 The first law of financial bubbles is that they are ubiquitous, and therefore they are banal. Wherever there exist markets and assets, there we will find hurting behavior, momentum investing, and prices decoupling from any relationship to cash flow past, present, or prospective. To begin with, this is what a bubble looks like. An accelerating increase in the price of the asset, followed by a precipitous decline. Here we have the share price of Radio Corporation of America from 1926 through 1932. Lined up with the share prices from 1997 through 2003 of the two enterprise software companies with which I myself was most deeply engaged, BEA Systems and Veritas Software. As you can see, their profiles bracket that of RCA. As proof of the value of studying history to understand context, I wrote my doctoral thesis on 1929 to 1931 around the great stock market crash and its consequences. So I knew the fate of RCA. As a result, Warburg-Pink has declared victory and exited its investments in Veritas and BEA before the bubble burst. Bubbles indeed are banal, but not all bubbles are alike. They can be characterized along two dimensions. The focus of speculation, is it productivity enhancing or not? And the locus of speculation, is it taking place in the liquid capital markets or in the banking system? In the upper right hand quadrant of this chart, the dot com telecom bubble of the late 1990s funded both the build out of the infrastructure of the internet and the first wave of Darwinian exploration of what it was good for. When it burst, the economic consequences were limited and contained within the scope of conventional policy. But now in the lower left hand quadrant, the credit bubble funded beach houses in the Nevada desert while infecting the core banking system that provides working capital to the regular real economy of production, employment, consumption. When it burst, the economic consequences were devastating. It is the degree of leverage that makes the difference about the locus of speculation. The extent to which assets are purchased and carried with debt. This recent and exhaustive empirical analysis of equity bubbles versus credit bubbles across the developed financial systems over more than 100 years sets the context for understanding the difference between 1998-2000 and 2004-2008. As for the focus of speculation, it seems it can be anything. Tulip bulbs, bitcoins, cannabis stocks, any tradable asset can be the focus of a bubble. Now we're going to take a rapid stroll through the history of bubbles including the productive ones. First when London was the central capital market of the world and then in the 20th century when New York supplanted it. Financial bubbles and financial innovation go all the way back together. Tulip futures were traded in Amsterdam in the 1630s preceding the equity derivatives that were traded in London in the 1690s during London's first full-born bubble. For some 300 years the South Sea bubble has been iconic. Behind a bubble we can usually find a plausible story. In the case of the South Sea bubble, it was the possibility of taking over trade to Latin America from the decaying Spanish Empire. A century later, the bubble of 1825 first defined the role of the central bank as lender of last resort as set forth in Badgerit's Lombard Street. With historic irony, this bubble 100 years later was substantially caused by the fact that the Spanish Empire did finally collapse. The model for productive bubbles, bubbles like 1998 to 2000 was established by the financing of Britain's railway network in two phases. The little railway mania of the 1830s followed by the great railway mania of the 1840s. The left-hand panel shows the index of stock prices of companies approved by parliament to build a railway. The right-hand panel shows the actual real investment in constructing the lines. Note how the real investment lags the speculative spike in the share prices. And when the bubble burst, many of those companies went bankrupt. Most of their investors lost money, but no one pulled up the railway tracks. London investors had barely recovered from the domestic railway mania when they had the chance to repeat the exercise across the Atlantic. In this instance, state subsidies, both debt guarantees and the gift of public lands to the railway promoters, supported the speculation. To the rest of the 19th century, a decade did not pass without a speculative bubble. 1863-4, 1871, Trollops, the way we live now, centered on this bubble in 1871, is the existence proof that literature is an essential assist to understanding history and extracting meaning from it. A decade later, the city of London had its first brush with the new technological frontier, electricity. Note the 133% first-day pop in the share of the Brush Electric Company. This is the subsidiary of an American company founded in Cleveland, which brought street lighting for the first time to the United Kingdom. That pop compares favorably with the hottest IPO of New York's tech bubble more than a century later. But London was burned and retreated back to basics, gold mines. Investment in deep mines did depend on innovative technology, but production of gold did nothing to stimulate productivity more generally. And when the first automobile companies went public in London, it turned out that they were led by a con man. David Kinniston, the great historian of the city of London, asks and answers the strategic question. Did it matter that by far the most important financial intermediary in the early history of the British motor car industry was a crook? The answer is surely yes. The analogy with the electrical industry following the catastrophic brush boom of the early 1880s is painfully obvious. At the same time as London was retreating from financing new technology, Wall Street was opening up to innovation. Autos, radio, aviation, electric utilities. Did this shift in the center of financial focus play a role in the transfer of leadership more generally from the UK to the US? I think it did. Radio is a key example of a mission-driven state initiative sponsoring new technological innovation. Under its activist cabinet secretary, Herbert Hoover, yes, the president blamed for doing nothing eight years later in the Great Depression, Herbert Hoover, Secretary of Commerce, joined with the United States Navy to create RCA as the American National Champion with access to a pool of US patents capable of competing effectively versus the UK's Marconi patents sponsored by the Royal Navy. But the most economically significant focus of speculation in the 1920s was electrification. Electrification depended upon a different sort of state intervention implemented in the United States at the state and local level. Given the enormous capital cost of constructing generating stations and distribution grids and the zero cost of the incremental electron, electric utilities needed sanctioned monopolies to protect themselves from competition that would drive price down to marginal cost and eliminate any possibility of profits. In return for those monopolies, the utilities accepted regulation on price and rate of return. All together, the scale of equity financing during the 1920s was unprecedented and remains unmatched. John Maynard Keynes provides a compelling summation of the productive economic effects of the 1920s boom. There can, I think, be no doubt, he wrote, that the world was enormously enriched by the constructions of the Quinquenium from 1925 to 1929. The expansion centered around building the electrification of the world and the associated enterprises of roads and motor cars. The capacity of the world to produce most of the staple foodstuffs and raw materials was greatly expanded. Machinery and new techniques directed by science greatly increased the output of all the metals, rubber, sugar, the chief cereals, etc. This was indeed a productive bubble. The roaring 20s was not all about gin and evading prohibition. Under the shadow of the 1929 crash through the Great Depression and the Second World War and for some 15 years afterwards, financial capitalism was repressed. The first signs of a rebirth of speculation was in response to the U.S. Defense Department's sponsorship of the newest high-tech industry, microelectronics. Sputnik generated a positive response in the financial markets as well as in the national security budget. The last quarter of the 20th century was dominated by what George Soros terms the super bubble that began with the defeat of oil price-driven inflation in 1982 and carried through to the global financial crisis of 2008. It was the combined result of three fundamentally new factors, big government, modern finance theory, and computerization. Let's start with big government. This chart shows the share of state spending and national income of the four most important economies. The two spikes are the two world wars. In 1929, and this is the most important point, the public sector's share ranged between only 5% to 15%. It was 7% in the U.S. Only 2% of that was at the federal level, far too small to cushion incomes and cash flows from the economic consequences of that financial crisis. In 2008, across the developed world, it ranged up to 50%. Even the U.S. public sector was 35% of the economy. Big government underwrote the increasing fragility of the private sector's finances and put a floor under the economy when the super bubble burst. Hyman Minsky is most closely associated with the financial fragility thesis. Financial institutions will evolve endogenously from robust to fragile condition, a thesis that was demonstrably proven in the run-up to 2008. But Minsky also correctly understood the profound economic significance of the revolution and the scale of the public sector. And as George Soros observed, every time the bubble was threatened, the authorities intervened, led by the Federal Reserve. This came to be known after then-fed chairman Alan Greenspan as the Greenspan put. The institutional transformation of the economy was complemented by a theoretical revolution that, ironically, was intended radically to constrain the potential for state intervention in markets. Modern finance theory asserted that efficient markets could not exhibit irrational ways of speculation and that financial innovations, a pyramid of derivatives priced in accord with theory, would enhance market efficiency by reducing and sharing risk. But any instrument for hedging risk is also, by construction, an instrument for gambling. Far from reducing speculation, derivatives amplified it. Modern finance theory would have been irrelevant without modern information technology. IT operationalized finance theory and made it possible to transform equations into derivative instruments, Warren Buffett's instruments of financial mass destruction. Embedded within the super bubble was our own most recent productive bubble in the late 1990s. The research and development boom that it funded was distinctly concentrated in seven high-tech sectors and accounted for almost entirely by companies that were less than 15 years old. Note here the eightfold increase of cash flow for these companies fueled by new share issues and devoted to research and development. Even after the bubble burst in 2000, the stock of cash raised cushioned the decline in R&D. Following the global financial crisis, the world's central banks adopted unprecedented aggressive monetary policy. The unicorn bubble generated by negative nominal returns on risk-free assets considered at some length in lecture three is illuminated by the words of Walter Badgett, longtime editor of The Economist in the second half of the 19th century, writing 150 years ago. The history of the trade cycle has taught me that a period of a low rate of return on investments inexorably leads towards irresponsible investment. People won't take 2% and cannot bear a loss of income instead they invest their careful savings in something impossible, a canal to Kamchatka, a railway to Watchit, a plan for animating the Dead Sea. The literature on the dynamic relationship between financial speculation and real investment is limited but growing. An early approach that struggles to stay within the confines of neoclassical assumptions is here. The authors conceptualize a game played between entrepreneurs and speculators akin to Keynes' beauty contest but extended to the real economy. Irrational exuberance is the term that Robert Schiller, Nobel Prize winner, applied to the growing tech bubble of the late 1990s, more than three years before it peaked. Note the authors' obeisance to the neoclassical gods represented by their repeated use of the phrase as if. They correctly assert that the effects are likely to be stronger during periods of intense change. But the information, the information is not dispersed as they put it up here. The information does not exist. Only after the realization, only after the investments are made, can the information be discovered. Ramana Nanda and Matt Rhodescroft have pursued a research program that focuses directly on the way that a bubble stimulates the financing of new innovative companies. They begin by adding to the set of risks facing every startup, technology risk, market risk, management risk, another one, financial risk. Nanda and Rhodescroft identify a potential coordination failure in time between the present and the future when procedurally rational investors are unable to collaborate effectively in mobilizing the capital needed to get to that promised land of positive cash flow. And the effect will be most negative on the more experimental startups. More than 15 years before their paper, I had played a leading role in constructing a way to escape the no-invest equilibrium that they had defined in theory. Our line of equity financing model proved hugely successful in funding what became half of Baratoss software and all of BEA systems. But its scope proved to be limited, conditioned on this rare coincidence, a plan to build a business by acquisition, a pool of available and relevant acquisitions, assured access to a known market for the technology that was acquired, very rough estimates of the capital needed to reach positive cash flow, including the cost of those key acquisitions to jumpstart the startup, and then price the whole thing up front, fund is required and as deserved. This was invented to fund IT startups in the 1990s when all of those conditions were available. Beyond information technology, this model proved to be especially relevant for oil and gas startups funding them to acquire crumbs from the table of the energy giants. The key practical implication of Nanda and Rhodescroft's theoretical model was that the coordination failure in time could be resolved by a bubble. When riskier ventures can be assured of adequate financing, money does not just chase deals, it changes the sort of deals that get financed. And so empirical analysis proved it to be the case. Startups funded in hot markets were both more likely to fail completely and more likely to be extremely successful and innovative. A recent paper confirms these empirical findings through a somewhat messy attempt to calculate overvaluation relative to fundamentals by using investment analyst forecasts and market movements driven by mutual fund redemptions. But the results are striking. Overvaluation induces greater investment in R&D, possibly amplified by spillovers. And as one would expect, the effects are highly non-linear. In other words, extreme overvaluation is associated with moonshots, projects that are exceptionally innovative. One final real-world example of the necessity of bubbles and the value of productive bubbles. Amazon took $2.2 billion to reach positive cash flow from operations, the last 672 million of which was raised weeks before the bubble burst. At the least, the great dot com telecom tech bubble accelerated the emergence of e-commerce by a decade or more. Recall the discussion of the unicorn bubble in lecture three. From this current excessive investment of speculative capital, it is likely that some, a few, significant sustainable businesses will emerge. Even as many that have no capacity for generating positive cash flow disappear. This is the process of exploring new economic space, the subject of the next lecture.