The Dot-Com Bubble: Valuation Narratives, Financing and the Crash
Table of Contents
The internet was reaching more American homes even as internet stocks began to collapse. The share of U.S. households with someone using the internet at home rose from 26.2% in December 1998 to 41.5% in August 2000. Yet the Nasdaq Composite fell 77.93% from its March 2000 closing peak to its October 2002 closing trough. Both facts matter: a technology can live up to its promise while an investment made at the wrong price disappoints. [1] [2]
The dot-com bubble is a case study in what happens when investors expect growth to produce future profits and companies need outside money to get there. The IPO of VA Linux shows who benefited from a soaring first-day price. Pets.com and Amazon show why the ability to raise cash and the ability to generate it from operations have to be evaluated separately.
What Was the Dot-Com Bubble?
The dot-com bubble was the late-1990s boom in shares of internet-related companies and the collapse that followed. Investors had a plausible reason to expect rapid adoption: more people were getting online and businesses could reach them directly. The harder judgment was how much future cash those businesses could earn and how much investors should pay for it. The Nasdaq Composite, a broad index of Nasdaq-listed stocks, is a useful record of the market episode. It is broader than internet stocks alone. [1] [2] [12]
Adoption tells you whether people use the technology. Business economics tell you whether the company can turn that demand into cash. Valuation asks how much of that future cash the share price already assumes.
How Investors Valued Internet Companies
For a company with negative earnings, a positive price-to-earnings multiple is unavailable. Analysts could compare the price-to-sales ratio or look earlier in the chain, at users, pageviews or orders. Every step toward cash required another assumption. A new visitor might see an ad, but only some views translate into inventory the site can sell, and the revenue must cover the costs of attracting and serving that audience.
Yahoo reported that daily page displays averaged about 207 million in December 1998 and 465 million in December 1999. The totals included its unconsolidated Yahoo Japan venture, and one person could generate several views. Yahoo earned most of its revenue from banner and sponsorship advertising, so attention was relevant to the business. The reported pageviews were neither a count of customers nor a direct measure of consolidated revenue. [3]
Researchers Brett Trueman, M. H. Franco Wong and Xiao-Jun Zhang studied 63 listed internet companies in 1998 and 1999. In their sample, web traffic helped explain differences in equity values beyond net income; the results depended on the metric, business model and financial data used. That tells us what was associated with market prices at the time. It does not tell us whether those prices were justified. [4]
The IPO Boom: Offer Price vs. First-Day Closing Price
An initial public offering (IPO) sets an offer price for investors who receive an allocation. Once trading starts, market buyers and sellers set a new price. The return from the offer to the first close is often called the IPO pop. The first table shows how unusual that gap became, and how sharply the number of completed offerings fell after 2000.
| Year | IPOs | Mean first-day return | Median first-day return |
|---|---|---|---|
| 1998 | 283 | 21.9% | 8.9% |
| 1999 | 476 | 71.2% | 37.5% |
| 2000 | 380 | 56.3% | 27.9% |
| 2001 | 80 | 14.0% | 10.2% |
| 2002 | 66 | 9.1% | 8.2% |
Ritter measures first-day return from the offer price to the first close and gives each IPO equal weight in the mean. The sample requires an offer price of at least $5 and screens out several categories of offerings. Its counts cover qualifying deals across the market; they do not isolate dot-com companies. The 2000 figures cover the whole calendar year, including offerings before and after the March market peak. [5]
VA Linux listed on December 9, 1999, at a $30 offer price and closed at $239.25. An investor allocated shares at $30 and able to sell at the close would earn ($239.25 − $30) ÷ $30 = 697.5%, before costs and taxes. An investor first buying at the close paid $239.25 and did not earn that gain. [6]
The company initially sold 4.4 million shares at $30, or 4.4 million × $30 = $132 million in gross proceeds on the base deal. This excludes fees and the later sale of additional shares under the underwriters’ option. The $239.25 close did not increase the cash VA Linux received for each share sold at $30. [6]
The first-day gain measures one purchase price against one sale price. A crossover investor who bought privately paid a different price and may still face a lockup when public trading begins.
The gap between the offer and the first close made hot allocations valuable. In a 2002 complaint, the SEC alleged that Credit Suisse First Boston directed some allocations to customers who returned part of their gains through commissions on other trades. The firm settled without admitting or denying the SEC’s allegations. This enforcement case illustrates the incentive to obtain allocations; it does not establish how common the practice was or why the broader market fell. [6]
How Far the Nasdaq Fell
The Nasdaq Composite was already down sharply by April 2000. The full decline took more than two years from closing peak to closing trough. The sequence matters because the peak, the selloff and the recession did not happen at the same time. [2] [11]
| Trading date | Closing level | What it marks |
|---|---|---|
| December 31, 1998 | 2,192.69 | Start of the 1999 calendar-year comparison |
| December 31, 1999 | 4,069.31 | Last close of 1999 |
| March 10, 2000 | 5,048.62 | Closing peak |
| April 14, 2000 | 3,321.29 | Early selloff, five weeks after the peak |
| December 29, 2000 | 2,470.52 | Last close of 2000 |
| December 31, 2001 | 1,950.40 | Last close of 2001 |
| October 9, 2002 | 1,114.11 | Closing trough |
| December 31, 2002 | 1,335.51 | Last close of 2002, above the trough |
Returning from the trough to the same peak level would take [(5,048.62 ÷ 1,114.11) − 1] × 100% = 353.15%, because the starting level is much lower. Neither number measures an individual stock’s return. The maximum drawdown guide explains how the observation frequency and measurement window affect that calculation. [2]
What the Fed and Recession Timeline Can Explain
The Federal Reserve ended its fall 1998 cuts with a 4.75% federal funds target and began raising the target in June 1999. It stood at 5.75% on March 10, 2000, when the Composite reached its closing peak. The Fed raised it again after the peak, to 6.00% on March 21 and 6.50% on May 16. It began cutting in January 2001 and ended that year at 1.75%. The U.S. business cycle as dated by the NBER peaked in March 2001 and reached its trough in November 2001. The stock market decline began before the recession and continued after it ended. [10] [11]
Higher required returns reduce the value today of a fixed future cash flow. For illustration, a single hypothetical $100 payment due in five years is worth $100 ÷ 1.085 = $68.06 at an 8% annual required return and $100 ÷ 1.125 = $56.74 at 12%. Those rates are valuation assumptions, not the Fed’s target. During the crash, investors could also change their estimates of the cash the company would earn and its ability to raise money before then.
Eli Ofek and Matthew Richardson proposed one explanation for the rise and fall of internet stocks. Their evidence suggests that optimistic traders could have more influence on prices when short selling was constrained, and that an increase in potential sellers when IPO lockups expired could help explain falling prices. A lockup expiration lets covered holders sell; it does not mean they all do. The study does not measure how much of the broader Nasdaq decline this explains. [12]
Glenn Rudebusch, writing for the San Francisco Fed in 2005, discussed the economic consequences of the bubble. He described overinvestment in fiber-optic cable and a subsequent drop in venture funding for new technology firms. His analysis did not estimate what drove the Nasdaq decline. [13]
Pets.com vs. Amazon: Raising Cash and Generating Cash
Both companies could raise money in early 2000. The terms of that financing and the cash their businesses generated afterward tell different stories.
Pets.com: an equity raise and a wind-down
- Financing: The February 2000 IPO sold 7.5 million shares at $11, so 7.5 million × $11 = $82.5 million gross. The March 2000 quarterly filing later reported about $75.3 million net of underwriting and offering costs. Its prospectus had already warned of operating losses and possible further funding needs. [7]
- Operations: For the quarter ended March 31, 2000, $7.651 million in sales less $12.515 million in cost of goods sold left a $4.864 million gross loss. The gross margin was −$4.864 million ÷ $7.651 million = −63.57%. Its quarterly filing attributed negative margins to promotional discounts and losses on shipping. The company also used $27.475 million in cash from operations that quarter, separate from the cash it raised in the offering. [7]
- Outcome: The board approved a liquidation plan in November 2000 after a search for financing or a buyer did not produce a deal. Shareholders approved the plan in January 2001. [8]
Amazon: convertible debt and an operating cash turnaround
- Financing: In February 2000, Amazon issued €690 million in principal of convertible subordinated notes. These were interest-bearing obligations with conversion terms, not proceeds from an IPO or cash earned from sales. [9]
- Operations: As reported in its 2002 Form 10-K, Amazon used $130.442 million in operating cash in 2000 and generated $174.291 million in 2002. It still reported a $149.132 million net loss in 2002. Operating cash flow and net income answer different questions. [9]
- Outcome: Amazon was still operating after the crash. Its financing and changed cash profile help explain why its trajectory differed from Pets.com’s, without showing that the notes alone caused its survival. [9]
Pets.com’s liquidation also differed from Webvan’s outcome. Webvan and three subsidiaries filed voluntary Chapter 11 petitions in Delaware in July 2001. These cases show why it helps to check both the income statement and cash flow statement, and to read what the company says about its need for more capital. They do not tell us how common each outcome was or what return a shareholder earned at a particular entry price. [8] [14]
How to Evaluate the Growth a Price Assumes
- Connect demand to revenue. Identify who pays and for what. A portal can sell advertising, a retailer earns sales from goods it sells, and a marketplace can earn transaction fees without receiving the full merchandise price as its own revenue.
- Connect revenue to cash. Estimate costs that rise with sales, such as shipping, and spending needed to build capacity. The Pets.com filing illustrates how rising sales can coexist with a gross loss: quarterly sales rose from $5.168 million in the quarter ended December 31, 1999, to $7.651 million in the quarter ended March 31, 2000 while cost of goods sold exceeded sales in both quarters. [7]
- Account for the cash gap. Compare cash available with operating needs, planned investment and debt payments. Ask how the company would close a funding shortfall if issuing shares became expensive or borrowing became unavailable. New equity can dilute an existing holding, and debt adds obligations.
- Work back from the price. For a plausible required return, ask what combination of future growth, margins and reinvestment could support today’s price. Repeat under less favorable assumptions. If the valuation requires revenue growth, margin improvement and cheap financing all at once, identify which assumption has the least evidence.
Common Mistakes
- Calling a first-day pop the issuer’s gain. The company receives the offer price on its new shares, less costs. Trading afterward changes the market price, while a sale by an existing holder sends proceeds to that holder.
- Explaining the whole crash with the May 2000 rate increase. The Composite had already peaked and suffered its April selloff. Rate changes can affect valuations without accounting for the whole decline. [2] [10]
- Assuming the winners were obvious. Amazon’s survival is observable after the fact. At the time, an investor still had to judge its financing needs, future cash flows and the price being paid. [9]
Limits of the Historical Evidence
The index records market prices; it does not independently measure fair value. Company filings show financing and operating results for particular issuers. Neither hindsight nor this small set of cases establishes a valuation threshold that identifies a bubble in real time. The useful comparison is between the outcomes a price requires and the evidence available about the business.
Frequently Asked Questions
Sources
- U.S. Census Bureau. Home Computers and Internet Use in the United States: August 2000 (September 2001), Figure 1 and Table A.
- Nasdaq, Inc. Nasdaq Composite daily closing data, via FRED. Nasdaq’s index fact sheet identifies the price and total-return tickers.
- Yahoo! Inc. 1999 Form 10-K, Item 7 and revenue-recognition notes.
- Brett Trueman, M. H. Franco Wong and Xiao-Jun Zhang. “The Eyeballs Have It: Searching for the Value in Internet Stocks” (Journal of Accounting Research, 2000), sections 3–5.
- Jay R. Ritter. Initial Public Offerings: Updated Statistics, Table 1, 1998–2002 rows and sample notes.
- SEC. Complaint in SEC v. Credit Suisse First Boston Corp., paragraph 54, and Litigation Release 17327 (January 22, 2002). VA Linux Systems, 2001 Form 10-K, Note 9.
- Pets.com. Final IPO prospectus (February 10, 2000) and Form 10-Q for the quarter ended March 31, 2000.
- IPET Holdings, formerly Pets.com. 2000 Form 10-K, Item 1, Review of Alternatives and Dissolution.
- Amazon.com. 2002 Form 10-K, statements of operations and cash flows and Note 6.
- Federal Reserve Board. Open Market Operations Archive, 1998–2002, and the FOMC statement of May 16, 2000.
- NBER Business Cycle Dating Committee. Announcement index, entries for the March 2001 peak and November 2001 trough.
- Eli Ofek and Matthew Richardson. DotCom Mania: The Rise and Fall of Internet Stock Prices (NBER Working Paper 8630, 2001; journal version published in 2003), abstract and lockup analysis.
- Glenn D. Rudebusch. “Monetary Policy and Asset Price Bubbles” (Federal Reserve Bank of San Francisco, August 5, 2005).
- Webvan Group. Form 8-K filed July 27, 2001, Item 3.
Disclaimer
This article is for educational purposes and does not constitute investment advice. The present-value example is hypothetical. Historical price changes and company outcomes do not predict future results.
