

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

New Ventures, Internationalization, and Asymmetric Grin Curve:

Analysis of Taiwan’s Big Data

Jwu-Rong Lin
1
, Chen-Jui Huang

2,*
, Ching-Yu Chen

1
, Huey-Ling Shiau

2
, Yan-Chen Yeh

1
Department of International Business, Tunghai University, Taichung, Taiwan, ROC

2
Department of Finance, Tunghai University, Taichung, Taiwan, ROC

Received 02 March 2018; received in revised form 18 July 2018; accepted 04 September 2018

Abstract

Under globalization, small and medium enterprises (SMEs) that predominate Taiwan’s economy have been

primarily original equipment manufacturers (OEMs) continuing to adjust operating strategies in order to extend

supply chains and enhance competitiveness. This paper adopts the big data composed of 104,377 Taiwanese

manufacturers from the 2011 Industry, Commerce, and Service Census to assess impact of the business life cycle,

brand revenue, R&D spending, and internationalization on value creation. Major findings are as follows. First, the

link of the firm’s operating years with value creation is characterized by a quadratic U-shaped curve where the

minimum point corresponds to 15 years of operation, suggesting a cost of lower value added for new ventures at the

early stage of development. Second, a reversed U-shaped curve of value creation is found as regards brand revenue

and R&D spending, with the greater impact of the latter. Third, the impact of overseas investment and export

expansion is also captured by a reversed U-shaped curve, with greater impact for the former. Fourth, an asymmetric

grin curve rather than a smile curve is found in Taiwan’s manufacturers, whose value creation can be strengthened by

strategies that focus on the learning curve, internationalization, internet, operating scale, and capital intensity.

Keywords: new venture, internationalization, grin curve, value creation

1. Introduction

The emergence of the Regional Comprehensive Economic Partnership (RCEP), which enlarges the trade link of the ten

ASEAN members and six Asian countries, marks a new challenge as regards the cost advantage that most original equipment

manufacturers (OEMs) in Taiwan have benefited over past decades. These firms are now forced to develop strategies that

transform the current cost-based industry into the one with high value added through new ventures, brand benefit, R&D

spending, and internationalization.

However, the 2011 Industry, Commerce, and Service Census conducted by Taiwan’s Directorate-General of Budget,

Accounting, and Statistics shows that the business operating years ranges from 1 to 100 and averages at 18 among the 104,377

firms surveyed. This seems to imply significant discrepancy in terms of the business life cycle. In addition, the number of firms

that have been engaged in brand, R&D, overseas investment, and export activities are only 12,440, 10,062, 14,326, and 17,206,

respectively accounting for 11.92%, 9.64%, 13.73%, and 16.49% of the total sample and leaving the ratio of the value added to

total revenue at 36% only. This paper intends to deepen the issue on insufficient value creation observed in most small and

medium enterprises (SMEs) that dominate Taiwan’s manufacturing sectors and discuss relevant strategies for transformation

in the business model.

Corresponding author. E-mail address: cjhuang@thu.edu.tw.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

The remainder of this paper is structured as follows. Section 2 succinctly reviews theoretic foundations and relevant

empirical evidence. Section 3 proposes the empirical model and hypotheses to be tested. Section 4 analyzes primary regression

results and checks robustness of the empirical model. Section 5 concludes with discussion on research limitations and

suggestion for future studies.

2. Theoretic Foundations and Literature Review

The smile curve characterizes the relation between value creation and supply chains. The value added is expressed on the

vertical axis, whereas the supply chains from the upstream to the downstream are expressed from the left to the right on the

horizontal axis. Over the supply chains, upstream, midstream, and downstream firms respectively play the role of original

design manufacturers (ODMs), original equipment manufacturers (OEMs), and own branding manufacturers (OBMs). In Less

Developing Countries (LDCs), the smile curve appears reversed, called the forced smile curve. In between, the grin curve

raises the two ends of the smile curve and appears flatter to reflect the evolution in industry competition. The typical smile

curve, forced smile curve, and grin curve are illustrated in Fig. 1.

Fig. 1 Smile curve, forced smile curve, and grin curve

New ventures have grown rapidly over the past decades. Lussier [1] defines the new ventures as businesses established

within 10 years. In Taiwan, the Start-Up and Incubation Center of the Small and Medium Enterprise Administration of the

Ministry of Economic Affairs adopts instead a definition of 5 years, also applied to associated start-up loans open to the youth.

But its Business Start-Up Award targets businesses of an age of not more than 3 years. Across relevant studies, businesses

established within 3 to 14 years exhibit essential characteristics of new ventures. As this research employs data from the

Industry, Commerce, and Service Census conducted every 5 years by Taiwan’s Directorate-General of Budget, Accounting,

and Statistics, subsequent analysis will define new ventures by an age of not more than five years. Businesses that have

continuously operated over more than 5 years are regarded as survivors.

Marco et al. [2] examine a sample of Italian companies between 1982 and 1992. Firm characteristics before and after the

public initial offerings (IPO) are compared in order to find determinants for listing on exchanges. The authors adopt the

competitive theory to distinguish cost-side and benefit-side determinants and classify new ventures and survivors by firm age.

The empirical evidence indicates lower capital demand reflected by a lower debt ratio for new ventures as these businesses

have shorter operations and/or funds provided by establishing shareholders are able to meet the demand. Erel et al. [3] turn to a

larger sample of European companies involved with mergers and acquisitions from 2001 to 2008 and investigate the sensitivity

of cash holdings and investment to cash flows. They find that it is the survivor rather than the new venture which significantly

increase cash holdings one year after the IPO. In addition, positive sensitivity of investment to cash flows is found among new

ventures, which implies that investment by financially constrained firms mainly relies on own funds as post-IPO firm

characteristics cannot not be changed immediately.

ODM OEM OBM

Smile Curve

Value Added

Grin Curve

Forced Smile Curve

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

Relevant studies such as Maine et al. [4] discuss value creation by startup- and science-based businesses such as

biotechnology firms and deepen the specifics inherent in value creation by companies producing high-end materials. The

authors apply hierarchical clustering to differentiate various types of material technology and observe differences in value

creation driven by decision incorporating uncertainty, commercialization of products, target-market-tailored R&D, and value

chain integration across the sample firms. Chemmanur et al. [5] advance by connecting the venture capital firm with corporate

innovation. The authors substantiate a higher degree of innovation in venture capital firms established and supported by parent

firms than in those which operate independently, suggesting the key role played by industrial knowledge and greater room for

failure in value creation achieved by external venture capital. Yang et al. [6] continue the studies in corporate growth and find

a U-shaped relation between diversification by venture capital businesses and value creation of new ventures. Quentier [7]

deepens with analysis of self-employment by the unemployed through creation of new ventures and find that government

subsidies in the form of loans better increase quality and selection of new ventures than direct subsidies. The loans also serve to

solve financial constraint faced by these new ventures.

Fig. 2 synthesizes the theoretical foundations inherent in the technology gap theory suggested by Posner [8] with

illustration of strategies by the OEM, ODM, and OBM. The firm is not established over the period of T0~T1 for lack of

competitiveness and starts to imitate other firms by operation in the form of OEM regarding conditions such as the economies

of scale, production standardization, and initial capacity over the period of T1~T2. Then the firm gradually raises operation

performance (OP) along the path of T1A under the imitation lag. However, the firm which keeps the OEM strategy may face

stagnant or decreasing OP and enjoy the marginal profit only under the agency problem caused by high concentration of buyers

or an increase in competitors as suggested by the transaction cost theory. To avoid such a strategic problem for OEM, the firm

should adopt the ODM strategy by an increase in spending on research and development to gain patents at the initial expense of

the design lag and failure in the early state of innovation, making OP fall along the path of AB over the period of T2~T3.

Beyond T3, the firm’s OP gradually rises along the path of BC with advantages from technological improvement and creation

of new product. If other OEMs also experience imitation and design lags to catch the firm’s ODM strategy, the firm’s OP may

become stagnant or declining beyond T4 under the love for varieties by consumers. As suggested in the product differentiation

theory by Krugman [9] and attribute differential theory by Lancaster [10], the ODM’s gross margin will decrease with mass

customization. To increase OP, the OBM strategy should be adopted, but the firm has to experience the demand lag and

substantial branding investment over the period T4~T5 where OP falls along the path of CD. Only as the capability for product

differentiation, operational digitalization, investment in intangible assets, and internationalization reach a stable level, the

firm’s OP grows along the path of DE.

Fig. 2 Strategies by OEM, ODM, and OBM Fig. 3 Business life cycle, internationalization, and value creation

In literature, Dunning [11], Wagner [12], Bhide [13], Zahra et al. [14], Kuo and Li [15], Head and Ries [16], and

Hmieleski and Baron [17] discuss from various dimensions the strategies for business evolution and expansion in the context

of international competition. Kumar and Siddharthan [18] place focus on the Indian enterprises, whereas Chen [19], Lundquist

[20], Lin [21], and Lin et al. [22] examine businesses in Taiwan. This paper attempts to evaluate whether the firm’s operating

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

years (business life cycle) and internationalization (overseas investment and export) shift the smile curve up to strengthen

value creation, illustrated by Fig. 3. Subsequent analysis is to construct a regression model of value creation and assess the

impact of the operating years, brand, R&D, internationalization, and other control variables of management on the value added

for Taiwan’s manufacturers in addition to proposing appropriate strategies.

3. Empirical Model and Hypotheses

3.1. Data

The empirical analysis adopts the 2011 Industry, Commerce, and Service Census conducted by Taiwan’s

Directorate-General of Budget, Accounting, and Statistics. The original sample covers 167,840 firms. With the deletion of

firms whose variable values are erroneous or omitted, the final sample contains 104,377 firms regrouped into ten classes. Table

1 recapitulates variable definitions.

Table 1 Variable Definitions

Variable Definition Remark

Panel A: Value Creation

LVA Value Added In Logarithm

Panel B: Business Life Cycle

AGE Operating Years

AGE
2

Squared AGE

Panel C: Asymmetric Grin Curve

BR Brand Revenue to Total Revenue In Percentage

RD R&D Spending to Total Spending In Percentage

FI Overseas Investment to Net Fixed Assets In Percentage

EX Export Revenue to Total Revenue In Percentage

Panel D: Operation Digitalization

EC1 Information Disclosure on Internet Dummy

EC21 Purchases on Internet Dummy

EC22 Internet Purchases to Total Spending In Percentage

EC31 Sales on Internet Dummy

EC32 Internet Sales to Total Revenue In Percentage

Panel E: Control Variable

OPR Manufacturing as Main Business Dummy

LSIZE Net Assets In Logarithm

KLR Net Fixed Assets to Number of Employees Capital-Labor Ratio

Panel F: Industry Class

INDA Food and Drink Dummy

INDB Textile, Clothing, and Leather Dummy

INDC Paper and Printing Dummy

INDD Oil, Coal, and Chemistry Dummy

INDE Rubber and Plastics Dummy

INDF Metal Dummy

INDG Electronics and Computer Dummy

INDH Machinery Dummy

INDI Car and Other Vehicle Dummy

INDJ Furniture Dummy

3.2. Model and hypotheses

To measure the impact of innovation, brand revenue, R&D spending, and internationalization on value creation for

Taiwan’s manufacturers, the model of Eq. (1) is constructed by the following regression equation.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

LVA = a0 + a1AGE + a2AGE
2
+ a3BR + a4BR

2
+ a5RD + a6RD

2
+ a7FI + a8FI

2
+ a9EX + a10EX

2
+ a11EC1

+ a12EC21+ a13EC22+ a14EC31+ a15EC32 + a16OPR + a17LSIZE + a18KLR + a19KLR
2
+ a20INDB

+ a21INDC + a22INDD + a23INDE + a24INDF + a25INDG+ a26INDH + a27INDI + a28INDJ + e

(1)

LVA is taken in logarithm and measure value creation. The parameters a1 and a2 serve to assess the role played by the

business life cycle in the firm’s value added, whereas the parameters from a3 to a6 respectively focus on the impact of OBMs

(a3 and a4) and ODMs (a5 and a6). The parameters from a7 to a10 serve to assess the role played by internationalization in terms

of overseas investment (a7 and a8) and export (a9 and a10). The parameters from a11 to a15 serve to evaluate the contribution of

operation digitalization through internet to value creation. Finally, the influence of the control variables and industry class

dummies are captured by the parameters from a16 and a28. The last term e represents the error term.

The Model of Eq. (1) is preliminarily estimated by the ordinary least squares method. The White test suggests significant

heteroscedasticity in residuals, with a chi-squared value of 58,305. Hence, subsequent analysis substitutes the robust standard

errors for standard errors estimated by the ordinary least squares method. Besides, results from parameter estimation of the

Model of Eq. (1) are used as foundations for testing the seven core hypotheses summarized from Eqs. on (2) to (8).

H1: Value creation is a U-shaped quadratic function of business life cycle,

implying increasing marginal benefit, or a1 < 0 and a2 > 0.
(2)

H2: Value creation is a reversed U-shaped quadratic function of brand revenue,

implying decreasing marginal benefit, or a3 > 0 and a4 < 0.
(3)

H3: Value creation is a reversed U-shaped quadratic function of R&D spending,

implying decreasing marginal benefit, or a5 > 0 and a6 < 0.
(4)

H4: The impact of brand revenue and R&D spending on value creation is

asymmetric, or a3 + 2a4BR  a5 + 2a6RD.
(5)

H5: Value creation is a reversed U-shaped quadratic function of overseas

investment, implying decreasing marginal benefit, or a7 > 0 and a8 < 0.
(6)

H6: Value creation is a reversed U-shaped quadratic function of export,

implying decreasing marginal benefit, or a9 < 0 and a10 < 0.
(7)

H7: The impact of overseas investment and export on value creation is

asymmetric, or a7 + 2a8FI  a9 + 2a10EX.
(8)

4. Empirical Results

4.1. Descriptive statistics

Table 2 summarizes descriptive statistics for the value creation, business life cycle, brand, R&D, internationalization,

operation digitalization, and other control variables of management across 104,377 manufacturers in Taiwan in 2011.

For value creation (VA), the mean for the original values is at $39,043 thousand new Taiwan dollars across surveyed

manufactures in Taiwan. The gap between the maximum value and the minimum value appears substantial, confirmed by a

high level of its standard deviation. A similar pattern is observed in the business life cycle (AGE) ranging from new ventures (1

year) to sustainable operating firms (100 years). The average of AGE is close to 18 years.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

The ratios of brand revenue to total revenue (BR), R&D spending to total spending (RD), overseas investment to net fixed

asset (FI), and export to total revenue (EX) are respectively 8.8812%, 0.3251%, 0.0511%, and 8.3778%, showing strong bias

to OEMs and insufficient internationalization for most manufacturers in Taiwan.

As regards operation digitalization, the ratios of internet purchases to total spending (EC22) and internet sales to total

revenue (EC32) are 1.3210% and 1.3337% only, suggesting that purchases and sales through internet remain at its early stage

of development in Taiwan’s manufacturing industry.

In terms of control variables, the dispersion for the value of net assets (LSIZE) that reflects the firm’s operating scale

remains significant in logarithm. The capital-to-labor ratio (KLR) measured by net fixed assets over the number of employees

is averaged at 1.5335.

Table 2 Descriptive Statistics

Variable Mean Maximum Minimum Std. Dev.

Panel A: Value Creation

LVA 8.5416 19.4878 5.0626 1.3276

Panel B: Business Life Cycle

AGE 17.5632 100 1 0.2784

Panel C: Asymmetric Grin Curve

BR 8.8812 100 0 25.6504

RD 0.3251 74.5755 0 2253.051

FI 0.0511 224.4135 0 2.0569

EX 8.3778 100 0 80.3330

Panel D: Operation Digitalization

EC1 0.6122 1 0 1.3542

EC21 0.1136 1 0 227.8531

EC22 1.3209 99.5421 0 22.1068

EC31 0.0847 1 0 1774.022

EC32 1.3337 99.9995 0 0.4872

Panel E: Control Variable

OPR 0.8308 1 0 0.3173

SIZE 248,994 1.53E+09 169 6.4962

KLR 1.5335 63.8248 0.0016 0.3749

Panel F: Industry Class

INDA 0.0382 1 0 0.1917

INDB 0.0681 1 0 0.2519

INDC 0.0964 1 0 0.2951

INDD 0.0338 1 0 0.1806

INDE 0.0882 1 0 0.2836

INDF 0.2933 1 0 0.4553

INDG 0.0755 1 0 0.2642

INDH 0.2135 1 0 0.4098

INDI 0.0464 1 0 0.2103

INDJ 0.0466 1 0 0.2107

4.2. Estimation results

Table 3 below summarizes major results from estimation of Equation (1) with adjustment in heteroscedasticity. With

additional regression for independent variables, the variance inflation factor (VIF) appears low except for AGE, AGE2, BR,

BR2, EX, and EX2 whose VIF exceeds 10. Therefore, there is, overall, no serious problem of collinearity. The adjusted R2 is

around 0.6533, substantiating sufficient explanatory power of the model specified. The chi-squared value for the White test is

at 58,305, implying significant heteroscedasticity in residuals. Therefore, the t-values reported in Table 3 are adjusted by

robust standard errors.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

Table 3 Estimation Results

Variable LVA t-Value VIF

Constant 3.2936*** 104.3353 NA

AGE -0.0029*** -3.4095 11.8816

AGE
2

1.01E-0.4*** 4.7657 11.7598

BR 0.0041*** 6.0753 40.9094

BR
2
-3.08E-05*** -4.0718 39.8211

RD 0.0805*** 22.0287 3.8738

RD
2

-0.0016*** -12.6025 3.4283

FI 0.0255*** 3.5577 4.3578

FI
2

-1.28E-04*** -3.4249 4.2128

EX 0.0119*** 26.6641 13.6124

EX
2

-9.36E-05*** -17.0069 12.4328

EC1 0.0393*** 7.5535 1.1698

EC21 0.1310*** 10.1139 2.6106

EC22 -4.72E-05 -0.0358 10.4591

EC222 -3.83E-05* -1.8291 7.8524

EC31 0.0175 1.1186 2.8294

EC32 -7.47E-04 -0.6356 11.0646

EC322 8.26E-06 0.5897 7.9876

OPR -0.0441*** -6.4857 1.0724

KLR -0.0942*** -16.0990 4.8555

KLR
2
1.72E-04*** 5.5128 1.4425

LSIZE 0.5627*** 139.6625 5.6422

INDB 0.0771*** 4.4743 3.0369

INDC 0.1214*** 7.5717 4.5435

INDD 0.1636*** 7.5286 1.7445

INDE 0.1141*** 6.9995 3.9985

INDF 0.1983*** 13.2852 8.7908

INDG 0.1771*** 9.7818 2.8244

INDH 0.0903*** 5.9271 6.8611

INDI 0.2896*** 15.6108 2.5161

INDJ -0.0217 -1.2003 2.5268

Adjusted R
2

0.6533

White Test 58,394***

Note: ***, **, * for significance at the 1%, 5%, 10% level.

Fig. 4 Marginal impact of business life cycle Fig. 5 Marginal impact of brand revenue

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

Next, the seven hypotheses summarized from (2.1) to (2.7) are each examined. For H1, the estimated coefficient for AGE

is significantly lower than zero and that for AGE2 is significantly greater than zero. This confirms that the relation betwee n

value creation (LVA) and the business life cycle can be portrayed by a U-shaped curve which decreases first before increasing.

This link can be illustrated by Fig. 4.

In Fig. 4, the minimum point of the LVA curve corresponds to around 15 years of the business life cycle. Hence, new

ventures have to pay a cost of lower value creation at the early stage of corporate development, supporting H1 (2.1).

For H2, the marginal impact of BR on LVA is observed from a reversed U-shaped curve that rises first before falling,

illustrated by Fig. 5. The maximum point of the LVA curve is at 66.5584%, far above the sample mean (8.8812%). Therefore,

the role of branding in value creation is captured by a grin curve and there seems great room for improvement in Taiwan’s

current OBMs, supporting H2 (2.2).

As regards H3, the coefficient for RD is positive whereas that for RD2 is negative. This can be illustrated by Fig. 6, where

the relation between R&D spending and value creation is presented by a reversed U-shaped curve. The maximum point of the

LVA curve is at 25.1563%, far above the mean for RD (0.3251%). Hence, on the left side of supply chains, the strategies for

ODMs are subject to a rising-then-falling grin curve, supporting H3 (2.3). Comparing the marginal impact of brand (Fig. 5) and

the marginal impact of R&D (Fig. 6), it is observed that the former impact is smaller than the latter impact. In other words, we

observe an asymmetric grin curve whose left-side peak is higher than the right-side peak, supporting H4 (2.4).

With respect to H5, the estimated coefficients for FI and FI2 also supports H5 (2.5), which implies a reversed U-shaped

LVA curve illustrated by Fig. 7. The maximum point of this curve corresponds to a ratio of overseas investment to net fixed

assets at 99.6094%, far above the industry mean (0.0511%), suggesting current underinvestment overseas by most SMEs in

Taiwan’s manufacturing industry.

The finding for H6 is analogous to that for H5. As illustrated by Fig. 8, the optimal export ratio is at 63.5684% and also far

above the industry mean (8.3788%). Hence, Taiwan’s manufacturers which aim to strengthen the value creation need to

substantially raise the weight of export in total revenue. Comparing Fig. 7 and Fig. 8, the marginal impact of overseas

investment is also higher than that of the export, implying an asymmetric grin curve and substantiating H7 (2.7).

Fig. 6 Marginal impact of R&D

spending

Fig. 7 Marginal impact of overseas

investment

Fig. 8 Marginal impact of export

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

In terms of operation digitalization, the positive coefficients for the two dummies, EC1 and EC21, suggest that

information disclosure and purchases on the internet serve to improve value creation by significant cost cut. However, the

coefficients for EC22 significant negative and the coefficient for EC 31 and EC 32 are found insignificant. Therefore, at least at

the current stage, internet activities may not effectively contribute to value creation in Taiwan’s businesses in the

manufacturing sector.

Finally, the empirical findings with respect to control variables can be analyzed from five aspects. First, the coefficient for

OPR is significantly negative, implying that whether the firm’s main activity involves manufacturing or not cannot effectively

increase the level of value creation. Second, the positive sign for the firm’s size (LSIZE) confirms that value creation rises with

corporate expansion. Third, the capital-to-labor ratio (KLR) exerts a U-shaped impact on value creation. Capital intensity

hence plays an important role in the firm’s value added. Four, as regards the dummies for the various classes in Taiwan’s

manufacturing industry, the value added appears higher among INDB～INDI as we adopt Food and Drink (INDA) as the

benchmark industry class for comparison.

Fig. 9 integrates previous findings and analysis illustrated by Fig. 5 to Fig. 8 and presents an overall grin curve for

Taiwan’s manufacturing industry. Figure 9 suggests that the grin curve shifts down as the firm’s operating years are less than

15 years. Only as the firm’s life cycle exceeds 15 years will value creation be raised.

Fig. 9 Asymmetric grin curve

Moreover, the marginal impact of R&D on value creation is higher than that of the brand, leading to an asymmetric grin

curve where the left-side peak is higher than the right-side peak. The marginal impact of overseas investment is greater than

that of exports, too. Overall, the seven hypotheses stated in (2.1) to (2.7) are in line with Fig. 9, which highlights graphic

asymmetry in the grin curve for Taiwan’s manufacturers.

4.3. Robustness check

To check the robustness of the previous analysis, we conduct three additional tests. Table 4 summarizes results of testing

the differences in means for major variables between new venture and survivors on the basis of two definitions for new

ventures: defined by 12 years against definition by 15 years. Overall, all variables except for EC31, EC32, INDF, and INDI

exhibit significant differences between new ventures and survivors regardless of the definition for new ventures adopted,

substantiating robustness of our analysis framework and results.

Table 5 compares coefficients respectively estimated with linear, quadratic, and cubic forms for selected variables for

regression of LVA. The estimation results appear consistent across the three regression models, further confirming the

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

robustness of our model specification. More specifically, the adjusted R2 remains around a similar level across the three

models, implying that the respectively effect of key variables included in Table 3 on value creation (LVA) all holds regardless

of the functional form adopted. For variables such as AGE, BR, RD, FI, and EX, the impact on LVA is consistently significant.

In contract, the role for dummies such as E31 and E32 keeps absent.

Table 4 Differences in Means between New Ventures and Survivors

New Venture Defined by 12 Years New Venture Defined by 15 Years

Variable Age≤12 Age≥12 Age≤15 Age≥15
LVA 8.3191 8.6497 8.3599 8.6698

(1443.683***) (1397.573***)

AGE 6.3419 23.0156 7.7371 24.5002

(134293.5***) (175364***)

BR 7.2344 9.6813 23.6322 26.9382

(209.4597***) (233.1375***)

RD 0.3677 0.3045 0.3831 0.28426

(21.6962***) (58.4488***)

FI 0.0210 0.0657 0.0244 0.0700

(25.0707***) (28.7915***)

EX 6.4475 9.3157 6.7972 9.4936

(388.1113***) (378.0316***)

OPR 0.7816 0.8547 0.7869 0.8619

(880.1822***) (1022.868***)

LSIZE 8.8140 9.3984 8.8743 9.4423

(2377.543***) (2477.660***)

KLR 1.4360 1.5809 1.4419 1.5982

(32.2502***) (41.4143***)

EC1 0.5523 0.6414 0.5705 0.6417

(773.5952***) (543.8418***)

EC21 0.1175 0.1117 0.1191 0.1097

(7.5348***) (22.3491***)

EC22 1.3690 1.2976 1.3943 1.2691

(2.7802*) (9.3998***)

EC31 0.0853 0.0843 0.0864 0.0834

(0.2861) (2.8233*)

EC32 7.6027 7.4995 0.0864 1.3224

(0.1274) (0.1645)

INDA 0.2150 0.1791 0.0459 0.0328

(148.7985***) (119.8543***)

INDB 0.0588 0.0726 0.0609 0.0732

(68.5690***) (60.6506***)

INDC 0.0882 0.1003 0.0898 0.1010

(38.5585***) (36.6144***)

INDD 0.0319 0.0344 0.0305 0.0361

(5.5848**) (23.6399***)

INDE 0.0712 0.0965 0.0725 0.0993

(183.2514***) (227.4697***)

INDF 0.4542 0.4558 0.2954 0.2918

(1.3431) (1.5731)

INDG 0.0942 0.0664 0.0936 0.0627

(254.5745***) (347.7844***)

INDH 0.2275 0.2067 0.2228 0.2070

(58.7679***) (37.4631***)

INDI 0.0465 0.0463 0.0463 0.0465

(0.0275) (0.0216)

INDJ 0.0420 0.0487 0.0423 0.0496

(23.2437***) (30.7220***)

Observations 34,132 70,245 43,194 61,183

Note: ***, **, * for significance of the F-value at the 1%, 5%, 10% level.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

Table 5 Robustness Test by Function Form

Linear Quadratic Cubic

Variable LVA LVA LVA

Constant
3.3533***

(53.0204)

3.2936***

(104.3353)

3.1809***

(108.8993)

AGE
0.0014***

(5.0475)

-0.0029***

(-3.4095)

0.0019

(1.3231)

AGE
2
NA

0.0001***

(4.7657)

-0.0001**

(-1.9651)

AGE
3
NA NA

2.48E-06***

(3.0442)

BR
0.0019***

(14.3199)

0.0041***

(6.0753)

2.00E-05

(0.0090)

BR
2
NA

-3.08E-05***

(-4.0718)

5.57E-05

(0.9852)

BR
3
NA NA

-4.84E-07

(-1.3533)

RD
0.0296***

(10.5057)

0.0805***

(22.0287)

0.1573***

(33.8342)

RD
2
NA

-0.0016***

(-12.6025)

-0.0077***

(-23.9350)

RD
3
NA NA

7.84E-05***

(18.7086)

FI
0.0081**

(2.3060)

0.0255***

(3.5577)

0.0594***

(6.1528)

FI
2
NA

-0.0001***

(-3.4249)

-0.0012***

(-5.3520)

FI
3
NA NA

4.38E-06***

(5.0506)

EX
0.0052***

(29.7733)

0.0119***

(26.6641)

0.0159***

(13.3506)

EX
2
NA

-9.36E-05***

(-17.0069)

-0.0003***

(-6.9366)

EX
3
NA NA

1.17E-06***

(4.5717)

EC1
0.0496***

(8.6773)

0.0393***

(7.5535)

0.0306***

(6.0460)

EC21
0.1578***

(13.3077)

0.1310***

(10.1139)

0.1059***

(7.3308)

EC22
-0.0024***

(-3.9812)

-4.72E-05

(-0.0358)

0.0064***

(2.7088)

EC22
2
NA

-3.83E-05*

(-1.8291)

-0.0003***

(-3.7794)

EC22
3
NA NA

2.29E-06***

(3.5345)

EC31
0.0137

(0.9719)

0.0175

(1.1186)

0.0178

(1.0281)

EC32
0.0005

(0.9216)

-0.0007

(-0.6356)

-0.0003

(-0.1436)

EC32
2
NA

8.26E-06

(7.9876)

-1.58E-05

(-0.2124)

EC32
3
NA NA

2.38E-07

(0.4314)

OPR
-0.0362***

(-0.0588)

-0.0441***

(1.0724)

-0.0438***

(-6.5869)

KLR
-0.0497***

(-3.4393)

-0.0942***

(4.8555)

-0.1414***

(-21.8759)

KLR
2
NA

0.0002***

(1.4425)

0.0009***

(6.3332)

KLR
3
NA NA

-1.13E-06***

(-5.2410)

LSIZE
0.5423***

(52.5408)

0.5627***

(5.6422)

0.5823***

(161.4039)

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

Table 5 Robustness Test by Function Form (Continued)

Linear Quadratic Cubic

Variable LVA LVA LVA

INDB
0.0965***

(4.9269)

0.0771***

(3.0369)

0.0556***

(3.2904)

INDC
0.1321***

(7.7247)

0.1214***

(4.5435)

0.1075***

(6.8314)

INDD
0.1857***

(8.6562)

0.1636***

(1.7445)

0.1520***

(7.0760)

INDE
0.1360***

(7.6700)

0.1141***

(3.9985)

0.0943***

(5.9018)

INDF
0.2177***

(13.2065)

0.1983***

(8.7908)

0.1779***

(12.1368)

INDG
0.2373***

(11.3780)

0.1771***

(2.8244)

0.1301***

(7.3822)

INDH
0.1175***

(6.9732)

0.0903***

(6.8611)

0.0637***

(4.2661)

INDI
0.3201***

(15.6212)

0.2896***

(2.5161)

0.2529***

(13.9078)

INDJ
-0.0113***

(-0.5899)

-0.0217

(2.5268)

-0.0414**

(-2.3481)

Adj. R
2

0.6332 0.6533 0.6696

White Test 90341.73***

58,394*** 34,574***

Note: ***, **, * for significance at the 1%, 5%, 10% level.

5. Conclusions

This paper empirically adopts the big data obtained from the 2011 Industry, Commerce, and Service Census conducted by

Taiwan’s Directorate-General of Budget, Accounting, and Statistics and examines the link between the business life cycle,

brand, R&D, internationalization, and value creation for Taiwan’s manufacturers. Regression results based on 104,377

observations in the prescreened sample can be recapitulated in seven points.

(1) A relatively low level of the value added, brand revenue, R&D spending, and internationalization suggests that Taiwan’s

manufacturing industry is currently subject to a business environment dominated by OEMs mainly oriented to the local

market.

(2) New ventures whose business life is less than 15 years have to face a low level of value creation. Only beyond 15 years will

value creation be strengthened by greater efficiency in operation.

(3) The marginal impact of brand revenue and R&D spending on value creation is captured by a reversed U-shaped curve. The

decreasing marginal effect is higher for R&D than for brand revenue.

(4) The marginal impact of the two gauges for internationalization (overseas investment and export) on value creation is

captured by a reversed U-shaped curve, too. The decreasing marginal effect is higher for overseas investment than for

export.

(5) Operation digitalization remains at a low level for Taiwan’s manufacturers and its impact on value creation appears

insignificant as of 2011.

(6) The average business life cycle is around 18 years in the sample. Under intensifying competition from globalization and

e-business, the business life cycle is anticipated to be further shortened, creating more challenges to Taiwan’s

manufacturing industry.

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

(7) Businesses in the manufacturing industry in Taiwan face an asymmetric grin curve rather than a smile curve. Therefore,

value creation can be strengthened through enhancement in the business life cycle, internationalization, internet activity,

operating scale, and capital intensity.

This study is conditioned on a few limitations, though. The big data obtained from the government census are based on

five-year surveys. The problem associated with potentially lagged information appears unavoidable. Besides, discontinuity in

our data makes it difficult to analyze the stock value and lagged effect for activities associated with the brand, R&D, and

overseas investment. Availability of a more complete dataset will benefit future research, which can also extend analysis to

issues such as income distribution and equality under intensified globalization.

Acknowledgement

The authors would like to acknowledge a research grant (MOST 103-2632-H-029-002-MY2) from the Ministry Science

and Technology in Taiwan, ROC.

References

[1] R. N. Lussier, “A nonfinancial business success versus failure prediction model for young firms,” Journal of Small

Business Management, vol. 33, no. 1, pp. 8-20, 1995.

[2] P. Marco, P. Fabio, and Z. Luigi, “Why do companies go public,” The Journal of Finance, vol. 53, no. 1, pp. 27-64,

February 1998.

[3] I. Erel, Y. Jang, and M. S. Weisbach, “Do acquisitions relieve target firm’s financial constraints,” Journal of Finance, vol.

70, no. 1, pp. 289-328, 2015.

[4] E. Maine, S. Lubik, and E. Garnsey, “Value creation strategies for science-based business: a study of advanced materials

ventures,” Innovation: Management, vol. 15, no. 1, pp. 35-51, December 2014.

[5] T. J. Chemmanur, E. Loutskina, and X. Tian, “Corporate venture capital, value creation, and innovation,” The Review of

Financial Studies, vol. 27, no. 8, pp. 2434-2473, August 2014.

[6] Y. Yang, V. Narayanan, and D. De Carolis, “The relationship between portfolio diversification and firm value: the

evidence from corporate venture capital activity,” Strategic Management Journal, vol. 35, no. 13, pp. 1993-2011,

December 2014.

[7] J. M. Quentier, “Self-employment start-ups and value creation: an empirical analysis of German micro data,” Advances in

Competitiveness Research, vol. 20, no. 1-2, pp. 37-57, 2012.

[8] M. V. Posner, “International trade and technical change,” Oxford Economic papers, vol. 13, no. 3, pp. 323-341, 1961.

[9] P. R. Krugman, “Increasing returns, imperfect competition, and international trade,” Journal of International Economics,

vol. 9, no. 4, pp. 467-479, 1979.

[10] K. Lancaster, “Intra-industry trade under perfect monopolistic competition,” Journal of International Economics, vol. 10,

no. 2, pp. 151-175, May 1980.

[11] J. H. Dunning, “Toward an eclectic theory of international production: some empirical test,” Journal of International

Business Studies, vol. 11, no. 1, pp. 9-31, 1980.

[12] J. Wagner, “Exports, firm size and firm dynamics,” Small Business Economics, vol. 7, no. 1, pp. 29-39, February 1995.

[13] A. Bhide, The origin and evolution of new business. New York: Oxford University Press, 2000.

[14] S. A. Zahra, R. D. Ireland, and M. A. Hill, “International expansion by performance,” Academy of Management Journal,

vol. 43, no. 5, pp. 925-950, 2000.

[15] H. C. Kuo and Y. Li, “A dynamic decision model of SME’s FDI,” Small Business Economics, vol. 20, no. 3, pp. 219-231,

May 2003.

[16] K. Head and J. Ries, “FDI as an outcome of the market for corporate control: theory and evidence,” Journal of

International Economics, vol. 74, no. 1, pp. 2-20, January 2008.

[17] K. M. Hmieleski and R. A. Baron, “Entrepreneurs’ optimism and new venture performance: a social cognitive perspective,”

The Academy of Management Review, vol. 52, no. 3, pp. 473-488, 2009.

[18] N. Kumar and N. S. Siddharthan, “Technology, firm size and export behavior in developing countries: the case of Indian

enterprises,” Journal of Development Studies, vol. 31, no. 2, pp. 289-309, December 1994.

https://econpapers.repec.org/article/blastratm/

Advances in Technology Innovation, vol. 4, no. 1, 2019, pp. 44 - 57

[19] A. Chen, “Taiwan’s paradigm shift: industrial engineers are shaping a nation,” Industrial Engineer, vol. 40, no. 10, pp.

31-34, October 2008.

[20] E. Lundquist, “Shih’s curve can bring smiles,” eWeek, vol. 56, September 2007.

[21] F. J. Lin, “The determinants of foreign direct investment in China: the case of Taiwanese firms in the IT industry,” Journal

of Business Research, vol. 63, no. 5, pp. 479-485, May 2010.

[22] J. R. Lin, C. J. Huang, and C. H. Hsieh, “Internationalization, domestic employment, and overstatement of export

contribution,” Taipei Economic Inquiry, vol. 51, no. 1, pp. 135-169, 2015.

under the terms and conditions of the Creative Commons Attribution (CC BY-NC) license

(https://creativecommons.org/licenses/by-nc/4.0/).