DOI 10.1007/s11135-011-9562-6
Does intellectual capital matter? Assessing the
profitability and marketability of IC design companies
Mei-Huan Kuo· Chyan Yang
Published online: 16 October 2011
© Springer Science+Business Media B.V. 2011
Abstract Under a highly competitive market and a dynamic industrial environment, how to evaluate and enhance an integrated circuit (IC) design company’s good performance is important. This paper develops a two-stage data envelopment analysis (DEA) combined intellectual capital theory through financial and non-financial data to evaluate a performance process on the IC design company. It adopts a new slacks-based measure (SBM) to obtain a more accurate performance estimation and rank between companies. This paper further uses the Simar and Wilson procedure with a truncated regression to explore the impact of intellectual capital variables on performance and competitive advantage. From the study we suggest to the company in how to enhance precisely its performance to create company value and success.
Keywords Data envelopment analysis· Slacks-based measure · Performance measurement· Intellectual capital
1 Introduction
Taiwan’s integrated circuit (IC) design industry started in 1975 with government sponsor-ship and achieved growth in 1990. Its industry revenue is now the second largest in the world, and the numbers and scale of companies are only behind the United States. Taiwan now has a complete industrial chain and advanced manufacturing capacity, becoming an important global IC industry supply center. Because an IC design company has high profit with less investment and because consumption of electronics products have increased, more and more IC design companies have been set up worldwide. Accompanying this increase in the numbers and scale of IC design companies globally especially in China and India, competition has become more and more fierce. Under a highly competitive market and
M.-H. Kuo (
B
)· C. YangInstitute of Information Management, National Chiao Tung University, 1001 Ta Hsueh Road, 300 Hsinchu, Taiwan, ROC
dynamic industrial environment, how to maintain an IC design company’s good efficiency is very important (Chang and Tsai 2002). The performance estimation of IC design compa-nies has been taken seriously quite recently (Chang et al. 2003;Hu and Li 2004;Chou and
Liu 2005;Kuo and Shen 2005;Wu and Ho 2007), because they help lift Taiwan’s overall
economic efficiency. Many previously good performing management companies have had difficulty to maintain their past superior performance, and therefore an IC design company’s accurate performance management is a subject worth discussing.
Market values of firms are estimated by the market, depending on several factors, such as book value, profit, economic outlook, speculation, and creating value ability. There exists gap between the book and market values such as the market value of Google was about U$50 billion in 2005, while its P/B ratio was about seventeen, which meant that the market value was seventeen times bigger than the book values of other firms. Such differences show that the existence of intellectual capital in firms.
Wernerfelt(1984) found that firms gain superior performance through acquiring, holding,
and using strategic assets (namely, tangible and intangible assets) from a resource-based view. Intellectual capital represents knowledge-related intangible assets embedded in an organization. Its importance is highlighted in the era of a knowledge-based economy, where intellectual capital, instead of traditional tangible assets, is the dominant property driver for an enterprise (Edvinsson and Malone 1997).Van Buren(1999) proposed a model to illustrate the impact of intellectual capital on firm performance. Therefore, the literature shows that intellectual capital affects largely the value creation and performance of a company to bring competitive advantage (Amir and Lev 1996;Edvinsson and Malone 1997;Stewart 1997;
Bontis 1999,2001;Sullivan 2000;Juma and Payne 2004).
The IC design industry is a fabless sector with less fixed asset investment, and there are 270 IC design companies in Taiwan in 2007. The core value of an IC design company is its workers’ high education and research innovation capability. Being a knowledge-intensive industry, intellectual capital is precisely the core essential factor for an IC design industry to maintain its competitive position and future living (Hung et al. 2004). Therefore, we investi-gate the relationship between intellectual capital and IC design firms’ performance in order to advise managers to pay more attention to intellectual capital and to promote competitive ability.
The current study uses data envelopment analysis (DEA) as the tool for assessing the company performance. DEA can measure regarding multiple-inputs and multiple-outputs among various firm performance. It has been proven effective in multiple performance mea-sures and also estimates the empirical efficient frontier from the observations while it does not require a priori information about the relationship among multiple performance measures. But traditional DEA models is the neglect of linking activities (intermediate products) or non-zero slack in inputs and outputs. To overcome the DEA methodological shortfalls referred to above when evaluating the performance of IC design firms in Taiwan, we adopt advanced DEA techniques, slacks-based measures (SBM) and slacks-based measures of super effi-ciency (super-effieffi-ciency-SBM), as respectively proposed byTone(2001,2002) to integrate the profitability as well as the efficiency of marketability to evaluate the IC design firms’ performance based on Seiford and Zhu’s model (1999).
The Tobit model was used to find intellectual capital factors which affect firms’ per-formance, untilSimar and Wilson(2007) demonstrated that it was inappropriate. Instead, they proved a truncated-regression approach with a bootstrap has satisfactory performance in Monte Carlo experiments. The adequacy of the functional form to the data is a prevalent problem and a common critique on the stochastic frontier models (Khumbakar and Lovell
2000). Here, we employ theSimar and Wilson(2007) approach to account for intellectual capital factors that might affect a firm’s performance.
The remainder of this paper is organized as follows. The next section provides analysis on the IC design industry. In Sect.3we develop the two-stage value-creating process of an IC design firm. Section4offers the data selection and description. In Sect.5we present our methodology model. Section6includes our empirical research results and analysis, and Sect.7concludes with remarks.
2 IC design industry analysis
Taiwan’s IC design industry is a very knowledge-intensive industry with less investment, shorter productive time, well-educated manpower, and shorter new product release time, with market demand satisfaction that is its competitive advantage. In 2007 Taiwan’s IC design companies saw NT$ 3,997 billion in total revenue, for a growth of 23.6%. Taiwan’s IC design industry has a 26.5% market share globally, second largest behind the U.S. Taiwan has a complete industrial chain and advanced manufacturing capacity, becoming an important global IC industry supply center. In 2007, IC design companies in Taiwan numbered 270, with competition becoming become more and more fierce. Thus, operation performance is something worth paying significant attention to.
The IC industry’s technology advancement is rather quick, with the product frequency cycle decided by the market demand, new products being published, and meeting consum-ers’ inexhaustible needs. For Taiwan’s IC industry in 2007, the best performance was in LCD monitors and consumer products-related domain, as other products saw low gross profits. In 2007, the highest proportion went into product application, information application, and communication application. Under countries’ national policy, strong capital support, and the global IC industry migrating to the Asia-Pacific area, China’s own IC design industry has grown rapidly. Due to a highly competitive market and dynamic industrial environment, how to maintain good efficiency for an IC design company is very important. An IC design com-pany in Taiwan needs to improve its performance with advanced design capability in order to face China’s low capital and huge human resource advantage.
3 Two-stage value-creating process of an IC design firm
Since a company’s performance is a complex phenomenon that cannot be characterized by just a single criterion, some studies have argued that a multi-factor performance measurement model may be used (Bagozzi and Phillips 1982;Chakravarthy 1986;Seiford and Zhu 1999;
Zhu 2000;Luo 2003;Lu and Hung 2009). This study adopts Seiford and Zhu’s two-stage
profitability and marketability model of the top 55 U.S. banks (1999) to design two per-formance models (Fig.1)—namely, a profitability performance model and a marketability performance model.
Figure1shows that the profitability performance model measures an IC design company’s efficiency by three inputs (equity, liability and employees) and two outputs (revenues and intangible assets). The marketability performance model that measures an IC design com-pany’s market value by using two inputs (revenues and intangible assets) and two outputs (outstanding shares and market value). Previous studies use asset as the input parameter, where the question is from the accounting aspect: asset= equity + liability. Therefore, both material data have overlaps. This paper uses the liability representing the external resources
Profitability Marketability Revenue Intangible Assets Liability Equity Employees Outstandin Shares Market Value
Value-Creating Process of IC Design Firm
Stage-1 Stage-2
Fig. 1 Profitability and marketability efficiency models for IC design firms
and the equity representing the internal resources as input parameters replacing asset and equity.
The output and input factors (seven financial measures) used in this study are defined as follows.
• Equity, the ownership interest of shareholders in a corporation, is the residual interest in the assets of an entity after its liabilities have been deducted at the company’s year-end. • Liability is probable future sacrifices of economic benefits arising from present
obligations.
• Employees are all staff members in an IC design company.
• Revenues, the entire amount of income before any deductions are made, include those of consolidated subsidiaries and exclude excise taxes.
• Intangible assets are assets that are saleable though not material or physical.
• Market value, a price which is likely to be paid for something, is obtained by multiplying the number of common shares outstanding by the price per common share at the last exchange date end of the year.
• Outstanding shares are stock currently held by investors, including restricted shares owned by the company’s officers and insiders, as well as those held by the public.
4 Data selection and description
Equity, liability, number of employees, revenue, intangible assets, market value, and out-standing shares data of the companies in 2007 are collected from the database of Taiwan Economic Journal. The companies included in the sample are all publicly listed and the information is complete and easy to collect. This paper uses 38 IC design companies in Taiwan as sample data. Each of these IC design companies is treated as a decision making unit (DMU) in the DEA analysis. Table1presents the descriptive statistics for our dataset in 2007. From this table, we observe that the deviations in the variables used are quite large, because of the various sizes of the IC design companies.
The DEA technique presumes the correlation coefficient relationship among the input and output data as performed in Table2. It can be observed that most input factors are highly correlated with output factors under a score larger than 0.7, implying that the IC design com-panies that employ more input resources will increase their revenue and market value. In the DEA model, the number of IC design companies should be at least twice the total number of input and output factors consideredGolany and Roll(1989). In this study’s profitability model the number of IC design companies is 38, which is at least twice the five factors
Table 1 Descriptive statistics for IC design firms in 2007
Variables Mean Minimum Maximum Std. dev.
Equity (NT$ thousand) 5,910,113 165,638 85,937,057 13,734,434 Liabilities (NT$ thousand) 1,593,343 48,970 11,353,281 2,491,581
Employees (persons) 358 55 1,817 392
Revenue (NT$ thousand) 6,879,256 223,536 74,778,579 13,001,787 Intangible assets (NT$ thousand) 16,312,737 13,711 352,261,943 55,492,651
Market value (NT$ million) 22,223 1,084 438,199 68,930
Outstanding shares (million) 245 27 1,409 324
Table 2 Correlation coefficients among inputs and outputs in 2007
Equity Liability Employees Revenues Intangible assets Market value Outstanding shares Equity 1.000 Liability 0.802 1.000 Employees 0.782 0.847 1.000 Revenues 0.933 0.917 0.773 1.000 Intangible assets 0.973 0.729 0.696 0.910 1.000 Market value 0.983 0.747 0.717 0.919 0.999 1.000 Outstanding shares 0.605 0.686 0.753 0.554 0.445 0.479 1.000
selected. Hence, the DEA model developed based on the profitability performance model has met construct validity requirement. By following the same rules, the marketability model in this study is also found on a required validity issue.
5 Methodology
The methodology procedures adopted are as follows. First, we use Tone’s SBM model (Tone 2001) to evaluate the performance of the IC design companies in the current period. Then use the super-efficiency-SBM model (Tone 2002) to rank the best performers from those exhibiting an efficiency score of one. Finally, the truncated regression model (Simar and Wilson 2007) is employed to account for intellectual capital factors that might affect a firm’s performance.
5.1 Slacks-based measure model
There are various DEA models that can be categorized into two forms (Cooper et al. 2000). The first form is the radial models including the CCR model byCharnes et al.(1978) and the BCC model byBanker et al.(1984). The second form is the non-radial models that are the additive model (Charnes et al. 1985), the Russell measure (Russell 1985), the range-adjusted measure (Aida et al. 1998), and the SBM model (Tone 2001).
We choose the SBM model as the appropriate version of DEA for investigating the effi-ciency of the process of converting multiple inputs into multiple outputs. The SBM model possesses more suitable features include: (1) processes directly with the input excesses and the output shortfalls of the firms concerned; (2) unit invariant and monotone decreasing with respect to input excesses and output shortfalls and affected by consulting the reference set of the DMUs not the whole dataset; (3) it is crucial to deal with negative outputs in the evaluation of efficiency and is closely connected to the other measures proposed, e.g., the CCR, BCC and the Russell measures.
The non-oriented SBM model gets the efficiency of the target DMUo(o = 1, . . . , n) by solving the following fractional programs:
Minηo= 1− 1 m m i=1 si−/xi o 1+1 s s r=1 sr+/yr o s.t. xi o= n j=1 xi jλj+ si−, i = 1, . . . , m, yr o= n j=1 yr jλj− sr+, r = 1, . . . , s, n j=1 λj= 1, λj≥ 0, si−≥ 0, sr+≥ 0. (1)
Here, n is the number of firms each firm produces s different outputs, using m different inputs; xi j > 0 and yr j > 0 are the level of the ith input and rthoutput, respectively, at the
jth firm; andλj is the weight of the firm j , the firm being evaluated is set as the target firm. The sum of the weights must be equal to one in Program (1), suggesting that the constructed best practice frontier exhibits variable returns to scale technology, so that, the frontier per-mits increasing, constant, and decreasing returns to scale. The efficiency score calculated by Program (1) reflects the target firm’s current scale of operations. It is referred to as “pure” technical efficiency, for representing the ability of management to transform inputs in order to produce outputs, whenη∗o = 1. The value of λj indicates that the firm j is an exemplar for the target firm can learn.
Program1can be transformed into the program with a positive scalar variable t (Charnes
and Cooper 1962). Minτo= t − 1 m m i=1 tsi−xi o s.t. 1= t +1 s s r=1 tsr+yr o (2) xi o = n j=1 xi jλj+ si−, i = 1, . . . , m,
yr o= n j=1 yr jλj− sr+, r = 1, . . . , s, n j=1 λj = 1, λj ≥ 0, si−≥ 0, sr+ ≥ 0, t > 0. Now let us define:
Si−= tsi−, S+= tsr+, = tλj.
Program2then is transformed into the following linear program in t, Si−, Sr+, and:
Minτo= t − 1 m m i=1 Si−/xi o s.t. 1= t + 1 s s r=1 Sr+/yr o t xi o= n j=1 xi jj+ Si−, i = 1, . . . , m, (3) t yr o= n j=1 yr jj− Sr+, r = 1, . . . , s, n j=1 j= t, j ≥ 0, Si−≥ 0, Sr+≥ 0, t > 0.
Note that t> 0 by virtue of the first constraint, the transformation is reversible. Thus, let an optimal solution of Program3be:
τ∗
o, t∗, ∗, Si−∗, Sr+∗
.
We then have an optimal solution of Program1defined by
η∗
o = τo∗, λ∗j = ∗j/t∗, si−∗= Si−∗/t∗, sr+∗= Sr+∗/t∗. 5.2 Super-efficiency-SBM model
The best performers share the fully efficient status denoted by a score of one such as multiple firms usually exhibit ‘efficient’ status. The super-efficiency model can rank these efficient firms to distinguish real benchmarks. The efficient observed firm is taken out from the pro-duction possibility set (PPS) by measuring the distance from the observed firm to the point located on the remaining PPS. If the distance is small, the super-efficiency of the firm is lower than the other firms. By contrast, if the distance is large, then the super-efficiency of the firm is higher than the remaining firms. Hence, we rank the efficient firms based on the super-SMB scores obtained.
Tone(2002) proposed the non-oriented super-SBM model for getting the super-efficiency of the observed D MUo(xi o, yr o) by solving the following fractional programs:
Mi nπo = 1 m m i=1 xi/xi o 1 s s r=1 yr/yr o s.t. xi ≥ n j=1, j=o xi jλj, i = 1, . . . , m, yr ≤ n j=1, j=o yr jλj, r= 1, . . . , s, (4) n j=1 λj= 1, x≥ xi o, y ≤ yr o, y ≥ 0, λj ≥ 0. xi, yr
is located on the remaining PPS. A weighted l1distance from xi oto xi(≥ xi o), is an average expansion rate of xi oto xiof the point
xi, yr
. A weighted l1distance from yr oto
yr(≤ yr o), is an average reduction rate for yr oto yrof the point
xi, yr
. Inversely an index of the distance from yr oto yr. Hence,πois a product of two indices: one is the distance in the input space, and the other is the distance in the output space.
We defineφ ∈ Rmandθ ∈ Rssuch that xi = xi o(1 + φi) and yr = yr o(1 − θr). It can be equivalently stated in terms ofφi, θr, andλjas follows:
Mi nπo= 1+m1 mi=1φi 1−1s sr=1θr s.t. n j=1, j=o xi jλj− xi oφi ≤ xi o, i = 1, . . . , m, n j=1, j=o yr jλj+ yr oθr ≥ yr o, r = 1, . . . , s, (5) n j=1 λj = 1, φi ≥ 0, θr ≥ 0, λj ≥ 0.
We use a positive scalar variable t (Charnes and Cooper 1962).
Mi nπo= t + 1 m m i=1 tφi s.t. t−1 s s r=1 tθr = 1 n j=1, j=o xi jλj− xi oφi ≤ xi o, i = 1, . . . , m, (6)
n j=1, j=o yr jλj+ yr oθr ≥ yr o, r = 1, . . . , s, n j=1 λj = 1, φi ≥ 0, θr ≥ 0, λj ≥ 0, t > 0. By defining the following:
tφi = i, tλj = j, tθr = r
Program6transforms into the following linear program in i, j, and r.
Mi nδo= t + 1 m m i=1 i s.t. t−1 s s r=1 r = 1 n j=1, j=o xi jj− xi o i ≤ txi o, i = 1, . . . , m, (7) n j=1, j=o yr jj+ yr o r ≥ tyr o, r = 1, . . . , s, n j=1 j = t, i ≥ 0, r ≥ 0, j ≥ 0, t > 0. We have an optimal solution of Program7be
δ∗ o, ∗i, ∗r, ∗j, t∗ . We express an optimal solution of Program5 π∗ o = δ0∗, λ∗j = ∗j/t∗, φi∗= i∗/t∗, θr∗= r∗/t∗ the optimal solution of Program4is expressed by:
x∗i o= xi o
1+ φi∗ and yr o∗ = yr o
1− θr∗. 5.3 Truncated Regression Model
The paper assumes and testsSimar and Wilson(2007) truncated regression model as the following
T E= Zβ + ε, (8)
here T E is a vector(n × 1) of technical efficiency; the (d × 1) vector β is unknown param-eters to be calculated; Z is an matrix(n × d) of environmental variables; and ε is an (n × 1) vector of statistical noise.
Previous DEA literature use Tobit model to show that intellectual capital factors on firms’ performance, until it was demonstrated inappropriate bySimar and Wilson(2007). They proved a truncated-regression approach with a bootstrap has satisfactory performance in
Monte Carlo experiments.Khumbakar and Lovell(2000) said that it is a prevalent problem and a common critique for the adequacy of the functional form to the data. We use theSimar
and Wilson(2007) approach to prove intellectual capital factors on a firm’s performance.
The distribution ofε is restricted by the condition ε ≥ 1 − α − Zβ, we followSimar
and Wilson(2007) and assume that this distribution is truncated normal with zero mean,
unknown variance, and left truncation point. Formally, our model is expressed by:
T E= Zβ + ε, ε ∼ N(0, σε2), such that ε ≥ 1 − Zβ, (9) here we maximize the corresponding likelihood function to estimate, with respect to(β, σε2), on our data. With asymptotic theory, normal tables are used to construct confidence when our regressands are not true variables, and their estimates are likely to be dependent on observed variables the construction can be more precise with a bootstrap. The bootstrap confidence intervals for the estimates of parameters(β, σε2) are constructed by using the parametric bootstrap for regression with incorporate information on the parametric structure and distributional assumption.
6 Empirical results and analysis
6.1 Measuring profitability and marketability performances
The evaluation of the IC design companies’ profitability and marketability efficiencies is conducted for the year 2007. The non-oriented SBM model is applied to assess the relative performances of the 38 Taiwanese IC design companies. To distinguish those efficient IC design companies that can be treated as real benchmarks, the super-SBM model is used as a ranking measure. All of the results are shown in Table3. The order (No.) of the IC design company is coded based on the respective sizes of the total assets.
The average scores computed from the SBM models based on the profitability and mar-ketability models are 0.533 and 0.499, respectively. The results show that eight of the IC design companies are efficient with scores all equal to one in Table3in the profitability performance field. Eight of the thirty-eight IC design companies with efficiency scores of one in the marketability performance field.
To make comparisons among the IC design companies, we calculate the mean values of their profitability and marketability efficiency scores based on the main focus as shown in Table4. An examination of the IC design companies’ profitability performance reveals that analog IC design companies (with a mean value of 0.653) operate better than digital IC design companies (with a mean value of 0.506). This result shows that analog IC design companies are more likely to generate relatively higher profit. Those IC design companies whose main focus is analog IC design on average outperform the other types. The reason for this is that the demand for analog IC design has recently grown rapidly in Taiwan’s market. By examining the performance of marketability, it is found that digital IC design companies operate better on average than analog ones, which can be explained by the finding that digital IC design companies can more easily attract the attention of investors with the trend towards digital circuit development.
6.2 Identification of benchmarks
Distinguishing among these efficient IC design companies and identifying the benchmarks have become interesting research subjects. Several authors have proposed methods for
Table 3 SBM-efficiency and SBM-super-efficiency for IC design firms in 2007
IC design firms Code Cast Profitability model Marketability model SBM-Eff. Super-SBM
Eff.
SBM-Eff. Super-SBM Eff. Mediatek Incorporation F01 C1 1.000 3.205 1.000 1.836
Novatek Microelectronics Corp. F02 C1 1.000 1.115 0.367 0.367 Realtek Semiconductor Corp. F03 C1 0.386 0.386 0.409 0.409 Richtek Technology Corp. F04 C2 1.000 1.305 0.458 0.458 Sunplus Technology Co., Ltd. F05 C1 0.288 0.288 0.468 0.468 Faraday Technology Corp. F06 C1 0.414 0.414 0.723 0.723 VIA Technologies, Inc. F07 C1 0.246 0.246 0.531 0.531 Elan Microelectronics Corp. F08 C1 0.638 0.638 0.708 0.708 Phison Electronics Corp. F09 C1 1.000 1.160 0.050 0.050
Ali Corp. F10 C1 0.506 0.506 0.656 0.656
Silicon Integrated Systems Corp. F11 C1 0.039 0.039 1.000 4.773
Elite Semiconductor Memory Technology In
F12 C1 0.386 0.386 0.152 0.152 Sonix Technology Co., Ltd. F13 C1 0.559 0.559 0.250 0.250 Sitronix Technology Corp. F14 C1 0.472 0.472 0.091 0.091
ITE Tech. Inc. F15 C1 0.734 0.734 0.290 0.290
Etron Technology, Inc. F16 C1 0.147 0.147 0.162 0.162 Holtek Semiconductor Inc. F17 C1 0.367 0.367 0.267 0.267 Weltrend Semiconductor, Inc. F18 C1 0.426 0.426 0.384 0.384 Anpec Electronics Corp. F19 C2 0.516 0.516 0.156 0.156
Springsoft Inc. F20 C1 0.258 0.258 1.000 1.058
Alcor Micro Corp. F21 C1 0.775 0.775 0.191 0.191
Advanced Power Electronics Corp. F22 C2 0.702 0.702 0.161 0.161 Princeton Technology Corp. F23 C1 0.170 0.170 0.310 0.310 Syntek Semiconductor Co., Ltd. F24 C1 0.275 0.275 1.000 1.862
Genesys Logic, Inc. F25 C1 0.246 0.246 0.248 0.248 Prolific Technology Inc. F26 C1 0.355 0.355 0.548 0.548 Integrated Service Technology Inc. F27 C1 0.159 0.159 0.372 0.372
Ame Inc. F28 C2 0.566 0.566 0.328 0.328
CoAsia Microelectronics Corp. F29 C1 1.000 1.540 1.000 1.777
Service & Quality Technology Co., Ltd. F30 C1 1.000 1.596 0.523 0.523 Aimtron Technology Co., Ltd. F31 C2 1.000 1.270 0.310 0.310 Higher Way Electronic Co., Ltd. F32 C1 0.603 0.603 0.222 0.222 Amic Technology Corp. F33 C1 0.283 0.283 0.396 0.396
E-Cmos Corp. F34 C2 0.478 0.478 0.999 0.999
Chip Hope Co., Ltd. F35 C1 0.433 0.433 0.245 0.245 Avid Electronics Corp. F36 C1 0.520 0.520 1.000 1.142
Table 3 continued
IC design firms Code Cast Profitability model Marketability model SBM-Eff. Super-SBM
Eff.
SBM-Eff. Super-SBM Eff. HiMark Technology Inc. F38 C1 1.000 1.789 1.000 1.532
Mean 0.533 0.499
C1 digital IC design company; C2 analog IC design company
Table 4 Summary statistics: TE of cast for IC design firms in 2007
Category Number Profitability performance Marketability performance Mean Test ( p-value) Mean Test ( p-value) Cast
Digital IC design company 31 0.506 0.1452 0.502 0.6912
Analog IC design company 7 0.653 0.487
ranking the best performers, including Andersen and Petersen (1993), Doyle and Green (1994),Tofallis(1996),Seiford and Zhu(1999),Zhu(2001), andTone(2002). We refer to this problem as the ‘super-efficiency’ problem. The super-SBM model first proposed byTone (2002) is an appropriate version of DEA for ranking the efficient IC design companies in this study. Several characteristics of the super-SBM model have been discussed before, espe-cially its ability to cope with a small number of DMUs compared to the number of evaluation criteria.
The IC design company with higher super-SBM efficiencies reveals itself to be differ-ent in the input/output space, and thus can be either referenced by very few DMUs, or just by itself. From managerial implication, the IC design company is a self-evaluator for niche player (Charnes and Cooper 1962). Table3reports the super-SBM efficiencies for both prof-itability and marketability models. There are eight technically efficient IC design companies under the SBM model for the profitability stage. The order of ranking in descending order is Mediatek Incorporation (F01), HiMark Technology Inc. (F38), Service & Quality Tech-nology Co, Ltd. (F30), CoAsia Microelectronics Corp. (F29), Richtek TechTech-nology Corp. (F04), Aimtron Technology Co, Ltd (F31), Phison Electronics Corp. (F09), and Novatek Microelectronic Corp. (F02). In the profitability model, Mediatek Incorporation (F01) is a niche player in the digital group, and Richtek Technology Corp. (F04) is a niche player in the analog group. The super-SBM efficiency for the eight technically efficient IC design com-panies in the marketability performance model is also reported herein. The order of ranking is Silicon Integrated Systems Corp. (F11), Syntek Semiconductor Co, Ltd. (F24), Mediatek Incorporation (F01), CoAsia Microelectronics Corp. (F29), HiMark Technology Inc. (F38), Avid Electronic Corp. (F36), Springsoft Inc. (F20), and Analog Integrations Corp. (F37). In the marketability model, Silicon Integrated Systems Corp. (F11) is a niche player in the digital group, and Analog Integrations Corp. (F37) is a niche player in the analog group. We therefore rank the efficient IC design companies from the highest to the lowest in order to rank the best performers based on the resulting list to determine niche players.
6.3 The relationship between intellectual capital and performance
The concept of intellectual capital was first proposed by economistGalbraith(1969) and is used as intangible assets to explain the differences between a company’s market value and book value that are not reflected on the balance sheet. It represents the intangible assets of a firm, including any that may increase the organization value, promote the organization competitive advantage, exceed the book value, and have important effects on achieving com-pany profit and competitive advantage in the market.Stewart(1994) thought that intellectual capital will become the value driver and the most advantageous competition power of an American company.Edvinsson and Malone(1997) also believed that in the knowledge econ-omy era, intellectual capital management may transform knowledge as property. Thus, the literature argues that intellectual capital largely affects the value creation and performance of a company to bring competitive advantage (Amir and Lev 1996;Edvinsson and Malone
1997;Stewart 1997;Bontis 1999,2001;Sullivan 2000;Juma and Payne 2004).
The IC design industry is a knowledge-intensive industry, which has many intangible assets. These intangible assets are of great importance for such an industry contained in the knowledge industry scope. Therefore, the many valuable intangible assets in the IC design industry have been the key area for its survival superiority and growth achievement. Thus, intellectual capital appears as a value in an IC design company. IC design companies, as such, gain competitive advantage and superior performance through acquiring and holding intangible assets.
Intellectual capital of a company, as an intangible asset, is developed from knowledge management, including expertise of knowledge, experiences, organizational technologies, customer relationships, and specialized skills. The core value of an IC design company is the education quality of its workforce and research capability. In IC design companies, valuable assets reside in the intellectual capital their employees produce, not in the tangible assets the company possesses. Many companies are sold far in excess of their book value. Previous lit-erature indicates that successful Taiwanese IC firms have better effective intellectual capital performance management, but there is little consensus on how intellectual capital can best be conceptualized and measured. Furthermore, little empirical research has specifically exam-ined the relationship between intellectual capital and IC design companies’ performance. This paper focuses on the impact that intellectual capital has on an IC design company’s performance and also on the management strategy it may have on this relationship.
The literature has a convergent view that a general concept of intellectual capital is divided into three elements: human capital, customer capital, and structure capital (Edvinsson and
Malone 1997; Stewart 1997; Bontis 1999; Sullivan 2000;Choo and Bontis 2002). This
paper classify intellectual capital into human capital, structure capital, and customer capi-tal as three essential factors, in order to discuss the relatedness of intellectual capicapi-tal, each construction surface, and an IC design company’s performance. Staff who receive a higher education from the best universities are assumed to have more specialized knowledge and to have high potential intellectual capabilities to learn and accumulate new or tacit knowledge
(D’Aveni and Kesner 1993). In the human capital variable selection part, we propose one
human capital variable as antecedent distribution (above master) based onEdvinsson and
Malone(1997). Structure capital consists of intellectual assets such as number of patents,
research expenses and management expenses. Customer capital includes earnings growth rate.
We use DEA to obtain the efficiency value and take the explanation variable by the intellectual capital’s three target variables, making the regression analysis. We adopt Simar and Wilson’s (2007) procedure to explore the impact of intellectual capital on performance
Table 5 Results of regression in profitability model
Coefficient Std. error z-Statistic p-Value
Intercept 0.379241 0.109369 3.46752 0.001444
Structure capital −0.000583 0.000416 −1.40349 0.169538
Human capital 0.001197 0.000313 3.82491 0.000533∗
Customer capital 0.006680 0.002524 2.64683 0.012222∗
R2 0.418
* Statistically significant at the 0.05 level
and competitive advantage with a truncated regression. The estimated specification is the following:
θi = α0+ α1· Structurei+ α2· Humani+ α3· Customeri, j+ εi, (10) hereθ represents the efficient score.
The truncated regression with a bootstrap model appears to fit the data well, with positive
t-statistics, which are statistically significant for all parameters. The estimations generally
conform to a priori expectations. It is observed from Table5that efficiency increases in the human and customer capital factors. The empirical results imply that human capital has an obvious positive effect on performance. Through better management of intellectual capital, firms can improve competitive advantages. Human capital is the primary factor that affects the firm performance, and management should put the most effort on it.
6.4 Analysis of managerial decision-making matrix
In the previous section the IC design companies’ benchmarks are identified either by their profitability or marketability performance. An overview analysis including these two perfor-mance models has not yet been discussed. This section presents a decision making matrix in Fig.2to further illustrate the difference between profitability and marketability. In Fig.2 classify the IC design companies, which fall into four quadrants—star, cow, sleepers, and dogs—which are similar to the classification done by the Boston Consulting Group. The IC design companies have been split subjectively into four groups plotted respectively in the zones of stars, sleepers, dogs, and cows. The IC design companies in each group are summarized as follows.
The zone of stars: In this zone, these IC design companies have high efficiency in both
prof-itability and marketability efficiency dimensions. Three IC design companies are included here: Mediatek Incorporation (01), Coasia Microelectronics (29), and HiMark Technology (38), These companies have good performance and can be reference companies for others. From the intellectual capital factors analysis based on our collected data, Mediatek Incorpo-ration has paid much attention in the antecedent distribution (above master) of human capital. Coasia Microelectronics should regard human capital and customer capital as important fac-tors for maintaining its good performance.
The zone of sleepers: In this zone, these IC design companies have efficient marketability
performance, but inefficient profitability performance. Five IC design companies are included in this zone: Silicon (11), Springsoft (20), Syntek (24), Avid (36), and Analog integrations (37). IC design companies in this zone should improve their profitability performance by putting more efforts to increase their revenue and intangible asset. From the intellectual
Efficient E ffi cient Stars Sleepers Cows Dogs Profit a bilit y Marketability Inefficient Inefficient Mediatek Incorporation(01) CoAsia Microelectronics(29) HiMark Technology(38) Novatek (02) Richtek (04) Phison (09) Service & Quality (30) Aimtron(31) Realtek (03) Sunplus (05) Faraday (06) VIA (07) Elan (08) Ali (10) Elite (12) Sonix (13) Sitronix (14) ITE (15) Etron (16) Holtek (17) Weltrend (18) Anpec (19) Alcor (21) Advanced (22) Princeton (23) Genesys (25) Prolific (26) Integrated service (27) Ame (28) Higher Way (32) Amic (33) E-Cmos (34) Chip Hope (35) Silicon (11) Springsoft (20) Syntek (24) Avid (36) Analog integrations (37)
Fig. 2 Profitability/marketability efficiency cross-tabulation
capital factors analysis, these companies can achieve progress by paying more effort to raise their earnings growth rate of customer capital.
The zone of dogs: In this zone, these IC design companies are inefficient both in
profitability efficiency and marketability efficiency. Twenty-five IC design companies are here: Realtek (03), Sunplus (05), Faraday (06), VIA (07), Elan (08), Ali (10), Elite (12), Sonix (13), Sitonix (14), ITE (15), Eton (16), Holtek (17), Weltrend (18), Anpec (19), Alcor (21), Advanced (22), Princeton (23), Genesys (25), Prolific (26), Integrated Services (27), Ame (28), Higher Way (32), Amic (33), E-Cmos (34), and Chip Hope (35). They should improve their performance in creating profit and market. For instance, Higher Way can pay more effort on human capital and customer capital to improve its performance.
The zone of cows: In this zone, these IC design companies have efficient profitability but
inefficient marketability performance. There are five IC design companies: Novatek (02), Richtek (04), Phison (09), Service & Quality (30), and Aimtron (31). Phison could empha-size human capital investment in the antecedent distribution (above master) and average period of service.
The IC design companies classified into the zone of stars are almost all digital IC design companies. This shows that digital IC design companies have better competitive power than analog IC design firms.
7 Concluding remarks
This paper has adopted a two-stage efficiency estimation model for using SBM DEA method on an IC design company. While DEA has been widely adopted in the literature on IC design company efficiency studies, it has limitations on deeply evaluating the intellectual capital variables’ effect on performance. The paper uses the Simar–Wilson procedure with a trun-cated regression to find out the obvious intellectual capital factors affecting an IC design company’s performance. Thus, the contribution of this paper to the literature is: to adopt a new two-stage DEA evaluation process by the SBM model in order to evaluate and rank
the performance; by combining DEA measurement with a recently developed Simar–Wilson method and a truncated regression, to verify that the intellectual capital effect on DEA effi-ciency performance. The findings are that human capital has a positive impact on IC design companies’ performance. Furthermore, the paper suggests to an inefficient company on how to manage its intellectual capital, arguing that human capital and customer parameters can create and enhance a company’s efficiency and success.
Acknowledgment The authors want to express their gratitude to Dr. Wen-min Lu for his discussion and assistance in DEA software usage.
References
Aida, K., Cooper, W.W., Pastor, J.T., Sueyoshi, T.: Evaluating water supply services in Japan with RAM: a range-adjusted measure of inefficiency. Omega Int. J. Manag. Sci. 26(2), 232–307 (1998)
Amir, E., Lev, B.: Value-relevance of non-financial information: the wireless communication industry. J. Account. Econ. 22, 3–30 (1996)
Andersen, P., Petersen, N.C.: A procedure for ranking efficient units in data envelopment analysis. Manag. Sci. 39(10), 1261–1264 (1993)
Bagozzi, R.P., Phillips, L.W.: Representing and testing organizational theories: a holistic construal. Adm. Sci. Q. 17, 459–489 (1982)
Banker, R.D., Charnes, A., Cooper, W.W.: Models for estimating technical and scale efficiencies in DEA. Eur. J. Oper. Res. 30(9), 1078–1092 (1984)
Bontis, N.: Managing organizational knowledge by diagnosing intellectual capital: framing and advancing the state of the field. Int. J. Technol. Manag. 18(5–8), 433–462 (1999)
Bontis, N.: Assessing knowledge assets: a review of the models used to measure intellectual capital. Int. J. Manag. Rev. 3(1), 41–60 (2001)
Chakravarthy, B.S.: Measuring strategic performance. Strateg. Manag. J. 7, 437–458 (1986)
Chang, P.L., Tsai, C.T.: Finding the niche position—competition strategy of Taiwan’s IC design industry. Technovation 22, 101–111 (2002)
Chang, S.C., Lee, Z.Y., Yu, H.C.: Performance evaluation of IC design companies listed in TSE. Ind. Forum
5(1) (2003)
Charnes, A., Cooper, W.W.: Programming with linear fractional functionals. Nav. Res. Logist. Q. 15, 333– 334 (1962)
Charnes, A., Cooper, W.W., Rhodes, E.: Measuring the efficiency of decision making units. Eur. J. Oper. Res. 2(6), 429–444 (1978)
Charnes, A., Clark, C.T., Cooper, W.W., Golany, B.: A developmental study of data envelopment analysis in measuring the efficiency of maintenance units in the US air force. Ann. Oper. Res. 2, 95–112 (1985) Chou, P.Y., Liu, M.H.: The empirical study of the relationship between the intellectual capital and profitability
in the fabless industry in taiwan and america. Institute of Management Science, National Chiao Tung University, Hsinchu (2005)
Choo, C.W., Bontis, N.: The strategic management of intellectual capital and organizational knowledge. Oxford University Press, New York (2002)
Cooper, W.W., Seiford, L.M., Tone, K.: Data envelopment analysis: a comprehensive text with models applications references and DEA-Solver software. Kluwer Academic Publishers, Boston (2000) D’Aveni, R.A., Kesner, I.F.: Top managerial prestige, power and tender offer response: a study of elite social
networks and target firm cooperation during takeovers. Organ. Sci. 4, 123–151 (1993)
Doyle, J., Green, R.: Efficiency and cross-efficiency in DEA: derivation, meanings and uses. J. Oper. Res. Soc. 45(5), 567–578 (1994)
Edvinsson, L.M.S. Malone: Intellectual capital: realizing your company’s true value by finding its hidden brainpower. Harper Collins, New York (1997)
Galbraith, J.K.: The New Industrial State. Penguin, Harmondsworth (1969)
Golany, B., Roll, Y.: An application procedure for data envelopment analysis. Omega Int. J. Manag. Sci. 3, 237– 250 (1989)
Hu, C.C., Li, H.L.: A company performance measurement by ROIC and DEA—an example of Taiwan IC design house. J. Chin. Inst. Ind. Eng. 21(4), 369–383 (2004)
Hung, S.W. et al.: Vertical disintegration of Taiwan’s semiconductor industries: price and non-price factors. J. Rev. Pac. Basin Financial Mark. Policies 7(4), 547–569 (2004)
Juma, N., Payne, G.T.: Intellectual capital and performance of new venture high-tech firms. Int. J. Innov. Manag. 8, 297–318 (2004)
Khumbakar, C., Lovell, C.A.K.: Stochastic frontier analysis. Cambridge University Press, New York (2000) Kuo, Y.H., Shen, H.L.: An empirical study on measuring performance for global top fifteen IC design
companies—application of data envelopment analysis. Institute of Business Administration, Soochow University, Taipei (2005)
Lu, W.H., Hung, S.W.: Evaluating the profitability and marketability of Taiwan’s IC fabless firms: an DEA approach. J. Sci. Ind. Res. 68, 851–857 (2009)
Luo, X.: Evaluating the profitability and marketability efficiency of large banks: an application of data envel-opment analysis. J. Bus. Res. 56, 627–635 (2003)
Russell, R.R.: Measures of technical efficiency. J. Econ. Theory 35, 109–126 (1985)
Seiford, L.M., Zhu, J.: Infeasibility of super-efficiency data envelopment analysis. INFOR 37(2), 174–187 (1999)
Simar, L., Wilson, P.W.: Estimation and inference in two stage, semi-parametric models of productive effi-ciency. J. Econom. 136, 31–64 (2007)
Stewart, T.A.: Your company’s most valuable asset: intellectual capital. Fortune 130(7), 28–33 (1994) Stewart, T.A.: Intellectual Capital: the new wealth of organizations. Doubleday Dell Publishing Group,
New York (1997)
Sullivan, P.H.: Value-driven intellectual capital: how to convert intangible corporate assets into market value. Wiley, New York (2000)
Tofallis, C.: Improving discernment in DEA using profiling. Omega Int. J. Manag. Sci. 24(3), 361–364 (1996) Tone, K.: A slacks-based measure of efficiency in data envelopment analysis. Eur. J. Oper. Res. 130(3), 498–
509 (2001)
Tone, K.: A slacks-based measure of super-efficiency in data envelopment analysis. Eur. J. Oper. Res.
143(1), 32–41 (2002)
Van Buren, M.E.: A yardstick for knowledge management. Train. Dev. 53(5), 71–78 (1999) Wernerfelt, B.: A resource-based view of the firm. Strateg. Manag. J. 5(2), 171–180 (1984)
Wu, D.D., Ho, C.-T.B.: Productivity and efficiency analysis of Taiwan’s integrated circuit industry. Int. J. Prod. Perform. Manag. 56(8), 715–730 (2007)
Zhu, J.: Multi-factor performance measure model with an application to fortune 500 companies. Eur. J. Oper. Res. 123(1), 105–124 (2000)
