The investment decision with technological and market uncertainties

ABSTRACT

This paper examines how technological uncertainty affects current investment; specifically, what is the impact on a firm’s investment in an existing technology when an improved technology might arrive in the future. The firm can invest in the current technology and upgrade to the new technology after its arrival (sequential investing), or it can bypass the current technology and invest directly in the new technology (leapfrogging). The main result is that, in the presence of market risk, future technological uncertainty has a non-monotonic effect on investment, with the investment trigger being a U-shaped function of the expected speed of arrival of the new technology. In this U-shaped relationship, the investment trigger starts rising later if the new technology is more attractive and also when volatility and interest rate are high and growth rate low; thus, technological uncertainty is more likely to have a positive effect on investment under these conditions. Finally, we apply the model to the sequential versus leapfrog investment decision, and find that leapfrogging becomes more attractive relative to sequential investment when interest rate and new technology earnings enhancement are higher, and when market volatility, growth rate and new technology investment cost are lower.

KEYWORDS:

• Continuous-time model
• real option
• investment trigger
• technological uncertainty
• sequential investment
• leapfrog

JEL CLASSIFICATION:

• G31

Acknowledgements

We would like to acknowledge helpful comments and feedback from two anonymous referees, as well as the Editor Chris Adcock, which resulted in improvement of the paper. We are responsible for all remaining errors.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. We use the term ‘improved technology’ broadly, to mean anything that results in a significant jump or discrete increase in the earnings stream; this could, of course, include an explicit technological innovation.

2. There are other types of uncertainty associated with technological innovation, such as uncertainty regarding cost of adopting new technology, size of the benefit from new technology, etc. We do not discuss these, because we want to highlight the role of the two types of uncertainty mentioned above.

3. Some papers (such as Jaimungal, de Souza, and Zubelli [2013]) consider mean-reverting process when the project involved is linked to commodities.

4. In their model, technological uncertainty is captured by a lognormal process X, and the future innovation arrives when X reaches an upper boundary X_h. Since X can move in both directions (up and down), this is an unsatisfactory way (in our opinion) to model technological progress, which should be able to move in only one direction (innovations can only improve the best-available technology, not worsen it). We therefore model technological uncertainty by a one-sided jump process, as in Murto (2007), Alvarez and Stenbacka (2001), etc.

5. The new technology could increase earnings through various channels, such as higher output rate, lower costs, etc. We do not model these channels, but take a reduced-form approach by assuming just that modernization causes earnings to rise from x to δx.

6. Since we are interested in sequential investment, we want x_i < x_m; therefore, we assume I₂ > (δ − 1)I₁.

7. We thank an anonymous referee for pointing this out.

8. Note that the parameter δ is chosen to be consistent with the assumption in endnote 4.

9. Kort, Murto, and Pawlina (2010) also has similar setting by assuming the cost of lumpy investment is smaller than the cost of sequential investment to trade off the flexibility of economic scale. However, their model is for capacity expansion and they did not consider the role of technological uncertainty.

10. Note that we consider x_{i(λ = ∞)} rather than x_i because the new technology has already arrived, which is equivalent to λ = ∞ (see Section 4.2).

11. This is called the ‘laggard’ strategy by Grenadier and Weiss (1997).

12. Grenadier and Weiss (1997) use the term ‘compulsive strategy’ to describe our sequential strategy.

13. Financial Times (2013 Sept) ‘Intel intends to leapfrog smartphone era with Quark chips’.

14. There have been several papers using real-option models for concrete valuation of firms, such as Trigeorgis and Ioulianou (2013).

Additional information

Funding

Chuanqian Zhang thanks Summer Research Fellow of Cotsakos College of Business at William Paterson University, National Natural Science Foundation of China[grant number 71702013] and National Social Science Foundation of China [grant number 17BGL007] for their financial support.