Co-integration Analysis of the Relationship between New Product Output and R&D Investment in China’s Pharmaceutical Industry

Li Gan,Wang Su,Chen Yuwen

(School of Business Administration,Shenyang Pharmaceutical University,Shenyang 110016,China)

Abstract Objective To study the possible relationship between the output of new products in China’s pharmaceutical industry and the investment in research and development (R&D),and to provide a theoretical basis for the decision-making of relevant enterprises and institutions.Methods The econometric software Stata 14 was used to perform unit root test on the relevant data.Then,a co-integration regression equation was established after stabilization,which was analyzed through co-integration test (E-G two-step method).Results and Conclusion There is a long-term equilibrium and short-term error correction relationship between the output of new products and the investment of R&D funds in China’s pharmaceutical industry.During the lagging periods from 1 to 6,R&D investment is the Granger reason for the output of new products.The investment of R&D funds has a positive effect on the output of new products and the effect is significant.Therefore,more investment should be made in R&D to enhance the output of new products.

Keywords: pharmaceutical production;R&D input;new product output;co-integration theory

As a high-tech industry,the most obvious feature of pharmaceutical industry is the intensiveness of knowledge and technology.The key to its future development lies in the cultivation and improvement of technological innovation capabilities.Research and development (R&D) input is an important indicator to measure the resource input capacity of technological innovation,and the output of new products is the indicator to show the output capacity of technological innovation.These two indicators are playing a vital role in improving the competitiveness of China’s pharmaceutical industry.Therefore,studying the interaction between the output of new products and R&D input is of great significance to enhance technological innovation capabilities of China’s pharmaceutical industry.

1 Literature review

The research on the correlation between output and R&D input at home and abroad has always been a hot topic for scholars.It is necessary to review the previous research results,and the relevant content is displayed as follows.

1.1 Foreign research

WN Leonard (1971)[1]found that R&D investment would have an impact on the company’s sales and profit growth in the second year,and this impact would continue to rise steadily for the next nine years through his research on American manufacturing.AG Hu (2001)[2]also found that scientific and technological investment had a significant impact on enterprise performance through the study of Chinese industrial enterprises.H Grabowski,et al.(2002)[3]studied the input and output of new drugs and found that product sales revenue would stimulate the R&D investment of enterprises.

1.2 Domestic research

Kuang Licong,et al.(2015)[4],based on the relevant data of the R&D input and output of China’s pharmaceutical industry from 1998 to 2012,established multiple regression and the ALMON polynomial distribution lag model,and concluded that the R&D input had a significant positive effect on output.Besides,its impact had a time lag.Fu Shuyong,et al.(2014)[5]used co-integration regression and other econometric methods to conclude that there was a long-term equilibrium relationship between pharmaceutical innovation input and economic development in Liaoning Province,and its economic development was the Granger causality of pharmaceutical innovation input.Liu Qiang,et al.(2015)[6]used the co-integration theory to analyze relevant data from the “Statistical Yearbook of China’s High-tech Industry” from 1995 to 2013,and found that there was a long-term equilibrium and shortterm error correction relationship between the R&D investment and the sales revenue of new products of large enterprises in China’s pharmaceutical industry,and new product sales revenue was the Granger reason for R&D investment.Yang Xin,et al.(2019)[7],based on the data from the “Statistical Yearbook of China’s High-tech Industry” released by the National Bureau of Statistics from 1995 to 2015,studied the relationship between the sales revenue of new products and the investment in R&D by using the cointegration theory,and obtained the economic output of new products might have a positive impact on the investment of R&D,but when the effect is not significant,the investment of R&D funds could affect the economic output of new products,but there was no benign relationship between them.

Based on the above studies,it can be found that foreign studies on the relationship between sales revenue and R&D investment mostly concentrated on enterprises,which rarely involved the pharmaceutical industry.Although some domestic studies involved the pharmaceutical industry,they often focused on the economic output,and the sales revenue was used as a measure.Besides,these research conclusions were not consistent.

The output capability of technological innovation is an important indicator for evaluating the technological innovation capability of an enterprise.The output capability of technological innovation usually includes patent output and new product output.New product output is usually measured by new product sales revenue or the proportion of new product sales revenue to main business income[8].Through literature analysis,it is found that the sales revenue of new products is often used to measure the output of new products,and the sales revenue of new products will stimulate the R&D investment of enterprises[9,10].Furthermore,taking the proportion of new product sales revenue (main business revenue) as a measure of new product output can not only exclude the impact of enterprise size,but also explain the impact of new product sales revenue on R&D investment[11].Therefore,the data of the pharmaceutical industry from 1995 to 2018 are selected as a sample.The ratio of new product sales revenue and main business revenue is used as a measure of new product output,and internal R&D expenditures are used as a measure of R&D investment to study the relationship between the output of new products and R&D inputs.Besides,empirical research on long-term equilibrium and short-term error correction is conducted to provide a theoretical basis for the decision-making of the government and related enterprises.

2 Research design

2.1 Indicators and data

2.1.1 Indicators

The explanatory variable is the output of new products,which is measured by new product sales revenue/ main business revenue.Since the ratio can exclude the influence of scale,it can also explain the relationship between new product sales revenue and R&D input,which is more convincing.

The interpreted variable is R&D investment,which is measured by the internal expenditure of R&D.According to the interpretation of indicators provided by “Statistical Yearbook of China’s High-tech Industry”,the internal expenditure of R&D refers to the actual spending of the survey unit for internal R&D activities in the reporting year[12].Therefore,the use of internal expenditures of R&D can show the real situation clearly.

2.1.2 Data

Pharmaceutical industry is selected as the object in this paper to conduct research on the relationship between R&D input and new product output from the overall industry level.In view of the availability of data,the relevant data in the “Statistical Yearbook of China’s High-tech Industry” from 1995 to 2018 are selected.Since the data come from the government,the data of 2017 is missing,and the data missing is not serious.In order to maintain the sample size,the method of linear interpolation is used to fill in the missing data[13].Excel and Stata14 software is used for data analysis.

The output of new products in China’s pharmaceutical industry (new product sales revenue/main business revenue) is represented byx,and the internal expenditure of R&D is represented byy.In order to eliminate the influence of heteroscedasticity of the original variables,the variablesxandyare logarithmically processed,and they are marked as lnxand lny.The correlation between variables after logarithmic processing will not change.Related data are shown in Table 1.The time trend diagram of lnxand lnyobtained through software is shown in Fig.1.It can be seen from the figure that the trend of new product output and R&D input changing with time is roughly consistent,which means the original variables are non-stationary.

Fig.1 Time trend of new product output and internal R&D expenditure (100 million yuan) from 1995 to 2018

Table 1 Data of new product sales revenue/ main business revenue (x) and internal R&D expenditure (y)from 1995 to 2018 in China’s pharmaceutical industry

(to be continued)

Continued Table 1

2.2 Research methods

The new product output and R&D input data in this article have a time trend.In view of the non-stationary nature of the data,if the traditional least squares method is used for regression analysis,it may produce pseudo-regression and cause inaccurate results.For non-stationary series,the co-integration theory and method proposed by Engle and Granger in 1987 provides a way to analyze this non-stationary time series[14].According to the definition of co-integration theory,although some economic variables are non-stationary series,some of their linear combinations (such as the combination of differential series) may be stable.This linear combination reflects the long-term stable proportional relationship between the variables,which is called the co-integration relationship[15].Therefore,co-integration theory is used to analyze the relationship between new product output and R&D input.

The analysis of co-integration theory mainly includes unit root test,co-integration test,establishing an error correction model and Granger causality test.The analysis of time series must ensure that the sequence is stationary first.Generally,the ADF test(augment Dickey Fuller test) is used to ensure the stability of the sequence.For the two variables in this article,the Engle-Granger test (EG test) is often used.The EG test is divided into two steps.First,the cointegration regression equation is established,and then the stationarity test is performed on the residual term of the equation.The stationarity test also uses the ADF test.After the residual items pass the stationarity test,it indicates that there is a co-integration relationship between the variables.Error correction model (ECM)is used to describe the short-term dynamic equilibrium between variables.ECM can reflect the short-term deviation of the variables to correct the long-term equilibrium and make up for the shortcomings in the long-term equilibrium model.According to Granger’s principle,there must be an ECM between variables that have a co-integration relationship.Finally,Granger test is used to determine the mutual influence of variables and statistical causality.

3 Empirical research

3.1 Unit root test

The ADF unit root test method and the software Stata 14 are used to test the stationarity of variables lnxand lnyin this paper.The results are shown in Table 2.Through the test,it can be found that both lnxand lnyhave unit roots at the 5% significance level,which are non-stationary series.After the original sequence is processed by the first-order difference,they are recorded as Δlnxand Δlny,respectively.The test results show that Δlnxand Δlnyreject the null hypothesis that there is a unit root at the 5%significance level.The sequence does not have a unit root after the first-order difference,and the sequence is a stationary first-order single integer sequence,which satisfies the condition of co-integration test.

Table 2 ADF unit root test results of lnx and lny

3.2 Co-integration test

The ADF test shows that the first-order difference of the original sequence is stable.Therefore,the E-G two-step method can be used to perform the co-integration test on the variables,which is specifically divided into the following two steps.

3.2.1 Establishing co-integration regression equation

The co-integration regression equation is obtained after analysis by the software Stata 14.

lny=4.244 5 lnx+12.790 0,where the goodness of fit (also known as the coefficient of determination)R2=0.940 4,the adjustedR2=0.937 8,andF=374.59,indicating that the regression equation fits well.Then,the residual of equation s is tested.

3.2.2 Stationarity test of residual items

Given that the regression equation is lny=4.244 5 lnx+12.790 0,when RESID=ECM,ECM=lny-4.244 5 lnx-12.790 0,the ADF test is carried out for the residual term without the constant term and trend term.The results are shown in Table 3.It can be seen from the table that when the significance level is at 5%,the residual sequence is stationary.Therefore,lnxand lnyhave a long-term equilibrium relationship,which means there is a co-integration relationship.From the coefficient of lnxin the equation of 4.244 5,it can be seen that per SD change in the output of new products in China’s pharmaceutical industry will increase the R&D input by 4.244 5 SD.

Table 3 ADF test results of residual items

3.3 Error correction model

Due to the co-integration relationship between the variables,the coefficients are obtained through the software,and the error correction model is as follows:Δlny=1.018 8 Δlnx-0.151 6 ECM (-1)+0.099 3.

It can be seen from the formula that the error correction coefficient is -0.1516 ＜ 0,which is in line with the reverse correction mechanism,indicating that for per SD change in the output of new products in China’s pharmaceutical industry in the short term,the R&D input will have 0.151 6 SD reversely.However,this reverse change is smaller than the long-term regression coefficient,indicating that the long-term impact is more significant.

3.4 Granger causality test

Granger causality test is performed on the lagging periods from 1 to 6 through the software.The test results are shown in Table 4.It can be seen from the table that all lagging periods from 1 to 6 accept the null hypothesis that ln x is not the Granger cause of ln y,and they all reject the null hypothesis that lnyis not the Granger cause of lnx,indicating that lnyis the Granger cause of lnxwith a long-term stability.

Table 4 Granger causality test of lnx and lny

4 Results and conclusion

In this article,the relevant data on the output of new products and the internal expenditure of R&D in the “Statistical Yearbook of China’s Hightech Industry” from 1995 to 2018and the software Stata 14 are used to make an empirical study on the relationship between new product output and R&D capital investment through co-integration tests,error correction models,and Granger causality.

4.1 Results

(1) From a long-term perspective,there is a fixed equilibrium relationship between the output of new products and the investment in R&D,and their elasticity coefficient is 4.244 5.For per SD change in the output of new products in China’s pharmaceutical industry,R&D investment will increase by 4.244 5 SD.In the short term,for per SD change in the output of new products in China’s pharmaceutical industry,R&D input will change by 0.151 6 SD reversely.

(2) The Granger causality test shows that lnxis not the Granger cause of lnyduring the lagging stage from 1 to 6.On the contrary,lnyis the Granger cause of lnx.

4.2 Conclusion

(1) There is a significant positive correlation between the output of new products and R&D input in China’s pharmaceutical industry,and R&D input will significantly promote the output of new products in the next year.

(2) The impact of R&D input on the output of new products has a significant impact when it lags from period 1 to period 6,indicating that continuous R&D investment will have a long-term impact on the output of new products.

5 Analysis and recommendations

R&D investment should be continuous and stable.There is a stable long-term equilibrium relationship and short-term error correction mechanism between the output of new products and R&D input in China’s pharmaceutical industry.Stable R&D input has an important impact on the output of new products.At the same time,stable capital investment can provide the support of advanced technology,equipment and talents for drug R&D activities.Therefore,it is necessary for a pharmaceutical enterprise to attach great importance to R&D investment.From a long-term perspective,sustained and stable R&D investment can promote the improvement of pharmaceutical enterprises’ new product output capabilities.

A multi-level,multi-channel and diversified R&D investment system should be established for pharmaceutical enterprises.The investment in corporate R&D funds at the macro national level should be increased.At the same time,it is necessary to strengthen the enthusiasm of enterprises,social groups and private capital to invest in China’s pharmaceutical enterprises.

The efficiency of R&D activities should be improved to accelerate the output of research results.R&D cooperation has a direct impact on the economic output of enterprises[16].Therefore,pharmaceutical enterprises should take measures such as cooperation to improve the efficiency of R&D.Firstly,they can cooperate with universities or other scientific research institutions to establish a research and development model that combines production,education and research.Universities and scientific research institutions should give full play to their technological advantages.Meanwhile,enterprises should make good use of their advantages in capital and market competition to complement between R&D activities and achievement transformation.Secondly,strategic alliances with other international pharmaceutical companies should be formed to conduct cooperative research and development,which will enhance their R&D and innovation capabilities.This can accelerate the transformation and upgrading of the industry[17].Through the above-mentioned cooperation methods,the R&D capabilities of pharmaceutical enterprises in China can be improved greatly.