by Robert Litan
More recently, and more damaging, was the highly public failure of the leading prediction market, Intrade, in 2013. The Commodities Futures Trading Commission (CFTC) filed lawsuits against the company for facilitating gambling and violating the securities laws. There were also disclosures that Intrade’s former, late CEO had taken customer funds for himself.5 Before these events, Intrade had racked up some notable successes, such as out-predicting polling data in the 2004 presidential election and for many statewide political races in 2008. The prediction market also performed well despite some very large, and ultimately wrong, bets placed on the 2012 election that appeared to prop up the election prospects of former Governor Mitt Romney.6
With the closure of Intrade, economists who had supported prediction markets have become decidedly pessimistic about their use in the future. One leading prediction market expert, Koleman Strumpf of the University of Kansas, pretty much summed it up when he concluded that the “CFTC’s intransigent interpretation of the law” (about what is a permissible futures contract and what is impermissible betting) made it unlikely that such markets would be used on a large scale again.7
Even if Strumpf’s forecast about future regulatory hostility toward general-purpose prediction markets proves accurate, I believe it is premature to write off the use of prediction markets in other, business-related contexts. Already, a number of private companies have used internal prediction markets among their employees to predict future sales, or which projects in the research pipeline are most likely to be commercially successful. The failure of Intrade should not dampen the use of internal prediction markets by other companies in the future. The core insight of the wisdom of crowds, after all, is too irresistible to ignore.
In addition, although the financial crisis unleashed criticism of perhaps the most widely known example of a prediction market at work—the credit default swap or CDS market—that market continues to exist, and in my opinion will grow over the long run, thanks in part to regulatory reforms enacted in the wake of the crisis.
A CDS is like an insurance contract in which the underwriter of the swap, most typically a bank (but also a hedge fund or, most infamously, an affiliate of an insurance company, like AIG), guarantees to make good on a loan default. Initially, banks bought CDS protection as a way of hedging their loan portfolios. Structured mortgage securities that were not guaranteed by Fannie Mae and Freddie Mac became marketable once bankers (J.P. Morgan was the first) figured out that they could convince credit rating agencies of the securities’ safety if they were protected by a CDS.8
The CDS market was heavily criticized in the wake of the financial crisis, mostly for two reasons. First, unlike conventional insurance, which protects things that people or firms already own, purchasers of CDS buy them without having an insurable interest in the underlying loans or securities they protect. In other words, buyers of CDSs can gamble or speculate on a loan or a security going into default, the kind of criticism that the CFTC leveled at Intrade (though the cash shortfall from the diversion of funds by its former CEO was the more immediate cause of the platform’s demise). Second, because a CDS contract can be sold to any qualified buyer (virtually all of the buyers are institutions) without the purchaser needing the insurance, theoretically there is no limit to how many CDS contracts underwriters can sell and purchasers can buy. For this reason, an explosion of CDS contracts that are not honored can expose an entire financial system to collapse.
Both these criticisms were reflected in the behavior of AIG’s structured products subsidiary, which sold over $400 billion in CDS contracts on securities backed by subprime mortgages, measured by the total face value of the securities.9 Because the parent of this subsidiary, AIG, was so large and well capitalized, the ratings agencies gave the company their top rating, AAA. That was enough for the buyers of the CDS sold by AIG, who otherwise may have insisted that AIG post some margin or collateral both to protect the buyers in the event the seller couldn’t pay, and to limit how many contracts AIG’s affiliate could write.
Most readers know the end of this story: Despite its sterling credit rating, AIG did not have enough margin or reserves set aside to fully honor all of its CDS commitments when they came due in fall 2008. Because so many of AIG’s counterparties (those who bought the CDS contracts from the AIG affiliate) could have been short of cash had AIG been unable to pay off, the Federal Reserve took the unprecedented step of rescuing the creditors of the entire company, putting in loans and capital infusions that eventually totaled nearly $200 billion in return for almost 80 percent of AIG’s common stock. Eventually, AIG recovered, and the Fed got most of its money back, though the net cost of the bailout remains in dispute.10 The rescue damaged both the Fed’s credibility and the glamor status of the CDS contract.
The CDS market remains very much alive, however, since the crisis. Although the notional value of those contracts—the total face amount of the loans and securities underlying the contracts—peaked at nearly $60 trillion just before the crisis, there were still almost $25 trillion in over-the-counter CDS in notional value outstanding at mid-2013.11
A major reason the CDS exists is that it still performs two useful functions, only one of which is insurance. The other function is that the CDS market acts as the functional equivalent of a prediction market in loan or bond defaults. In principle, investors could look to the markets where these debt securities are traded—not only for mortgage-backed securities but for bonds issued by private companies and even sovereign governments—to see how investors assess the likelihood of default.12 But bonds and individual loans are traded intermittently in the marketplace, so any market prices, if they exist, are likely to be infrequent and dated. A more liquid market for CDS thus provides a timely market-based indicator of credit quality, one that would not exist or be as effective as it is, without both hedgers and speculators participating in the market.13
The Dodd-Frank Act has led to fundamental reform and regulation of the CDS market, most importantly by requiring that standardized CDSs (well over half the market) that are readily traded must be “cleared” through centralized clearinghouses, analogous to those for banks. What this means in plain English is that instead of sellers and buyers of these standardized contracts dealing directly (or bilaterally) with each other, they are now required to contract with a clearinghouse. Thus, buyers pay the clearinghouse, while sellers receive funds from the clearinghouse, which requires its counterparties to post margin (or collateral). Margin provides the clearinghouse at least some financial protection in case any counterparty reneges on its deal. The clearinghouse also must have capital and liquid assets of its own (subject to regulation) to cover any shortfalls arising from the failure of margin to cover payments that may be required.
It is not clear yet whether the clearinghouse function is a natural monopoly or whether multiple clearinghouses can coexist and through their competition encourage innovation. The key fact, however, is that because the clearinghouse is the counterparty to the swaps rather than the seller, then in principle another AIG cannot threaten to bring down the financial system—as long as the clearinghouse itself is regulated, as Dodd–Frank requires (and because these clearinghouses are implicitly or explicitly backed by the Federal Reserve in another financial crisis).
The experiences in the CDS market underscore an essential element to the effective workings of any prediction market: Those running the market or the rules governing it must assure that those making predictions through their bets collect their winnings when their bets prove accurate (those losing will have already paid either by buying a ticket up front, as in the case of the Iowa presidential market or internal private sector markets, or buying a swaps contract in the case of CDS). The fact that regulators and policy makers do not appear willing to sanction some kinds of bets does not mean that prediction markets are dead. In fact, they are very much alive in some business contexts, and my prediction is that they will become more widely used over time.
Potentially
Good Financial Innovations
The financial crisis demonstrated all too clearly that not all financial innovation is good. But as I hope to have convinced you in Chapters 8 and 12, many financial innovations are.
Just as in other sectors of the economy that will surely continue to experience innovation, finance is unlikely to be left out. One of the leading thinkers about future financial innovation, 2013 Nobel winner Robert Shiller who was profiled in Chapter 8, has already set out in his frequent writings an extensive menu of economic ideas to be taken up. The common theme among all of them is that each is designed to shelter individuals and institutions from uncertainty and change.14 The following ideas are worth mention.
The first, and perhaps most obvious suggestion, made especially relevant by the financial crisis, is to enable homebuyers to buy financial protection against future declines in the value of their homes. These contracts, or long-term put options, would have greater attractiveness and marketability if they were based on city-wide price indices (Shiller, with his collaborator Karl Case, has created them for 20 of the nation’s largest cities) rather than customized to the values of individual homes (which would make them like traditional insurance contracts which are not tradable). Despite the crisis, however, sufficient interest has not yet developed among either sellers or buyers of such instruments, so the market for them does not yet exist. But that doesn’t mean the idea eventually won’t take off at some point, either as an insurance product or a security, or both.
Second, wage insurance is another idea that would seem to be attractive given the large wage cuts that millions of Americans have been forced to accept in the wake of the Great Recession triggered by the financial crisis. Shiller views this as a product that insurance companies can sell to individuals in case they involuntarily take a new job that pays them less than their previous one, by compensating them for some portion of the wage difference for some limited period of time.
With various co-authors, I have long championed this idea, but have argued that it must instead be a government program funded by a small levy on earnings, analogous to unemployment insurance.15 I highlight this policy idea in this chapter on economic ideas that may provide future business opportunities because government-provided wage insurance is a close intellectual cousin of Shiller’s private sector proposal.
The reason for making wage insurance a government program is that otherwise the market for wage insurance will not develop, because of “adverse selection.” Private insurers will be reluctant to sell such insurance because of the fear that only those most likely to be laid off will buy the insurance. Unemployment insurance, for example, was not provided by the private market until the government stepped in to do it for the same reason.
Nonetheless, there are potentially significant societal benefits to a program of wage insurance for individuals who are involuntarily displaced. Knowing the government will make up some portion of any wage loss and that the offer extends only for a set time once unemployment begins gives those who are forced into unemployment strong incentives to accept new jobs, even if they are lower-paying. For employers, the limited wage insurance paid to employees can be viewed as a kind of subsidy for training them. These outcomes should be especially appealing in the wake of the Great Recession, after which long-term unemployment as a share of total unemployment spiked to the 40 percent range, far higher than it has been for at least the past four decades.16
Shiller, and later with colleagues, has proposed a third financial innovation: a security whose payoffs are linked to national GDP. In principle, as many such securities could be designed and sold as there are countries, or roughly 200.17 Shiller explains why many investors would want to own GDP-linked securities as means of diversifying their portfolios and also to hedge against risks posed by other financial instruments they may hold.
There are good reasons why these instruments do not yet exist, however. On the buy side, investors can already buy the market, as it were, by buying mutual funds or exchange-traded funds linked to the performance of broad market indices. It is not clear how much additional diversification benefits GDP-linked securities would provide.
On the sell side, the impediments seem even more insurmountable. Who would issue these securities? Shiller envisions institutions offering them, much as they now sell puts or calls on individual stocks or baskets of securities. In these latter instances, sellers can cover or hedge the risks of these contracts by holding the underlying security. But there is no underlying GDP-linked security that can hedge against the risk of selling it, except another GDP-linked instrument issued by another financial institution. But some institution must start the process, by taking an unhedged risk in selling a GDP-linked instrument. That has not happened. Perhaps this problem will be solved by clever entrepreneurs in the future, and perhaps there will be a ready market for these securities that existing index-linked instruments do not yet provide. At the very least, Shiller has planted the seed for the concept.
A fourth idea, one not developed by Shiller or his colleagues, that awaits implementation is the application of financial engineering to the financing of drug discoveries, cancer drugs initially, then drugs for all kinds of diseases.
Drug research costs a lot of money and outcomes are highly risky. Drugs must pass three progressively stiffer rounds of testing with the Food and Drug Administration before they can be sold to the public. Pharmaceutical companies are shying away from such risky endeavors. Venture capitalists, which once were quite interested in funding life science startups, have backed away in light of poor returns.
Andrew Lo, one of the world’s leading financial economists, and his colleagues at MIT, have come up with a different financing model for drug research, modeled on the way that mortgage-backed securities are structured.18 The notion is for securities firms to assemble and issue research-backed securities, in which the assets are drugs in testing trials. If the securities are backed by a sufficiently large number of trials—say 100 or more—then all it takes to compensate investors for the risk is for several of the trials to produce home-run drugs whose royalties more than make up for all the strikeouts in the fund. A computer simulation suggests that a $5 to 15 billion mega-fund would return equity investors 9 to 12 percent and bond investors 5 to 8 percent. These are attractive returns for institutional investors like pension funds.
At this writing, no such research-backed securities have been issued to my knowledge, perhaps because there are many details still to be worked out, and possibly because institutional investors are still traumatized by the financial crisis and the role that asset-backed securities in general played in it. But this risk aversion should fade, and eventually I believe the time will come when we will see a major market in these innovative securities, and even more importantly, new life-saving drugs that these securities will help make possible.
A final financial innovation—virtual currency—is already here, and though it was not thought up by economists, but rather by technologists, it has attracted much spirited commentary about its future from economists (and others).
The famous (or infamous depending on your point of view) example at this writing is Bitcoin, a virtual peer-to-peer currency developed by someone named Satoshi Nakamato, who may or may not be Japanese or the code name for more than one software developer. Whoever it is, he, she, or the group left the project in 2010, a year after the currency was launched in 2009. The Bitcoin system operates with publicly available software, which any developer can review.19
Since Bitcoin’s launch, its popularity has had its ups and downs while its price in terms of dollars has fluctuated wildly. By the time you read this, it may not even be around, or it could be more popular than ever. But even if Bitcoin fails for any of the reasons discussed shortly, its short history (with the discussion below reworded to be in the past tense) is worth knowing because the notion of a virtual currency is not likely to die.
As recently as 2011, a Bitcoin could be purchased for less than one doll
ar. As 2014 opened, it was trading for over $1000, although the price fell back later in the year after disclosure that the vice-chairman of the Bitcoin Foundation was being prosecuted for money laundering, and a number of countries, including Russia and China were cracking down on its use. The closure of one of the most popular Bitcoin exchanges, Mt. Gox in Japan, in February 2014, also rocked confidence in the currency. In the future, Bitcoin’s price could either collapse, as some notable economists have predicted (more about them soon), or continue to climb, although not without continued price fluctuations.
The huge increase and variability in Bitcoin’s value conflict with the objectives of its founder and promoters, which is to make Bitcoin a new kind of money, used as a means of payment. If a commodity, even a virtual one, becomes more valuable to hold for speculative purposes than to use, then it’s not really functioning as a means of payment. In March 2014 the Internal Revenue Service ruled that capital gains taxes were owed every time Bitcoin was sold, on the theory that Bitcoin was “property” and not money. This ruling should effectively rule out Bitcoin from replacing official currency, at least in the United States, although if Bitcoin’s price settles down, some users may find it more attractive than conventional money for some transactions for two reasons.
