Digital information is cheap to copy. At the same time it can be expensive to create the first instance, whether it’s a piece of software, a movie, a medicine or a design.
Today, in our Closed world we mainly use intellectual property monopoly rights such as patents and copyrights as a way of incentivising private innovation and creativity and paying back innovators and creators for their investment in creating that first instance.
In an Open world we need an alternative open-compatible mechanism for doing the same thing that can replace intellectual property monopoly rights.
The solution is to innovate legally and introduce a new kind of property-like right: the Remuneration Right.1
With Remuneration Rights, creators and innovators would would receive a “remuneration right” instead of a patent or copyright they receive today. The remuneration right would entitle their owner to payment from a central Remuneration Rights fund. The payment would be proportional to the usage and value generated by their creation. For example, with a medical remuneration rights payment would be linked to the estimated health benefits. With music remuneration rights it would be based on total plays.
In return for the remuneration right, the innovation is open and freely available for anyone to use, build on and share.
Below are detailed case studies of the application of Remuneration Rights in specific domains. We also address key further questions such as: how to handle reuse; how to address international free-riding and the case where only one or a few countries have transitioned to remuneration rights etc.
In setting out a blueprint for an open world there are two main questions to consider. First, how do we decide how much we want to spend on information creation; and second, who pays for it and by what mechanism?
The basic question of how much of our wealth we should allocate to creating new information goods is hard to answer. Ultimately it depends upon an assessment of the value we derive from new medicines, new apps and new movies and how that compares to other things we need or like such as schools and roads or cars and holidays.
But this challenge is neither new nor specific to information goods. Societies face the same problem whenever they decide how much to spend on parks, schools or fighter planes. It is a difficulty inherent in any choice to spend money on one activity rather than another. To do this properly we need to quantify the marginal value of spending, that is, the value derived from spending the next dollar of public money in this way or that – for example, on a new school versus a new park versus new fighter planes. Unsurprisingly, this is very hard to do. For example, a fighter plane may be needed in war, so its value therefore depends strongly on the likelihood of future wars, which in itself may be affected by the deterrent effect of fighter planes.
To help us here, one might look to the market and its prices. Unfortunately, here the market is of little help. Markets only give “good” prices that reflect underlying value in certain conditions. And here many of those conditions are absent. First, public goods such as national defence, parks, education and information all have significant fixed costs combined with low marginal costs. For example, once we have spent the money to have an army, protecting additional citizens comes at no cost. Similarly, once a park is made, the cost of an additional visitor walking over the lawns is nothing – at least up to the congestion point.
Pricing such goods in the market is hard: for instance, without turnstiles, how can the market determine the value people place on Hyde Park? It’s like asking how much we value the seaside or a summer’s day. In addition, the presence of “externalities” or “spillovers” is also problematic for market pricing. In information production positive externalities are significant: new innovations in one area often inspire innovations or ideas in other areas – often quite unrelated – and the nature of knowledge makes spillovers ubiquitous and hard to capture. If we sit next to each other at work, or chat in a cafe, we’ll be learning something from each other. If we tried to own every thought and idea it would be nightmare: how would we distinguish what I’d come up with and what I’d got from you? And the more we try to capture and own information the more complex it would become: should we be paying our school teachers for every idea we come up with because it was, in some small way, inspired by a lesson they gave us thirty years ago?
Finally, at a fundamental level, true market pricing of information goods is impossible. Information goods, unlike land for instance, are not naturally exclusive. Just as my enjoyment of a summer’s day does not prevent you enjoying it, so my watching a film does not prevent you watching to it too. It requires the state to create artificial monopolies over data (copyright on films, patents on software and so forth) for a prices to be attached to them. But the very act of creating these monopolies undermines the pricing mechanism of the market, because it allows the producer of the information to set the price. Although there may be other films to watch and rival software, none of them is exactly equivalent, so there is no proper competition, the pricing mechanism is distorted and resources are not optimally allocated.
A monopoly rights system for information, such as we have at present, does have the advantage of allowing differential pricing of information goods, which gives some indication of differences in value. The holder of the patent on one drug, for instance, may charge a lot more for it than the holder of the patent on a competitor drug. In an open system, by contrast, all information would be priced at the cost of reproduction, which for digital information is uniformly low (even zero), so there would be no indication from pricing of the value that people put on the information. However, market pricing for information covered by monopoly rights has many flaws. Most obviously, the ability or willingness of people to pay may have limited relation to its value: a rich person may pay more to see a film than a poor person can pay for a life-saving drug. This does not help us to compare the relative value of entertainment and medicines – it is like comparing apples with oranges.
Moreover, our current system of monopoly rights already faces many of the questions about how money would be allocated to the creation of information goods. For example, the length and form of monopoly rights that society grants in a given area depends upon our lawmakers’ sense of the value of the information goods that are created. Should patents for new life-saving drugs last for twenty years or ten years or fifty years? Governments have to decide, too, how much to spend on scientific research, which is particularly hard to set a value on. Basic biological research, for instance, may (or may not) lead to breakthrough treatments, but the time-lag may be decades. Legislating for monopoly rights of sufficient length and scope to reward work over such spans (with patents of perhaps 50 years, say) would have hugely detrimental effects and create huge deadweight losses. Innovation would be damaged by the number of spurious patents arising from increased rent-seeking, the incompleteness of contracts, patent examination, and so on.
Monopoly rights for information goods face other specific problems for pricing absent from an open world. Information goods are mostly “experience goods” which have to be accessed and used before they can be evaluated. Second, they often reuse many previous goods, which makes pricing them complex and prone to patent thickets, errors and disputes. Third, information goods often do not deliver immediate end-user value: their benefits may be far in the future.
In summary, appeals to the market will not help us here, because the reasons usually advanced in arguing that markets lead to good allocation of resources break down in the face of the special nature of information.
There can be no final answer to these “how much” questions, but it is possible to suggest a general approach to them that offers a reasonable way forward towards an open world. To begin with, we are not starting from scratch, and in each creative field the current level of investment offers a starting point for the new model. Second, we can at least put nominal values upon our information goods in the open world, using existing tools and the large amount of data available about the levels and forms of usage of digital goods. In the absence of differential pricing, we have several powerful techniques to assess value, and so to compare – albeit imperfectly – musical apples with medicinal oranges. Sampling techniques, for instance, give a very good idea of the overall usage of information goods, be they songs, software or medicines. This, of course, only tells us levels of use, not the value put on it, for which a wider range of economists’ techniques can be employed, ranging from willingness-to-pay surveys to hedonic pricing, which takes into account both internal characteristics and external factors.2
The second of our overarching questions is who should pay to develop the new drugs, new software, new films and so on that our economy and society depend upon and want – and by what mechanisms?
Front and centre here is the problem of the “free rider”. If everyone can use information freely, how do we ensure that people help to pay for its creation, and sufficiently for it to go on being produced on the scale that we want? We are all familiar with the free-rider problem, though usually with regard to shared physical resources rather than the informational commons. We see it in many forms, from the person who drops litter on the street and the housemate who leaves the dishes for someone else to wash up, right through to global issues such as climate change and overfishing.
Of course, digital information is different from physical goods because it does not wear out through use. No matter how often we read Shakespeare, he will still be there, and the formula for Ibuprofen and the source code for Linux aren’t threatened by over-use.3 Instead, the issue with informational goods is the mirror image of this: if there is no commercial incentive, we might not create enough in the first place.
Resources for the creation of new information may come from various sources including government and individual private entities (both for-profit and non-profit), but revenues controlled by the state are likely to predominate. Just as we pay taxes, on an equitable basis, for the provision of shared public goods such as defence and the road network, so we can pay them to fund information goods too. For it is only the state as our collective representative that is in a position to raise funds equitably from all of us and to solve the free-rider issues. The greater the non-state funding that is available the better, of course, because this minimizes the politicization of these issues, but such sources can largely be left to run themselves. It is likely to be up to the state to organise the collection of resources for disbursement to information-creators, and to put in place mechanisms for their allocation – activities that can largely be kept separate.
Collecting the monies would involve a combination of general taxation and levies similar to existing value-added taxes. These levies would be best attached to specific goods or services, usually with a strong informational aspect, for three reasons. First, dedicating a particular levy to the creation of a specific type of information – an internet levy going to fund music and movies, for instance – feels instinctively fair and appropriate. Second, open information makes digital devices such as computers and internet services more valuable, because the more information they can access, the more valuable they are. Third, demand for these services and devices is probably quite inelastic, so a tax would be efficient, in that it would be unlikely to distort behaviour and so create deadweight losses.
But relying upon the state to collect the funds does not mean that we must depend on it to disburse them. Distributed market mechanisms can be used instead to allocate all or part of the money – ensuring, for example, that the choice of which pop music to fund does not lie in the hands of an elderly bureaucrat whose taste is strictly for Gregorian chant.
Within each area of activity, there is competition for resources, and health is one of the largest, so let us consider health research as a paradigm case.
Medical research includes both basic research on human biology and psychology, and efforts to create medicines and treatments that target a particular disease. The example of drug-development, taken here as an instance, applies to all medical innovation.
At the top level, an open world for medicine would look both similar and quite different from ours. Here are its main features:
For ordinary users of medical information such as doctors or patients the surface change would be slight – which is a good thing. Hospitals and clinics would function as they do today, and for users of the healthcare system things would be much the same (especially in countries which already have good universal health insurance). But medicines would be substantially cheaper and more readily available, and – more important still – the development and dissemination of new treatments would be more rapid.
On the producer side, the changes would be greater, though there would still be substantial continuity with existing practice. For example, much existing research is publicly-funded and “open” and we already have many “generic” medicines which can be manufactured by anyone because their patents have expired. In addition, remuneration rights would function a lot like patents – but without the monopoly aspects.
The resourcing of new medical information and drugs would follow this two-step process:
Determining funding levels is hard, as we have seen, but there are methods that yield, if not precise answers, at least some guidance. To begin with, we are not starting from scratch. We know about current funding levels for medical research, for instance, and we can project these forward. In the US, for instance, current levels of public and private research funding together amount to $100bn a year (depending exactly how medical research is demarcated). In future, this will need to be increased, and we could begin by using the basic yardstick of inflation. However, this may be in itself be too low, as demand for healthcare has been growing faster than inflation. Instead, we might look at other areas of healthcare spending that are not information-based, such as spending on doctors and hospital beds, and gear the extra spending on research to the growth of those costs.
When it comes to disbursing research money, the state would use a combination of three mechanisms, each of these has different benefits: academic funding, prizes and remuneration rights.
Traditional state funding of academic research pays for effort, not specific results. It allows researchers to determine their avenues of study partially or entirely for themselves and does not condition remuneration of achieving stated goals. It is particularly useful for funding basic research, which can be very long-term indeed, and high-risk research, which may be worthwhile even though it produces negative results. This kind of funding often envisages a combination of research with teaching, which is desirable for both the teachers and the taught, and is itself a form of long-term investment. The results are unpredictable, and very hard to quantify, but can be spectacular, and have included many of the great achievements of every academic discipline.
The second mechanism, using prizes, is almost the opposite model, for its aims are very strictly defined. A desired outcome is specified, and a prize is offered to the first group to achieve it. For example, the outcome might be “a treatment for malaria resulting in 80% recovery among treated patients” with a prize of, say, “$50m for the first treatment achieving this result as measured in a clinical trial with more than 1000 people”. This approach is valuable when the outcome is clearly specifiable and assessable – often the case in medicine – and there is a benefit in encouraging competition.
The third funding mechanism offers payment according to the impact of the work, in the form of “remuneration rights”. Inventors of, say, new medical treatments would apply for these remuneration rights, just as they would apply for patents today, and would then be entitled to a share of the remuneration rights funding pool. The payment to the rights holder would be proportional to the usage and health benefits of the treatment covered by the remuneration right. So rather than paying specifically for research, or even for a particular outcome, this model pays according to the ultimate impact on health. This impact might be measured in, for instance, “quality-adjusted life-years” (known as “Qualys”).
Adjusting for the actual impact and not just tracking plain usage is important. We cannot simply compare raw usage of Viagra to a life-saving respiratory drug. The fact that only 10,000 people use the respiratory drug but 100,000 people use Viagra should not mean we allocate ten times as much money to Viagra. Allocation based on use must also incorporate some assessment of the value of that use: Viagra may be useful but it isn’t life-saving. That is where Qualys come in: Viagra’s impact on your quality of life is less than that of a drug that keeps you alive. This assessment of value is a key part of managing remuneration rights. For instance, suppose in a given year the national benefit from all treatments covered by remuneration rights is assessed at 100 million Qualys. At the same time, the total funding pool is $100 billion. Then, if the benefit last year from the new cancer drug WonderCure was 1 million Qualys, this would be 1% of the total, so the rights holder, WonderDrug, would receive $1 billion.
In this way, remuneration rights have a strong market-like aspect: payment is determined by individual usage and its benefits. In some respects they are similar to patents, but without the disadvantage of creating monopolies. And just as patents have a limited term, remuneration rights would eventually expire. This would ensure that money continued to flow towards newly invented treatments, and would reduce the need to keep track of the impact of older treatments, which might be infrequently used or simply superseded.
These three mechanisms, then, have different strengths and suitabilities, so the balance between them is important. The aim must be to allocate funds between them so as to maximise the overall impact on human well-being. This, however, is hard to determine, not least because the impact of long-term basic research is indirect and may not be felt for decades. Basic cancer research in the 1960s and 1970s, for instance, led to drug developments twenty or thirty years later.
A suggested initial allocation might be: academic funding 40%, prizes 10%, and remuneration rights 50%. This is based on current funding, which in the US is approximately half and half, private patent-based funding and public funding. Remuneration rights are a natural replacement for current patent-based payments, and current public funding is largely by the academic model. So a 40:10:50 split would be close to the current provision, while assigning some money to prizes.
For the future, a general process for managing and regulating these basic allocations would be needed, and this could be done by establishing an independent organization of experts. This would then also be responsible for setting research prize priorities, for assessing health outcomes (using measures such as Qualys), and managing the disbursement of funds. No doubt other bodies would contribute their expertise, and some of the decision-making would be delegated and distributed. For example, decisions about which research projects should be funded under the academic model would be delegated – as they are now – to panels of scholars in each field. Many countries already have bodies performing such functions. In the UK, the National Institute for Clinical Excellence (NICE) makes decisions about which treatments the NHS will offer, based on a combination of their cost and impact on health (measured in Qualys), while in the United States, the National Institute of Health (NIH) manages much of the academic funding.
The shape of future research and development would probably be much as it is today: a mixture of non-profit bodies on one side and for-profit companies on the other – the latter including both startups as well as larger, more mature companies. Traditional academic and other programs would, of course, continue, and would be funded much as they are today, either by direct government investment or by out-sourcing of research by private enterprise. Non-profits might also engage in research with its sights on the prizes for specific breakthroughs, and could apply for remuneration rights in any new discoveries.
As with patents today, however, the nature of the two risk/reward models – prizes and remuneration rights – makes them principally suited to commercial enterprises. Both are outcome-based, and failure to develop, say, a new treatment would mean the research brought no revenue. On the other hand, success could bring blockbuster returns, which fits them perfectly for a certain kind of investor. Companies would continue to raise money from traditional investment sources including venture capitalists, banks and equity markets, and this involvement of private money has the advantage of harnessing the selection and monitoring skills of investors in determining which projects take priority.
A few differences should be evident, though. First, there ought to be more competition and innovation in the industry, with a growth of medium-sized firms and less concentration of effort in a few massive concerns. This hope is based on the historical experience of countries such as Italy, Argentina and India, which for many years did not have patents on drugs yet had very vibrant, innovative and dynamic pharmaceuticals industries with substantial innovation in both processes and products. However, with the introduction of product patents their industries saw substantial concentration, often coupled with a decline of local industry and the associated innovation.4
A mixed funding programme, then, with a substantial role for the state, could sustain research and development within a country and solve the free-rider problem. But what about free-riding between countries? What is to stop Canada free-riding on the medical research in the United States, for instance?
This problem is not limited to the open model. It exists whatever approach one takes to paying for the production of information, including monopoly rights such as copyright or patents. For example, if a country does not recognize the patents or copyrights of its neighbour, it can benefit from the innovative and creative efforts next door without contributing to the cost: they use the products but pay a lower price that does not include a component for the rights-holder.
The United States offers a historical example. For most of the nineteenth century, it offered copyright on works created and published within the US, but did not recognize foreign copyrights. British authors such as Dickens were not covered by copyright in the US and were freely published by American houses, which made no payments to the author. The books were cheap and very widely available, and since the great bulk of writing in English was done in Britain, this was highly advantageous for the free-riding Americans. Naturally, English publishers and authors were very unhappy and repeated efforts were made to persuade the US to recognize foreign, and especially British, copyright. The eventual solution was an international treaty that gave reciprocal recognition, with the US and Britain recognizing one another’s copyrights.
The solution to the problem of international free-riding in an open world is very similar: international agreements under which countries bind themselves to minimum levels of medical research funding. At its simplest, each country would agree to allocate, say, 0.5% of GDP, but it is more likely that the level would differ between countries, with richer countries committing themselves to higher proportions. Countries might also agree to, for instance, reciprocal recognition of remuneration rights, so that a remuneration right registered in one country could draw upon the remuneration rights funds in other countries too.5
Research is a cumulative process and the ideas and innovations that yield new medical treatments usually build on and incorporate previous work. It is important to think about how this is handled in the open model because otherwise the incentives could become heavily distorted, with resources going to the wrong people for the wrong things.
Consider the problem of “me too” drugs that arises under the current patent system.6 Imagine, in the open world, that a researchers at company WorkLongAndHard create a new drug for diabetes called Diax. They apply for and receive a remuneration right. Then another company, MeToo, makes a slight tweak to Diax and produces a variant called Diox, and this too is granted a remuneration right. But Diox is slightly cheaper to manufacture than the original Diax, and 7 as a result, everyone uses Diox. In the simple model of remuneration rights, all the rewards would go to the predator company, and WorkLongAndHard, which did the pioneering work, would get nothing. This clearly reduces the incentives to innovation. What is needed is a way to apportion a part of MeToo’s remuneration right income back to WorkLongAndHard.
The solution here is very similar to the current system for patents: first, licensing for remuneration rights, and, second a dispute resolution system. Specifically, where a new piece of information being commercially exploited makes use of other information which is subject to a remuneration right, it would require a licence with the rights-holder. So MeToo would make an agreement with WorkLongAndHard, granting it a proportion of any income from the derivative drug Diox.
What, though, if MeToo does not make such an agreement, either because it neglects to ask for one or because the parties cannot agree on a fair rate? In this case, either company could apply for arbitration, either to the courts or (depending on how the system is structured), to an arbitration panel created for the purpose. This is much like the system of patents today, with the one major difference that lack of a license would not prevent MeToo from engaging in research or releasing its drug (though it would run the risk that later arbitration might award a lot of any future income to WorkLongAndHard). Patents, by contrast, are usually seen as providing complete exclusion: without a license, reusers liable for damages and can do nothing, for if they go ahead in the knowledge that they may be infringing the patent, they risk more severe damages for wilful infringement.
Recently, the so-called “bright line” demarcating patents has become a little blurred, at least in the software industries, where there is a lot of reuse. Judicial decisions suggest a move away from an approach based on absolute proprietary rights towards a model of “fair remuneration” similar to that proposed here. This is especially noteworthy in the US, which has otherwise been one of the leading proponents of more and stronger patents (it was the first country to grant patents on genes, for instance, and to extend protection to software and even business processes). Recent decisions signal a possible retreat from this trend, perhaps reflecting a growing awareness of how poorly the patent system functions in sectors such as software, where innovation is cheap and reuse is ubiquitous and massive, since a new technology may build on thousands or tens of thousands of prior innovations. In such circumstances, exclusive patent rights can rapidly create “patent thickets” which are hard or impossible to navigate. This generates uncertainty and risk, especially for smaller players, and acts as a restraint on innovation. In addition, it has given rise to what is called “patent trolling”, in which patents of dubious validity are acquired by tiny companies who then use them to threaten very large ones such as Microsoft, Apple or IBM with litigation in the hope of obtaining a settlement. There are now huge numbers of such dubious patents, because patent offices have been overwhelmed by applications and have granted large number of patents that are either trivial or not novel. Though previously foreseen, this problem has grown massively in the past twenty years as the US courts have become increasingly aggressive in patent enforcement. The “trolling” is done mainly by small companies with no other operations and nothing to lose, which protects them from retaliation. Previously, large companies kept stockpiles of patents and, to avoid mutual stymying, would bargain among themselves to reach reasonable terms, much as the remuneration rights model proposes.
So much for remuneration rights. What about reuse of information or techniques that have been paid for by direct academic funding or by prizes? Probably in the case of direct support for research, reuse would be handled in the traditional manner, with no specific licensing but clear acknowledgement and crediting (citations etc). This makes sense: up-front funded research should be able to build on prior work without any need to license and, in turn, this research may be built on completely freely.8 In the case of prizes, there are already precedents from instances where final success by one team has been largely dependent upon earlier work by another team, and the prize committee has accordingly divided the prize-pot between the two in proportion to their respective contributions.
The funding arrangements discussed so far all rely on state co-ordination, but in medical research especially there are other sources of funds, such as very substantial private philanthropy. Unesco estimated that private, non-profit financing of research and development in the US in 2012 amounted to just under $15bn. This includes both the “mega-philanthrophy” of super-rich individuals and the aggregate contributions of many small donors, such as the various charities that focus upon particular diseases such as cancer.
Some innovations within commercial settings are openly shared, with revenues being generated from complementary goods, such as related services or consultancy (known as the “fries and ketchup” approach: give away the fries and sell the ketchup or vice-versa). This is the basis for a good deal of the innovation in, for instance, surgical techniques: doctors frequently develop new methods and eagerly share them, knowing that this will bring them professional credit and more patients (and one should also not overlook the simple desire to do good and improve human well-being). Here, the new idea is shared freely with the reward being indirect.
This kind of innovation is much more common than we may at first suppose. Research by Eric von Hippel and his colleagues has shown that it extends far beyond medicine, and is ubiquitous – from Michelin chefs to the chemical industry. In fact, von Hippel estimates that the majority of all production innovations are made by practitioners, with a great many of whom do not seek or need exclusivity to justify or resource their efforts.9
The open model of resourcing, with much of it led by government, is entirely compatible with the continuance of these other means of encouraging innovation. In fact, it can enhance them. For example, the existence of remuneration rights would provide new avenues for private philanthropic support – philanthropists could pledge their money directly into the remuneration rights pool, for instance – and would also support traditional “open-compatible” innovation by offering the opportunity to complement traditional revenue streams with new ones such as prizes or remuneration rights.
An open world would be partly funded from private sources. Some of these would anticipate no direct return, including large philanthropic bodies with existing funds, and pro-bono contributions individual motivated either by specific interests, personal need (such as user-driven open innovation) or a concern for the general welfare. Wikipedia is an example of all three. Other non-government bodies would contribute to an open world in expectation of a private financial return, usually from selling some kind of complementary goods. Classic examples are open-source software, consulting and bespoke commissions.
But for several reasons the principal funds for the creation of new information would come from government. First, only the state has the power to solve the free-rider problem, by acting on the behalf of its people; it can its authority to ensure universal and equitable contributions, largely through taxation. It also has the collective decision-making systems for agreeing contribution levels and for allocating the funds.
For much cultural content, the complementary goods option does not offer much promise in an open world: for instance, what can a film-maker charge for, that will bring in revenues on anything like the scale needed to pay for the cost of making a film? Even in software, which already has a vibrant open source economy supported by a complementary goods, this business model can support only a fraction of the desired activity.10 Even a large open-source project such as Wikipedia, with its hugely impressive collaborative and voluntary effort, is largely dedicated to collecting and collating information that had already been produced elsewhere, the great bulk either in academia, funded by government, or in commercial contexts. In fact, a good deal of open-source material available to us today has received implicit or explicit support from government, as well as indirect support from venture capitalists backing closed-source companies (such as Firefox). In an open world, this investment would no longer happen, because closed-source companies would not exist.11
Similar arguments apply to the private no-return option. Over time, funding this way might grow as society grows wealthier, meaning that more great fortunes could be turned to philanthropy and many more individuals would have time to dedicate to knowledge creation. Nevertheless, these are speculative – perhaps idealistic – projections, and in the foreseeable future such funding would probably remain small relative to government funding. Moreover, much of most lavish philanthropy arises precisely from monopoly capitalists – a group that might be smaller in future as a result of political efforts to combat inequality, as well as the direct impact of open information on the shape of the economy. There might be fewer billionaires like Bill Gates.
An understandable concern about government resourcing is that it might not be reliable. One year the government might provide $10bn for research, but in the face of budgetary pressures it might cut this to $7bn the following year. Such unpredictable variations would introduce major uncertainty and add to the risks for innovators. Now, it is true that such risks exist under almost any system, including monopoly rights, since any law can, ultimately, be changed by the government if it so chooses – but some rules are harder to change than others. For example, there may be constitutional protections against “expropriation” that prevent government changing existing contractual arrangements without compensation (this often the case for real property). Governments have also established arm’s-length agencies, such as the BBC or Transport for London in the UK or the Triborough Bridge Authority in New York, which by standing between the state and particular interests, help to guarantee stability.
Similar methods could be adopted for remuneration rights and prizes. Government could set up independent legal bodies to hold the monies for these funds, so ensuring that these monies were ring-fenced from the state. In addition, the government could make the long-term guarantees of funding levels to the electorate and the independent agencies. These bodies might then issue bonds to raise capital to cover remuneration rights arising, for the duration of their life. The annual government funding would then pay the interest on the bonds and repay principal. Since the bonds, held by private investors, would be guaranteed by the state, it would be extremely difficult for any government to default.12
Finally: how do we get from here to there?
Obviously the simplest, but least plausible, is a global “big bang”, with every country worldwide abruptly shifting to this new approach on the same day. Existing monopoly rights such as patents and copyrights would be abolished, new remuneration rights would be available, and the new, open, model would go into operation. The complications, though, are enormous. How, for example, would all the existing patents and copyrights be handled? Would they be converted into remuneration rights – and who would calculate the values? Or would there be compensation, and if so, how would it be paid? But even these huge questions are not the major obstacle. The biggest problem would be how to achieve global agreement for a synchronized change. This would be extremely hard to achieve and negotiations would stretch out to the crack of doom.
Incremental adoption, with individual nations or groups of nations trying out the open approach in full or in part is much more likely to succeed, with the pioneers helping to improve the model, and inspiring the laggards. So let us imagine a flexible transition of this kind, and make it as politically plausible by ensuring that major interest groups do not lose out from the transition. This may involve explicit compensation.
Consider, then, as an example, the medical sector (for which the data is relatively abundant and harmonized), and how the process might look in the US. First, academic-style funding for research could continue largely unchanged. One small change would be to bring the management of clinical trials under the National Institutes for Health. Clinical trials would still be conducted by doctors and hospitals, of course, but the supervision and reporting, and the disbursement of funds, would move over – though trials could still be commissioned and paid for by private enterprises.13 This change might well save money, but its principal purpose would be to address many of the current quality and transparency issues with clinical trials.
Second, legislation would create remuneration rights and abolish monopoly rights. This involves an crucial choice: between abolition of prospective rights only, and retrospective abolition as well. Either way, new medical discoveries would be given remuneration rights, rather than monopoly rights. The question is whether existing monopoly rights would be permitted to continue until their expiry, or whether these would be converted into remunerations rights too. Conversion would have the advantage of simplicity, but the remuneration rights might need to be supplemented by compensation.14
Implementation of remuneration rights would be central to the transition. Remuneration rights would be examined and issued much like patents and could be done by the existing patent office. A new agency (or department, for instance, of the FDA) would be established to carry out the healthcare impact assessment that would be the basis for the annual award of funds. A funding body would be set up to manage the remuneration rights funds. It should be legally and managerially independent of government but with clear contractual agreements with government. Annual funding levels would be allocated on a five-year prospective basis so as to provide security to holders of remuneration rights. Initially, the levels could be based on a formula allowing for various factors including demand and the number of drugs under patent (if the legal change were prospective-only, many drugs would remain under patent, so the remuneration rights funding would start low and grow as the patents expired). In budgetary terms, the funding would come from savings in Medicaid/Medicare and a small tax increase reflecting the lower cost of private healthcare (this tax might even be in the form of a value-added tax on pharmaceuticals).
The key to political feasibility here is keeping existing interests such as big pharma reasonably happy. The social gains from the change would be large, but given the record of the industry, it might find them distasteful, so from the beginning remuneration rights funding would need to be at a high level.
Lastly, international obligations. The US and many other countries are signatories to treaties such as the World Trade Organization’s TRIPs agreement (on trade-related aspects of intellectual property rights), which require provision and recognition of monopoly rights. How is this obstacle to be negotiated? First, experience shows that countries, especially large ones such as the US, can often change or work around their international obligations with surprising rapidity if it is in their interest to do so. Next, domestic companies can continue to obtain patents and other monopoly rights in countries that continue to recognize them. In return, the US would permit foreign firms to obtain remuneration rights on the same basis as domestic companies. But what if it proved impossible to abolish monopoly rights fully? In this case, remuneration rights could existing in parallel to monopoly rights, with innovators having to choose one or the other. Remuneration rights could be made highly attractive relative to monopoly rights in a variety of ways. First, one would create a large remuneration rights fund which would mean the rewards under remuneration rights are high. But this is not the only approach. Government spending on patented drugs could be sharply limited. For example, Medicaid and Medicare Part D drug reimbursements could be limited to remuneration-right drugs only. There could also be direct taxes on income from patents – sharply limiting their attractiveness whilst maintaining compatibility with international agreements like TRIPs. Whilst this opt-in approach is less attractive than a full transition, it might be useful in cases where abolition of monopoly rights is not feasible in the short term, for political or legal reasons.
Music may seem quite a different proposition from medicines, but in the case of music as information (usually digital) – in the form of compositions, recordings and music videos – the overall model is very similar. (Live performances are not “information” in the sense used here, although now they are often a major part of the business, perhaps accounting for a greater share of the artists’ income than recording sales.)
As for medicines, initial funding levels could begin with current ones. Total revenues for the music industry worldwide are around $15bn.15 However, only a small fraction of this is actually going to artists. Just as only a small proportion of pharmaceutical revenues end up funding innovation – with the larger part spent on marketing, profits and administration - so many of the music industry’s costs are superfluous. Roughly a quarter of industry revenues are invested in marketing and promotion, 16 whereas about 13% goes to the artists themselves.17 In an open world, artists would probably pay something to market their own music and we would also imagine them generally getting more. Thus, rather than 13% we will take a figure of 40% of today’s revenues for the amount to go to artists. On that basis, the funding level necessary to take the place of income from monopoly rights would be around $6bn a year globally, with the lion’s share coming from the US (around a third of the global music market), the EU (35%), and Japan (18%).
There are several methods by which government could raise this revenue. Most obvious is general taxation. The sums involved would be quite low, less than a dollar a month per person in the US. However, if general taxation is considered unsuitable, there are other options, of which the most obvious is a levy on related non-information goods: for example, a small levy on internet usage or on digital devices that play or store music. Again the amounts would be modest: for example, with 264 million internet and mobile data plans in the US in 2015, the cost on your internet or data plan would be $0.62 per month.
A more novel option would be to tax online advertising. Many major online businesses, especially those that use content are heavily funded by advertising. For example, Google makes almost all of its money from advertising and relies heavily on the use of free content from others: for example, YouTube consists of videos and music provided by others, and its search engine would be of little value without access to all of the content on the web, via Google. Similarly, Facebook’s revenue comes almost entirely from advertising, but its audience is attracted by content from the users themselves. At present, only a tiny proportion of these monies are paid to the creators. For example, almost none of Google’s search engine advertising revenue reaches them. Even on YouTube, which has a revenue-sharing agreement, the portion paid to rightsholders is low. So a tax on online advertising revenues to fund open information goods is attractive. It would benefit artists and rightsholders by increasing the monies flowing to them, and in a transparent way. The impact on Google and Facebook would be limited, since an advertising tax would be passed on to their advertising clients. It would also have benefits to them in the long-term as it would help secure the supply of high-quality creative work which is essential to their businesses.
There are two further benefits to this kind of tax. First, the major players in online advertising such as Google and Facebook appear to enjoy near monopolies, so such a tax would have less of distortionary impact and act more like a tax on monopoly profits. Second, any potential reduction in advertising brought about by taxation would probably be welcome. Advertising tends to invade our mental space and to prey on our behavioural weaknesses. And its competitive nature tends towards oversupply: rather than improving our choices, advertising merely alters them; for example persuading you to buy a BMW rather than a Mercedes does not increase the value of the BMW to you, or improve the quality of your choice.
However the funds were raised, the mechanisms for allocating them would be similar to those for medical goods. Given the greater importance of individual taste in artistic judgments, they would be weighted differently. For instance, here is a suggested allocation:
Expert-selected (“academic”) funding – say 10% – would be allocated up-front to particular artists or organizations to create new pieces and recordings. This would be very similar to the work of existing public arts programs around the world. It would not, however, include funding for, say, opera houses or concert halls simply for performances as performances are not covered. A performance could only be covered if it was producing “new” valuable information.
Prize-based funding would be quite modest – say 1 or 2% – but might be relevant in cases such as competitions to compose an anthem, a hymn, or a Song for Europe, where the required result could be specified.
Remuneration rights – say 80% of the funding for music – would be issued for both compositions and recordings and would entitle the holder to a share of the remuneration right fund. This would be allocated in approximate proportion to usage of the works. Such a legal and administrative framework is already in general use in the industry. For example, composers already provide automatic fixed fee licences to recording artists, and collecting societies already administer collective licensing for performances to commercial users such as retail stores, bars and nightclubs. In the open world as under the existing monopoly right, the overwhelming bulk of funding would be distributed in this way. One possible change, which has been pioneered by some collecting societies already, would be to make distribution progressive, by reducing the proportion paid to the very biggest stars so as to pay more to those earning less, in order to subsidize the up-and-coming and experimental.
User-choice (the “kickstarter” or “X-Factor” model) – say 8% – would allow some active consumer-choice in the allocation of funding to particular artists, projects or even general policies (supporting blues artists, for instance). Artists would propose projects, such as an album or new song, with a budget. Citizens would each be allocated “voting dollars” with which they could support such projects (with unused dollars being allocated proportionally). This gives the public some control over up-front funding, and has similarities to crowdfunding schemes such as Kickstarter or audience-voting on shows such as X-Factor.
As in medicine, reuse is frequent and important in music. Performers play works composed by others, and composers themselves borrow and elaborate the work of previous artists – and this borrowing is increasingly direct, with the growth of sampling in genres like hip-hop. Such musical reuse fits naturally within the open framework, just as medical reuse does. Here too, anyone would be free to build on the work of others, but would then be liable to pay a proportion of their own remuneration rights payments (or other revenues) to those whose work they build on. These proportions might be standardized for simple cases such as recording a song by someone else. For more complex cases, the two parties could negotiate, with ultimate recourse to the courts if no mutually acceptable solution were found.
In many ways this would be like copyright today the major difference would be that the original creator would not have an absolute right to prohibit the reuse as they do today under copyright but would rather have the right only to “equitable remuneration” (as the lawyers put it). In an age where reuse is ubiquitous this could have a major impact on creativity by making it easier for new creators to build on and incorporate previous work.18
One last point. Current copyright is not just about artists being paid. In addition to the economic rights, there are important moral rights that provide for credit and recognition (attribution). This would be fully retained in this model: artists would have the right to have their work credited wherever it was used or reused.19
Just as with medicine, how to transition to an world of open music is an important question. The main principles are the same: in short, focus upon change within a single country, and upon change that has a good base of support among existing rightsholders. Up-front funding, by the National Endowment for the Humanities, for instance, or the Arts Council in Britain, would continue unchanged. Monopoly rights such as copyright would be phased out and replaced with remuneration rights. Full conversion that is retrospective, would be preferable, with existing monopoly rights converted to remuneration rights from the start. The reason is that copyright term are so long that waiting for them to expire would mean a mixed system persisting for decades. In addition, the amounts involved in “buying out” existing monopoly rights are relatively low.
A new body would be set up to manage remuneration rights, specifically registration, tracking of usage and allocation of funds. Some countries already have public bodies suitable for some or all of these functions. In the US for example, the Copyright Office would be the natural registrar, and in the UK the Copyright Council already adjudicates compensation issues. Many countries also have collecting societies managing copyright in a manner that anticipates remuneration rights: copyright owners pool their rights in the society (which are private entities), and users of copyrighted music, such as bars and restaurants, purchase a single license allowing them to use as much as they wish for a fixed fee. The division of income is also almost identical to the way remuneration rights would work: the society maintains a registry of its members and their copyrights, surveys usage, and then divides up the income in proportion.
These existing institutions and mechanisms could be readily adapted to the implementation of remuneration rights. One approach would be to centralise their management in a single public institution, but alternatively it would be possible to build on the existing collecting societies. There might even be an advantage in allowing these to remain private entities, with regulatory oversight, because many countries have two or three collecting societies, and maintaining this diversity would allow creators Breaking with the Berne Convention will sound to many like returning to a situation akin to the US book piracy years. to choose who to be represented by, so promoting competition, good practice and low administration fees.
Finally, there is the question of how a country can make the transition from copyright when it has existing international obligations. For example, almost all countries are signatories of the Berne Convention, which guarantees the existence and term of copyright for a wide range of works. As with medicines, this could be circumvented by creating a framework that gave rightsholders strong incentives to exchange their copyrights for remuneration rights voluntarily. Copyright and remuneration rights might, for instance, be made mutually exclusive, so that creators had to choose one or the other; remuneration rights could then be made much the more attractive by funding them generously, along with measures such as a high consumption tax on copyrighted music (which would generate revenue as well as reducing sales of copyrighted material).
So far our focus has been on information. But what about information infrastructure, such as fibre-optic networks and server farms which are essential to host, transport and produce information. This question is pressing because in the online world many companies are hybrids, not only providing information or infrastructure but building online services powered by a combination of the two. Google and Facebook are both hybrids in this sense.
Google’s primary offering is a search service accessed online. Behind that is a mass of algorithms, software and databases, but none of that information is ever supplied directly (although intriguingly, in the very early days Google did offer a version of their software to buy as a “Google box”, for companies to install in their systems and use internally). Instead, Google’s own information and its infrastructure are used to provide users with access to the primary webpages they want. Similarly, at the heart of Facebook are software and data, but what users see is an online service, combining that information with infrastructure to connect them to thousands of other users. And many other “internet companies”, such as Netflix, operate on this basis. In fact, even traditional software companies such as Microsoft and Oracle are making a move to the “cloud” – hosted online services that clients utilize over the internet – and away from providing software that clients buy and run for themselves. So, from buying information, there is now a general move to buying information-plus-infrastructure.20
Does this change our general argument? Do we need open infrastructure as well as open information? In short: no. For one thing, infrastructure is unlike information because it is physical and “rival”. Computers and cables have large and growing capacities, but their capacity is limited, and a computer or network being fully used to do one thing cannot do another. And although information and infrastructure are complementary, they are distinct – like cars and roads – and can be developed independently.
Some of the greatest concentrations of power and wealth in the information age are in online services such as Google and Facebook. A big driver of this concentration is economies of scale in information creation and infrastructure provision (all those server farms).
A significant argument of this book is that this concentration is problematic – and that openness of information can combat it. If information and infrastructure were locked together this task would be much greater, but information can be converted to the open model quite independently of infrastructure. And furthermore, by opening up information we would greatly enhance competition in infrastructure, making it less prone to monopoly or oligopoly.
Take Google. Imagine that its core algorithms and software were open. Google would still be a substantial enterprise, thanks to economies of scale and risk-sharing, but it would be far less dominant than today. Similarly, imagine that Facebook provided an open protocol that allowed other parties to integrate their social networks seamlessly with Facebook, so that a user on another network could connect with and share with anyone on Facebook (rather as the internet itself permits anyone to connect to anyone else). Again economies of scale might mean that Facebook was still sizeable, but it would almost certainly not be so utterly dominant: many other social networks would be interconnecting and competing with it.
The infrastructure that powers this digital world is still stuff made of steel and copper and silicon. Infrastructure and information can be separated, and the openness of information needn’t affect the infrastructure – just as the roads being open to anyone does not mean that anyone can use your car. Furthermore, though the two are distinct, openness of information supports and encourages competition and efficiency in infrastructure. For evidence of this we need look no further than the internet itself. Thanks to its open protocols and software, much of the internet’s infrastructure facilities – such as web hosting and internet service provision – are commodity businesses, which are highly competitive and very efficient.
The basic concept of remuneration rights has been around in different forms for many years. For example, it is very similar to the way collecting socities operate. Proposals like it have come up the term “Alternative Compensation System”, “Cultural Flat Rate”, “Prize Fund”. However, there is no standard agreed term for this model. ↩︎
Hedonic pricing uses prices we can see to estimate prices – and potentially values – for things where there is no observable price. For example, what is the value of having a beautiful view? There is no “market” with prices for beautiful views. However, we do observe the prices people pay for houses. Suppose that, in addition to those prices, we also have information on which houses have beautiful views and which do not – along with other factors like size, location etc. Then we can use that information to tease out the implied value of a beautiful view. For example, suppose two otherwise identical neighbouring houses differ only in their view with house A having a great view and house B having none. If house A sells for $10,000 more than house B then we could say that $10,000 must be the price for the view. Closer to home, we can look at the value of life. People rarely buy life directly, even in medicine. But we can look at what people will pay to avoid a small increase in the risk of mortality – for example, what do we have to pay people extra to take on dangerous jobs like window cleaning skyscrapers? These kind of techniques would allow us the value of an additional year of life using the implied price embedded in other transactions. ↩︎
Some resources are needed simply to maintain the availability of information, keeping libraries open and organized, for instance, but in the digital age the costs involved are miniscule compared to the cost of the development of new knowledge. For our knowledge to be useful, however, something more is needed: the education and expertise to use it: there’s little value in the recipe for Ibuprofen without the knowledge of chemistry to understand the formula and the skill needed to manufacture the drug. (This is a trope of many dystopian fantasies, in which the informational wreckage of the present survives but is misunderstood, discarded or even worshipped in some future dark age.) ↩︎
See Boldrin and Levine, Against Intellectual Property. ↩︎
This is almost exactly the approach proposed by in the Medical Innovation Convention proposed by Jamie Love and others and discussed earlier. ↩︎
See, for example, [@hollis:me-too:2004] http://www.who.int/intellectualproperty/topics/ip/Me-tooDrugs_Hollis1.pdf and [@angell:truth:2004] Angell, Marcia, 2004, The truth about the drug companies. New York: Random House. ↩︎
If the drug were exactly there would still be an issue since Diox would still reduce the market for Diax. Yet the remuneration rights model is better than the patent model here. Under the patent model, the predator takes market share and reduces prices, so undermining the profits for the innovator in two ways. Moreover, even if Diox is inferior, the predator can take a share of the market under the patent model by offering its drug at a lower price. Remuneration rights, on the other hand, make it much less likely that a predator will offer a cut-price product. ↩︎
In cases where an organization is receiving up-front funding for only part of its costs, it might apply for a remuneration right as well, but this would be accordingly scaled down. ↩︎
See the work of Eric von Hippel in this area. ↩︎
Using this model, two things have to be paid for – the open-source good and the complementary good – but only the latter generates revenue. So if the open-source good is costly, the revenue will be insufficient. ↩︎
Attention itself can be seen as a complementary good. Google provide me with useful search results and in exchange get my attention which they sell to advertisers. In these cases we also have what are called two or multi-sided markets (Google is a three-sided markets acting as intermediary between users, content providers and advertisers). ↩︎
For a more detailed treatment of the question of commitment looking at the more challenging case of straight-up funding of data within government see [@pollock:models:2008]. ↩︎
This suggestion is inspired by [@angell:truth:2004] where it is one of the main policy recommendations for improving the pharmaceutical industry. ↩︎
One option in valuing the difference between existing monopoly rights and new remuneration rights would be an auction process similar to that described in Kremer. ↩︎
Figure from the IFPI’s “Record Industry in Numbers”, 2015 ↩︎
[@richardson&stahler:2016:uniform] p. 59 cite estimates from Elberse & Ofek (2007) and the IFPI (2012, p7) to give this figure. ↩︎
This figure comes from estimates made by various industry figures, found in media reports for which they were questioned on the subject. I am including here some of the monies that go not to artists but to record labels on the logic that some of these monies do, in fact, go indirectly to artists via the backing that labels give to unsuccessful artists (with the successful artists, labels recoup their up-front support to an artist by deducting those costs from the artists share). At the same time, artists have long complained that record labels charge unfairly and the oligopolistic nature of the recording industry with just four major labels might give credence to concerns over unequal bargaining. If labels do “overcharge”, with that money going to things like profits or bloated management, then we should reduce our estimated percentage of the amount going to artists further. ↩︎
This is a point made eloquently by Lawrence Lessig in his discussions of “remix culture” in his book [@lessig:free:2004]. ↩︎
There is also integrity which provides, in theory, for some control over the use of work – not to use the work in a way that is detrimental to the “integrity” and artistic purpose of the work. As integrity can be used to block others, including other artists, in using and reusing a work we are reluctant to retain integrity in its full form. (For example, we have seen copyright and integrity arguments used to prevent a performance of Samuel Beckett’s Waiting for Godot in which the lead roles are taken by women). Instead of the right to forbid a use, we would suggest that artists have the ability to ensure that those who reuse their work explicitly mark it as being different so that there is no confusion between their original and the derived work created by someone else. ↩︎
The move to the cloud is driven largely by continuing improvements in network reliability, bandwidth and latency – which provide a good user experience when connecting over the network – and the economies of scale in hosting and deployment. Whether these trends will continue is an open question, and it is especially unclear whether the current increase in the coupling of software and service will continue. ↩︎