339–374.
Harford, Jarrad (2005). What drives merger waves? Journal of Financial Economics, 77, 529–560.
Heeley, Michael B., King, David R., & Covin, Jeffrey G. (2006). Effects of firm R&D investment and environment on acquisition likelihood. Journal of Management Studies,
43(7), 1513–1535.
Higgins, Matthew J., & Rodriguez, Daniel (2006). The outsourcing of R&D through acquisitions in the pharmaceutical industry. Journal of Financial Economics, 80,
351–383.
Ho, Yew Kee, Tjahjapranata, Mira, & Yap, Chee Meng (2006). Size, leverage, concentration, and R&D investment in generating growth opportunities. Journal of Business,
79(2), 851–876.
Israel, Ronen (1991). Capital structure and the market for corporate control: The defensive role of debt financing. Journal of Finance, 46(4), 1391–1409.
Lantz, Jean-Sebastien, & Sahut, Jean-Michel (2005). R&D investment and the financial performance of technological firms. International Journal of Business, 10(3),
251–270.
Massa, Massimo, & Zhang, Lei (2009). Cosmetic mergers: The effect of style investing on the market for corporate control. Journal of Financial Economics, 93(3),
400–427.
Oliver, Joseph R. (2003). Accounting and tax treatment of R&D: An update. CPA Journal, 73(7), 46–49.
Oriani, Raffaele, & Sobrero, Maurizio (2008). Uncertainty and the market valuation of R&D within a real options logic. Strategic Management Journal, 29, 343–361.
Palepu, Krishna G. (1986). Predicting takeover markets: A methodological and empirical analysis. Journal of Accounting Economics, 8, 3–35.
Powell, Ronan G. (1997). Modelling takeover likelihood. Journal of Business Finance & Accounting, 24(7), 1009–1030.
Powell, Ronan G., & Yawson, Alfred (2005). Industry aspects of takeovers and diverstitures: Evidence from the UK. Journal of Banking and Finance, 29, 3015–3040.
Trigeorgis, L. (1993). Real options and interactions with financial flexibility. Financial Management, 22(3), 202–224.
Tsai, Kuen-Hung, & Wang, Jiann-Chyuan (2005). Does R&D performance decline with firm size? A re-examination of terms of elasticity. Research Policy, 34, 966–976.