The need to protect innovation with patents, which used to be holy writ for economists, is still seen as a cornerstone of capitalism by many lawyers and business analysts. But that cornerstone is weaker than you may think. We urge you to suspend judgment and read on.

The case for patents is familiar to most of us. Imagine a world in which patents didn't exist. In that world, an innovator must still decide whether to innovate or not. If the decision is no, nothing is gained (or lost) by the innovator or by society as a whole. If the decision is yes, resources (time, energy, capital) must be expended in getting the innovation to market.

Potential rivals must then choose to imitate or not. If nobody imitates, the innovator effectively becomes a monopolist and reaps the bounty – though, of course, society gets some benefit as well, since something offered at a high price is better than nothing at a lower price. If, however, an imitator does decide to take the plunge in a patentless world, the resulting competition will likely prevent the innovator from even recovering its costs. Society still reaps the benefit of the innovation, and at a low price. But, unfortunately (as the standard story goes), the societal gains are pyrrhic because potential innovators will lose the incentive to innovate.

The moral of the story: no patents, no party. If we want to live in a world where new products, services and production technologies are born from private initiative, we must accept the idea that successful innovators will reap large profits at the near-term expense of consumers.

But typically, this conclusion is unwarranted. Even in the absence of patent protection, imitators' prospects for profit in competition with innovators are at least as problematic as those of the innovator. In the previous story, if the innovator cannot recoup its costs, neither can imitators. Foreseeing this, potential rivals often don't enter the market, leaving the innovator with the opportunity to earn profits commensurate with the risky investment.

Drilling Deeper

What is going on, then? The conclusions reached by the familiar case for patents generally rests on two key assumptions:

• Imitators can enter the market at the same time as the innovator, or at least very shortly afterward, thereby denying the innovator first-mover advantages.
• Once they enter the market, both the innovator and the imitator will have no difficulty scaling up to meet demand, thereby limiting either's ability to profit from initial shortages.

These assumptions rarely pass muster. Yet, in one form or another, they have endured, even in the sophisticated economics literature on the sources of growth, which have stressed the role of patents and the resulting monopoly power as the engines of innovation and growth.

In our own research, we start with the polar opposite assumptions – namely, that imitation generally requires significant amounts of time and resources, and that both innovators and imitators initially face constraints on production capacity. These assumptions, we believe, are closer to reality.

To see why incentives for innovation survive in the absence of patents, look more closely at how an innovative industry evolves. As in a world of patent protection, at the start there is still the innovator, who can enter the market at considerable expense, developing the new product and building some manufacturing capacity. Imitators may or may not enter right away. But even in the case in which imitators can break in quickly, they also do so with only limited capacity.

This implies that the industry's total output will be limited by the innovator's and imitator's capacity, allowing both companies to sell the new good for more than its marginal cost. They thereby reap what economists call competitive rents – revenues above direct production costs – and what pretty much everyone else calls handsome profits. As more imitators are attracted and the industry's total capacity expands, these rents will be reduced. But only in the most extreme circumstances will capacity reach the point at which prices are driven down to marginal cost.

It follows that these competitive rents are collected for a considerable length of time by the innovator and imitator alike, and that both have the opportunity to amortize their initial investments. Indeed, since the innovator gets to market first, or establishes a superior reputation, or both, it will generally earn higher rents for a longer time than the imitators.

Smartphones are a perfect example of this dynamic, but far from the only one. Indeed, this is what has happened in virtually all innovative industries, with or without patents to protect the innovator. It has been the case for the chemical industry since its inception, and was, even for the pharmaceutical industry, until the second half of the 20th century when enforceable patents became the norm for drugs. The same has been true for cars, for agriculture and for steel, and even now for the software industry. Tesla, the fabled market leader in electric cars, acknowledged this reality when it decided to dispense with all patent protection.

In fact, the past 200 years of technology-driven economic growth has never been dependent on patents. Technological progress has followed a simple pattern. A new industry is created by one or more innovators in the absence of patents. The first few companies grow even as they imitate one another – what amounts to spontaneous cooperation that is disciplined by the incentives of competition and that generates major cost reductions and product improvements. Patents take center stage only after the industry matures and a few dominant firms emerge. And those patents are used defensively, either to prevent the entrance of new firms or to limit their market share.

Microsoft is a fine example. Early on, it was opposed to software patents, but now it uses patent protection to protect its dominance in its piece of the software market.

The story of innovation, of course, is not always rosy. Market entrants sometimes overestimate demand, which leads to excessive entry. In that case, prices don't cover research and development costs, and a subsequent shake-up in which productive capacity ceases to outrun growth in demand become inevitable. But this doesn't invalidate the case against patents. It only implies that, in designing a system of intellectual-property protection, we should acknowledge there is a trade-off between the societal cost of legal monopoly protected by patents and that of excessive production driven by excessive entry.

Recast the issue, then. Both theory and historical evidence show convincingly that innovations are generally possible even with competition between innovators and imitators. Further, a plethora of studies covering the past few decades shows that the adoption of patents (as in software and genetically modified plants) or their strengthening (as in pharmaceuticals and biotech) do not lead to higher rates of innovation. These studies suggest instead that the greater the intensity of competition, the more rapid the growth in productivity. Finally, research in economics, management and law provide strong evidence that the costs of maintaining the patent system have been rising rapidly as producers increasingly use patents as strategic weapons to retard competition.

The bottom line here is simple: while, on balance, patents have generated little in the way of societal benefits, they do exact sizeable (and apparently growing) societal costs.

There is thus no reason to maintain a system in which genuine innovations are automatically protected. Rather, if we believe there are cases in which the absence of patents may either impede innovation or bring about costly industrial dislocation, the appropriate response is to confer monopoly rights only in those cases. But which cases actually merit protection – and how should cases be adjudicated?

Start with some basic principles. Standard economic theory justifies some form of intervention to protect intellectual property when the following circumstances lead innovators to conclude that they couldn't otherwise expect to make a profit by going forward.

• The fixed costs needed to innovate are especially large relative to production costs.
• Once an innovation appears, imitation is cheap and accomplished quickly.
• Productive capacity can be built up rapidly.
• Even a moderate increase in total sales would lead to a substantial decrease in price.

How often all four criteria are met is an empirical matter, but not one beyond the capacity of reasonable people to discover. Government agencies often make such judgments in, for example, antitrust cases. It's unclear why a similar approach could not be applied to cases in which an innovation is at stake.

Consider, too, that the extra incentive to innovate need not be in the form of a patent. It could be that a cash prize, financial subsidies or a guaranteed market for successful innovation, which may yield greater incentives per dollar spent than implicit cost of granting monopoly rights through a patent. And in our view, protection should be granted only when the innovation passes the four-step test above.

It should also be noted that opening the door to protection just a crack raises the strong possibility that special-interest lobbying will push it wide open. Hence, on balance, we believe that the better policy choice is to eliminate patents entirely, sacrificing some innovation in return for a system that is surely better than the one we have now.

Drug Dilemma

In any event, let's consider one candidate for special-case status – arguably the most plausible candidate – the pharmaceutical industry.

Note that pharmaceuticals has been among the most steadily profitable industrial sectors for decades, despite a long-recognized crisis in the pace of innovation. The number of really new drugs reaching the U.S. market has decreased from around 45 per year in the late 1990s to only 30 in 2011. Note, too, that the industry has become increasingly concentrated, with a relatively small number of multinationals acting as the gatekeepers of innovation and imitation by engaging in endless and costly patent infringement battles. On the periphery are a volatile group of innovative firms lacking the resources to test and distribute drugs, which leaves them with only one practical business strategy: obtain patents on drugs, then sell the rights to Big Pharma.

In 2012, the FDA did approve 39 new drugs – a seeming reversal of this depressing trend. Among them, however, 11 were marketed at prices exceeding $100,000 a year.

Ordinarily, exceptionally high prices are seen as a reason to suspect excessive market power. But the pharmaceuticals makers have managed to turn this suspicion on its head. High prices, they argue, follow from the high cost of testing, layered on top of R&D costs. Without the promise of patents for successful drugs that leads to multi-billion dollar payoffs, they wouldn't innovate in the first place. Once more: no patents, no party.

The problem here is that it is hard to swallow the idea that the primary cause of drug-price escalation is the cost of research and development. While the proprietary nature of the relevant data makes it impossible to come up with a completely reliable number, industry-sponsored studies place the average total cost of bringing a truly new drug to the market at around $800 million to $1 billion.

Look more closely, though, at that billion-dollar price tag. If we subtract both the publicly funded portion of the research and development process and the cost of clinical trials, estimates of the residual technological cost are a more affordable $100–200 million. It's true that drug companies must invest vast sums in testing with highly uncertain outcomes. It's also true they would probably balk if they lacked patent protection (and the market power it gives). But the premise of the case is flawed. There is no good economic reason to ask the drug companies to shoulder the costs of testing in the first place.

The knowledge gained from a clinical trial is what economists call a public good, in this case giving every potential consumer information on the safety and effectiveness of the drug in question. It may be convenient for Washington to dump the costs of these public goods on private corporations (hence, on patients rather than on taxpayers). But convenience comes at a high price – the need to give monopoly patent rights to drug makers.

Think back to the four criteria justifying patent protection. Many drugs may, indeed, be easy to imitate; producers may not be subject to significant capacity constraints; demand may be very sensitive to price. But the high fixed costs of developing drugs is largely the product of the institutional arrangements for testing – arrangements that could be changed even if one insisted on allowing business rather than government to do the actual testing. For example, once the government determined that a new drug was sufficiently promising to merit advanced testing, it could use competitive bidding to minimize the cost of hiring private contractors to do the work.

Minding Change and Changing Minds

It's not hard to convince people that patents (and other sorts of intellectual-property protection) cost them money. Who doesn't balk at the idea of paying hundreds of dollars for software that only costs a fraction of a penny to distribute via the Internet? It's equally easy to convince them that billions are wasted in battles over property-rights enforcement – think, for example, of the titanic legal struggle between Apple and Samsung. The hard part is explaining that in most cases the inherent production and marketing advantages to being the innovator offer plenty of room for profit even without the benefit of patent protection.

Might we manage some sort of reform, anyway? There is some movement in that direction – in particular in Silicon Valley, where patent protection is increasingly viewed as a major barrier to progress with uncertain short-term benefits even for the Microsofts and Apples of the digital world. And strikingly, there is even some movement in this direction in pharmaceuticals, where the idea of publicly funded testing is no longer entirely anathema to policymakers, and drug makers are quietly licensing emerging-market producers at low rates in the belief that some revenue is better than none.

The patent system as we know it will not fall because do-gooders and ivory tower economists deem it to be counterproductive. But it is vulnerable to rational calculations by corporate patent holders, which plainly have less and less to gain from protection-as-usual.