Forecasting Foreign Direct Investment with Modified Diffusion Model

Prior research has not effectively investigated how the profitability of Chinese branches affect FDIs in China [1, 2], so this study for the first time incorporates realistic earnings information to systematically investigate effects of innovation, imitation, and profit factors of FDI diffusions from Taiwan to China. Our nonlinear least square (NLS) model, which incorporates earnings factors, forms a nonlinear ordinary differential equation (ODE) in numerical simulation programs. The model parameters are obtained through a genetic algorithms (GA) technique and then optimized with the collected data for the best accuracy. Particularly, Taiwanese regulatory FDI restrictions are also considered in our modified model to meet the realistic conditions. To validate the model-s effectiveness, this investigation compares the prediction accuracy of modified model with the conventional diffusion model, which does not take account of the profitability factors. The results clearly demonstrate the internal influence to be positive, as early FDI adopters- consistent praises of FDI attract potential firms to make the same move. The former erects a behavior model for the latter to imitate their foreign investment decision. Particularly, the results of modified diffusion models show that the earnings from Chinese branches are positively related to the internal influence. In general, the imitating tendency of potential consumers is substantially hindered by the losses in the Chinese branches, and these firms would invest less into China. The FDI inflow extension depends on earnings of Chinese branches, and companies will adjust their FDI strategies based on the returns. Since this research has proved that earning is an influential factor on FDI dynamics, our revised model explicitly performs superior in prediction ability than conventional diffusion model.

Authors:

• Bi-Huei Tsai

Keywords:

• diffusion model
• genetic algorithms
• nonlinear leastsquares (NLS) model
• prediction error.

References:

[1] Tsai, B.-H., 2009. Analysis of dynamic growth model in foreign direct
investment of Taiwan IC industry into China, International Journal of
Accounting Studies, 49, 73-94. 2
[2] Tsai, B.-H. 2010. Application of dynamic diffusion theory in foreign
direct investment of Taiwan IC industry into China. International Journal
of Computational Science and Engineering, 5(1), 2-8. 3
[3] Smith, D. M. 1981. Industrial location: An economic geographical
analysis. 2nd edition. New York: John Wiley. 4
[4] Bass, F. M. 1969. A modified product growth model for consumer
durable. Management Science, 15, 215-227. 5
[5] Mahajan, V., E. Muller, and F. M. Bass. 1990. New product diffusion
models in marketing: A review and directions for future research. Journal
of Marketing, 54, 1-26. 6
[6] Rogers, E. M. 2003. Diffusion of Innovations, 5th edition. New York:
Free Press. 7
[7] Bass, F. M. 2004. Comments on a new product growth for model
consumer durables. Management Science, 50 (12), 1833-1840. 8
[8] Meade, N., and T. Islam. 2006. Modelling and forecasting the diffusion
of innovation: A 25-year review. International Journal of Forecasting, 22,
519-545. 9
[9] Ruiz, E., and Leeflang, P. S. H. 2006. Diffusion of franchising as an
innovation of managerial organization. Marketing JRM, 65-75. 10
[10] Mesak, H., and Berg, W. 1995. Incorporating price and replacement
purchases in modified products diffusion models for consumer durables.
Decision Science, 26, 425-449. 11
[11] Mesak, H. (1996). Incorporating price, advertising and distribution in
diffusion models of innovation: some theoretical and empirical results
expectations in diffusion models. Computers and Operations Research,
23, 1007-1023. 12
[12] Parker, P. (1992). Price elasticity dynamics over the adoption life cycle.
Journal of Marketing Research, 9, 358-367. 13
[13] Tsai, B.-H., Li, Y. and Lee, G.-H., 2010, Forecasting global adoption
of crystal display televisions with modified product diffusion model,
Computers and Industrial Engineering 58 (4), 553-562. 14
[14] Berkoune, D., Mesghouni, K and Rabenasolo, B. 2006. Lower bounds
for the scheduling problem with uncertain demands. International Journal
of Applied Mathematics and Computer Science, 16 (2), 263-269. 15
[15] Czekalski, P. 2006. Evolution-fuzzy rule based system with parameterized
consequences. International Journal of Applied Mathematics and
Computer Science, 16 (3), 373-385. 16
[16] McCall, J. 2005. Genetic algorithms for modelling and optimization, J.
Comput. Appl. Math. 184, 205-222. 17
[17] Venkatesan, R., T. V. Krishnan and V. Kumar 2004, Evolutionary
Estimation of Macro-Level Diffusion Models Using Genetic Algorithms:
An Alternative to Nonlinear Least Squares, Marketing Science 23, 451-
464. 18
[18] Goldberg, D.E. 1989. Genetic Algorithms in Search, Optimization and
Machine Learning, Addison-Wesley. 19
[19] Kalish, S., and S. K. Sen. 1986. Diffusion models and the marketing
mix for single products. In Innovation Diffusion Models of New Product
Acceptance, edited by V. Mahajan, and Y. Wind, 87-115. Cambridge,
MA: Ballinger.