Investment in High-Tech Industries : An Example from the LCD Industry

This paper considers a representative firm taking investment decisions in a high-tech environment where different generations of products are invented over time. First, we develop a real options investment model in which, according to standard practice, the sales price and the unit production cost both satisfy a geometric Brownian motion (GBM) process. However, from real life data of the LCD industry it follows that output prices behave according to a crystal cycle that does not match a GBM. We proceed by conducting a thorough econometric analysis, leading to the conclusion that a vector autoregressive model (V AR) provides the best fit. Integrating this model with the real options machinery, we find that (i) at the moment of investment the increased production capacity goes along with increasing production cost and decreasing price, (ii) a management effect is present in the sense that a price drop is followed by a cost decrease due to management pushing harder on cost decreasing programs, and (iii) investing can be optimal while at the same time a GBM yields a negative net present value (NPV). We also find that investment decisions taken in practice are better supported by our V AR model than by the standard real options model based on GBM.