Ec261-5-SP MANAGEMENT OF NEW TECHNOLOGY 2019/2020
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Ec261-5-SP
Second Year Examinations Summer 2019/2020
MANAGEMENT OF NEW TECHNOLOGY
1. Answer all parts (a), (b), and (c) of this question.
a. [20 marks] Does a market economy provide too little or too much incentive to undertake innovation? Explain, identifying the criterion you are using to answer. Identify the roles that competition and appropriability play in your answer.
Sketch of answer: We would normally use social surplus to address this, although consumer surplus may be thefocus of some answers and should get some credit. Static welfare loss compared to dynamic welfare gain would be included in a good answer. We argued in the module that becausefirms do not capture consumer surplus (unless they price discriminate)for completely new products and generally positive information externalities exist even with veryfirm patent protection or other appropriability mechanisms such asfirst mover advantages, there is an underincentive to invest. This can be seen in some of the growth models we discussed early in term. On the other hand, R&D races and exclusion considerations may generate an overincentive to invest. A good answer would identify the argumentfor both points of view and would also cite empirical evidence to support the answer. A very good answer would give additional detail on differences in incentivefor different market structures, identifying replacement effectsfor monopoly/concentrated markets, and would distinguish between different types of innovation.
b. [20 marks] Does empirical evidence show that patents increase innovation rates or not? Explain, citing specific evidence as appropriate.
Sketch of answer: Students should cite at least one source takenfrom historical or current data to answer this question. A good answer would either give considerable detail on one or would review several with less detail. A very good answer would do both. Another aspect of a very good answer is to include some critique of existing evidencefrom a methodological or other perspective.
c. [10 marks] Compare the static and dynamic welfare effects of using patents to incentivise innovation or, alternatively, using prizes to incentivise innovation. Point out both advantages and disadvantages of each method.
Sketch of answer: Patents will tend to result in a static welfare loss in exchangefor a dynamic welfare gain. Prizes do not have this drawback if we ignore thefinancing of the prize. On the other hand, “picking winners” may result in poor direction of research efforts at where demand actually lies; although patents will tend to result in directing research awayfrom potentially socially useful but not very profitable innovations (think of orphan drugs or vaccines to be used in relatively poor countries). A very good answer could go into more detail on other benefits of patents that create innovation indirectly (for example, byfacilitating diffusion of the underlying knowledge, allowing a degree ofspillover…which could be tied back to innovation incentives using the Cohen and Levinthal reading). Prizes and patents can be complementary systems, with prizes used where patents tend tofail and vice versa, as well as using prizes in cases where the direction of socially desirable research is clear. A very good answer could draw in some other comparisons including mentioning additional tools to incentivise innovation.
2. Answer all parts (a), (b), and (c) of this question.
a. [10 marks] Define and explain the concepts of moral hazard and adverse selection. Illustrate each concept with two examples: one in the context of new technology sale, the other in the context of new technology funding.
Sketch of answer: moral hazard and adverse selection are bothforms of opportunism present under asymmetric information. In moral hazard, the asymmetric information is on the actions taken by the informedparty (in particular once the contract is signed) and these actions affect the value of the trade; whereas in adverse selection, the asymmetric information is on a characteristic (of the informedparty or of other conditions of the trade) –and similarly these characteristics affect the value of the trade.
New technology sales may be characterised by both and infact by various manifestations. For example, it may not be clear that the technology that is sold is used in production even if it is valuable. For example, a buyer may purchase an exclusive license in order to “bin” the technology if it potentially eclipses the existing technology of the buyer. If the license isfor a royalty that is conditioned upon use, this will reduce the earnings of the licensor. On the other hand, a licensor may be selling a “bum” technology that is not as good as is claimed. This may harm the licensee, particularly if an upfrontfixedfee is used as the pricefor the sale.
Thefinancing of new innovation also potentially allowsfor both types of AI. We reviewed this in the context of a bank loaning money to an innovator. Where there is no cost of default, this loan can result in the innovator choosing to exercise suboptimal effort (inventing a bad technology may actually be more profitablefor the borrower since the lender suffers morefrom the downside of the project than the borrower) or can result in a borrower that is a “good researcher” funding innovation internallyfrom retained earnings and only going to the lending marketfor bad technologies –for the same reason.
b. [20 marks] Explain why asymmetric information may cause a market failure in the markets for selling new technology and the market for funding new technological development.
Sketch of answer: Thisfollows on to the answer to (a), and could involve a numerical example illustrating that moral hazard and adverse selection both occur and affect the value of the less informed party to an extent that trade will not occur, as changing the cost of borrowing does not make the problem better (indeed it makes it worse). A good answer would include both the numerical example and intuition of this. Similarly,for new technology sales, developing a case of selling with afixedfee only under adverse selection and sellingfor a royalty onlyfor moral hazard could be developed with a numerical example and intuition. The example need not be highly developed. If one of the two examples is with a well-developed numerical basis and the other is relatively briefthat’s enough.
c. [20 marks] Propose and defend a solution to these market failures in each of the two markets (technology sale and new technological development funding). Point out any disadvantages as well as any advantages of your solution.
Sketch of answer: students could review any institution that resolves this problem, but we spent considerable time on venture capital as a method of monitoring and equalising information betweenfunder andfunded, pointing out various mechanisms
for this to occur so a good answer would include some channels of equalising
information, notjust the statement that VC does this. Technology sale issues can be dealt with using demonstrations (with appropriate confidentiality agreements, although we are sceptical of these in lecture), changing the contractform to spread the risk, signalling via the contractform, or any of a number of other solutions. Again, a good answer would include detail on one of these and some mention of the
other without a great deal of detail. A very good answer would include a developed example of both.
3. Answer all parts (a) to (c) of this question.
a. [10 marks] Illustrate and explain what is meant by a “double marginalisation problem” in technology licensing.
Sketch of answer: A graph showing that the price-cost margin of an input can
generate inefficient production should be accompanied by a definition: double marginalisation refers to a double markup whereby a supplier marks up an input, which then increases the marginal cost of the downstreamfirm and so raises the final markup of the output. In technology licensing, usually thefirst markup isfor an input that is sold using a per unit royalty. A good answer would relate the definition to a well labelled graph and illustrate the loss to both consumers and thefirms of this double markup compared to the case where no additional markup is used (ie, vertical integration).
b. [20 marks] A single (monopoly) upstream licensor proposes to supply an input under license to a single (monopoly) downstream licensee. This input has the effect of reducing the marginal cost of production of the downstream firm. The licensor proposes a contract whereby the licensee pays an up-front fixed fee but no recurrent royalty for the input. Does this solve the double marginalisation problem? Would both firms accept such a contract? Does your answer depend on whether there is full information on the quality of the input? Explain.
Sketch of answer: A purefixedfee contract would solve the problem, as no second margin is introduced. Bothfirms gain – taken together - compared to the double marginalisation case, but thefixedfee must be set so that each individualfirm does better than the alternativefor both to accept the contract. As the “pie” effectively increases in size in this scenario, such a middle road should be available to thefirms. If there is asymmetric information on the quality of the input, then of course there may be an adverse selection problem that gets in the way of market exchange. A very good answer would include a graphical explanation along with the intuition.
c. [20 marks] An alternative to the license in part (b) is to merge the two firms together in order to bring the technology transfer inside the firm. Does this solve the double marginalisation problem? Explain, relating your results to those of part (b) as appropriate.
Sketch of answer: Integration solves the problem, yes. Again, this could be illustrated in a very good answer. The intuition is that the mergedfirm effectively trades the input at marginal cost when this is an internal transaction. In addition, there normally should be no asymmetric information problem if this is an internal transfer.
4. Answer all parts (a)- (c) of this question.
a. [20 marks] Define and explain the term “technological progress” . Illustrate your definition as appropriate. How can technological progress be measured? Illustrate your answer with two examples.
Sketch of answer: Technological progress was defined in words as well as in graphical terms, using shifts in the productionfunction and including biased technological change. A good answer would review these. Technological progress is
normally measured by an increase in outputs per unit input over time, and this is the way we measured it in lecture. Students could describe technological progress in any two examples, but we reviewed details of improvements inflight speed, computational speed, watch making, and several other areas, so some precision should be expected of a good answer. A very good answer would relate the answer to growth models and discuss the way to “accountfor growth” between technological progress and other sources, such as changes in capital intensity. Empirical results in this area include notjust Solow’s basic results but also Dennison’s more granular accountingfor growth including economies of scale and improvements in human capital.
b. [20 marks, 5 per point below] Explain how economic growth can be related to technological progress. Which theoretical model best applies to each of the facts outlined in points (i) – (iv), below?
i. Real wages were at the same level in England in the year 1280 as in the early 1600s. Real wages rose between these periods, as population fell due to disease.
ii. The percentage growth in per capita GDP in the period 1950-1987 for five advanced economies approximately matched the percentage growth in technology, adjusted for a capital factor.
iii. Watch production benefitted from division of labour and economies of specialisation in England from the late 17th century until 1815.
iv. Waves of innovation have increased computational speed over the past 150 years. The identity of firms that have been technological leaders in these waves has shifted across firms.
Sketch of answer: We covered all these cases in lecture and associated them each with a different type of growth model. We associated thefirst with Malthusian growth, the second with exogenous growth models, the third with the classical model, and thefourth with Schumpeterian growth models. We presented the roles of technology in each, pointing to the underlying assumption of relatively little (compared to population increases) technological progress in the Malthusian view and the resultant “cap” on real wages; we pointed to the central role of technology in exogenous growth models, includingformalism related to balanced growth and accountingfor growth; the primacy of economies of specialisation underlying the classical approach; and the changes in technological progress expected as industry concentration and intellectual property protection change in Schumpeterian models. The underlying intuition should be givenfor each of these models. A very good answer would includeformalism in the exogenous growth model. It is unlikely that anyformalism would be presentedfor the others. A very good answer may also critique the models as policy-making tools.
(c) [10 marks] The UK Government’s Industrial Strategy document (2017) proposes to make the UK the world’s most innovative economy by raising R&D investment to 2.4% of GDP by 2027, increase the R&D tax credit to 12%, and invest £725Mm in government funding targeting three broad industrial areas.
Critique this strategy. Do you agree that these policies are likely to make the UK the world’s most innovative economy? Outline other changes you think should be considered as part of your answer.
Sketch of answer: The answer could refer back to the role of technological progress in the previous answersfor a general argument behind government intervention, and could discuss the sources of underinvestment in R&Dfrom the private market
(although with a caveat where there is competitionfor innovations). These measures are all part of the general toolkit of a government to promote technological change, but measures are not limited to these. Students could outline other possible measures that could be included as part of the mix, including improving legal protection of intellectual property (not necessarily *increasing* it), stimulation of alternative means offunding such as venture capital, or any of a number of other prize/tax credit/RJV or competition policy interventions. The list is long and need not be exhaustivefor a good answer. They need to show some breadth of knowledge, though.
In terms of critiquing this set of suggestions, they are a good start, but may not get the UK to the position that is the aspiration of the document. For example, we reviewed general levels offunding of R&D and 2.4% is modest compared to several other countries. Students could discuss this, again returning to the role of technology in growth models. Tax credits are a good idea, but are not the only possible measure. Here they could bring in alternatives. In terms offunding specific industrial areas, we discussed “picking winners” as one would have in a prize system versus allowing a “bottom up” approach as with the patent system. Picking winners depends on having those winners actually well chosen. Students could discuss this, as there are pros and cons. All that is requiredfor a good answer is an informed and intelligent discussion. No particular view is “better” as an answer.
5. Answer both parts (a) and (b) of this question.
a. [25 marks] What is meant by the term “wage polarisation”? Define skills biased technological change and explain how this concept helps your explanation.
b. [25 marks] Discuss recent evidence for wage polarisation. Have improvements in computing caused wage polarisation? Explain, outlining the arguments and evidence thoroughly.
Sketch of answer: Skills biased technological change is the case where technology tends to complement skilled workers, increasing their productivity. One could depict this on a graph, but it is not necessaryfor a good answer. A very good answer should depict it, though. If technological change is skills biased, then itfavours people with more training and experience, making them relatively more productive and so increases demandfor skilled labour. If this outstrips supply, then wagesfor this group can rise compared to wagesfor the unskilled group (again, if the demand decrease is not accompanied by a significant decrease in supply of this group). This can also tend to increase the productivity of certain sectors that are able to exploit skilled labour.
Empirical evidence: We see such a split starting in the 1980s in the US, despite a
doubling of the university educated populationfrom 1960 to 1980. Splitting by females and males we see a marked increaseforfemales in both skills classes, but a
relative stagnationfor less experienced males. Measured as a rising degree premium, this is also evident even in recent years.
Wage polarisation refers to the divergence of cumulative (log change in) weekly earnings of different income groups (high, medium, low). This is occurring even though the overall university premium is also increasing in this time. Indeed, the middle ofthe income distribution is being “hollowed out”, and the suggestion is that this is due to a high level of substitutability of computersfor middle income tasks accomplished before by (for example) clerks. Breaking down tasks into routine and non-routine; manual and non-manual, computerisation is a good substitutefor the routine tasks but actually has little effect on manual non-routine tasks. The main
substitution isfor non-manual routine tasks (like clerks). This is the root of the hollowing out. Hence, the most educated demand rises (as they perform highly non- routine tasks), the demandfor middle educatedfalls (the clerks, the demandfor the non-routine manual workers is the same as it always was pretty much butfor routine manual tasks, it alsofalls. It’s unclear ifthis is computerisation (think of manufacturing) or simple substitution of other types of capital (digging).
The estimate of the number ofjobs at risk varies a lot, so the magnitude of this effect is hard tojudge.
A good answer would give the gist of the data and show some breadth of knowledge of the sources. Progressively better answers would have more and more detail of the data.
2022-05-24