Devil Take the Hindmost: Lessons from 400 Years of Speculative Manias
Revised once · last on
- — about 1,900 words added, 4,000 removed
- First published
USD30 for Edward Chancellor’s Devil Take the Hindmost: A History of Financial Speculation felt a little pricey, so I got ChatGPT to help me explore the book’s themes and the history of market crashes. These are my notes from that exercise, with links to the historical sources.
What interests me is how often speculation begins with a reasonable observation. Railways would transform commerce; the Internet would change how people communicated and did business. Investors who recognised these developments understood something important about the future. Their mistake was assuming that this understanding justified the price they were being asked to pay.
To follow how these booms developed, it helps to distinguish the opportunity from the investment. A technology can be useful while the companies building it earn poor returns, and a good business can still be bought at a price that leaves little for its shareholders. The financing adds another consideration: even an investment that eventually pays off can ruin someone who cannot meet their obligations while waiting. These distinctions recur throughout the episodes below.
Tulipmania, 1636–1637
Rare tulips were luxury goods in seventeenth-century Holland. Some varieties were difficult to reproduce, and buyers valued their beauty and scarcity. A bulb could be expensive for much the same reasons a collectible is expensive today.
The market changed as people began trading contracts for future delivery. Buyers and sellers could enter transactions without having the cash or bulbs ready, expecting instead to settle the difference between the contract price and a later price. It became possible to speculate on tulips without much interest in owning one. (Cambridge University Press)
If you buy something mainly because you expect to sell it for more, a rising price appears to confirm your judgement. It also attracts people who have watched you make money. But the arrangement depends on someone else being willing to pay an even higher price.
In early February 1637, buyers stopped appearing and the market broke. One explanation links the timing to tulip shoots becoming visible, revealing information about supply that had been hidden while the bulbs were underground. That is a proposed explanation for the timing, rather than a complete account of the collapse. The same research cautions against the popular story that Tulipmania ruined the entire Dutch economy; the wider economic damage appears to have been limited. (Cambridge University Press)
The South Sea Bubble, 1720
The South Sea Company offered the promise of trade with Spanish-controlled South America, although the actual opportunities were limited. Its more consequential business involved government debt. In 1720, government creditors could exchange their claims for shares in the company.
A rising share price made this arrangement more attractive. Creditors wanted the shares, and the company could use its highly valued equity to support further debt conversions. Instalment subscriptions also allowed investors to participate without paying the full purchase price immediately. (Bank of England)
The trouble was that the share price had become part of what kept the arrangement going. Once it fell, investors who still owed instalments needed cash, and those who had bought simply to resell had little reason to stay. The reversal brought losses and investigations into bribery. The broader economic effects, however, should not be assumed to have been equivalent to those of later banking crises. (Bank of England archive)
This is a recurring problem: an investment can appear financially stronger because its price has risen. If that price is also helping it raise money, a decline can damage the business or financing arrangement itself.
Britain’s Railway Mania, 1840s
Railways presented a particularly convincing investment case. The trains worked, people used them, and early ventures had made money. It was reasonable to expect more railways to be built. It was much harder to know which routes would justify their cost.
British railway share prices more than doubled between 1843 and autumn 1845. By autumn 1845, 562 proposed schemes had been submitted to Parliament. Participation extended beyond the established financial elite to the growing middle classes. (Cambridge University Press)
Many shares were part-paid. An investor could subscribe for a £100 share by paying only part of the amount initially, while remaining liable for later calls. This let people commit to more projects than they could have funded in full at the outset. As construction progressed, the companies called for the remaining money, putting pressure on investors and share prices. (Cambridge University Press)
The initial payment could make the commitment feel smaller than it was. Even if a railway eventually became useful and profitable, its investors still needed enough cash to meet the calls along the way. The value of the finished network did not remove the risk of financing its construction.
America’s railways and the Panic of 1873
America faced a related problem after the Civil War. Railways opened land and connected markets, but construction required large amounts of money before a line could earn revenue. Investment could run ahead of the demand needed to support it.
European investors held substantial American railway securities. Financial disturbances in Vienna in 1873 contributed to a retreat from American assets, making railway financing more difficult. (Federal Reserve History)
Jay Cooke & Co., which was heavily involved in financing the Northern Pacific Railway, could no longer sell enough securities to meet the project’s needs. The firm failed on September 18, 1873. Other failures and withdrawals followed, and the New York Stock Exchange suspended trading on September 20. (Library of Congress)
A half-built railway cannot necessarily pay its way to completion. Investors needed to judge both the eventual business and whether enough funding would be available to get it there.
The crash of 1929
The American economy changed substantially during the 1920s. Mass production, cars, radio, and electrification gave investors good reasons to be optimistic. The Dow rose from about 63 in August 1921 to 381 in September 1929. Some investors bought on margin, in certain cases putting down roughly 10% of the purchase price and borrowing the rest. (Federal Reserve History)
Suppose you buy a $100 stock with $10 of your own money and $90 borrowed. If it rises to $120, your equity becomes $30, before interest and other costs. A 20% rise in the stock has produced a 200% gain on your money. But if the stock falls to $90, your equity is gone.
Lenders can demand more collateral or repayment as prices fall. Investors may then have to sell regardless of what they think the business is worth. Their selling pushes prices lower, creating more pressure on other borrowers.
The Dow fell nearly 13% on October 28, 1929, and almost another 12% the following day. By July 1932, it was roughly 89% below its 1929 peak. (Federal Reserve History)
The crash alone does not explain the Great Depression. The economy had already begun contracting, and subsequent banking panics, monetary contraction, policy mistakes, and international strains made the downturn much worse. This matters when comparing crashes: losses in a market can have very different consequences depending on how they affect banks and the supply of credit. (Federal Reserve History)
Junk bonds and leveraged buyouts
High-yield bonds allowed companies with below-investment-grade credit ratings to borrow from investors willing to accept more risk for a higher yield. That can be a sensible transaction. The question is whether the interest compensates for the risk and whether the borrower can service the debt.
The American high-yield market reached nearly $200 billion by the end of the 1980s. Michael Milken and Drexel Burnham Lambert were central to its growth. (Federal Reserve Bank of Cleveland)
These bonds also helped finance leveraged buyouts. An acquirer could borrow much of the purchase price and use the acquired company’s cash flow to service the debt. This could produce a large return on a relatively small equity investment, provided the company earned enough to meet its obligations.
By 1989, highly indebted companies were running into trouble. Defaults increased and financing became harder to obtain. (Federal Reserve Bank of New York) Drexel failed in February 1990, raising concerns about how the market would function without its largest market maker. Investors depended on it for trading and liquidity as well as new issues. (Federal Reserve Bank of New York)
The usefulness of high-yield debt did not make every debt-financed acquisition sensible. A business still needed enough cash to service what had been borrowed, including when conditions became less favourable.
Japan’s asset bubble
Japan’s industrial success provided another convincing basis for optimism. During the late 1980s, rising asset prices were accompanied by strong economic activity and expanding money and credit. Bank lending to real estate helped support the rise in property prices. (Bank of Japan)
Property also served as collateral. When land values rose, borrowers appeared able to support larger loans. More lending could fund further asset purchases, pushing prices higher again. The assets used to justify lending were themselves being bid up with borrowed money.
The Nikkei nearly tripled between the end of 1985 and the end of 1989. After the bubble broke, borrowers still owed their debts while the assets backing them were worth less. Banks and businesses spent years dealing with the damage, and economic growth weakened. (IMF)
Here, falling prices affected much more than investors’ portfolios. They weakened lenders and constrained the credit available to the rest of the economy.
The Internet bubble
The Internet investment thesis was broadly correct about the technology’s importance. It transformed communication, commerce, and advertising. But that left several questions unanswered: which companies would profit, how much they would earn, and what those earnings were worth at the time.
Investors poured money into technology companies while measures such as traffic and market share took precedence over profits. Rising share prices helped fund more start-ups, which in turn bought advertising, equipment, and services from other technology companies. Some of the revenue growth used to justify valuations therefore depended on continued investor funding.
The Nasdaq rose 170% between September 1998 and March 2000. From its closing peak on 10 March 2000 to its trough on 9 October 2002, about two and a half years later, it fell about 78%. An IMF study estimates that approximately $5 trillion in market wealth was lost from peak to trough. (IMF)
The Internet continued to develop after many of those investments failed. Investors could be right about the direction of technological change and still pay far too much for the companies expected to benefit.
What I take from these episodes
Across these episodes, investors repeatedly faced three different questions: was the underlying opportunity real, did the price leave room for an adequate return, and could the investment survive the way it was financed? Confidence in the first often displaced attention to the other two.
Railways and the Internet illustrate the problem with that reasoning. Recognising their importance did not tell investors which companies would profit or how much to pay for them. As prices rose, early gains made the investment thesis appear increasingly credible. The apparent evidence of success drew in more buyers, even as higher prices reduced what those buyers could reasonably expect to earn.
Financing then shaped the consequences of a reversal. An investor who owns an asset outright may be able to wait through a decline. A borrower may have to sell to meet a demand for repayment. A railway still under construction may need fresh capital before it can earn revenue at all. These are different vulnerabilities, but each can force a decision at precisely the point when waiting would be preferable.
The wider damage also depends on who has lent the money and what they are relying on for repayment. Tulipmania’s limited economic effects and Japan’s prolonged difficulties caution against treating every bubble as the same event. Where falling asset prices weaken banks and restrict credit, losses can affect businesses and households far removed from the original speculation.
I think this makes judging a speculative market more difficult than looking for obviously ridiculous ideas. The opportunity may be real, and the enthusiasm may have a sound basis. We still need to work out how much of that promising future is already reflected in the price, and what happens to our investment if it takes longer to arrive than we expect.
A convincing account of how the world will change is a starting point for that work. It is easy to mistake it for a completed investment case.