Economists and historians wish to understand the causes of the Industrial Revolution. In particular why it happened where it did, in the 'West' and more specifically in Britain, and why it happened when it did.
What marks out the Industrial Revolution? What are its defining characteristics? This is an area of keen debate but a consensus still remains: the Industrial Revolution marks a transition, more or less abrupt, to an era of quantitatively and qualitatively different economic progress. The debate, in its crudest terms, remains focused on increases in output1.
At the same time attention is drawn to the qualitative nature of this growth and particularly to innovation (technological progress2) and its role in producing growth. As Mokyr states: It is appropriate to think about the Industrial Revolution primarily in terms of accelerating and unprecedented technological change. In the words of T. S. Ashton's (1948, p.42) famous schoolboy, it was first and foremost a "a wave of gadgets" that swept over Britain after 1760, a string of novel ideas and insights ...
[mokyr_1990:82] (emphasis added). Therefore understanding and explaining innovation and its causes is doubly important: in the first place innovation in itself it is a major feature of the Industrial Revolution and, in the second place, it is also plays a central part in explaining growth.
We seek to narrow the field of enquiry further by focusing on one particular institutional determinant of innovation: intellectual property rights in the form of patents. Of course, to gauge the effect of one factor, such as patents, will require us to take into account the many other influences on innovation since otherwise how can we ascribe a given change to patents rather than some other cause.
To analyze the effect of patents on innovation we first introduce a simple theoretical framework to analyze how patents influence innovation. We then examine the state of the patent system, how it worked, and how it developed over the 18th century. Finally using the theoretical framework and the evidence available an assessment is made of the influence of the patent system on innovation.
(E1) u = u(payoff(i),cost(i))
Increasing in the first argument and decreasing in the second. Cost of innovation would include explicit costs such as materials and labour, as well as items such as opportunity cost of the innovator's time etc. However as patents affect the payoff it is this that will be examined in more detail. Let us write:(E2) payoff(i) = payoff(x(i),y(i),Z(i))
x = income from innovation
y = indirect income from innovation (via consultancy etc)
Z = vector of other variables (e.g. status, alignment of outlook with surrounding culture, pleasure of discovery etc)
(E3) x(i) = x(i,app)
app = level of appropriability of returns from invention = app(pat,mark,trans,learn)
pat = level of patent protection
mark = marketing and sales efforts
trans = ease of imitation and transmissibility of knowledge (level of tacit knowledge in innovation and also level of potential secrecy)
learn = learning curve effects for this innovation
(again app is an increasing function of all of its arguments)
Assuming reasonable behaviour for payoff (eventually diminishing returns) and cost (increasing or constant) we can 'invert' to get:
(E4) i = i(app,...) = i(pat,O)
Where O is set of input variables which includes those already mentioned above and others such as current technological possibility set (state of science and level of current technology), the level of demand, etc. Having derived this 'innovative' function at the individual level, we can aggregate, in theory, by simple addition (it is, after all, a production function)4. Let v be one other explanatory variables in O, in particular one of y,Z,mark,trans,learn. It follows from the assumed signs of the derivatives above that:
iapp,v(i,v,...) < 0
Where this denotes the mixed partial derivative with respect to app and v. To put this into words: increases in other variables that incentivize innovation will reduce the (positive) effect of patents on innovation. This implies that alternative methods of appropriating returns from innovation such as reputation rents, secrecy, associated tacit knowledge etc, will all serve to diminish the effects of patents on the level of innovation. Similarly a reduction in the c(i) (cost of innovation) or an increase in y(i) (indirect sources of income) will lead both to increase in innovation and a reduced role for patents in stimulating innovation.
The modern patent system's origin is in the Statute of Monopolies passed in 1624. Intended primarily to curtail the crown's abuse of its dispensing power, the Statute's role as the legal basis for the patent system was a curious side-effect, a quirk of history. It exempted letters patent for new invention from the general proscription.
[macleod_1989:15] Henceforth, in theory, the monopoly powers (for a term of 14 years) of letters patent would be available to true innovators alone.
Nevertheless the tradition of patents as royal grant, given not to reward innovation so much as to assist and reward friends and supporters, died hard. As Macleod notes, The Restoration patent system ... was still a creature of courtly circles
[macleod_1989:40]. It was only early in the 18th century that the system emerged recognizably in its modern form. Reaction to the abuse of patents in the financial speculations of 1691-4 and 1717-20 led to new restrictions on the system and more sober use of patents themselves. This combined with declining crown power, and new market forces and opportunities [had the result that] ... patents ceased to be the perquisite of courtiers, office-holders and immigrant tradesmen. They began to assume a distinct and recognizable form as instruments of protection and competition among native inventors and entrepreneurs ...
[macleod:40]. From this point on change continued to occur but at a much more gradual rate and within the framework already established. It was mediated via the actions of patent examiners, the decisions of judges and the actions of Parliament (an Act of parliament was needed if a patent were to be extended beyond its natural term of 14 years). It was not to be until the 19th century that Parliament was again to address the patent system and by then the culture and system of innovation had been well established.
How then did the system actually work? Unlike today, and unlike other contemporaries at the time such as France and the Netherlands, the English system was one of simple registration. That is, patents were not and were not expected to be examined very thoroughly either on the basis of novelty or value. Moreover the patentee was not obligated to make public the exact nature of his invention (the introduction of specification in the 18th century made little difference; patents could be sealed before any details were revealed). A final feature worth noting is that, at least from 1720 onwards, the number of shares in a patent was limited to five (this was intended to prevent the damaging speculation associated with patents that had occurred over 1717-1720).
For our purposes we are concerned with changes in the system that would affect the value of patenting and thereby the incentive to innovate. There are three factors of potential importance:... the British Patent Office did not alter its standards throughout the industrial revolution[dutton_1984:111]. The system was one of registration. If an inventor paid the fees the patent was normally granted as a matter of course.
financial costs were ruinous, while the outcome seemed random and too often dependent on a technicality[macleod_1989:73]. But standards (however random) do not appear to have altered systematically before the 1830s. The proportion of contested patents was only 1.5% 1770-1799 (and by the 1840s had only risen to 2.8%) [dutton_1984:180] - though it is difficult to draw any firm conclusion as to what this indicates about judicial decision making.
Hence during the eighteenth-century the patent system emerged in something like its modern form. It provided, in theory, a monopoly position to an inventor for a period of 14 years. Moreover the level of protection afforded by patents and the costs of obtaining this protection did not alter greatly over the period.
i = b0 * pp + b * Z + e
Where i is (log) level of innovation, pp is a variable measuring the amount of patent protection and Z is a vectorofother explanatory variables . Drawing on the previous theoretical analysis below are listed some examples of the variables contained in vector Z chosen on the basis of their prominence in the literature5:... far from ruining him [Rennie] as Watt predicted, established his reputation and led to a flood of commissions.[macleod_1989:104]
Unfortunately it is difficult to quantify the importance of each of these items and therefore hard to formally estimate an equation such as the one above. Thus we cannot derive any precise numerical estimate for the importance of patent protection for innovation. Nevertheless, even the brief survey above of alternative factors influencing innovation would suggest that patents were unlikely to have been of central importance. We now turn to other, more circumstantial, forms of evidence.
[T]o estimate the extent to which inventions were patented would require a monumental search and a good deal of fortune[dutton_1984:110] goes on to assure the reader (without adducing a great deal of evidence) that
although patents and invention were not synonymous, a large number of inventions were patented during the industrial revolution[dutton_1984:112]
On the other hand Macleod dedicates an entire chapter (chapter 6) to 'Invention Outside the Patent System'. This demonstrates that there was a significant amount of unpatented innovation. Moreover innovation was not spread across industries in any direct relation to the size of that industry in the economy. This allows us to infer some level of unpatented innovation: take agriculture for example. There was a very low level of patenting in agriculture yet agriculture formed a very large part of the economy and witnessed significant productivity improvement during this period. Therefore it is very likely that there was significant unpatented innovation in this area8. However as Macleod, in a typically cautious manner, informs us [W]hile it has been shown that there was much innovation outside the purview of the patent system, it remains impossible to quantify the proportions patented and unpatented.
[macleod_1989:114]
One study which provides some quantitative data finds that for the period 1700-1800 only 44% of textile inventions were patented [griffiths_1992]. However this global figure conceals a sharp change between the mid-century, when most inventions were unpatented, with the end of the period when patented inventions represented a majority9. Their data also indicate the importance of other forms of institutional support for patenting, in particular, for the case of textiles the role of the Society of Arts which awarded prizes and actively promoted and disseminated innovation10.
We also have the implications of the raw quantitative evidence on the number of patents enrolled. Table 1 lists the numbers of patents granted in each year from 1720 to 1799. As can be calculated from that table, the entire period from 1720 to 1799 saw only 1942 patents enrolled, which is an average of approximately 25 a year. Even in the final decade the average was only 65 patents a year. Now, if it were the case that a large proportion of innovations were being patented then these numbers would imply an implausibly low level of total innovation. This would be true even if we confined our attention to major advances; moreover we should bear in mind that some proportion of patents were for trivial or economically worthless inventions. These figures therefore suggest that a significant amount of innovation was not patented.
Thus the evidence available indicates that, for the early part of the industrial revolution studied here (up to 1800), a substantial amount of innovation occurred outside of the patent system. In fact, it seems likely that the majority of innovation, even if reweighted by importance - of course a very difficult task, occurred outside the patent system.
The second direct piece of evidence we possess is information on the financial value of patents to innovators11. This can be obtained in several ways. First by self-reportage. Second by price paid for patents. Empirically the second of these is strongly preferred given the error and bias present in self-reported figures. At the same time there are several problems with the second form. Theoretically in a competitive market the price paid would accurately reflect the value of the patent. However in practice the market was highly imperfect, and, rather than selling at arms length via licences or the like, many innovators tended to form partnerships, or set themselves up to exploit the innovation themselves. Moreover given the selection bias inherent in the data (only innovations that were successfully sold, and then usually only those sold for substantial amounts, enter the record) it is hard to know how much inference we may draw regarding patents as a whole.
We also encounter many of the same issues already discussed in the previous section. For example, the payment for a patent does not allow us to value the marginal value of the patent. A patent purchase or license did not simply confer the right to make such and such an object but provided whole set of other services such as legitimation, tacit knowledge and other support from the innovator, marketing benefits, etc. These other services were not protected by the patent's monopoly and would have therefore been purchased anyway. Thus, of a patent sold for a thousand pounds how much do we ascribe to the patent itself? Lacking this information, it is difficult to start on a utility calculation (using the framework set out in the section on theory above) that would allow us to work out the contribution of patents to the level of innovation.
Can we now answer the central question: how important were patents for innovation? The analysis above has shown how difficult it is to draw conclusions on this issue. However perhaps the complexity itself lends support to a negative answer. With so many other factors, both major and minor, that influence the level of innovation it is unlikely that patents themselves had any impact, positive or negative, other than a minor one. Moreover, how do we distinguish patents from the culture they were part of? A society which awards patents (in the modern sense) is one that must be conscious of innovation, possesses a culture of empiricism and scientific enquiry, and, perhaps most significantly, has advanced to a level of considerable legal and economic sophistication. In this sense patents are a very strong signal for advance into the modern age of economic progress, so strongly based as it is on technological progress. However, interpreted narrowly, as a method by which innovators are incentivized by the ability to appropriate greater returns from their investment, patents are only one part, and probably a rather insignificant one, of the explanation for rise in innovation over the period studied.
Also omitted is consideration of alternative methods of reward. For example the government could state that they would compensate all innovators appropriately and in return the innovation must be made public and freely available. This solution suffers from all the usual problems of a central planner solution: 1. no market mechanism to aggregate private information on the worth of a invention (via its price); 2. government likely to be less efficient than market (a complex bureaucracy would be required to assess patents); 3. incentive alignment between central planner and population at large not guaranteed; etc. However empirically we do observe that most economies, both in the past and even more so today, expend very considerable funds on supporting public research especially of the basic, 'blue sky' type.
The verdict on the importance of the patent system in explaining technological creativity is thus decidedly mixed. Given the private and social costs of patenting and patent litigation, and given the many alternatives to patenting, its impact on the technological creativity of societies is far from clear.[mokyr_1990:251]
We have also ignored 'cross terms' to capture interaction between variables. For example does competition increaseor diminish the value of a patent and thereby its effect on innovation? Consideration of such issues would complicate further an already difficult analysis and are likely to be of second order. Thus, in general, they will be ignored.
.... it must be noticed that the partly tacit nature of innovative knowledge and its characteristics of partial private appropriability makes imitation, as well as innovation, a creative process .... and which is economically expensive - sometimes even more expensive than the original innovation (...[cites] ...) This applies to both patented and non-patented innovations.[dosi_1988:1140]See also [macleod_1989:105] on lock manufacture by Bramah.
For many of them [patentees], exclusivity was not a major preoccupation: they sought enabling rather than monopolizing power, defence rather than offence, kudos rather than control. And where control was deemed important and secrecy not a viable option, a potentially weak patent was the only alternative to no protection at all.[macleod_1984:74]
for more evidence that patent registrations clearly could reflect competitive necessity as much as genuine creativity.[griffiths_1992:889]
... the sample does not constitute a comprehensive record of innovative activity in the relevant years. Many and perhaps most inventions continued to go unrecorded. However, a more extensive investigation or 'private' (unsponsored) innovation would merely serve to confirm that the majority of additions to the stock of productive knowledge over the eighteenth century, even in a technologically sophisticated industry such as textiles, were not patented.[griffiths_1992:885-886]