BPP Field Exam 2009 Answers

From edegan.com
Jump to navigation Jump to search

The answers on this page are provided by Ed Egan, unless otherwise noted. They are not necessarily complete.

Answers to the 2009 Field Exam

Answer A1: R&D Integration (Mowery)

Part A

a.) What factors led U.S. firms to integrate into R&D in the late 19th/early 20th centuries and how if at all did firms exploit intrafirm and external sources of technology during the 1900-1940 period?

By the 1870s railways and other infrastructure investments had lead to the rise of national markets. By 1890 less than 5% of patent assignments used partial geographic rights; national rights to a single exclusive licensee where now the norm (Lamoreaux and Sokoloff). At the same time, the cost of R&D were rising and would continue to rise. However the late 19th century was the era of the independent inventor, with inventors who held more than 10 patents accounting for more than a third of all patenting activity (Ibid). These inventors were mobile and specialized in invention, most usually selling the rights to each invention outright.

However, by the turn of the century the barrier to entry were rising, in part because formal education, particularly in the sciences and engineering, became more necessary, and formal education could practically gaurantee an inventor a position with a large firm, with easy access to capital markets, that had the resources necessary to undertake innovative activity. The number of assignees began to drop, and assignments to patents early in an inventors career became more difficult; suggesting that inventors needed to prove themselves in order to mobilize capital. Furthermore, the dominance of exclusive rights of assignment meant that inventors who wanted to commercialize their products would have to become principals of their own firms. The Supreme Court's 1904 Northern Securities decision made mergers and acquisitions difficult, so large firms could no longer buy R&D in this manner.

This lead to an increased usage and growth of in-house R&D facilities (Mowery 2009). At the same time, reforms to patent policy strengthened patent rights allowing firms to appropriate greater returns from innovations that could be patented, and there was a wave of key patents expiration for leading firms including AT&T and GE, which lead to greater competition (Mowery 2009). With patents increasingly held by firms, and a continued rise in the interdependence of patents, one might imagine that inventors were increasingly forced to commercial their products through large firms that held cospecialized assets (Teece 1986).

Human capital was one external source of technology following this period, as the independent inventor declined. The infrastructure developments had assisted in communication of ideas and labor mobility and several clusters of innovative activity developed. Cleveland may have been an early example of Silicon Valley, and the Mid-Atlantic region was particularly associated with innovative firms (Lamoreaux and Sokoloff). However, pharmaceutical firms, in particular, were located close to universities and apparently actively engaged in funding research and catching spillovers.

Part B

b.) What factors explain the asserted "externalization" of corporate R&D in the 1980s and 1990s? How do these factors operate across different industries?

In the late 1970 there was a structural change, with large corporations reducing or eliminating their in house R&D. There was a change in the interpretation of anti-trust policy which meant that vertical acquisitions were not punished. In addition there was a strengthening in patent policy, including the following changes:

1980 Diamond vs. Chakrabarty - which upheld validity of broad patenting in biotech (including life forms) 1980 Bayh-Dole Patent and Trademark Amendments Act 1982 creation of the CAFC (Court of Appeals for the Federal Circuit) - which upheld 80% of cases argued before it (as compared with 30%) 1986 Federal Transfer Act - authorized federal labs to conduct commercial research and development agreements (CRADAs) with private firms 1986 Polaroid vs. Kodak judgement awards billion dollar damages for patent infringement

By 1984 the growth rate of patents exceeded 7%. This was accompanied by a period of vertical specialization (that had begun the late 1960's in the semi-conductor industry), and was also evident in the new biotechnology industry that arose from university research in the 1970s. By the 1990's some knowledge intensive activities were commonly outsourced and there was again a rise in smaller firms doing R&D. It is noteworthy that the venture capital industry, that begun the late 1950's, finally reached maturity in the early 1980's and then rose to become synonomous with innovation financing by the 1990's. The ready access to capital for nascent start-ups, and the development, and later prevalence, of the personal computer which reduced the costs of research in many fields, both lead to a resurgence in the independent invention, and in commercialization by inventors.

Two industry of particular note where ICT (information and communications technology) and Biotechnology. In the former case, new firms rose to display incumbents, as well as filling vertical specialization roles, whereas in the later case, the new entrants did not for the most part (Chiron and Genentech being exceptions) manage to displace incumbent pharmaceutical firms. They did suceed in vertically specializing in many of the key positions in the value chain though, which suggests that barriers to entry have been and are being reduced. Another important difference is in patent policy, which has grown and strengthened in both areas. Software became patentable, as did electronically conducted business methods, to support the ICT sector, whereas life forms and other bio-inventions became patentable in biotech. University technology transfer became dominated by biotech, where the licensing of inventions was straight forward and lucrative.

Part C

c.) What indicators support or undermine the argument that U.S. firms have indeed "outsourced" their R&D and other innovation-related activities?

Indicators that support this notion include:

  • R&D dollars spent abroad by US firms has increased
  • R&D dollars spent in the US by foreign firms has increased
  • Patent inventor and assignee locations - there is a trend of broadening assignment consistent with outsourcing and globalization.
  • Particularly in biotech and IT there has been an increase in alliances between firms in the US, and with international firms
  • The US share of US filed patents is falling
  • The number of patents assigned to Universities is increasing
  • More universities began patenting
  • University industry collaboration rose (in dollar terms) dramatically

Indicators that undermine this notion:

  • The number of NCRA alliances filed peaked in the mid-1990's and has then fallen dramatically

Based on your answers to (a) - (c) develop a theoretical framework to predict the evolution of infrafirm and external R&D within the innovative strategies of large U.S. firms during the next 10 years (2008-2018).

I believe that there is strong evidence that policy materially drives innovation, with anti-trust policy, patent policy, and other R&D policy of foremost importance. I would anticipate that anti-trust policy will not materially change, and that in fact the interpretation of policy in terms of economic efficiency will become ever more sophisticated, favouring innovation enhancing strategies. Patent policy is unlikely to change. The dominant change will come through other R&D policy, particularly the spending on R&D by the federal government. With a period of prolonged recession likely, spending on R&D by government will rise relative to spending on R&D by private firms, and capture the spillovers of this spending, perhaps through increased alliances with universities and government labs, perhaps through private research conducted with federal investment, will become increasingly more important. Obama's campaign included a promise of a $10b green technology investment fund, to be operated along the principals of venture capital. This policy would not only fund research, but would do so in the most effective (to date) method of forcing commercialization. Firms will respond by spinning out green ventures, and by buying up promising green companies. This will be particularly prevalent in the oil industry, as both the supply of oil dwinddles and as oil dependence (because of foreign policy and environmental policy) becomes ever less attractive.


The answer to the last part should have included a discussion of ICT and biotech. Biotech seems likely to continue to growth at an increasing rate. With the barriers to entry to the industry reduced, we should also see a number of new entrant suceeding in becoming incumbents. University output of biotech invention shows no sign of slowing in the short term to medium term, and universities have the cospecialized assets for research, which will only grow in importance. In ICT, it is possible that the anti-commons effect will start to become material. In this sector, open source was born and continues to grow. Furthermore, as closed source technology becomes ever more dependent on other closed source technology, and the tangled web of agreements needed for development increases and is ever more widely held (causing hold-up problems as in Ziedonis (2004)), we might expect open source to become ever more attractive, and for ICT firms to specialize further - that is we will continue to see a increased seperation of hardware and software developement, and even within software, development on top of open source platforms will increase.


Mowery, D.C. (2009), "Plus ca change: Industrial R&D in the Third Industrial Revolution", forthcoming, Industrial and Corporate Change.

Lamoreaux, N. and K. Sokoloff (2005), "Decline of the Independent Inventor: A Schumpeterian Story", NBER working paper 11654

Teece, D.J. (1986), "Profiting from technological innovation: Implications for integration, collaboration, licensing, and public policy," Research Policy.

Ziedonis, R.H. (2004), "Don't fence me in: Fragmented markets for technology and the patent acquisition strategies of firms", Management Science.

Answer A2: Competition, Innovation and Antitrust (Dal Bo)

Part I

Whether the innovation is cheap or not is not first order germane. The important question is the appropriability regime. I will assume that the cheap innovation (making the phone red) is appropriable (i.e. there are no property rights associated with the colour red), and that the battery life innovation is patentable and that strong (perfect) IPR protection exists.

Then the red phone would indicate an introduction of differing qualities that are imitable at some (possibly zero) cost. This gives a Hotelling (1929) or Salop (1979) like model, assuming that customers have some preferences over quality. Perhaps a line from White to Red could indicate the distribution of customer preferences, the cost of innovation could be the cost of entry on the line, and the transport cost could represent the lost utility of a customer away from thier ideal colour. We begin the game with two firms situated on White, one of which gets to 'spawn' another firm that locates on Red. The White-only competitor may spawn an entrant or other new entrants may enter the market.

Absent new entry, if there was only one firm at White and one at Red we would expect them to price to compete against each other exactly in the Hotelling model (given these location choices). However, the Bertrand competition at the White end of the line reduces price to marginal cost, and only customer whose valuation of a White above marginal cost can buy from them. This gives the player at Red, potentially at least, the rest of the market without competiting with players at White - the customer whose valuation of White is marginal cost would have to have a valuation of Red greater than or equal to marginal cost for the entire market to be served. However, as the Red player is a monopolist over this area he would rather price above marginal cost and earn positive profits which he does.

As the cost of entry is extremely low, I will assume it to be effectively zero. Then firms will enter and while there are a finite number of firm they locate so that they have space around themselves on the line, charge prices above marginal cost and reap profits above zero. Once an infinite number of firms have entered the line will be saturated all firms will charge marginal cost and earn zero profits.

Given that the battery innovation has perfect property right protection around it (for 20 years) by assumption, we expect the that firm will have a monopoly on it. Thus, if there is no effect between the markets, Bertrand competition will continue for the White phones which will be priced at marginal cost and earn their firms zero profits, and the battery life phones will be monopoly priced and earn monopoly profits. If there is an effect between the markets, that is because of a taste for variety a change in the Battery phone's price will have an effect on not only the demand for Battery phones but also for White phones, then we have a Dixit-Stiglitz (1977) type model, except that we do not have a free entry assumption. With free entry firms would earn zero profits, but with costly entry (i.e. the cost of coming up with a comparable innovation to the Battery innovation), the Battery firm will earn positive profits. However, because of the demand effect, the battery firm will not earn monopoly profits, but something less. With free entry the number of firms will increase but remain finite in this model - firms enter until the customer's taste for variety is satiated. Likewise with costly entry but excess taste for variety we could expect more entry and more positive profits (to cover the cost of entry) until satiation.

Part II

Shaked and Sutton (1982) provide an explanation for an observation of a duopolist, in conjunction with both positive profits and zero entry costs. In their model there is price competition, but they provide grounds for an objection to the investigation. In this model the consumers taste for product variation allows two different quality entrants to enter the market, compete on prices, and earn positive profits. The entry of a third firm (if the parameters are as described in their model) would reduce profits to zero for all firms, and reduce welfare - as customers would then only have one choice of product when they have a taste for variety. Thus the grounds would have to that the firms are producing different qualities, and that the higher quality firm is making the greater profit - this would be entire consistent with the model. For different parameters more than two firms could be in the market even and even with a free entry condition could be earning positive profits - in this case the profits should still be quality ordered. Profits here do not indicate anticompetitive behaviour or barriers to entry - they indicate vertical differentiation. In fact zero profits would indicate too much competition, and too little provision of variety.


In Part I the Red phone firm would not have wanted to give up its operations at the White location (and so allow the White phone firm to earn positive profits) if there were economies of scale to manufacturing so that the marginal cost would be lower when keeping White. This might also allow it to displace the White player entirely, or at least earn positive profits from the White location.

The last answer (Part II) might also have considered price discrimination more generally - price discrimination is not inherently bad (actually it can be welfare improving) and may lead to positive profits.

After consulting Ernesto about this answer the following corrections should be made:

  • The battery life extension (from 4hr to 6hr) was supposed to be a vertical differentiation question. The application of a horizontal model would be wrong. There was ordering of quality.
  • The Red/White problem should have assumed that the innovator made their phone red (i.e. did not keep white).
    • It is not clear that given the option to keep white would be exercised even if there where economies to scope. The economies would have to be sufficiently large to overcome the problem of competiting with yourself!
  • The horizontal model should have used Hotelling for prices and profits, and the Salop for entry - Salop exists to study entry in Hotelling models.
  • The final part could also have been argued using a horizontal differentiation model
    • Differences in quality (or seperation in distance) would lead to positive profits and prices above marginal cost.
    • Differentiation may improve social welfare as in Dixit-Stiglitz where customers have a taste for variety.
    • Anti-trust with a duopoly would require collusion of some sort.


  1. Dixit, A. and J. Stiglitz (1977), "Monopolistic competition and optimum product diversity", American Economic Review 67, 297-308. pdf (Class Slides)
  2. Hotelling, H. (1929), "Stability in competition", Economic Journal 39, 41-57. pdf
  3. Salop (1979), "Monopolistic competition with outside goods", Bell Journal of Economics 10, 141-156. pdf
  4. Shaked, A. and J. Sutton (1982), "Relaxing price competition through product differentiation", Review of Economic Studies 49, 3-13. pdf

Answer B2: Opportunistic sellers

Suggested answer here


Answer C1: Bargaining over Power

Proposed solution here

Answer D: Soft Drink Organization (Spiller)

Bottlers have to make relational specific investments under incomplete contracts, which results in a classic hold-up problem (Williamson 1971). Contracts are inherently incomplete because they were granted perpetually (to which I will return) and market conditions, over the lifetime of the assets needed for effective production and distribution, are impossible to forecast. The cost of writing a contingent claims (i.e. Arrow-Debreu) contract would be prohibitively high. Relational specific investments were also inherently necessary: The bottler needed specialized equipment to manufacture the product that would have little value outside of the relationship, and would be investing in brand-specific capital that was only applicable to the CM. Likewise, the CM was dependent on the bottler for production and distribution. Brand capital is long lived, perhaps never ending, and specific plant investments might have (say) 20 year depreciation periods.

Thus given such long-term relational specific investments and incomplete contracts one might expect vertical integration. However, this was not optimal for the CMs. In the early 20th century, markets were local, and bottles were returnable. Returnable bottles suggests local manufacturing to lower transportation costs, and local markets suggest that inventive problems that would arise from a lack of automous adaptation (under vertical integration) would be more than offset by the gains from cooperative adaptation (Williamson 1991). With local markets there is little need to cooperate on branding across markets. Thus a hybrid form of governance is suggested - and essentially this is what arose: bottlers acted as local franchisees.

Long term investments in relational specific investments suggest long term contracts with renegotiation. Perpetual contracts are much stronger and might have been implemented for a number of reasons. Firstly, the investment in brand capital might result in a perpetual asset, and the duration of the asset specificity should be matched to the duration of the contract. Without perpetual contracts, the CM could expropriate the gains of the bottler by entering (or even threatening to enter) the market at the renegotiation stage. The perpetual brand capital is a sunk investment for the bottler with no salvage value; an appropriable quasi rent that can be taken by the CM (as per Crawford, Klein and Alchian 1978). Secondly, renegotiation places a cost burden on the bottler. With a large number of bottlers we can assume near competive conditions. Aside from the fact that at the renegotiation stage, a new potential bottler could context the market and expropriate the existing brand capital, the legal cost burden would place a strain on a competitive bottler, and the incurrance of such a cost burden would be inefficient if the contract was stable anyway. Historically, the seperation of the roles of the bottlers and the CM was clearly demarked and both parties had their own incentives to live up to their responsibilities within the bounds of the contract making renegotiation unnecessary. The stablity of the contract was due to the fact that major disruptions that would require renegotiation were very rare.

The granting of exclusive bottling and distribution territories followed naturally from the localness of the markets. Each bottler conducted local campaigns that would not have large effects on neighboring territories, that is there were little spillovers, and the requirement for the return of the bottles meant local distribution and collection was efficient. There was little need for coordination across territories and the bottler's major threat was of expropriation of relational specific investments within their territory, either from the CM who might set up a captive bottler, or new entrants. At the time economies of scale were met locally, and economies of scope allowed bottlers to sell other, not directly competing, products to their local markets.

To test this emprically, we should note that Coca-cola's fountain drinks were always provided by the CM and not the independent bottlers. In the early 20th century, we would expect this to be inefficient, and so that Coca-cola's fountain drinks will be higher priced (and less profitable) than Pepsi's, which were provided by independent bottlers. Thus the unit of observation should be local markets, and a dummy for Coke versus Pepsi should be positive on the price of fountain drinks.

While marketing was predominantly local it was efficient to have market coordination done by the bottlers. However, because marketing established a relational specific asset (i.e. the brand), it is not transferable to another purpose and so financial support from the CM was necessary for the bottlers to undertaken efficient investment. Furthermore, as the industry and advertising as a whole developed, the degree of spillovers from one bottler to another increased. Thus when advertising in a medium that extended beyond their territory, a bottler would pose a positive externality on adjacent bottlers, which would encourage freeriding. As the private benefits of marketing to the bottlers are less than the private benefits to the CM (who captures the full benefit including the externality), the CM should subsidize marketing efforts. Beyond a certain point, however, the advertising became increasingly national, which not only reduced the benefits to local distributors from the same expense, but also required coordination on timing and execution to conduct a successful campaign. Therefore more coordination was needed, and to begin with this could be accomodated within the franchisee arrangement, but later it began to favour vertical integration (covered below).

Consider, for example the CM's need to conduct a national campaign with independent bottlers, or to introduce a new product. The bottlers would not only be hard to coordinate, each requiring seperate negotations, but each bottler is capable of hold up. It may not be clear to a bottler that a campaign will have positive advantages (many innovation, including marketing ones) fail, and new products may not fair well in their market. The bottler would prefer to wait until the campaign or product has been proven, so the CM needs at least some bottlers that it can order (by fiat) to introduce the campaign. Furthermore, a bottler with an exclusive territory may hold out against implementing a campaign, reducing its effectiveness. Centralized bottling facilities that are vertically integrated into the CM become increasingly favoured.

Overall, my arguments above suggest that there should have been a move towards more captive distribution over the last century. As national marketing grows, as there is less need for return of bottles, as there are greater economies of scale and scope to be had by centralized distribution, we would anticipate that first order economizing would predict a change in governance structure from franchising to vertical integration, with the vertical integration better able to accomodate the cooperative adaptations that arise from national marketing campaigns and the introduction of new products, and at the same time as the need for autonomous adaptation, such as from adaptation to local conditions, to decline. To support the later argument, it should be noted that contemporaneously with the growth of national marketing channels, there was a rise of national retailers, particularly supermarket chains and fast food firms that are the major vendors. These chains used warehousing and their own distribution networks to serve their establishments, and conducted their own national (cooperative adaptation relavent) marketing campaigns. It would be more efficient for a CM to negotiate with these retailers centrally, and distribute the product as the retailer wishes than for the individual bottlers to negotiate. Apart from the savings on transaction costs, there is bargaining power to be considered. Furthermore, even if the CM negotiates, if it then has to negotiate again with its own bottlers, there are more transaction costs with the potential for hold up. Thus again, a vertically intergrated CM is favoured.

With first order economizing we should expect gains in efficiency that translate into lower prices and higher profits. Therefore I propose a empirical test along these line. The unit of observation would a sales at a retail branch at a point in time, and we would expect that a coefficient on whether that retail branch was supplied by a vertically integrated bottler to follow to be positive following the structural changes in the market (and negative before the change). To simplify the analysis we could aggregate retail branches.