On the case against intellectual property

Or, why I hope to never read ‘it seems to us that there’s no hard empirical evidence to disagree with our foundational assumption’ again.

The papers and articles I am commenting on are listed here;
–https://www.researchgate.net/publication/227390419_Against_Intellectual_Monopoly?fbclid=IwAR0ScE35P7EgoHHu1akh21_n1yRab-yrg9wkW43_4XQgnEh193HjK1lQQPU
–https://onlinelibrary.wiley.com/doi/full/10.1111/j.1468-2354.2004.00128.x?fbclid=IwAR0ScE35P7EgoHHu1akh21_n1yRab-yrg9wkW43_4XQgnEh193HjK1lQQPU
– http://www.dklevine.com/archive/refs4786969000000000990.pdf

Boldrine and Levine’s work on the abolition of intellectual property is a good read, engaging the reader with interesting thought experiments and appealing to his more sado-mathomatical side with a speckling of equilibrium statements.

I believe that the main problems with this work are the assumption of perfect information as well as the assumption that private businesses won’t try and maximise their own profit within the given constraints.

I will begin by deconstructing the arguments presented and then facing them in turn to show why the theory is not self-consistent and requires too many assumptions for it to be applicable to the current markets. While I do think it is a reasonable model for long-term economic planning, my personal conclusion is that the optimal solution for innovation and public welfare is a time-limited intellectual-property monopoly with a legal framework to release the IP after a time in which it is expected that the firm can recover the initial investment costs.

I aim to build a mathematical formalism for the concept of cascading advancement within a highly controlled environment using a dynamic model including price fluctuation based on monotonically increasing market competitors, which I will link when I manage to get around to  it.

Forewarning: either in my ignorance or simple abstract mindset, I either do not understand the numerical assumptions produced (such as net revenue from all sales over the entirety of eternity being less than the initial cost of producing first batch of the product being a given assumption with a hand-waving explanation) or outright disagree with (in the case of the subtitle, where the authors back up the fundamental point of their work with an egregious and stupefying ‘well we aint seen a black swan yet’ instead of a good set of reasoning). So if anyone would like to offer an explanation of what the authors were thinking then you can probably find my email and send me it.

Article Summary

The authors begin their work trying to decipher how intellectual property (IP) in its current form limits social welfare as opposed to a freer model. The arguments for this position can be summed up as; ‘current IP law is arbitrary’, which is certainly true when it comes to IP laws relating to the movie-industry, and ‘eternal monopolies of IP would be devastating for social welfare’, which is true but not realistic (unless Disney gets its way).

Section two then tries to explain how the mechanisms behind copy-proofing software etc is unreliable and costly in the long-term. What the authors don’t acknowledge is that in a system with no IP rights whatsoever, the developer of the property won’t just install copy-proof measures, they will also prevent the user from accessing any of the information within the system, with built-in fail-safes that could wreck the product.

In the scenario of developing an innovative product, the easiest way of increasing initial profits is by not allowing others to see the inner workings of your product forcing them to develop it themselves, thereby you may use your capital to innovate to another product that is fully integrable with the first system, meanwhile the next product might be more innovative and increase social welfare, but it will not be accepted because of back-compatibility issues.

The third section introduces us to some of the fundamental assumptions in their model in a mathematical way;

• The time-based preference discount is always less than one. One value of one unit consumed today is strictly less than the expected value of consuming one-unit tomorrow.
• This is a request for help with a concept. I do not understand why the current pricing is determined by the infinite sum of the exponetiated discount rate multiplied by future consumption. This should instead, I think, be modelled by the current rental rate times by the initial quantity. The current rental rate could be measured by the fraction of (sum u’(t)*discount^t *LT^(-t))/(sum_(tt=1:inf)(sum_(t=tt:inf)u’(t)*discount^t*LT^(-t))) where LT measures the life-time of a product, where LT can be strictly increasing, decreasing or constant.
• Assuming their expression is correct, their next assumption for the model to work is for the value of P_0 > initial costs of production, this statement is produced thrice in separate editions of the publications with no supporting evidence or reasoning.
•  The social-welfare optimum cannot be realised under framework if and only 1. The indivisibility of the product in comparison to the initial stock is too great, and 2. The innovative potential of the IP vastly overtakes its current value.

The conclusion of the article is that in the right conditions, with the right capital requirements and in a system that preserves the Pareto equilibrium, the innovations caused by freely available reproduction and innovation, total social welfare is increased. But the initial costs have to be covered by initial profit and by releasing secondary products such as training in its use.

Criticism

Principles

In the article, they set the scene with a team of 12 researchers producing a unique and innovative drug. They argue that the formula for the drug, the IP, is not the only essential information and that the researchers are a foundational aspect of actually producing the drug.

And here the main criticism of their analysis comes in: they have redefined IP without acknowledging it. You can’t just count the patent as the sole determinant of the production of the good- you need to have all the necessary capital to produce the good, intellectual and practical.

In this lens, the researchers are licensable distributors of the IP, and the concept of IP in the absence of IP laws simply shifts to the necessary means of producing it. This is not a novel concept, but it is a dangerous one as people themselves become seen as capital to be held on to at any and preferably minimal cost. Currently these people are protected from this view by the existing IP laws.

Moreover, they consider the product to be perfectly replicable. This is becoming more true in the age of industry 4.0 (more on that later) with intelligent algorithms able to test hypothesis and make intelligent leaps in logic that would take their human counterparts many hundreds of times as much time to make. But this also ignores the ability of technology to provide ‘trap-doors’ that are un-openable if the user doesn’t have the key.

Let us now consider an ideal product within their framework- 1. a minimally divisible capital requirement, 2. a strictly increasing marginal-utility function, and 3.  A higher initial value of reproducing the good over making a new innovative product from it.

1. The capital requirements mentioned, as I have pointed out, are in fact the operators who know how to produce the good. Assuming this and their minimality in comparison to the initial output, or stock, of the good then the capital must actually be seen either as a future competitor or as an inbuilt type of capital such as a very inexpensive piece of automated lab-equipment. In this  case, the initial profits can be easily recovered by the firm in question if and only if the initial rent covers the cost of research and production. In this case, there is an incentive for the human-capital, or IP licensers, to become competitors and challenge the product within a short period of time. Using their thought experiment, this would be the researchers all quitting and setting up their own lab, becoming competitors in their own right, decreasing the welfare of the company involved as they operate with the same efficiency but with no overhead costs of research. In this case, the company would, as profit-seeking rational actors, reinforce their ability to make a profit by calculating the expected value of returns for the time limit of contracts of those researchers, or simply hold them until the initial costs are covered by their work plus whatever future profit they can expect to make in a potentially saturated market. Either of these options seems abhorrent as the limitation of the individuals freedoms to quit- with which the company has no assurances that the costs will be covered unless the initially available stock is enough to cover the research costs. This is unlikely as after one cycle, all competitors with the correct IP will be operating at the marginal cost rate plus capital, only an industry with a discount rate approaching zero would be covered such as life-saving drug manufacturing. Using inbuilt capital makes it easier to ascertain the ‘loyalty’ of the capital requirements, however this also introduce the problem of keeping its operation as secretive as possible, enforcing IP through means of ‘black-boxing’ production, as goes the apocryphal story of the owners of coca-cola. This would also not lead to the innovation sought by the authors, enforcing IP-protection through means other than the legal system. I will come back to this later.
2. With a strictly increasing marginal utility, the rental rate of the item is always increasing, and so social welfare is constantly increasing (and I assume the argument is) as does the private welfare of the developing-firm. This again cannot necessarily be guaranteed when the free-to-use IP has no costs in its transmission (e.g. easily usable CAD software in designing 3D printed objects, or has been mentioned, Microsoft’s SDOS for $1) and the price equilibrium is determined by the perfectly elastic supply of the good moderated by marginal running costs. Effectively, after the first (or rather a finite number of) cycle the firm which is able to not necessarily innovate but streamline production gains the eventual monopoly on the good, for reasons which will be explained. Within a system that has the current labour-regulation, the potential for IP in the form of human capital to leak from the firm and produce its own more streamlined version can bog the market down with cheaper editions of that good, without the sunk cost of research tying down its profit margin, even with a highly innovative good. This can prevent the medium-term returns of the good to the investing firm to cover their initial costs. Essentially, while the rental rate goes up, the supply is exaggerated by its relative elasticity and the price per unit can dip below what is necessary to make up for the investment. 
3. Tin can openers were invented roughly fifty years after tin cans were invented. The ability to innovate them is severely limited by their crude nature. Therefore they have a much higher value to reproduce and sell than they do to innovate by sellotaping a potato peeler to the top of one. However, if people were free to copy the design of the tin-can opener, the initial costs in design optimisation and production-cost reduction are just dead-weights. In this scenario, it is optimal for the firm to pump out the cheapest design they can as fast as they can to try and secure initial market dominance. This is neither social-welfare optimal (unpriced spill-overs in opportunity cost for the land and resources used in producing an over-manufactured tin can opener) nor very practical for the people who suddenly have a working but less-than-optimal can opener. The argument of the authors would be that the first company to innovate at the smallest cost would then hold market dominance, but with a short production-cycle and limited information inherent through the market, the company that did this would have to make a massive innovation (already discounted from consideration) or have such a reduction in marginal costs that the first cycle can cover the time gone into development of the product, which is not guaranteed.

The Market Today

This paper was released in 2002, well before the concept of industry 4.0 was borne. They can be forgiven, then, for not fore-seeing the technological developments and the main derivations of value in the modern economy.

In today’s society, we see an increasing quantity of open-use technology. The most interesting example of the potential of open-source licensing is Microsoft’s Kinect system.

Launched in 2012, the initial concerns for Microsoft were to protect user’s data and not to stop users from using the output of its technology in innovative ways. That’s why, after a short tumult, the company decided to allow open-source interpretations of its software, since the information being passed from Kinect to the xbox was available through a USB-3 port and was freely interpretable.

The open-sourcing of their hardware outputs has been fantastic for the field of volumetric scanning, with an entire website for open-source projects for hobbyists. But Microsoft has since decided to cut support for the Kinect and are trying to increase its rent (to make up for the opportunity cost of lost revenue in terms of in-house software) by increasing the cost of its connectors and plugs.

This is an example of exactly how the market would operate under such a system without IP. Microsoft, unable to sell its own Kinect based software because a minimal-cost/free use alternative exists are ‘black-boxing’ their innovation and stopping developers from making full use of it through inbuilt arcane-mechanisms.

Apple, by contrast, have always done this with their hardware and will not accept any modifications or repairs outside of their stores (getting around the law which Windows is currently falling fowl of, by simply not having replacement parts). It is worth remembering that even in this system, a method of insurance, i.e. warranty, is always possible at a competitive price- so not all is lost!

But by instigating high barriers to entry, or in other words, a large capital IP requirement, the tech industry is one step ahead of freelancers looking to innovate. Within a free-IP system, it is assumed that the product can be reproduced with the right capital requirements. If it is profitable for people to gain those requirements, they will do so and remove the advantage the firm has. This is okay if the product can make back its initial costs in the first n-cycles where n is the length of time required to gain that minimal capital, but in theory you can extend this time to infinity through artificial means.

This can be done quite easily with the Kinect example. Public RSA keys are openly available and free to use, turning the problem of internet security into a labyrinthine nightmare of solving enough equations in the shortest amount of time to gain the right information to open the right trap-door before it shuts and the keys re-align. The addition of a layer of real-time cryptography to the interface with an encoder and decoder with tamper-proof guarantees is all that is required for the Kinect information to be lost to the developer forever, and the IP to be protected without a law being necessary for such.

In a highly innovative product with an increasing marginal utility and high potential for innovation, such as Kinect, the extra capital requirements are minimal compared to the added profits. Using a ‘loyal’ piece of hardware does not preclude IP in the form of labour from leaking and spreading state-secrets, but it does stop them from interfacing the Kinect, requiring a whole new product development cycle which is costly and does not guarantee returns since the original firm may have developed a loyal customer base, as I will explain below.

The point of this section is that, with industries (admittedly covered by the author’s disclaimer of ‘high potential for future innovation’) with such a large potential, you see the closing of markets and stagnation in innovation that comes with increased fear over stolen ideas as the concept of IP is changed from the legal definition to that of a labour and capital-IP.

It is reminiscent of the fate of all the pigs in Germany after the first world war. There was a huge pork shortage, causing the price to sky-rocket. In response to the problems this caused, it was decided that all the pigs in Germany be slaughtered. This reduced its cost to reasonable prices at first- and only private and paranoid swine-farmers ate well thereafter.

(The pigs are innovative IP’s, the government the government and consumers, consumers in this analogy)

The Road to Monopoly

Why do people use their phone as their wallet, their social media machines, as their health quantifiers and their personal assistant? It is easy to do so.

If you had a product for each of those tasks, even reduced to the minimal scale to match that of a smart-phone’s functionality, then the ability to develop a product separately and successfully is severely hampered.

 Currently, apple and Windows allow developers to access their respective app-stores on the basis of one-time/yearly fees, allowing them to access their source code and state secrets. It’s even possible to develop apps for apple on a Microsoft platform. This is quite neat, as it allows many users to benefit from the existing capital of those companies while being empowered to make money themselves. The fact that there is some unnecessary price-gauging in these figures only goes to show which of these companies believes it has a greater monopolistic stake. But the costs, while potentially prohibitive, are not excessively so.

In a system that allowed the IP of apple to be openly redistributed and copied, the capital which makes those companies money (to a lesser extent in the future with apple’s changing business model) would be produced within a cycle or two at a cheaper rate, losing them long term profit and the ability to invest massive sums into innovative products. Apple would, I believe in this environment, seek to increase its own protection by black-boxing the goods and the source-code, charging steep fees so that its two distinguishable products- software and hardware, can remain profitable in the long run. The market for developers close and innovation is stifled as the only firm able to innovate is the original firm itself.

But the model allows for this, they will simply be overtaken by a firm that has the labour and time to invest in decoding, streamlining production of and eventual innovation and usurpation of the original product after ‘n’ cycles. The issue then becomes; 1. What if n is extraordinarily high, and 2. people choose ease over innovation?

If the initial product has been consumed for a long period before competitors can develop, the market forms a monopoly initially. But this is Ok in the short-term. But they also have a head-start on producing innovations on that tech. If the cycle time (which they influence through black-boxing or by swallowing up competitors) is n>>1, then the competitors really stand no chance, as their products are incompatible with the established firm’s products (by necessity of the original firm’s driving motives as rational and profit-seeking and easily implemented with a digital watermark that is unique and impossible to copy, such as is possible with crypto-chains such as bitcoin et al).

At the end of the long-term cycle, the company produces innovations that can outpace their competitors, or retain a large enough market share to not worry about any moderate innovations on their products. Moreover, since the IP is no longer protected for a limited period by law but by clunky machinery, there is no time-limit on when this becomes public knowledge and the established companies remain established.

In short, systems are developed to be opaque and incompatible and innovation rests on the shoulders of the monopoly powers simply through the all-too-human want for ease.

I have perhaps overused the tech giants in this comparison, as the reader might be thinking ‘sure, these scenarios apply in a capital intensive production environment- but what about more flexible production parameters within industry 4.0, which you said was the point?’.

Take Uber and Uber-eats as an example.

The minimal capital-requirements are minute, if you consider that the capital is fairly freely-available and with minimal barriers to entry (i.e.: own a vehicle). Therefore, development of alternate apps for the same purpose should in theory be very easy- just make the nicest most efficient Travelling-Sales-Person algorithm and voila, a weekend in your mother’s basement surrounded by the corpses of mountain-dew cans and doritos packets instantly makes you into a serious market competitor. But once Uber came out, even despite the relatively short production cycle of creating such an algorithm (or copying theirs if you can decrypt the information), the brand is still available. Even outcries over their supposed mistreating of workers didn’t stop the monolithic force of Uber and its edge on the market through advertising.

Let’s say that you then innovate marginally- you repurpose their algorithm for food delivery. You might cover the initial costs in the sudden flurry of business, but even if it would take more than n cycles for uber to do the same, they can copy your idea immediately (and, in a great benefit to them, you may have even based your algorithm on theirs, making it very easily implementable) and take over the emerging market. While your brand has increased functionality, with imperfect information it is unlikely you will ever take over the established Uber, and even less so with Uber eats incorporated within the original platform.

The point is that, no matter what IP laws you abolish, the original firm will have ways of backing up its informal IP, or capital-IP.

Conclusion

I have analysed the article(s) to present their ideas in a logical way and set up their foundational arguments. I have then discussed the merits of their assumptions, and provided reasons why they are not perfect assumptions to make. I then discussed the modern problems associated with the suggested policy and its limitations, both from a welfare viewpoint and from a labour-right viewpoint.

I cannot stress enough that industry 4.0 will create a mixed market of high-entry-cost and highly labour-intensive markets that would simply close the drawbridge on information if the policy were implemented, and the low-entry-cost, low starting-capital market that would be the software market that will over-centralise and create inefficient monopolies that rely on branding and ease of use over innovation. In some low-capital industries, this is possible in the current climate with open-use software available and driving innovation, and so I can understand the logos of applying this theory to the current climate- with packages such as Python being free, relatively easy to use, and with a high innovation-value. But this is mainly due to the altruism of the developers, which if we took as a reliable mechanism would mean taxes would be halted and philanthropy would be the given M.O. of government finance. This strays into the behavioural economics of scientists, and nobody really wants to take them as an exemplar for the population at large let alone the market.  

As I mentioned earlier, the company that can streamline production of multiple outputs in minimal time ensuring cross-product efficiency and reliance would take the lead of the market, with little to no legal recourse unless anti-monopoly laws were enforced (which I think goes against the grain of the article in the first place). You would see the rise in monopolies that are good at only one thing: protecting or simplifying. Neither of those is necessarily a bad thing, but the problem of finding the optimal social welfare remains, but now the companies which lead are not necessarily the ones with a research-led ethos’. Moreover, the actual IP licensers are now either dis-empowered workers with binding contracts or a minimal view to the actual machinations, or a ‘loyal’ operator such as an AI-decision system which would reinforce the monopolistic aspect that is the issue. And the status quo would be returned- at a lower social welfare.