Influence of intellectual capital on financial results of high-tech companies in russia and the bric countries
The non-linear influence of intellectual capital (IC) on financial performance of firms from IC-intensive industries in the BRIC countries. Using Almon Distributed Lag Model. The impact of company’s life-cycle stage on IC benefits for firm’s performance.
| Рубрика | Экономика и экономическая теория |
| Вид | дипломная работа |
| Язык | английский |
| Дата добавления | 27.09.2016 |
| Размер файла | 1,2 M |
Отправить свою хорошую работу в базу знаний просто. Используйте форму, расположенную ниже
Студенты, аспиранты, молодые ученые, использующие базу знаний в своей учебе и работе, будут вам очень благодарны.
Размещено на http://www.allbest.ru/
Размещено на http://www.allbest.ru/
National Research University Higher School of Economics
Faculty of Economic Sciences
Influence of Intellectual Capital on Financial Results of High-Tech Companies in Russia and the BRIC Countries
Master of Science in Finance Thesis
Master's Programme “Strategic Corporate Finance”
Pupshev Andrey
Reviewer Scientific Supervisor
Mariya Molodchik, Ph.D. Sergey Kuzubov, Doctor of Economic Sciences.
Moscow, 2016
Abstract
Purpose ? The main purpose of this paper is to examine the non-linear influence of intellectual capital (IC) on financial performance of firms from IC-intensive industries in the BRIC countries. The additional one is to find out whether firm's life-cycle stage important for the above-mentioned influence or not.
Design/methodology/approach ? The data include financial information of 318 firms from IC_intensive industries (IT, Healthcare, Chemical and Telecom) form the BRIC countries, observed over 16-year period of 2000 to 2015. We used R&D exp. as IC proxy as it is the most appropriate option for high-tech companies. The panel data regressions were conducted using Almon Distributed Lag Model with GMM estimator and Newey-West standard errors.
Findings ? Results support the hypothesis about non-linear relationship between R&D exp. and firm's performance. We found that R&D exp. has significant positive influence on firm's performance after five years and this impact has reversed U-shape with optimal influence in the range of 14.0% to 14.8% of R&D intensity (R&D exp. to Sales ratio). The hypothesis about higher performance increase from IC at mature firms was also confirmed.
Research limitations ? The lack of observations with more than six lags of R&D expenses, the acute absence of balance in the sample in terms of firms' countries are the main limitations of the study.
Practical implications ? Management of high-tech companies may use optimal R&D intensity and frontier of “overinvestment in IC” zone as tentative value in budget planning. Investors may use the findings on mature firms in investment planning.
Originality/value ? This study researches non-linear influence of IC on firm's performance in the BRIC countries, which wasn't researched before. In addition, it opens discussion of impact of company's life-cycle stage on IC benefits for firm's performance.
Keywords ? Intellectual capital, R&D, High-tech firms, BRIC.
Paper type ? Research paper.
financial capital lag intellectual
Contents
- Abstract
- Introduction
- Chapter 1: The literature review
- Classification
- Prior researches with similar methodology
- Prior researches on the BRIC countries data
- Chapter 2: Methodology and data
- Hypotheses
- Methodology: variables
- Dependent variables
- Independent variables
- Control variables
- Methodology: the models
- Data
- Chapter 3: Results and discussion
- Empirical results
- Robustness check
- Research limitations
- Conclusion
- Appendix
- The results of the robustness check
- Density histograms of variables
- Correlation matrixes
- Results of tests on econometric problems
- References
Introduction
Company's financial results are the most important indicators to investors, but they are affected by huge amount of internal and external factors. Nowadays, when economy is becoming more innovative, one of the most significant factors of companies and entire economies is intellectual capital (IC). According to research of Ocean Tomo (2015), up to 84% of firm's market value can be explained by IC rather than traditional sources of value, and this number has grown significantly from 17% in 1975. Since IC is being valued highly at high-tech companies, the problem of IC influence on enterprise's performance is the most important for this type of companies.
Following to growing role of IC in firms' value, scholars had been paying their attention on this phenomenon. IC was firstly mentioned in professional literature in 1925. Many researchers started exploring the IC influence on company performance just 25 years ago. Number of investigations has grown significantly since 1990 (Chart 1). Thus, IC is one of the most significant source of companies' value, and simultaneously it is intensively studied field of finance with significant number of unsolved problems.
Chart 1. The number of documents containing words “Intellectual capital” in Scopus database
Number of papers dedicated to IC in emerging markets is much lower than IC in developed economies. Taking into account that amount of developing countries is almost five times higher than number of developed economies, this gap seems to be enormous.
Speaking of BRIC countries, there are some papers on Brazil (Pitelli(2014)), on Russia (Tovstiga and Tulugurova(2007), Tovstiga and Tulugurova(2009), Andreeva and Garanina(2016)), on India (Kamath(2008), Ghosh and Mondal(2009), Pal and Soriya(2012), Vishnu and Gupta(2014)) and on China (Xinyu(2014), Wang et al. (2014)), and one on the BRICS countries (Ilyin(2014)). Nevertheless there is one paper on the BRIC (even BRICS) countries data, it can't close the gap in the literature because the sample in Ilyin(2014) is almost completely (on 93%) consists of Chinese data.
Albeit, there are some researches on emerging markets and even on the BRIC countries separately, there are no researches on emerging markets for:
· Company's life cycle stage as a factor of IC impact on company performance.
· Non-linear influence of IC on firm's performance.
To remove out the gap, we examine in this paper the relation between the investments in IC proxy (R&D intensity = R&D exp./Sales) with six lags in the first and second power, and firms' financial performance (Residual Income spread and EBITDA margin). We used financial data of 318 companies from the BRIC countries and made regression analysis using GMM estimator with Newey-West standard errors and Almon polynomial distributed lag model. After this regression analysis and robustness check, we found out that companies get significant increase in performance after 5 years since making R&D exp. and this influence is non-linear. This result is consistent with two prior researches of non-linear influence of IC (Huang and Liu (2005), Fredriksson and Wikberg (2015)). This result approved our first hypothesis.
We found the optimal R&D expenses in the range of 14.0% to 14.8% of sales. And we got the frontier of “overinvestment in IC” zone in the range of 28.0% to 29.7% of R&D exp./sales ratio. We believe that these numbers could be useful for management of high-tech companies in the BRIC countries as tentative values in budget planning.
Using the same data and methodology (but with slightly different model), we found that mature companies enjoy higher performance growth from R&D intensity than others. This approved our second hypothesis. This result can be useful in further research of IC and for investors who make decisions about investments in high-tech companies.
Chapter 1: the literature review
Classification
Researches of influence of IC on company's performance can be classified by two factors:
· Origin of data (developed or emerging market is under investigation)
· Nature of analyzed companies (Public companies or SMEs -- small and medium enterprises)
Such classification with appropriate research papers is on the Table 1. Looking on the table, we can see that authors mostly focused on public companies in developed markets. The first reason for that is the number of innovative companies in prosperous countries. The second reason is that these countries have a higher share of public companies due to highly developed financial markets. The share of investigation on SME's is relatively lower because of difficulties to get data (the survey must be done).
Table 1. Classification of research papers by originality of data and publicity of companies
|
Public companies |
SMEs |
||
|
Developed markets |
Lev and Sougiannis(1996), Deng et al.(1999), Chen et al.(2005), Huang and Liu(2005), Ng(2006), Tan et al.(2007), Kujansivu and Lцnnqvist(2007), Wang(2008), Liang and Lin(2008), Sriram(2008), Liu et al.(2009), Laing et al.(2010), Zeghal and Maaloul(2010), Maditinos et al.(2011), Clarke et al.(2011), Guo et al.(2012), Naidenova and Parshakov(2013), Tseng et al.(2015), Goebel(2015) |
Castro and Sбez(2008), Durst(2008), Tovstiga and Tulugurova(2009), Diez et al.(2010) |
|
|
Emerging markets |
Kamath(2008), Ghosh and Mondal(2009), Chu et al.(2011), Mehralian et al.(2012), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Pucar(2012), Kweh et al.(2013), Vishnu and Gupta(2014), Moriariu(2014), Pitelli(2014), Ilyin(2014), Dћenopoljac et al. (2016) |
Tovstiga and Tulugurova(2007), Tovstiga and Tulugurova(2009), F-Jardуn and Martos(2009), Steenkamp and Kashyap(2010), Boujelben and Fedhila(2011), Edige and Su(2014), Wang et al. (2014), Andreeva and Garanina (2016) |
Table 2. Classification of research papers by methodology. The main methodologies used in researches (Table 2) are:
|
Method |
Researches |
|
|
Case study |
Ng(2006), Laing et al.(2010) |
|
|
Lukert scale survey + analysis |
Castro and Sбez(2008), Durst(2008), Tovstiga and Tulugurova(2009), Diez et al.(2010), Tovstiga and Tulugurova(2007), F-Jardуn and Martos(2009), Steenkamp and Kashyap(2010), Boujelben and Fedhila(2011), Edige and Su(2014), Wang et al. (2014), Andreeva and Garanina (2016) |
|
|
Regression analysis of public data |
Lev and Sougiannis (1996), Deng et al.(1999), Chen et al.(2005), Huang and Liu(2005), Tan et al.(2007), Kujansivu and Lцnnqvist(2007), Wang(2008), Liang and Lin(2008), Sriram(2008), Kamath(2008), Ghosh and Mondal(2009), Liu et al.(2009), Zeghal and Maaloul(2010), Maditinos et al.(2011), Clarke et al.(2011), Chu et al.(2011), Mehralian et al.(2012), Komneniж and Pokrajиiж(2012), Guo et al.(2012), Pal and Soriya(2012), Pucar(2012), Kweh et al.(2013), Naidenova and Parshakov(2013), Vishnu and Gupta(2014), Moriariu(2014), Pitelli(2014), Ilyin(2014), Tseng et al.(2015), Goebel(2015), Fredriksson and Wikberg (2015), Dћenopoljac et al. (2016) |
· Case study. Comprehensive research of several companies.
· Lukert scale survey of SMEs with subsequent descriptive, correlation or regression analysis
· Regression analysis of publicly available information on companies
This research is dedicated to public companies on emerging markets and has “regression analysis of public data” methodology.
In literature, the most widely used regression model has the following form:
As a dependent variable the most frequently applicable variables (table 3) are:
· Market-to-book ratio
· Return on assets
· Return on equity
· Assets turnover (ATO=Revenue/total assets)
· Return on sales
· Growth of revenue
· Tobin's Q
· EBITDA margin
Table 3. Classification of research papers by dependent variable
|
Dependent variable |
Papers |
|
|
M/B ratio |
Goebel(2015), Mehralian et al.(2012), Chen et al.(2005), Zeghal and Maaloul(2010), Ghosh and Mondal(2009), Maditinos et al.(2011), Pal and Soriya(2012), Moriariu(2014), Liang and Lin(2008), Kamath(2008), Chu et al.(2011), Ilyin(2014) |
|
|
ROA |
Mehralian et al.(2012), Chen et al.(2005), Zeghal and Maaloul(2010), Huang and Liu(2005), Ghosh and Mondal(2009), Maditinos et al.(2011), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Vishnu and Gupta(2014), Kamath(2008), Chu et al.(2011), Clarke et al.(2011), Guo et al.(2012), Ilyin(2014), Dћenopoljac et al. (2016) |
|
|
ROE |
Chen et al.(2005), Tan et al.(2007), Maditinos et al.(2011), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Kweh et al.(2013), Moriariu(2014), Chu et al.(2011), Clarke et al.(2011), Dћenopoljac et al. (2016) |
|
|
ATO |
Mehralian et al.(2012), Ghosh and Mondal(2009), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Moriariu(2014), Kamath(2008), Chu et al.(2011), Dћenopoljac et al. (2016) |
|
|
ROS |
Huang and Liu(2005), Vishnu and Gupta(2014) |
|
|
Growth of Revenue (GR) |
Diez et al.(2010), Maditinos et al.(2011), Clarke et al.(2011), Fredriksson and Wikberg (2015) |
|
|
Tobin's Q |
Goebel(2015), Kweh et al.(2013), Pitelli(2014) |
|
|
EBITDA margin |
Lev and Sougiannis (1996), Fredriksson and Wikberg (2015) |
M/B ratio and Tobin's Q are less related to growth opportunities on emerging markets than on developed markets; ROA is vulnerable to accounting rules; ROE can be boost by using high leverage, therefore we use EBITDA margin as accounting measure of firm's performance and residual income as financial measure of firm's performance.
The most common IC proxy variables are:
· R&D intensity (R&D exp./sales)
· VAICTM (Value added intellectual capital coefficient)
· Intangible assets/Total assets
Table 4. Classification of research papers by IC proxy variables
|
Independent variable |
Papers |
|
|
R&D exp./sales |
Lev and Sougiannis (1996), Huang and Liu(2005), Guo et al.(2012), Ilyin(2014), Fredriksson and Wikberg (2015) |
|
|
VAICTM |
Mehralian et al.(2012), Chen et al.(2005), Zeghal and Maaloul(2010), Ghosh and Mondal(2009), Tan et al.(2007), Maditinos et al.(2011), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Pucar(2012), Kweh et al.(2013), Vishnu and Gupta(2014), Moriariu(2014), Pitelli(2014), Kamath(2008), Chu et al.(2011), Clarke et al.(2011), Xinyu(2014), Ilyin(2014), Dћenopoljac et al. (2016) |
|
|
Intangible assets/TA |
Goebel(2015) |
We chose to use R&D intensity because VAICTM is oversimplified measure of IC in our opinion, meanwhile Intangible assets is small part of IC. We think that R&D intensity (with lags), as investment in IC, is the best proxy of IC itself in High-Tech companies because the main performance driver for such firms is ability to create new technologies, which is consequence of R&D expenses made in previous periods.
The most frequently used control variables are:
· Size
· Leverage
· Industry dummy
Table 5. Classification of research papers by control variables
|
Control variable |
Papers |
|
|
Size |
Kamath(2008), Ghosh and Mondal(2009), Zeghal and Maaloul(2010), Chu et al.(2011), Komneniж and Pokrajиiж(2012), Pal and Soriya(2012), Guo et al.(2012), Moriariu(2014), Pitelli(2014), Xinyu(2014), Ilyin(2014), Goebel(2015), Fredriksson and Wikberg (2015), Dћenopoljac et al. (2016) |
|
|
Leverage |
Huang and Liu(2005), Kamath(2008), Ghosh and Mondal(2009), Zeghal and Maaloul(2010), Chu et al.(2011), Clarke et al.(2011), Pal and Soriya(2012), Pitelli(2014), Ilyin(2014), Goebel(2015), Dћenopoljac et al. (2016) |
|
|
Industry dummy |
Huang and Liu(2005), Clarke et al.(2011), Moriariu(2014) |
We use size, leverage, industry dummy and dummy for mature companies as control variables.
Prior researches with similar methodology
Residual income as a dependent variable was used by Ilyin (2014). He used this measure of corporate performance among with other dependent variables.
Guo et al.(2012) used R&D/total assets as an independent variable to research influence on companies performance using 279 US biotech companies during the period from 1995 to 2005 as a sample. They have found out non-stable (significant positive or significant negative, depending on dependent variable) impact.
Chen et. al.(2005) and Goebel (2015) made an attempt to find optimal number of IC-proxy's lags. They investigated how many lags had the higher explanatory power (one, two or three) and which lag length is suitable. Their researches in this field were resultless. Thus, we make an attempt to use the largest number of lags as possible according to our data.
It is quite rarely when authors use lags of independent variables to reflect influence on company performance. Lev and Sougiannis (1996) showed significant positive influence of R&D intensity on US firms' performance in 1976-1991, using 7 lags of R&D intensity. Boujelben and Fedhila (2011) used one lag of intangible investments (R&D, advertising and other) to show its impact on cash flow from operating activity of Tunisian companies. Ng (2006) used two lags of human, structural and innovation capital to find its impact on revenue of Canadian companies. All of these three papers revealed statistically significance of independent variables lags.
Non-linear influence of IC on companies' performance was investigated by Huang and Liu (2005). They used data of 1000 companies in Taiwan and searched influence of (R&D/Sales) and (R&D/Sales)2 on performance indicators. As a result, they have found significant non-linear impact on company performance. Another research of non-linear influence of IC on firm's performance was done by Fredriksson and Wikberg (2015) who investigated impact of first, second and even third power of R&D int. (averaged across 2008-2014) on performance of 209 public manufacturers of industrial equipment around the world. They found positive non-linear influence of R&D on firm's performance and their results showed insignificance of third power of R&D int. We used quite similar approach on BRIC countries data, but using lags of R&D intensity, which is more informative in our opinion. Following the results of Fredriksson and Wikberg (2015) we didn't use the third power of R&D int. in our models limiting to only first and second power.
Influence of life cycle stage is not being used in research on intellectual capital. One of rare papers in this field is Liang and Lin (2008) who tried to find what component of IC has higher explanatory power of company performance for each life cycle stage. Salehi et al.(2013) researched interrelation between company life cycle and influence of IC on company performance and revealed significant differences among different life cycle stages.
Prior researches on the BRIC countries data
Nowadays a few researches of IC on Brazil, Russia, India and China exist. Tovstiga and Tulugurova (2007) investigated influence of IC on performance of Russian SMEs in Saint-Petersburg. They have found out that IC is the most important company driver of competitive performance. Tovstiga and Tulugurova (2009) analyzed Russian, Danish, German and USA small and medium enterprises. According to their results, IC contribution in company performance in these countries is rather similar than different. Andreeva and Garanina (2016) analysed Influence of three parts of IC (structural, human and relational capital) on performance of 240 Russian manufacturing firms, and the found significant positive influence of structural and human capital, meanwhile relational capital turned out to be insignificant.
Pitelli et al. (2014) have made the only investigation on Brazilian data. They explored tangible intensive (real estate industry) public companies in Brazil. They found out that IC has significant negative influence on market value of companies.
India has the highest number of papers on IC among BRIC countries, however, all of them use data on pharmaceutical industry that may bias the results. Bharathi and Kamath (2008) researched largest Indian public pharmaceutical companies during the period from 1995 to 2006. They have found growing positive impact of IC on companies' profitability. Ghosh and Mondal (2009) investigated public software and pharmaceutical companies in India from 2002 to 2006. Also they obtained unstable results (significant positive and non-significant negative) on influence of IC on performance of companies. Pal and Soriya (2012) used data on Indian public intellectual-intensive (pharmaceutical) and tangible-intensive (textile) companies for 2000-2010 period. They have revealed unstable (significant and non-significant positive) interrelations of IC and companies' performance. Vishnu and Gupta (2014) explored large public pharmaceutical companies from 2005 till 2011. As a result, they have revealed significant positive effect of IC on companies' performance.
Xinyu (2014) investigated China's public pharmaceutical companies for period 2010-2012. He has showed up significant positive influence of financial and human capital and non-significant positive influence of structural capital. When exploring high-tech SMEs in China, Wang et al. (2014) have found positive effect of IC on company performance.
Having used data on 2481 companies from BRICS countries during the period from 2005 to 2012 as sample, Ilyin (2014) found positive effect of IC on company performance and on reducing company cost of equity. Though it was the only research on BRICS countries, it could not close the gap in literature because of majority (93% of sample) of Chinese companies.
Chapter 2: methodology and data
Hypotheses
H1: Ceteris paribus, the higher R&D of the company, the higher its performance, and this impact is non-linear.
Most of researches reflect positive (significant or insignificant) effect of intellectual capital on company performance (e.g. Wang (2008), Vishnu and Gupta (2014), Chu et al. (2011)). Thus, we expect the same impact in case of BRIC countries.
Huang and Liu (2005) tried to find out non-linearity for Taiwan companies. They show significance of second-power IC-proxy. We try to do similarly on BRIC countries.
H2: Companies on maturity stage have higher growth in performance from IC than companies on other stages of life cycle
We assume that mature companies have higher experience in working with IC and they have accumulated knowledge base obtained in previous years.
Rare investigation in this field was made by Salehi et al. (2013), who approved this idea and revealed that companies on maturity stage are more influenced by IC.
To identify company's life cycle stage, we use approach proposed by Dickinson (2011) because of its simplicity.
Methodology: variables
Dependent variables
To get reliable results, we use both accounting and financial measure of firm's performance. As financial measure, we use RI spread, which equals to difference between firm's ROIC and WACC. We consider this measure as the most appropriate financial metric, because it is related to company's value.
As accounting measure, we use EBITDA margin because it is related to FCFF (and thus to firm's value), not as volatile as FCFF itself, and less subject to difference in accounting rules than EBIT margin.
Independent variables
As independent variable, we use R&D intensity (R&D expenses/Sales), its square, and dummy equals one for mature companies and zero for other. We consider R&D intensity (with its lags), as investment in IC, is the best proxy of IC itself in High-Tech companies because the main performance driver for such firms is ability to create new technologies, which is consequence of R&D expenses made in previous periods.
To calculate maturity dummy we used methodology of Dickinson (2011) as it is simple in calculation.
Control variables
We used CapEx/Sales ratio as control variable, because firm's performance is dependent on its tangible assets and therefore investments in them. Other control variables were used following other papers (see for example Huang and Liu(2005), Pal and Soriya(2012)): size in form of ln(Sales), leverage (as E/(D+E)), and industry dummies.
The last control variable we used, was maturity dummy, because mature companies, in average, more profitable than firms on other life-cycle stages, which could bias our results if we did not use this variable.
Methodology: the models
To check up the first hypothesis, we use the following model:
We conduct the following tests:
H10: R&D expenses do not have an effect on companies' performance
H11: The effect of R&D expenses on companies' performance is linear
H12: The effect of R&D expenses on companies' performance is non-linear
To check up the second hypothesis, we use the model:
We conduct the following tests:
H20: Companies on maturity stage do not have any advantage in operating with IC
H21: Companies on maturity stage are more effective in working with IC
To ensure in our results, we made the following tests:
· On multicollinearity
· On normality of errors
· On heteroscedasticity
· On autocorrelation
· Robustness check
We revealed strong multicollinearity, non-normality of errors, heteroscedasticity and autocorrelation. To get appropriate results under such conditions, we use GMM estimator with Newey-West standard errors.
For models with lagged R&D expenses, strong multicollinearity is typical (Lev and Sougiannis (1996)). In our case we get VIF equal 24.1 for our variables, which means inappropriately strong correlation between independent variables. For weakening this problem, we use Almon distributed lag model (Almon (1965)) following Lev and Sougiannis (1996). This model allow to mitigate multicollinearity problem by making an assumption that dependence of lags among each other has polynomial form. Thus, we need to evaluate less number of parameters and therefore mitigate multicollinearity problem.
Data
We have got financial information from Bloomberg and Capital IQ databases on companies from the BRIC countries in 2000-2015 which belongs to 4 high-tech industries (table 7). Initial number of companies was 760, but after elimination of duplication and removing companies with insufficient number of R&D lags (we treated R&D exp.=0 as lack of the observation) and deletion of outliers (see table 6), the final sample contained 318 companies, 771 observations. The distribution among countries is on the table 8.
Table 6. List of criteria used to remove outliers.
|
Variable |
Outlier criteria |
|
|
R&D intensity |
<0% or >50% |
|
|
CapEx/Sales |
<0% or >100% |
|
|
EBITDA margin |
<-50% or >100% |
|
|
RI spread |
<-30% or >30% |
|
|
Leverage |
>80% |
Table 7. Breakdown of the data by industries.
|
Industry |
# of companies |
# of observations |
% of the sample |
|
|
IT |
152 |
369 |
47.86% |
|
|
Healthcare |
83 |
188 |
24.38% |
|
|
Chemicals |
73 |
184 |
23.87% |
|
|
Telecommunication |
10 |
30 |
3.89% |
Table 8. Breakdown of the data by countries.
|
Country |
# of companies |
# of observations |
% of the sample |
|
|
Brazil |
3 |
12 |
1.56% |
|
|
Russia |
2 |
3 |
0.39% |
|
|
India |
64 |
179 |
23.21% |
|
|
China |
249 |
577 |
74.84% |
Analysis of the data shows that the sample is more or less balanced by industry, but very unbalanced by countries in favor of China. It is not a good news, but it is objective state of affairs in the BRIC countries: Chinese economy is much bigger and its companies disclose R&D expenses much better than Russian and Brazilian companies do.
On the table 9 there is descriptive statistics on other variables.
Table 9. Descriptive statistics on non-dummy-variables.
|
# of obs. |
Mean |
St. Dev. |
Min |
1st quartile |
Median |
3rd quartile |
Max |
||
|
RI spread |
771 |
-1.52% |
7.61% |
-28.97% |
-21.17% |
-2.30% |
22.77% |
24.74% |
|
|
EBITDA margin |
775 |
16.39% |
13.51% |
-37.35% |
-27.16% |
15.92% |
63.17% |
65.02% |
|
|
R&D int. |
775 |
5.98% |
6.63% |
0.01% |
0.01% |
4.01% |
39.37% |
44.86% |
|
|
CapEx/Sales |
775 |
10.93% |
11.95% |
0.02% |
0.21% |
7.14% |
71.68% |
84.71% |
|
|
D/(D+E) |
775 |
11.86% |
17.06% |
0.00% |
0.00% |
4.16% |
74.25% |
79.80% |
|
|
Ln(Sales) |
775 |
5.54 |
1.64 |
1.43 |
2.04 |
5.26 |
10.57 |
10.88 |
Chapter 3: results and discussion
Empirical results
Table 10 presents the results of the research. The first two columns were made to check the first hypothesis. The zero lag of R&D int. significantly negative for both performance indicators, which is expected result, because R&D expenses as pre-tax investment worsen firm's performance in the current year for sure. The first, second, fourth and sixth lags turned out to be insignificant, while the third lag demonstrates significant influence only for one performance indicator. We treat such cases as random significance and consider these coefficients as insignificant. We observe stable significant positive influence of fifth lag of R&D intensity on firm's performance. This lag is significant at 1% level for both performance indicators.
Speaking about second power of R&D intensity, we see that there are some lags (zero, second and fourth) which significant only for one performance indicator, one lag (sixth) which insignificant for both dependent variables. We consider all of them as insignificant. There is one lag (the first) that is significant for both performance indicators, but only on 10% level. We treat this lag as insignificant due to low significance level and especially because the robustness check (see below) revealed that this significance is unstable. There is also the third lag, that significant at 5% level for both dependent variables, which seems to be solid, but the robustness check showed that this significance is unstable too. There is only one lag significant at 1% for both performance indicators and which pass the robustness check ? fifth one. This confirms our Hypothesis 1: impact of IC on firm's performance is non-linear.
The last two columns present the results of regressions for Hypothesis 2. We need to analyse coefficients of ((R&D int.)*Maturity) only, because this is the clue for the hypothesis. We see a lot of lags significant for one dependent variable, we consider them as insignificant. But there are two lags of ((R&D int.)*Maturity) with positive significant at 1% impact on firm's performance: fourth and fifth, and there is no any negative coefficient significant for both performance indicators. This obviously supports our Hypothesis 2: companies on mature life-cycle stage get higher performance benefit from IC.
Table 10. Empirical results.
|
Model 1 |
Model 2 |
||||
|
RI spread |
EBITDA margin |
RI spread |
EBITDA margin |
||
|
R&D int. |
-0.562*** |
-0.690** |
-0.282*** |
-0.798*** |
|
|
Lag 1 |
0.240 |
0.429 |
-0.094 |
-0.111 |
|
|
Lag 2 |
-0.028 |
-0.062 |
0.030 |
0.183** |
|
|
Lag 3 |
-0.255** |
-0.315 |
0.015 |
0.065 |
|
|
Lag 4 |
-0.016 |
0.273 |
-0.091 |
-0.241** |
|
|
Lag 5 |
0.435** |
1.054*** |
-0.120 |
-0.268 |
|
|
Lag 6 |
0.157 |
0.137 |
0.217** |
0.693*** |
|
|
(R&D int.)2 |
0.793* |
-0.342 |
|||
|
Lag 1 |
-0.873* |
-1.616* |
|||
|
Lag 2 |
0.340 |
0.928** |
|||
|
Lag 3 |
1.004*** |
1.492** |
|||
|
Lag 4 |
0.003 |
-1.174* |
|||
|
Lag 5 |
-1.474*** |
-3.767*** |
|||
|
Lag 6 |
0.072 |
1.561 |
|||
|
(R&D int.)*Maturity |
-0.130* |
-0.029 |
|||
|
Lag 1 |
0.115** |
0.107 |
|||
|
Lag 2 |
0.108*** |
0.057 |
|||
|
Lag 3 |
0.093** |
0.132 |
|||
|
Lag 4 |
0.152*** |
0.379*** |
|||
|
Lag 5 |
0.214*** |
0.585*** |
|||
|
Lag 6 |
0.046 |
0.277** |
|||
|
CapEx/Sales |
0.006 |
0.164*** |
-0.007 |
0.141*** |
|
|
Lag 1 |
-0.037** |
0.003 |
-0.036** |
0.025 |
|
|
Lag 2 |
-0.023* |
0.017 |
-0.025** |
0.017 |
|
|
Lag 3 |
0.002 |
0.066** |
-0.004 |
0.049* |
|
|
Lag 4 |
0.011 |
0.080*** |
0.010 |
0.075*** |
|
|
Lag 5 |
0.004 |
0.054 |
0.010 |
0.077** |
|
|
Lag 6 |
0.002 |
0.049 |
0.000 |
0.061 |
|
|
Leverage |
-0.189*** |
-0.211*** |
-0.182*** |
-0.206*** |
|
|
Ln(Sales) |
0.020*** |
0.028*** |
0.019*** |
0.025*** |
|
|
Healthcare |
0.042*** |
0.059*** |
0.039*** |
0.050*** |
|
|
Chemicals |
0.047*** |
0.023* |
0.046*** |
0.012 |
|
|
Telecom |
-0.004 |
-0.036* |
0.000 |
-0.022 |
|
|
Maturity |
0.017*** |
0.037*** |
0.017*** |
0.023** |
|
|
Constant |
-0.122*** |
-0.065*** |
-0.113*** |
-0.021 |
|
|
R2 |
40.15% |
38.60% |
43.36% |
40.70% |
|
|
# of observations |
765 |
769 |
731 |
735 |
Robustness check
To check our results, we made robustness checks for all our regressions. We made it by using different subsamples. We excluded approximately 20% of our sample from different regions of the sample. Thus, we have got 6 additional regression results on 6 subsamples.
Tables 11 and 12 (in Appendix) present results of such robustness check, where we can see that R&D intensity stable impacts on firm's performance in the current year (negative influence) and after five years (reversed U-shape influence, see figure 1), other lags are not stable significant, or even insignificant. Depending on the dependent variable used, we get optimum R&D intensity in the range of 14.0% to 14.8% and maximum useful R&D intensity in the range of 28.0% to 29.7%.
Figure 1. Reversed U-shape of fifth lag of R&D int. influence on firm's performance.
Prior researches of non-linear influence of IC on firm's performance found the same shape of influence. Huang and Liu (2005) revealed the optimum R&D int. for Taiwanese companies in the range of 6.3% to 6.4% of sales, meanwhile Fredriksson and Wikberg (2015) find out the same optimum for manufacturers of industrial equipment around the world in the range of 3.9% to 4.1% of sales. Of course, the maximum useful R&D int. in these two papers just two times higher than optimum one. Albeit, our results on optimal R&D int. and maximum useful R&D int. are significantly higher than in two above-mentioned papers, we think that it is because we study fast grew emerging markets, whereas Huang and Liu (2005) and Fredriksson and Wikberg (2015) explored mainly slower and more stable developed market.
Tables 13 and 14 (in Appendix) presents robustness check for the second hypothesis. There we can find stable positive influence of maturity life-cycle stage on getting benefits from R&D expenses.
Thus, we consider our results on the both hypotheses robust.
Research limitations
Unfortunately, our sample is very unbalanced in favour of China (almost 75% of our sample) and has too few observations on Russia and Brazil (fewer than 2% of the sample totally). Additionally, our sample has not so much companies with high number of continuous years with data on R&D expenses. We would like to make research with not less than ten lags of R&D expenses, but nowadays it is impossible.
Another limitation is related to R&D exp. as measure of IC investment. Although, in our opinion, R&D exp. is the best measure of IC investment for high-tech companies, but still it is not perfect one. Companies can (and they really do, of course) some investments in IC, which is not included in R&D exp.
Conclusion
This study approved that R&D expenses, as investment in IC, have non-linear influence on financial results of high-tech firms from the BRIC countries. We revealed that R&D exp. have significant effect on firm's performance after five years and this influence has reversed U-shape. We also got optimal R&D exp. in the range from 14.0% to 14.8% of sales, and we found overinvestment zone of R&D exp. starting from 28.0%-29.7% of sales and higher.
This paper also demonstrates that mature companies in the BRIC countries enjoy higher benefits from R&D exp. than firms on other life-cycle stages.
This research may be expanded to wider spectrum of countries to obtain the whole picture of non-linear influence of R&D exp. on firm's performance around the world. Other possible direction of further research is to replicate this paper in some years when there will be more data on R&D exp. in the BRIC countries, and include higher number of lags in the model.
References
1. Alain Daou Egide Karuranga Zhan Su , (2014),"Towards a better understanding of intellectual capital in Mexican SMEs", Journal of Intellectual Capital, Vol. 15 Iss 2 pp. 316 - 332
2. Almon, S. (1965). The distributed lag between capital appropriations and expenditures. Econometrica: Journal of the Econometric Society, 178-196.
3. Andreeva, T., Garanina, T. (2016). Do all elements of intellectual capital matter for organizational performance? Evidence from Russian context. Journal of Intellectual Capital, 17(2), 397-412
4. Andriessen, D. (2004), Making Sense of Intellectual Capital: Designing a Method for the Valuation of Intangibles, Elsevier Butterworth-Heinemann, Burlington, MA
5. Artie W. Ng, (2006),"Reporting intellectual capital flow in technology-based companies", Journal of Intellectual Capital, Vol. 7 Iss 4 pp. 492 - 510
6. Baruch Lev, Theodore Sougiannis (1999) "Penetrating the Book-to-Market Black Box: The R&D Effect", Journal of Business Finance & Accounting , Vol. 26, No. 3-4, pp. 419-449
7. Biserka Komneniж, Dragana Pokrajиiж, (2012),"Intellectual capital and corporate performance of MNCs in Serbia", Journal of Intellectual Capital, Vol. 13 Iss 1 pp. 106 - 119
8. Bontis, N., Keow, W. and Richardson, S. (2000), “Intellectual capital and business performance in Malaysian industries”, Journal of Intellectual Capital, Vol. 1 No. 1, pp. 85-100
9. Carlos Maria F-Jardуn Maria Susana Martos, (2009),"Intellectual capital and performance in wood industries of Argentina", Journal of Intellectual Capital, Vol. 10 Iss 4 pp. 600 - 616
10. Chen, M.-C., Cheng, S.-J. and Hwang, Y. (2005), “An empirical investigation of the relationship between intellectual capital and firms' market value and financial performance”, Journal of Intellectual Capital, Vol. 6 No. 2, pp. 159-176.
11. Cheng Jen Huang Chun Ju Liu, (2005),"Exploration for the relationship between innovation, IT and performance", Journal of Intellectual Capital, Vol. 6 Iss 2 pp. 237 - 252
12. Chiung-Ju Liang Ying-Li Lin, (2008),"Which IC is more important? A life-cycle perspective", Journal of Intellectual Capital, Vol. 9 Iss 1 pp. 62 - 76
13. Cristina Maria Morariu , (2014),"Intellectual capital performance in the case of Romanian public companies", Journal of Intellectual Capital, Vol. 15 Iss 3 pp. 392 - 410
14. Daniel Pitelli Britto Eliane Monetti Joao da Rocha Lima Jr , (2014),"Intellectual capital in tangible intensive firms: the case of Brazilian real estate companies", Journal of Intellectual Capital, Vol. 15 Iss 2 pp. 333 - 348
15. Daniel Zйghal Anis Maaloul, (2010),"Analysing value added as an indicator of intellectual capital and its consequences on company performance", Journal of Intellectual Capital, Vol. 11 Iss 1 pp. 39 - 60
16. Day-Yang Liu Kuo-An Tseng Szu-Wei Yen, (2009),"The incremental impact of intellectual capital on value creation", Journal of Intellectual Capital, Vol. 10 Iss 2 pp. 260 - 276
17. Dickinson, V. (2010), "Cash Flow Patterns as a Proxy for Firm Life Cycle", THE ACCOUNTING REVIEW, Vol. 86, No. 6, pp. 1969-1994
18. Dimitrios Maditinos Dimitrios Chatzoudes Charalampos Tsairidis Georgios Theriou, (2011),"The impact of intellectual capital on firms' market value and financial performance", Journal of Intellectual Capital, Vol. 12 Iss 1 pp. 132 - 151
19. Dћenopoljac, V., Janoљevic, S., & Bontis, N. (2016). Intellectual capital and financial performance in the serbian ICT industry. Journal of Intellectual Capital, 17(2), 373-396
20. Ferenhof, H.A., Durst , S., Bialecki, M.Z. and Selig P.M. (2015) "Intellectual capital dimensions: state of the art in 2014", Journal of Intellectual Capital, Vol. 16 No. 1, pp. 58-100
21. Fredriksson, N., Wikberg, J. (2015) "The Relationship Between R&D Spending and Firm Economic Performance" (Master thesis, Chalmers University of Technology, Gothenburg, Sweden). Retrieved from http://publications.lib.chalmers.se/records/fulltext/218108/ 218108.pdf
22. G. Bharathi Kamath, (2008),"Intellectual capital and corporate performance in Indian pharmaceutical industry", Journal of Intellectual Capital, Vol. 9 Iss 4 pp. 684 - 704
23. George Tovstiga Ekaterina Tulugurova, (2007),"Intellectual capital practices and performance in Russian enterprises", Journal of Intellectual Capital, Vol. 8 Iss 4 pp. 695 - 707
24. George Tovstiga Ekaterina Tulugurova, (2009),"Intellectual capital practices: a four-region comparative study", Journal of Intellectual Capital, Vol. 10 Iss 1 pp. 70 - 80
25. Gholamhossein Mehralian Ali Rajabzadeh Mohammad Reza Sadeh Hamid Reza Rasekh,(2012), "Intellectual capital and corporate performance in Iranian pharmaceutical industry", Journal of Intellectual Capital, Vol. 13 Iss 1 pp. 138 - 158
26. Gregorio Martнn de Castro Pedro Lуpez Sбez, (2008),"Intellectual capital in high-tech firms", Journal of Intellectual Capital, Vol. 9 Iss 1 pp. 25 - 36
27. Gregory Laing Jillian Dunn Susan Hughes-Lucas, (2010),"Applying the VAIC™ model to Australian hotels", Journal of Intellectual Capital, Vol. 11 Iss 3 pp. 269 - 283
28. Hong Pew Tan David Plowman Phil Hancock, (2007),"Intellectual capital and financial returns of companies", Journal of Intellectual Capital, Vol. 8 Iss 1 pp. 76 - 95
29. Ilyin D. S. (2014), "The Impact of Intellectual Capital on Companies' Performance: Evidence from Emerging Markets", Journal of Corporate Finance Research, Vol. 32, Iss. 4, pp. 46-66
30. Iuliia Naidenova Petr Parshakov , (2013),"Intellectual capital investments: evidence from panel VAR analysis", Journal of Intellectual Capital, Vol. 14 Iss 4 pp. 634 - 660
31. Jose Marнa Dнez Magda Lizet Ochoa M. Begoсa Prieto Alicia Santidriбn, (2010),"Intellectual capital and value creation in Spanish firms", Journal of Intellectual Capital, Vol. 11 Iss 3 pp. 348 - 367
32. Jui-Chi Wang, (2008),"Investigating market value and intellectual capital for S&P 500", Journal of Intellectual Capital, Vol. 9 Iss 4 pp. 546 - 563
33. Karam Pal Sushila Soriya, (2012),"IC performance of Indian pharmaceutical and textile industry", Journal of Intellectual Capital, Vol. 13 Iss 1 pp. 120 - 137
34. Lev, B., D. Nissim, and J. Thomas. 2008. On the Informational Usefulness of R&D Capitalization and Amortization. Visualising Intangibles: Measuring and Reporting in the Knowledge Economy.
35. Lev, B., Sougiannis, T. (1996) "The Capitalization, Amortization, and Va
36. lue-Relevance of R&D", Journal of Accounting and Economics, Vol. 21, pp. 107-138
37. Lev, Baruch. Intangibles: Management, Measurement, and Reporting. Washington, D.C.: Brookings Institution, 2001. Print.
38. Martin Clarke Dyna Seng Rosalind H. Whiting, (2011),"Intellectual capital and firm performance in Australia", Journal of Intellectual Capital, Vol. 12 Iss 4 pp. 505 - 530
39. Ming-Chin Chen Shu-Ju Cheng Yuhchang Hwang, (2005),"An empirical between intellectual capital and firms' market value and financial performance", Journal of Intellectualinvestigation of the relationship Capital, Vol. 6 Iss 2 pp. 159 - 176
40. "Muhammad Khalique Nick Bontis Jamal Abdul Nassir bin Shaari Abu Hassan Md. Isa , (2015),""Intellectual capital in small and medium enterprises in Pakistan"", Journal of Intellectual Capital, Vol. 16 Iss 1 pp. 224 - 238"
41. Natasja Steenkamp Varsha Kashyap, (2010),"Importance and contribution of intangible assets: SME managers' perceptions", Journal of Intellectual Capital, Vol. 11 Iss 3 pp. 368 - 390
42. Ocean Tomo (2015) "Intangible Assets Market Value Study"
43. Paula Kujansivu Antti Lцnnqvist, (2007),"Investigating the value and efficiency of intellectual capital", Journal of Intellectual Capital, Vol. 8 Iss 2 pp. 272 - 287
44. Pulic, A. (2000a), “VAIC - an accounting tool for IC management”, International Journal of Technology Management, Vol. 20 Nos 5-8, pp. 702-14.
45. Qian Long Kweh Yee Chuann Chan Irene Wei Kiong Ting, (2013),"Measuring intellectual capital efficiency in the Malaysian software sector", Journal of Intellectual Capital, Vol. 14 Iss 2 pp. 310 - 324
46. Ram S. Sriram, (2008),"Relevance of intangible assets to evaluate financial health", Journal of Intellectual Capital, Vol. 9 Iss 3 pp. 351 - 366
47. Salehi, M., Valipour, H., Moradi, J. (2013), "Moderating Effect on the Relationship Between a Companies's Life Cycle and the Relevance of Accounting Practices Intangible Assets", Asian Economic and Financial Review, 2013, 3(8):1096-1109
48. Samuel Kai Wah Chu Kin Hang Chan Wendy W.Y. Wu, (2011),"Charting intellectual capital performance of the gateway to China", Journal of Intellectual Capital, Vol. 12 Iss 2 pp. 249 - 276
49. Santanu Ghosh and Amitava Mondal, (2009),"Indian software and pharmaceutical sector IC and financial performance", Journal of Intellectual Capital, Vol. 10 Iss 3 pp. 369 - 388
50. Saoussen Boujelben Hassouna Fedhila, (2011),"The effects of intangible investments on future OCF", Journal of Intellectual Capital, Vol. 12 Iss 4 pp. 480 - 494
51. "Sirinuch Nimtrakoon , (2015),""The relationship between intellectual capital, firms' market value and
52. financial performance"", Journal of Intellectual Capital, Vol. 16 Iss 3 pp. 587 - 618"
53. Stevo Pucar, (2012),"The influence of intellectual capital on export performance", Journal of Intellectual Capital, Vol. 13 Iss 2 pp. 248 - 261
54. Sudi Sudarsanam Ghulam Sorwar Bernard Marr, (2006),"Real options and the impact of intellectual capital on corporate value", Journal of Intellectual Capital, Vol. 7 Iss 3 pp. 291 - 308
55. Susanne Durst, (2008),"The relevance of intangible assets in German SMEs", Journal of Intellectual Capital, Vol. 9 Iss 3 pp. 410 - 432
56. Tseng, K.-A., Ching-I, L. and Szu-Wei, Y. (2015) "Contingencies of intellectual capitals and financial capital on value creation. Moderation of business cycles", Journal of Intellectual Capital, Vol. 16 No. 1, pp. 156-173
57. Viktoria Goebel, (2015),"Estimating a measure of intellectual capital value to test its determinants", Journal of Intellectual Capital, Vol. 16 Iss 1 pp. 101 - 120
58. Vishnu Sriranga Kumar Gupta Vijay , (2014),"Intellectual capital and performance of pharmaceutical firms in India", Journal of Intellectual Capital, Vol. 15 Iss 1 pp. 83 - 99
59. Wen-Chung Guo , Shin-Rong Shiah-Hou & Wei-Jer Chien (2012) A study on intellectual capital and firm performance in biotech companies, Applied Economics Letters, 19:16, 1603-1608
60. Zhang Xinyu (2014), Journal of Chemical and Pharmaceutical Research, Vol. 6(4), pp. 999-1004
61. Zhen Deng, Baruch Lev, and Francis Narin (1999) "Science and Technology as Predictors of Stock Performance", Financial Analysts Journal, Vol. 55, No. 3, pp. 20-32
62. Zhen, D., Lev, B. (2006) "In-process R&D: To capitalize or expense?", Journal of engineering and technology management jet-m
63. Zhining Wang Nianxin Wang Huigang Liang , (2014),"Knowledge sharing, intellectual capital and firm performance", Management Decision, Vol. 52 Iss 2 pp. 230 - 258
Appendix
The results of the robustness check
Table 11. Robustness check of the regression RI spread = f(R&D int., (R&D int.)2, …).
|
RI spread |
RI spread |
RI spread |
RI spread |
RI spread |
RI spread |
RI spread |
||
|
R&D int. |
-0.562*** |
-0.560** |
-0.757*** |
-0.628*** |
-0.492*** |
-0.569*** |
-0.478** |
|
|
Lag 1 |
0.240 |
0.254 |
0.276 |
0.218 |
0.235 |
0.325** |
0.187 |
|
|
Lag 2 |
-0.028 |
0.013 |
0.006 |
-0.046 |
-0.011 |
-0.026 |
0.128 |
|
