During the year 2000, the Nasdaq index achieved historical highs of approximately 5,000 points. During the following two years, the index fell back to around 1,000. Economists have been trying ever since to find a reasonable explanation for the causes of such a huge increase and subsequent decline in the prices of information and communication technology (ICT) firms at the turn of the century. But most of the models for stock valuation they have constructed failed to predict the soaring prices achieved by the Nasdaq index. The high values at the start of this millennium were apparently unfounded. For some reason, investors were buying shares at prices well above their real value. This fact has been referred to as the “Nasdaq bubble”: an irrational overvaluation that was doomed to burst sooner or later.

In this paper, the authors cast doubt on the idea that the prices of dot-com firms were indeed experiencing a bubble at the end of the millennium. They argue that there could be a crucial missing element in past models for stock valuation: uncertainty about the future profitability of ICT firms. Taking into account this uncertainty, it is possible that the fundamental value of the stocks could have increased to achieve the observed high levels. The authors’ view is that investors were taking a chance that technology stocks would continue to soar, knowing that there was a risk that they could fall. According to the authors’ model, the chance that the stocks would gain value was worth the risk.

This is not the first study to construct a model seeking to rationalize the high prices observed in the Nasdaq index at the turn of the century. Several researchers have already constructed models to explain this phenomenon, but they differ from this model in at least two ways. First, most of the previous models attribute the bubble to other factors besides uncertainty about future firm profitability. Second and most importantly, none of the previous models could be credibly calibrated. In other words, the high prices of the 1999-2000 Nasdaq index could not be achieved with plausible levels of the models’ inputs.

The idea of including uncertainty in the stock valuation model is backed by empirical data. During the bubble period, the volatility of ICT stock prices increased significantly relative to the volatility of other stocks. ICT investors reacted sharply to earnings announcements during this time, showing that new information radically changed their minds about firm profitability. Finally, many of the ICT firms went public unusually early; these firms’ short earnings history made it particularly hard to figure out their future profitability, adding to the uncertainty.

The authors’ valuation model is built on the tradition of financial valuation models, to which they add the consideration of uncertainty about firms’ future profitability. In the model, firms’ value is measured by the ratio of the market valuation to the book value of equity. This ratio increases when uncertainty about the average future profitability of firms increases. They build this valuation model using an advanced methodology, called stochastic processes, that allows for the rigorous inclusion of random future changes in firms’ profitability.

The first result is that when uncertainty is high, the value of the firms increases. This finding could partly justify the high price levels observed for the Nasdaq firms during the late 1990s, since uncertainty about the future profitability of these firms was quite high during this period. Appealing as it might be, this result is only qualitative; it is insufficient to show that the Nasdaq prices were not experiencing a bubble during the end of the century.

The authors therefore take the next step and work with the quantitative predictions of their model. In economists’ jargon, they “calibrate” their model - they feed in all the necessary numerical inputs such as values for the market uncertainty, expected profitability and discount rates - and attempt to replicate the behavior of the Nasdaq index during the bubble. In fact, they manage to replicate the bubble quite satisfactorily and the numerical values required for achieving such replication are realistic. Specifically, they find that the value of uncertainty (as measured by the volatility of stock returns) needed to match the high values of the stocks is very near to what was experienced in reality.

To double-check their results, the authors repeat the same exercise with a number of individual ICT stocks such as Akamai, Amazon, Ciena, Cisco, Dell, eBay, Immunex, Intel, Microsoft, Priceline, Red Hat, and Yahoo!. The results are similar to those for the Nasdaq Index as a whole: the model is able to replicate the price trajectory followed by these stocks and the uncertainty implied by the model’s calibration is very plausible.

So far, the explanations presented by the authors have been convincing. There is, however, one missing link: if uncertainty was the sole cause of the excess pricing of ICT stocks, then what drove the prices down? Did the uncertainty disappear from one day to another? Their answer is no, and the argument they provide is simple and intuitive: investors observed the profitability of Nasdaq declining in 2000 and suddenly changed their minds about the expected profitability of ICT firms. This lower expected profitability led investors to want to get rid of their stocks, accounting for the sharp decrease in the market-to-book value of firms.

Did the investors behave irrationally and overvalue the ICT firms? By incorporating uncertainty of the future firm cash flows into a stock valuation model, the answer is not so obvious. We might therefore want to think twice before we call the Nasdaq phenomenon of the 1990s a bubble.

Report Source:
"Was There a Nasdaq Bubble in the Late 1990s?", a Working Paper by: Lubos Pastor (Graduate School of Business, University of Chicago, CEPR and NBER), Pietro Veronesi (Graduate School of Business, University of Chicago, CEPR and NBER).
June 2005, pp.52