We value liberty because liberty maintains the diversity that feeds the evolutionary process that grows an open society. We should value asset price bubbles in the economic system for the same reason.

When Ian Macfarlane at the Reserve Bank of Australia raised interest rates last year, at least one of his thoughts would have been that this might help bring the property bubble under control. He could even have entertained the notion that this might protect the gullible masses from their own irrational exuberance. This is certainly an acceptable thought, for most of us think like that too. It is a near article of faith, even among economists, that bubbles are bad and that central authorities should protect us. But this is wrong. Bubbles are a normal part of the market capitalist system when they emerge spontaneously from the market discovery process, and problematic only when artificially induced by ill-considered policy.  

This is not conventional wisdom. Few economists go long on bubbles.[1] Most tend to short them as rather obvious signs of market inefficiency or agent irrationality, as portents of looming disruption of the otherwise smooth functioning of financial markets in particular and of economic systems in general. It is widely held that bubbles are inconsistent with the theory of stable general equilibria in markets, and conventionally argued that bubble formation disrupts the natural efficiencies of market capitalism. Bubbles are a runaway replication of a bad idea, a deviant and amplified noise signal that market capitalism is, both in theory and reality, better-off without. This is why economists short bubbles.

Bubbles are what their metaphor says they are: gaseous pockets of frivolity and distraction; amusing when lightly done, but harmful when over-inflated. Bubbles are like clowns that turn bad, suddenly evil. Economic historians, economic theorists and central bankers alike instinctively believe this to be true. In the Newtonian physical universe there are no negative frictions and in the Newtonian market universe there are no bubbles. In so far as bubbles exist, they are exotica; fit for explaining irregularities, but not regular in themselves. Bubbles are not natural in modern economics.

However, I shall argue that bubbles are natural features of an economic system, and natural mechanisms of economic growth and evolution, when they arise spontaneously from the undistorted market process. I call this real bubbles theory.[2] It argues that, far from being a diabolical pathogen making otherwise stable growth paths unstable, bubbles are more in the manner of spontaneously occurring opportunities. They are the expectational, contractual and financial front end of the process by which economic growth accelerates to economic evolution, and ultimately a sign of system health and vigour, not of decadence and decay.

According to real bubbles theory the lowered cost of experimentation, both in terms of access to finance and market opportunity and to social acceptance of entrepreneurship, among other factors, works to increase variety and learning, to lower uncertainty, and eventually to accelerate the long run growth of the sector ‘under bubble’. Bubbles are a natural outcome of institutional frameworks that facilitate the open market discoveryprocess. Real bubbles are good for economic systems, if they can withstand them: indeed, as will be my argument, the more the better.

Disturbing bubbles?

A bubble occurs when speculative investors chase the market value of a financial asset well above its fundamental value. A bubble begins harmlessly enough as a spur of optimism about a new idea of uncertain value. Examples of subjects are a new financial instrument, organisational idea or interesting technology, or perhaps the prospects of a key financial asset, such as real estate, shareholdings or bonds. But sometimes, and only sometimes, this initial fillip will coalesce into the makings of a bubble when feedback occurs.

A bubble happens when investors change their behaviour, in a coordinated way, to switch from valuing an asset in terms of its yield (for example, dividends on stocks, rental on real estate, and so on) to instead valuing it on the expectation of capital growth. No one has yet given a satisfactory account of why this occurs, but, in any case, when it does, expectations become focused on speculations about the psychology of the market rather than about the fundamental value of the asset. Consequently, a bubble erupts when a herd of investors all begin to move rapidly in one direction. It need not matter why this started, nor where it is headed; what matters, from the individual perspective, is that the many cannot be wrong.

What happens next is usually described in terms of a frenzy or mania, as is popular in journalistic commentary. Yet a bubble is not necessarily an individually irrational process, for many investors do of course make a lot of money in a bubble and, as we shall see, microeconomic theorists have no problem with the idea of a rational bubble. Rather, the problems that a bubble may cause emerge largely from the systemic and unintended consequences of these behaviours when they begin to interact. A bubble is a phenomena of high-order complexity in economic systems.

Several things happen at once during a bubble. Optimistic expectations become increasingly synchronised, and for the same reasons, investment portfolios become increasingly correlated and debt structures become increasingly leveraged. The aggregate consequence of these three phenomena alone is that the economic system becomes dangerously fragile. When this happens, the network of economic coordination becomes highly vulnerable to any disruptive shock that might consequently trigger a rapid unfolding of leveraged positions on a suddenly illiquid market. This is a bad effect that without a subsequent infusion of liquidity, or some equivalent, may yet unleash the even worse situation of general financial meltdown. And this is a very bad outcome. It is why central bankers fear bubbles.

Asset price bubbles disturb the economy. Disruption occurs both as they rise, with the misallocation of capital, and after they burst, with the loss of liquidity and confidence in financial and related markets. From the perspective of economic well-being, the best to be hoped for is a quick correction of the initial market abnormality by rational investors. Unfortunately, however, the worst case scenario is seemingly the normal run of events; the bubble grows to a mass euphoria of avarice and amateurism that, eventually, crashes panic-stricken into a suffocating debt overhang.

Asset price bubbles disturb economists too, and they have ever sought to investigate them.[3] Recently, for example, microeconomists, have uncovered the rational nature of bubbles in a mathematical variant of what is popularly known as the ‘greater fool theory’.[4] This, in essence, states that if you believe that there are greater fools still, then it is rational to buy a bubble asset, even when you know it to be overpriced on fundamental grounds. But a ‘rational bubble’ in this way is only dubiously rational, for it is ultimately consequent on market imperfections that will eventually correct. A bubble cannot last forever, for something will eventually spook it to crash.

This is why both macroeconomists and central bankers are down on bubbles. To them, a bubble is a chaotic and destructive force that causes a surge of liquidity to rise and then dump painfully over the investor classes, so flattening otherwise stable growth paths with a turbulent wash of recriminations, not least of which is the loss of investor and consumer confidence and, with that, spending. As such, few economists go long on bubbles. Most tend to short them as an aberrant and ultimately transient instability, a kind of amplified noise that slows economic growth through its disruptive effects. If a bubble cannot last forever in the long run, then neither can it last for long in the short run.

 Reality bubbles

Yet two stylised empirical observations should give us pause.

First, asset price bubbles are an endemic feature of market capitalism and as old as open societies with property rights over financial assets.[5] In ancient Rome, there were bubbles in municipal property. In Mediaeval Europe, there were bubbles in government bonds. In the 17th century, there were bubbles in tulips, and, by the 20th century, there were bubbles in dot.com stocks. Bubbles have commandeered entire economic systems; France in the 1720s, the United States in 1929 (and again in 1998) and Japan in 1989, for example. The 21st century was rung in with the collapse of a global bubble in technology stocks and the arrival of a near global bubble in residential ‘investment’ property, of which Australians will certainly be aware. There are few asset classes, regions or times that have not at some point hosted bubbles. History may well document a striking antecedent proximity to subsequent advance, as is our second observation.

Second, then, when viewed from a long enough time period, bubbles do not seem overly associated with ruin. Now, it is well-known that bubbles have preceded economic slumps— for example, the US stock bubble in 1929 and the subsequent great depression, or Japan’s land bubble in the late 1980s and its subsequent troubled economic performance. These events are surely causally related; but a longer time scale is required in order to properly observe the effect of the bubble. The observable price dynamic effect of a bubble, what is normally assumed to be the entire bubble, is but the first phase of a real bubble. The performance of the subsequent periods more than makes up for the creative destruction of the initial bubble period.

England was wracked with bubbles in the early 19th century (for example, canal bubbles in the 1810s, railway bubbles in the 1840s, holding-company bubbles in the 1860s, and many others between). But by the end of that century England was the fastest growing industrial economy in the world. The US had railway bubbles in the 1860s, technology bubbles in the 1890s and, by the 1920s, was the new fastest growing economy in the industrial world. Australia had property bubbles in the early 1980s, financial bubbles by the late 1980s, and a few years later entered a decade long period of structural transformation and growth that exceeded the OECD average during that same period. The dot.com bubble burst over five years ago, and maybe its real effects are only just being felt. We now know, for example, a lot more about the value of software and new business models and, moreover, about how the internet works as economic infrastructure. The surge of diversity and lessons learnt from what didn’t work, as much as from what did, has cleared up a lot of uncertainties that previously were obstacles to real long-term investment in information technology.

A good example of this can be seen in the current property bubble in Australia. A number of factors have fuelled this bubble, not least of which is the arrival of important new financial technology allowing more flexible household balance-sheet management (for example, equity lending). The RBA has been publicly worried about the unsustainability of current prices and leverage, yet they seem to be ignoring precisely the financial innovation that the investment public has adopted. Furthermore, it is entirely possible that the current property bubble will leave Australia with a more entrepreneurial investor class, with rejuvenated inner city living and renovated housing, all of which may facilitate a more flexible, integrated and productive work/life balance. Questionable policy warped the bubble (tax breaks on negative-gearing and first homeowners subsidy) but that alone did not cause it. As well, and more importantly, there was an underlying market discovery process at work as investors experimented with the implications of the technological change in personal investment finance. Yet there seems nowhere in sight any recognition of the potential benefits that will naturally accrue from this.  

That something so coincident with times of great economic imagination and wealth creation could perpetually remain maligned should trouble the thinking person, economist or otherwise. If bubbles are so bad, why do they keep happening to good economies? Instead, maybe bubbles are part of what makes an economy good. Maybe we have systematically overlooked the main benefit of a bubble. This is certainly not an entirely new idea. Economists from the Austrian and Evolutionary schools of economic thought, such as Carl Menger, Friedrich von Hayek and Joseph Schumpeter, have long alluded to the positive side of mass disruptions to the economic system in promoting structural change.[6] Yet this point has never really been elaborated in any satisfactory way, and has mostly concentrated about explaining the residual benefits of recessions.

Bubbles are good because they promote variety and experimentation in an economic system. The bubble process facilitates the sort of structural change that economic growth always, in some form, requires. Economic systems, when they are open and therefore competitive, need bubbles to grow. So they require institutional systems and policy frameworks capable of (perhaps vigorous) interaction with bubbles.

A bubble works by concentrating financial liquidity and entrepreneurial attention onto an asset class and its forward prospects. Inside a bubble, the cost of experimentation, and therefore variety generation, is lowered and, by incentive effect, the process of structural change is accelerated. Access to finance is easy inside a bubble. Similarly, the cost of failure is reduced, and the uptake of novelty is high. The economy becomes energised around the bubble, as do the entrepreneurial spirits of agents who happen upon it. Learning is accelerated within a bubble, and radically new business ideas can get a start, as can radically new products. Real bubbles theory, then, is the idea that from a bubble environment there flows the incipient variety upon which the evolutionary economic process of enterprise and wealth creation feeds. Bubbles breed variety, and variety feeds economic evolution, and therefore growth.

From the perspective of evolutionary economics, asset price volatility in general and bubbles in particular are a consequence of experimentation and discovery in an open, evolving environment. Bubbles stem from the private vices of greed and fear, but their aggregate consequence, even accounting for the eventual disruption they bring, is not necessarily bad. The variety and experimental learning bubbles feed into the economic system in a highly-concentrated manner often shows up, sooner or later, as robust growth. Bubbles are a natural part of economic reality in an open society.

Liberty bubbles

One way of seeing bubbles is in terms of their disruptive effects, as the punishment for irrationality, fear and greed. Another way to see them is as creative opportunities bought by a confluence of optimism, liquidity and attention.

In the first view, bubbles create problems. In the second view, they focus attention and liquidity onto hard problems of investment coordination as normal processes of the market capitalist economic system under the guidance of liberal principles. In the first, bubbles are problems to be solved; in the second, they are solutions to a problem.

Liberals would be right to emphasise the latter mechanism as the natural counterpart of Friedrich von Hayek’s famous defence of market capitalism’s failings, namely the liquidationist thesis. The liquidationist thesis argues that recessions are times of structural cleansing because, during a recession, factors of production are cheap and therefore reorganisation is more easily achieved.[7] The implication is that recessions should be left alone so that they might perform the evolutionary function of market selection and structural change as efficiently and quickly as possible.[8]

Real bubbles theory argues that a similar argument can be made about a bubble. But, rather than factors such as capital and labour being cheap and pliant, it is the improved access to finance, and to the supply of and demand for technological novelty, that lowers the cost of structural change. Let us be clear about this: economic growth and evolution requires structural change and structural change is costly, and in many ways. The liquidationist thesis does not say that recessions are good, just that they lower the real cost of structural change, and this has a good effect on growth. Similarly, real bubbles theory does not say that all aspects of bubbles are good, just that they lower the cost of structural change and market discovery, and that this has a good effect on growth. It is cheaper to access factors of production during a recession and it is cheaper to make mistakes during a bubble. Factors are still costly during a recession and mistakes are still expensive during a bubble; the point is that they are relatively less so, and that this has positive incentive effects on the behaviour of investors and entrepreneurs alike. What the liquidationist thesis is to a recession, real bubbles theory is to a boom. In both cases, they are evolutionary mechanisms working to promote structural change and adaptation.

In the first case (the liquidationist thesis), selection and efficiency dominates; in the other case (real bubbles theory), variation and exploration dominates. The sum of the two is an open system evolutionary process.[9] Market capitalism requires recessions for structural change, as the liquidationist thesis argues, but it also requires booms and bubbles, as I argue here.

Bubbles are experiments in knowledge; a natural concomitant to Karl Popper’s notions of an open society and an evolutionary approach to the growth of knowledge,[10] as well as to Hayek’s emphasis on the importance of common law and federalism, as an institutional structure conducive to experimentation and to dealing with uncertainty. In the economy, as in both science and society, sometimes we learn from experiments. From this process, knowledge grows.[11]

Liberalism is a philosophy of openness to the virtues of learning and adaptive response subject, of course, to the boundary constraints of the environment. Liberalism should uphold the freedom of bubbles because they are an experimental endeavour, and therein central to the health of an economically open society. A bubble is a mass expression of liberalism, and liberals, at least, should comprehend this. The Hayekian view of economic liberty can be connected to the nature of economic growth and development with real bubbles theory. An experimentally organised economy does not simply grow; rather, it evolves. And it does so as a continuous process of variety generation, selection and differential replication. This is a process of self-organising markets and other forms of economic organisation as spaces where ideas are tested.[12]

The attention and liquidity that concentrates about a bubble asset leads to a fall in the real cost of economic endeavour, in both cognitive and financial terms, and so causes entrepreneurship and innovation to flourish. People become simultaneously more willing to demand and supply investment. This is not a bad thing. Indeed, the excitement and rich diversity of search acts to feed variety back into the economic system that, in turn, will be selected upon and differentially replicated into a new economic order.

This is how economic systems grow. A bubble is the first phase of evolutionary economic growth. And so the crash of the bubble is not the end of the story, but merely the end of the beginning of the story.

Bubbles and evolutionary economic policy

According to real bubbles theory, the bursting of a bubble, so often met with despair and recrimination, should instead be welcomed as the resolution of a problem. The nature of the problem is the forward investment coordination of an asset under uncertainty. The solution of this will eventually stabilise both expectations in general and speculative actions in particular and, moreover, announce the new rules of the new reality of the asset under focus.

Consequently, bubbles should be left alone as much as possible, both as they rise and after they burst. From the evolutionary perspective, agent and system resilience withstanding, the more powerful and disruptive a bubble is, the better. That is, the more complex the system is, the bigger the bubbles it can withstand, the more experimentation it can process, the more it can grow through evolution.

Without bubbles, insufficient liquidity and attention may arrest the developmental thrust of an asset in its evolutionary relation to the rest of the economic system. To seek to eliminate all losses caused by bubbles (by trying to prick or smooth bubbles, for example, or to legislatively exclude their formation) may be to destroy the dynamical benefits they bring in the first place. According to real bubbles theory, macroeconomic policy should not worry about bubbles; if anything, and where it is safe to do so, it should perhaps even encourage them.

It is important to distinguish between bubbles that originate spontaneously from the market process (what I have here called real bubbles), and those that are induced by regulatory distortions or other bad policy. A recent example of misguided policy response to bubbles was the aftermath of the 1997–98 emerging markets crisis. The origin of the bubble (in low grade government bonds) can in part be attributed to lending distortions induced by the Basle Accords as well as moral hazard emanating from official sector international financial institutions like the IMF. Yet the policy response was to view the market instability as pathological and to seek to restrict the ‘new international financial architecture’ with capital controls and other distortions on the underlying market discovery process. This is not an untypical sequence of events, and may yet be the fate of the Australian ‘investment property’ bubble. Nevertheless, where bad policy starts a bubble and more bad policy is used to end it, the overall bubble process is unlikely to be beneficial in the sense argued here.

Bubbles are good for economic performance when they are the spontaneous outcome of a market process, that is, when they are real bubbles. Real bubbles concentrate an energised mass of attention and liquidity onto a hard investment coordination problem. The bubble activity generates the increased variety that lowers the costs of experimental ventures that opens new territories of forward contracts into which real investment sometimes, and more often than not, flows. Through this evolutionary mechanism, asset price bubbles lead economic growth.

A bubble is good for growth because it creates a low cost environment for experimentation. The results of these experiments may continue to fuel the evolving economy for decades to come. Real bubbles cause long run growth in economic systems that can withstand them. They should be left alone to do so. This is why real liberals don’t worry about real bubbles, and nor should anyone else.

Jason Potts is a lecturer at the School of Economics, University of Queensland, j.potts@economics.uq.edu.au. He studies complexity theory and economic dynamics, and is author of The New Evolutionary Microeconomics, which won The Schumpeter Prize in 2000.