Construction of R&D Investment Model: An Application of Techno-Dynamics Model. to Japanese Companies

Size: px
Start display at page:

Download "Construction of R&D Investment Model: An Application of Techno-Dynamics Model. to Japanese Companies"

Transcription

1 Construction of R&D Investment Model: An Application of Techno-Dynamics Model to Japanese Companies Yoshiki NAKAMURA and Masashige TSUJI Department of Industrial and Systems Engineering, Aoyama Gakuin University, Setagaya-Ku, Tokyo Japan Abstract R & D has lately become a part of information system. However, it has primarily remained to be for individual activities. Consequently, private firms need to construct a separate strategy of R & D in future. Specially, enterprises are required to produce the investment strategy system in many areas of R & D policy. Against these backgrounds, one of the most important purposes of this study is, first, to create a useful model for R & D investment and activities that would become the basis for the information system. Second, through that model, the study tries to make sure that the R & D investment will influence the sales. In validating, the paper has simulated the model in seven different companies of Japan. It has calculated the Stock of Technological Knowledge for the fourteen-year period from 985 to 999. The model has also computed sales and checked the outcomes against the actual open data by referring to a statistical analysis. In the end, R & D Investment Model, which this paper proposes, would help to introduce the information system of investment strategy for different enterprises. It also assists business administrators or R & D managers to make accurate and appropriate decisions of their investments for fundamental technologies. Keywords: Research and Development, Investment Strategy, Techno-Dynamics Model Introduction

2 Research and Development (hereafter R & D) has lately become a part of information system. The system is contrived as a means to come up with a rapid research result, while it is also for improving the quality of technology. To put this in a different perspective, the system acts as a personal computer and helps facilitate the process of complicated technical calculation and simulation. It also contains a database that stores a vast volume of test results and information; likewise, the information system is akin to a search engine system of paper, report and literature from inside and outside of the enterprises []. These information systems are primarily for individual activities. Consequently, the information systems for different enterprises ought to be different. Private firms need to construct a separate strategy of R & D in future. These may be a project-planning system, project progress management system, etc. This study pays attention to the strategy of R & D investment in many areas of R & D policy. Of many questions, the present study hopes to raise a significant question as to how R & D funds should be allocated, given the many areas of research. In this undertaking, the emphasis is on the efficiency of R & D investment through the information system, even though the area of R & D has been regarded as a safe haven concealed in the enterprise in the past. Any study of this matter must examine a key question: what are the results of R & D activity and how are they defined? Brow [2] said that R & D output consists not only of patents or new products but also previously unknown knowledge and experience, which is widely valuable to the business organization. Following the flow of output activity from R & D to point of sale, research first creates knowledge, which later becomes tangible as papers or patents. Then both applied and development research attempt to apply this knowledge to merchandising, thus strengthening productivity and marketing. Finally, the output, as either product or service, contributes to sales. Consequently, the total amount of R & D investment should be based on what output will be generated. Many research studies have already been conducted on R & D investment. E. Geisler [3] proposed an integrated model of R & D evaluation, which links the cost of research to its various outputs. Jehiel Zif and Daniel J. McCarthy [4] explored the relationship between the total R & D budget and two major components, product and process R & D, examining a cyclical pattern (S-Curve) in the relationship between them. These studies primarily deal with a mathematical model between R & D investments and research results. However, they do not all classify output in the same way, as it is difficult to clearly specify the outcome of R & D. To resolve this difficulty, a concept has been developed that defines output as technical knowledge, and which can measure its value to business. Roger E. Bohn [5] proposes eight stages of technological knowledge, representing a spectrum from art to science. Akio Kameoka and Sei-ichi Takayanagi [6] suggest a Corporate Technology Stock Model, which provides guidelines for R & D investment. They present two categories of technological knowledge: Basic Technology Stock and Application Technology Stock. However, these approaches that

3 the Distribution Ratio in terms of R & D character R&D expenditure Sales Sales Stock of Basic Research obsolescence Stock of Applied Research obsolescence Stock of Development Research obsolescence Operating Fig. Fundamental frame of Techno-Dynamics categorize technological knowledge have a few issues. First, in real situations, R & D is multi-faceted, with dimensions such as basic research, applied research and so on, which are not independent, but interact with each other. Thus, a transport mechanism -- i.e., a mode of sharing technological knowledge among researches -- should be developed. Second, it is necessary to comprehend the effect of research activities on the output, which is production and sale. Generally, it is difficult to determine how research activities contribute to the benefit of enterprise, which characterizes the difficulty in deciding how R & D money should be invested. Therefore, any approach should explore the spiral of research outcomes that could result. As the above studies have implied, the actual construct of system of R & D investment strategy has still a lot to be considered and looks premature. For these reasons, the most important purpose of this study is, first, to create a useful model for R & D investment and activities. To accomplish that aim, it first focuses on the process between the investment aspect of R & D and its subsequent effect on some Japanese enterprises. Second, through that model, it is make sure how the R & D investment gives the influence to sales. It can support the investment information of enterprise by the form. In addition, it should be analyzed the sway of sales with the talent s fluidization, which has been become a serious problem even for R & D in recent years. Finally, it is thought that R & D investment model, which this paper propose, can help the enterprise to introduce the information system of investment strategy. The Fundamental Frame of the Techno-Dynamics Model

4 Research, which creates knowledge and technology, can be divided into various levels; it is, however, difficult to determine the spread of these levels, and their linkage to sales. To have universal application, any model used for discussion with R & D investors must be simple, with no more than five stages between R & D and sales: knowledge, technology, product development, production and finally volume of sales. This paper defines these layers as Stock of Basic Research, Stock of Applied Research, Stock of Development Research, Operation, and Sales. Here these layers offer a dynamic model, showing accumulation of technological knowledge, technological transfer, obsolescence, technological development and time lag in production. This paper called that dynamic model, Techno-Dynamics Model (Fig. ). The model first introduces the concept of technological knowledge as an accumulation of information about how to produce output -- papers, patents and goods -- which can be measured by R & D expenditure [7][8][9]. This accumulation is defined as Stock of Technological Knowledge (STK). Over time, STK characteristically decreases, reaching the point of obsolescence [8], when the old STK must be updated or replaced, as it no longer produces results, i.e. output. This study further defines two types of STK: non-competitive and competitive, the difference between them whether or not the Technological Transfer in STK is included. In this report, Technological Transfer flows from basic research to applied research to development research. Generally, non-competitive STK, (here, Stock of Basic Research without Technological Transfer) is less competitive, and therefore regularly increases, reaching obsolescence frequently. However, competitive STK is regularly traded, thus the Stock of Applied Research and Stock of Development Research are not only increased by R & D expenditures, but also augmented by STK from both Basic Research and Applied Research. Obsolescence will occur faster than it does in Basic Research, in proportion to the rate of Technological Transfer [8]. Operation signifies overall product activity: supplying and/or developing new raw material or technology by both applied and development research, the energy for the new product, production, inventory control, marketing, etc. This study contends that Stock of Applied Research and Development Research decide the scale of Operation. Sales is defined as the provision of a service or product to industry or the marketplace, for money. This paper argues that Stock of Development Research and Operation will create Sales. This multi-layered model is appropriate as either a support system or information tool for R & D investment. All that is required is to set the policy variable, which depends on Japanese industries or companies, of the total amount of R & D investment and the activity, time lag, and ratio of obsolescence, etc. The following definitions outline what this paper refers to as a Techno-Dynamics Model.

5 Construction of Mathematical Model The Techno-Dynamics Model is constructed in mathematical form. To simplify discussion, time is set as a fiscal year unit. First, R & D character from the total amount of R & D investment is calculated as follows: λ i ( t ) = Rc i ( t )/R( t ) () where R( t ) = Rc i ( t ) (2) λ i ( t ) (0 i i λ i ( t ) ) is the Distribution Ratio in terms of R & D character i, R( t ) is R & D expenditure in a t period, Rc i ( t ) is the character of i s R & D expenditure in a t period, and i ( i = B, A, D ) is R & D character (B: basic research, A: applied research, D: development research.) Expression of Stock of Basic Research In this paper, Stock of Basic Research S B signifies non-competitive technological knowledge. The expression is as follows: S i ( t ) = S i ( t - ) + ins i ( t ) - outs i ( t ) (3) ins i ( t ) = ϕ i Rc i ( t - m i ) (4) outs i ( t ) = ρ i S i ( t - ) (5) where S i ( t ) is Stock of Technological Knowledge in a t period, ins i ( t ) is inflow in a t period, outs i ( t ) is outflow in a t period, ϕ i (0 ϕ i ) is the changing probability from R & D expenditure to STK, m i is time lag, and ρ i (0 ρ i ) is the probability of Obsolescence. Expression (3) presents the accumulation of STK until t- period, and the increase and decrease of STK in a t period. Expression (4) clarifies that R & D expenditure changes to STK over time, by time lag m i. Generally speaking, m i represents the time between starting the research and putting it to such practical use as papers, patents or products. Basic research usually takes 5 years in Japan and 7.4 years in the US [0][]. Expression (5) shows the decrease in the STK in a t- period to be proportional to ρ i. D.L Bosworth [8] and Akira Goto [] measured the ρ i. Expression of Stock of Applied Research and Stock of Development Research Stock of Applied Research S A and Stock of Development Research S D signifies co mpetitive technological knowledge. The expression is as follows (in this case Technological Transfer takes place i s S TK to j s): S j ( t ) = S j ( t - ) + ins j ( t ) - outs j ( t ) (6) ins j ( t ) = ϕ j Rc j ( t - m j ) + δ ij S i ( t ) (7) outs j ( t ) = ( + δ ij )ρ j S j ( t - ) (8) where j is as well as parameter i that is R & D character, δ ij (0 δ ij ) is the rate of the Technological Transfer from i to j. Expression (7) explains the accumulation of j s STK by

6 R & D expenditure and Technological Transfer from i s. Expression (8) explicates that the decreasing speed of STK is in proportion to the rate of Technological Transfer. Further, the Technology Transfer in this paper is considered only in one way, as i to j. Expression of Obsolescence In the recent years, talent s fluidization is extremely activated in many industries []. The talent s fluidization means a halfway hiring, a changing job, headhunting, etc. Stock of Technological Knowledge is fundamentally depends on the talent capacity. Thus, in the recent situation in which the talent s fluidization promotes scarcely, STK will be reconstructed with fluctuation of talent s fluidization. As a consequence, in this study, it will set up the ratio of talent s fluidization. The ratio of talent s fluidization is a function of the obsolescence depended on the person s increase and decrease. The mathematical expression is as follow: ρ i = ρ i0 ( + ε ) (9) where ρ i0 is initial value of obsolescence of i s, and ε ( - < ε < ) is the ratio of talent s fluidization. In short, if some company has to be inflow the employee by a halfway hiring and/or headhunting, the ratio of talent s fluidization will be small number. Therefore, it is restrained the decrease of STK through the diminution the obsolescence. In the opposite, if some company has to be outflow by headhunted, the ratio of talent s fluidization will be large number. Consequently, the STK is growing larger decrease according to the huge reduce of obsolescence. Expression of Operation Based on the above-noted, the mathematical form of Operation with Stock of Applied Research and Stock of Development Research is as follows: P( t ) = α + β S A ( t ) + β 2 S D ( t ) (0) where P( t ) is operation in a t period, α is a fixed number, and β l ( β l 0, l =,2 ) is parameter. A fixed number α and parameter β l will be decided depending on the industry or enterprise. α explains the constant time frame of production and marketing without time progress. Expression of Sales Similarly, Operation, i.e., the mathematical form of Sales with Operation and Stock of Development Research is as follows: Q( t ) = α 2 + β 2 P( t ) + β 22 S D ( t ) () w here Q( t ) is sale in a t period, α 2 is a fixed number, and β 2l ( β 2l 0, l =,2 ) is parameter. A fixed number α 2 and parameter β 2l will be decided depending on the industry or enterprise. α 2 explains the constant time frame of sale without time progress.

7 Validity of the Model In an attempt to analyze the efficacy of the Techno-Dynamics Model, a simulation is created for Japanese companies: they are,,,,, Takeda Chemical Industries and. Theses enterprises are the easy to get the finical data like sales, R & D investment, etc. This is the reason they are selected. And they try to calculate the STK by using open data. Verification is based on comparing the measurement value of Operation and Sales with those of the calculated value. Also it tries to check the influence of Sales when the company has been done the talent s fluidization. The measuring period is 4 years (from 985 to 999), with 985 the benchmark. The Distribution Ratio in terms of R & D, λ i ( t ), and R & D expenditure, R( t ), use from their company s WWW [6] ~[7]. Time lag, m i, refers to [][2], personally recalculated. Obsolescence, ϕ i, refers to [] and is again personally recalculated. Operation is calculated by expression (0). To decide the fixed number, α, and parameter, β l, multiple regression analysis is used to give the measurement number with the above-noted, that the explanatory variables are the Stock of Applied Research and Stock of Development Research, and that the explained variable is Operation. As a result, (0) is: Pˆ ( t ) = a + bs A ( t ) + b2s D ( t ) (2) ( ) where ˆ t is P( t ) s value of calculation, and a and b ( l =,2 ) is partial regression P l coefficient. The estimating period is for each of the 8 years from 980 to 998. In the same manner, Sales is calculated by expression (). To decide the fixed number, α 2, and parameter, β 2l, multiple regression analysis again gives the measurement number above-noted, in which the explanatory variables are the Operation and Stock of Development Research, and the explained variable is Operation. As a result, () is: Q ˆ ( t ) = a 2 + b P( t ) + b S ( t ) (3) 2 ˆ 22 D where ˆ( t ) Q 2 2l is Q( t ) s value of calculation, and a and b ( l =,2 ) is partial regression coefficient. Similar to Operation, the estimating period is again for each of 8 years from 980 to 998. The Changing Probability from R & D expenditure to STK, ϕ i, and the rate of Technological

8 Fig.2 Transition of Stock of Basic Research Fig.3 Transition of Stock Applied Research Fig.4 Transition of Stock Development Research Fig.5 Transition of Operation Fig.6 Transition of Sales Tab. The Result of Multiple Regression (Operation) a b b 2 R 2 DW (0.9774) (0.782) (.4356) (0.5794) (0.6637) (0.42) (0.323) (0.484) (0.8543) (4.7787) (0.5337) (.98) Takeda Chemical Industries (.2865) (.3757) (2.66) (.935) (0.896) (0.098)

9 Transfer, δ ij, gives the fixed numbers and 0.0 each. Figures 2 to 4 show the transition of Stock of Basic Research (as non-competitive technological knowledge), Stock of Applied Research (as competitive technological knowledge) and Stock of Development Research (as competitive). This time, the rates of increase is shown, with 980 s stock as standard. In Figure 2, Stock of Basic Research in has the largest rate of increase from 980, while is the smallest. Because there had the accident at in 993, it would be required that R & D investment of that company should be diminished. Consequently, they had some influence for the Stock of Basic Research in 994. From figure 3, and is the decrease tendency. Against it, and is increase inclination. Figures 5 and 6 show the transition of Operation and Sales established in 980. Tables and 2 are the result of multiple regression analysis with t-value under the numerical numbers, coefficient of determination, R 2, and Durbin-Watson ratio, DW. Examining the relationships among R & D activity, Operation and Sales from Table and 2, it is apparent from these outputs that increasing R & D investment creates greater technological knowledge, thus intensifying Operation and generating increased Sales. Next, to analyze how talent s fluidization influences the production and sale with a fixed quantity, a sensitive analysis assesses how talent s fluidization sways Sales. Therefore in this paper, the ratio of talent s fluidization will be set. It will be given the Sin function for the ratio of talent s fluidization. ε ( x )= -γ * Sin x (4) where γ (γ > 0) and x (-90 < x < 90 ) is parameter. The parameters, γ and x, can set by own enterprise. However, to do the relative comparison among seven companies, the ratio of talent s fluidization give the same condition for seven enterprises. Employs of their companies will be increase and decrease among 5 people by talent s fluidization (see Table3). Figure 7 (only shows here) is the case that talent s fluidization influences to Sales in 999. When the talent s fluidization happens at Development Research, it can observe a big pressure on Sales. Against it, it cannot notice too much at Basic Research. In figure 8 to 0, it is the results of sensitive analysis of talent s fluidization depended on the enterprises. Comparison between Basic Research and Development Research, Basic Research is growing larger margin among the companies.

10 Next, it will be analyzed, if talent s fluidization will be happened at Basic Research, how the Sales be influence for the long term. Set up condition as follows: I) Talent s fluidization of decrease will be happened in 980. II) Three years later in 983, the ratio of talent s fluidization will recover to initial number, 0. Figure is a transition of the gap between no talent s fluidization of Stock of Basic Research and talent s fluidization (- to -5) of it. Figure 2 is same as figure that is the Sales of a transition of the gap. From these figures, even if it happened the talent s fluidization, it could recover in case of Stock of Basic Research when the time passes. Sales, however, would be grown lager difference. That it to say, it is understood that the influence of Sales through the talent s fluidization should be a larger damage on passing of time. Based on these figures, this paper concludes that the Techno-Dynamics Model is capable of explaining past R & D activity. Further, it can confirm the intensity of the set up data in the model about the influence of the talent s fluidization of basic research. Tab.2 The Result of Multiple Regression (Sales) a 2 b 2 b 22 R 2 DW (0.926) ( ) (0.86) (0.766) (0.5405) ( ) (0.3303) (0.3287) (0.28) ( ) (0.244) ( ) (0.4535) ( ) (0.7609) Takeda Chemical Industries (5.567) (0.3687) (0.54) (0.7593) (0.642) (0.327) Tab.3 The Relationship between the Number of Employs and the Ratio of Talent s Fluidization The number of employs x e( x )

11 Stock of Basic Research Stock of Applied Research Stock of Development Research (the right-axis) Fig.7 Sensitive Analysis of Talent s Fluidization () Fig.8 Sensitive Analysis of Talent s Fluidization at Basic Research Fig.9 Sensitive Analysis of Talent s Fluidization at Applied Research Fig.0 Sensitive Analysis of Talent s Fluidization at Development Research Fig. A Transition of the Stock of Basic Research s Gap () Fig.2 A Transition of the Sale s Gap () Conclusion Throughout this paper, the distinctive features of the multi-layered Techno-Dynamics Model examined R & D structure via simulation with a fixed quantity, from seven Japanese companies. The Techno-Dynamics Model can also provide a basis for R & D investment to business administrators or R & D managers. In addition, using the simulation, it can confirm the importance of the talent s fluidization. Through this model, it could be useful to construct the information system of R & D investment strategy.

12 The present study can lead to several areas for future research. In this paper, the Techno-Dynamics Model used only a linear model to propose Technological Transfer. It should be reconstructed with network models, etc. Further, it will be ensured the utility of Techno-Dynamics Model not only Japanese companies but also US and EU companies. Hopefully, the Techno-Dynamics Model will become universally applicable, a valuable utility for any R & D investment study. Reference [] Boer, F.Peter (999): The Valuation of Technology Business and Financial Issues in R & D, John Wiley & Sons, INC. [2] Mark G.Brown and Raynold A.Svenson (998): Measuring R & D Productivity, Research, Technology Management, Vol.4, No.6, pp [3] E Geisler (995): An Integrated Cost-Performance Model of Research and Development Evaluation, Omega, Vol.23, No.3, pp [4] Jehiel Zif and Daniel J.McCarthy (997): The R & D Cycle: The Influence of Product and Process R & D on Short-Term ROI, IEEE Transactions on Engineering Management, Vol.44, No.2, pp.4-23 [5] Roger E.Bohn (997): Measuring and Managing Technological Knowledge, IEEE Engineering Management Review, Vol.25, No.4, pp [6] Akio Kameoka and Sei-ichi Takayanagi (999): A Corporate Technology Stock Model: Financially Sustainable Research and Technology Development, Proceeding of the PICMET Portland, Oregon, USA, pp [7] Zvi Griliches (980): R & D and the Productivity Slowdown, American Economics Association, Vol.70, No.2, pp [8] D.L Bosworth (978): The Rate of Obsolescence of Technological Knowledge-A Note, The Journal of Industrial Economics, Vol.26, No.3, pp [9] Zvi Grilices, R & D and Productivity, The University of Chicago Press, 998 [0] Michiyuki UENOHARA and Daizaburou SHINODA (995): R & D and Technology Management, Corona Publication CO., LTD Japan [] Akira GOTO (993): Innovation and Industrial Organization in Japan, University of Tokyo Press [2] Gellman Research Associates (976): Indicator of International in Technological Innovation, NSF [3] Cheongho LEE and Kyoichi KIJIMA (997): A Viability-based Efficiency Concept of R & D Investment and its Indices, Journal of the Japan Society for Management Information, Vol.6, No.2, pp.5-3 [4] Japan Statistical Association, Japan Statistical yearbook, Statistics Bureau Management and Coordination Agency Government of Japan,

13 [5] Ministry of Education, Culture, Spots, Scientific Socialism, White Paper, Printing Bureau, Ministry of Finance, Japan, [6], [7], [8], [9], [20], [2], LTD, [22],