Strategic Management Journal Strat. Mgmt. J., 30: 323–347 (2009) Published online 11 December 2008 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.736 Received 8 May 2008; Final revision received 14 June 2008

BIG LOSSES IN ECOSYSTEM NICHES: HOW CORE FIRM DECISIONS DRIVE COMPLEMENTARY PRODUCT SHAKEOUTS LAMAR PIERCE* Olin Business School, Washington University in St. Louis, St. Louis, Missouri, U.S.A.

This study examines shakeouts in the context of business ecosystems. Market turbulence generated by core firm decisions in competing differentiated ecosystems can generate financial losses and exit for complementary niche market firms. I develop hypotheses predicting which niche markets will suffer larger losses and be more susceptible to shakeouts, and how core firm decisions will drive complementor performance and survival. I then apply these hypotheses to brand-based differentiated ecosystems in the automotive industry, where networks of suppliers, customers, and complementors surround car manufacturers. More specifically, I study the complementary niche market of automotive leasing, where manufacturers sway leasing markets through product change, entry, and subsidization. To test the hypotheses, I use a proprietary dataset of 200,000 individual car leases between 1997–2002 to identify how manufacturer product design and niche market entry drive complementor losses and exit. These data allow a unique opportunity to understand how the strategic choices of core firms can have substantial and often devastating effects on niche markets in their ecosystem. Further, the results suggest how the dynamic capabilities to adapt to core firm behavior might improve performance for certain niche market complementors. Copyright  2008 John Wiley & Sons, Ltd.

INTRODUCTION The dynamics and sources of industry shakeouts have long interested scholars of organizations and markets. Such shakeouts, where the majority of an industry’s firms exit or consolidate, have implications for market structure, industry evolution, and firm strategy. Organizational ecologists have typically modeled shakeouts as a systematic process whereby inflexible firms are selected from an evolving competitive environment. This work has focused primarily on how Keywords: business ecosystems; dynamic capabilities; leasing; automotive industry; forecasting *Correspondence to: Lamar Pierce, Olin Business School, Campus Box 1133, Washington University in St. Louis, 1 Brookings Drive, St. Louis, MO 63130, U.S.A. E-mail: [email protected]

Copyright  2008 John Wiley & Sons, Ltd.

competition and legitimation structure the relationship between population density and firm failure (Hannan and Carroll, 1992). Other sociologists have explored how a firm’s strategic choices can improve its own survival in these competitive environments (Romanelli, 1989; Sorenson, 2000). Work from the economics and management literatures demonstrates how pre-founding experience of new entrants can determine their performance (Helfat and Lieberman, 2002) and survival (Franco and Filson, 2000; Klepper, 2002; Phillips, 2002; Agarwal et al., 2004; Klepper and Sleeper, 2005). This work has largely focused on the role of technological change in driving shakeouts (Nelson and Winter, 1978; Metcalfe and Gibbons, 1989; Dasgupta and Stiglitz, 1988; Sutton, 1998), observing firm exit and consolidation broadly across industries and products (Gort and Klepper, 1982;

324

L. Pierce

Klepper and Graddy, 1990; Klepper and Simons, 2000). This broad body of work highlights the importance of environmental and firm characteristics in shakeouts, but it has left an important research area unexplored. The literature has not empirically examined how decisions by one large firm can drive financial losses and shakeouts in complementary markets even when technological change is minimal. The strategy and economics literature certainly acknowledges the interdependencies of firms in complementary markets (Katz and Shapiro, 1985; Farrell and Saloner, 1986; Henderson and Clark, 1990; Economides and Salop, 1992), and the organizational ecology literature has recently explored how niche firm survival can be influenced by an entire market (Dobrev, Kim, and Carroll, 2002, 2003). Similarly, Teece (1986) and Tripsas (1997) established the importance of complementary assets in driving performance in the face of radical technological change. Yet there is little evidence as to the impact of decisions by one core firm on the losses and survival of a network of its complementors, or complementary service providers.1 This study identifies how the entry and pricing decisions of one core firm can drive exit by complementary firms in a business ecosystem. The importance of complementary market shakeouts is demonstrated in Teece’s (2007) conceptualization of business ecosystems, where technological, product, and strategic changes by one enterprise have widespread implications for firm performance and survival. Large networks of firms can revolve around corporations, both through formal contracting and symbiotic relationships. In this framework, preceded by Rosenbloom and Christensen’s (1994) value network and work on organizational ecology,2 a core enterprise’s survival and performance depend heavily on ‘the community of organizations, institutions, and individuals that impact the enterprise and the enterprise’s customers and supplies’ (Teece, 2007: 1325). These core firms may control the technological architecture or brand that drives value in the ecosystem, but their long-term 1 The term ‘complementor,’ from Brandenburger and Nalebuff (1996), refers to firms producing complementary products or services. 2 See Amburgey and Rao (1996) for a review of early work in organizational ecology. It is also important to note that the concept of ‘business ecosystems’ was used previously by Moore (1993) and Iansiti and Levien (2004a, 2004b) in the managerially focused business literature.

Copyright  2008 John Wiley & Sons, Ltd.

performance nevertheless depends on the support of suppliers, retailers, and complementary goods and services. Similar to Albornoz and Yoguel’s (2004) characterization of core firms, these enterprises serve to coordinate suppliers and complementors through compatibility and direct network management. These networks include both suppliers and customers in the value chain, as well as vast numbers of complementors providing complementary goods and services. These firms fill niche markets within the ecosystem, which represent specialized functions tied to the core firm.3 Strategic and ad hoc cooperation, compatibility, and network externalities in the ecosystem tie the health and performance of many niche markets together. The decisions of the ecosystem’s core firm, as well as exogenous economic and technological shocks, can have wide-ranging implication throughout the network of firms (Garud and Kumaraswamy, 1993). Businesses occupying niche markets within the ecosystem must react to a dynamic landscape driven by factors beyond their control. These factors can include core firm adjustments in pricing, design, technology, market penetration, and marketing. Adaptation is critical to the performance and survival of firms in niche markets throughout the ecosystem, as they must react to the often highly idiosyncratic moves of core firm strategies. Successful or failed adaptation will be based in the know-how and routines of the organization, many of which are deeply rooted in the firm’s history (Nelson and Winter, 1982; Tushman and Anderson, 1986; Henderson and Clark, 1990). These routines are the critical basis for dynamic capabilities, the capacity for a firm to create, recognize, and exploit emerging opportunities and threats through the reconfiguration of competences (Teece, Pisano, and Shuen, 1997; Eisenhardt and Martin, 2000; Helfat et al., 2007). When the core market is highly differentiated or dominated by just a few firms, niche market firms must develop dynamic capabilities so as to identify the need for change as well as accomplish it. The frequency and severity of product changes by the core firm increase risks for complementors through reduced compatibility, which reduces the 3 Greve (2000) has an extensive discussion on the literature on market niches. It is important to note, however, that firms in niche markets need not be small, as they may be large multiproduct firms differentiated in many niche markets.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches value of the complementary product to the customer. Investments by complementors may also lose value due to their reduced importance to the altered product. When these changes are widespread or sufficiently severe, losses by complementors will result in shakeouts in sections of the ecosystem. Furthermore, competition by core firms across ecosystems may lead to much broader shakeouts. A niche firm that does not understand the widespread implications of core firm actions will be susceptible to frequent changes in that firm’s activities. This study develops hypotheses predicting which complementary niche markets will be more susceptible to shakeouts. I then apply these hypotheses to brand-based ecosystems in the automotive industry, where networks of suppliers, customers, and complementors surround large car manufacturers. More specifically, I apply this framework to the niche markets of automotive leasing, where manufacturers shock complementary leasing markets through product change, entry and exit, and subsidization.4 This setting is unique and important in that the actions of car manufacturers have had dramatic effects on losses and market exit for equally large financial institutions. In 1999, these lessors lost $2,600 on the average lease returned vehicle, with industry-wide losses from residual values reaching $11 billion in 2000 and $10 billion in 2001 (Rauschenberg, 2001). To test the hypotheses, I use a proprietary dataset of nearly 200,000 individual car leases for the period 1997–2002. These data, when combined with aggregate sales, redesign, durability, and financial institution data, provide a detailed description of the leases written by lessors and the markets in which they competed. The lease transaction data are paired with used car price data, allowing direct measurement of performance on nearly 200,000 lease contracts by comparing residual value forecasts with realized lease-end car values. These data provide a unique opportunity to understand how the decisions of core firms can have substantial and often devastating effects on the financial performance and survival of niche market complementors. I find evidence that the decisions of manufacturers have effects on the exit and performance 4 Note that my conceptualization of a niche market in a brandbased ecosystem is different from the representation of niche automotive manufacturers in Dobrev et al. (2002, 2003).

Copyright  2008 John Wiley & Sons, Ltd.

325

of lessors in model-specific niche markets. More specifically, product redesign severity by manufacturers increases hazards for other lessors and reduces their forecasting performance. I also find that entry by manufacturers into these niche markets has a similar deleterious effect, particularly when the manufacturers are subsidizing leases. The findings suggest that the actions of core firms in business ecosystems can have widespread and severe effects on complementors, and that monitoring and understanding the actions of these core firms must be of primary importance to managers of firms in niche markets. These results also suggest that the key to survival and performance in many ecosystems may be constant adaptation to core firms, and that the importance of dynamic capabilities is not limited to the sphere of technological innovation.

THEORY AND HYPOTHESES Complementary niche market shakeouts in the ecosystem framework A traditional industry or market may contain numerous overlapping ecosystems revolving around core firms. The concept of business ecosystems is based in the organizational ecology literature, which studies the financial health of communities of interrelated firms that collectively form ecosystems (Hannan and Freeman, 1977, 1989; Carroll and Hannan, 2000). While individual actors in this ecosystem may seek personal gain, the competitive environment selects out those firms that hurt the community’s economic health. Adaptation for such firms is severely constrained and rare (Carroll, 1988) due to structural inertia, and firm survival is based primarily on how well firm characteristics match the changing competitive environment (Freeman and Hannan, 1983). In the strategy literature, these ecosystems typically revolve around core firms that define a technological architecture, as Rosenbloom and Christensen (1994) and Teece (2007) emphasize, but they also may be based on brand, geography, or product characteristics. Suppliers, customers, and complementors fill niche markets, similar to Hannan and Freeman’s (1977) characterization of market niches that support one form of organization. As Greve (2000) and others explain, while niches may in actuality exist in continuous space, they Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

326

L. Pierce

are best defined in discrete units, whether by geography, products, or services (Freeman and Lomi, 1994; Sorenson and Audia, 2000). Unlike the organizational ecology, the strategy literature emphasizes organizational adaptation as the critical component for firm survival, with some firms holding the operational or dynamic capabilities to overcome organizational inertia (Collis, 1994; Eisenhardt and Martin, 2000; Helfat and Peteraf, 2003; Helfat et al., 2007). High-level operational capabilities may allow adaptation of a specific task or output (Winter, 2000, 2003) while dynamic capabilities coordinate change across the organization (Teece et al., 1997).5 Multiple ecosystems may compete for the same customers, with each core firm differentiating their products or services. This differentiation generates incompatibility across ecosystems by making components and complements cospecialized, with each ecosystem having unique suppliers and customers as well as complementors specialized to the core firm (Teece, 1986). But they will also share many members, using the same retail outlets, components,

and complementary services. Figure 1 provides a representation of two competing ecosystems in a differentiated market. Similar to both the organizational ecology and dynamic capabilities literature, the risk to niche market firms in business ecosystems is primarily based on levels of turbulence and change. This change may be exogenous—from regulatory, technological, or other sources. But it may also be motivated by the actions of the core firm, whose pricing, product design, and niche entry decisions can greatly impact niche firm viability. Core firms may provide little consideration for niche market companies when making decisions. Their decisions, such as vertical or horizontal integration, product changes, pricing, or innovation, may include minimal concern for ecosystem-wide stability and profitability, as the core firm focuses on competition with firms in rival ecosystems. Competition with rival ecosystems may drive core firm strategies, whose effects will spill over to niche market firms in each ecosystem. Suppliers, buyers, and complementors must react in order to survive these ripple effects, and since some firms will fill niches in multiple ecosystems, they must monitor the actions of several core firms. If a core firm

5 See Winter (2003) or Helfat and Peteraf (2003) for extensive discussion of operational and dynamic capabilities.

Ecosystem

Suppliers

Niche

Niche

Ecosystem

Niche

Niche

Niche

Niche

Suppliers

Niche

Niche

Core firm

Complementary markets

Core firm

Niche Niche

Buyers

Niche

Figure 1. Copyright  2008 John Wiley & Sons, Ltd.

Niche

Niche

Niche

Buyers

Competing differentiated ecosystems Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches frequently generates serious hazards for firms in niche markets, niche market firms can at some cost reallocate resources and focus into other related ecosystems. A supplier, customer, or complementor may shift between differentiated ecosystems based on expected profitability, so long as the sunk costs from investments cospecialized to the core firm are not prohibitive. We should expect both the frequency and magnitude of strategic change by the core firm to directly influence niche market performance and survival. One important dimension for this change is product design, where uncertainty about core product characteristics, popularity, obsolescence, or deterioration creates hazards for complementors. Products whose popularity wanes quickly will reduce the future value of complementary products, and will drive customers toward competing products. This popularity loss will spill over to the performance and survival of complementary product and service providers unless their products are transferable or compatible with the replacement product. Technological obsolescence will also necessitate the consumer’s replacement of the core product, again posing a risk for any complementor who has made investments specific to the initial core product (Tripsas, 1997). As core firms alter product design and technology, those niche market firms unable to adapt their own activities will suffer financially and potentially exit the ecosystem, a pattern potentially exacerbated by herd-like mimetic exit (DiMaggio and Powell, 1983; Haunschild, 1994). Consequently, these core firm decisions may result in financial losses and widespread niche market shakeouts. Hypothesis 1a: A higher frequency of product design change by core firms will reduce the financial performance and increase the hazard of exit for firms in complementary niche markets. Hypothesis 1b: More radical product design changes by core firms will reduce financial performance and increase the hazard of exit for firms in complementary niche markets. Similar to product redesign, low durability in core firm products will yield hazards to firms in complementary niche markets. Durable core products ensure a longer life with customers and generate a lengthy stream of payments to complementors. Low-durability core products instead generate Copyright  2008 John Wiley & Sons, Ltd.

327

more rapid replacement and uncertainty for complementors. Low durability may stem from lack of design or manufacturing capabilities (Handfield et al., 1999), from strategic cost reduction, or from planned obsolescence (Bulow, 1986; Waldman, 1993). Manufacturers may choose to produce low-quality products to save costs, or they may intentionally sabotage the product’s durability to ensure replacement purchases. Hypothesis 1c: Lower durability core products will reduce financial performance and increase the hazard of exit for firms in complementary niche markets. Another shakeout risk stems from core firm entry into niche markets. As core firms integrate into complementary markets, their involvement competes with existing firms (Economides and Salop, 1992; Heeb, 2003). This effect can be severe when the core firm accrues benefits from this integration in its competition with other ecosystems. Such benefits could come from vertical externalities (Tirole, 1988), vertical foreclosure (Aghion and Bolton, 1987), knowledge creation and transfer (Nickerson and Zenger, 2004), lowering transaction costs (Williamson, 1985), technology appropriability (Teece, 1986), or reducing moral hazard (Alchian and Demsetz, 1972). These benefits may create incentives for the core firm to price the complementary product or service below a complementor’s cost, making the latter uncompetitive. Complementors must thus compete with a core firm that intentionally loses money in the niche market in order to profit from core products. When a core firm identifies a complementary product or service as strategically important, its entry will crowd out existing niche market firms. Hypothesis 2a: Increased penetration by core firms into complementary products or services will reduce financial performance and increase the hazard of exit for complementors. Hypothesis 2b: A core firm pricing a complementary good or service below cost will reduce financial performance and increase the hazard of exit for complementors. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

328

L. Pierce

RESEARCH SETTING This study examines complementor shakeouts in differentiated consumer automobile ecosystems. In the automotive industry, ecosystems revolve around specific brands and models, where thousands of firms may be involved in the manufacturing, marketing, distribution, and sales of a single product. Numerous complementary niche markets revolve around car makes and models, including financing, insurance, fuel, after-market parts, service, media, and others. Leasing, where the lessee pays the lessor for the right to use the vehicle, is an important complementary niche in automotive ecosystems where core manufacturing firms compete directly with independent lessors through so-called ‘captive lessor’ subsidiaries. Leases typically originate at car dealerships, where a lessor might offer customers leases of term length T months for a given car model.6 These leases, by federal law,7 must include explicit values for the contracted residual value r, the money factor f ,8 the capitalized cost c, and the monthly payment p. The residual value r reflects the lessor’s forecast for the lease-end resale value of the car net of transaction and refurbishment costs.9 The capitalized cost c is the new vehicle price negotiated between the dealer and the customer.10 The money factor f reflects the lessors calculations of lessee default risk, with high risk 6 Examples of car models include the Honda Civic, Ford Focus, and Lexus ES300. 7 Since January, 1998, the Federal Consumer Leasing Act, Regulation M, has required the following specific information in each lease contract: the capitalized cost (or new car price), residual value, lease charges, monthly payments, and finance charges. It has also required that this information be uniformly displayed. 8 The money factor is derived from the annual percentage rate (APR) of interest and reflects default risk. The money factor calculated as APR × (1/2) × (1/12) × (1/100) or APR/2400, where 1/2 averages c and r, 1/12 transforms annual to monthly interest, and 1/100 presents interest rates as actual fractional values. 9 The lessor typically also charges an origination fee and will often charge a lease-end fee on returned vehicles to cover refurbishment and transaction costs for repossession. 10 The new vehicle price directly influences the monthly lease payment, but is negotiated by the dealership instead of the lessor. While the dealership and customer may negotiate over monthly payment, the dealer’s inability to adjust residual value and lower the money factor make this negotiation inherently one on the capitalized cost. Dealers can often mark up the money factor, but cannot decrease it below the buyrate, or the interest received by the lessor. While many customers may negotiate on monthly payment, dealers are effectively negotiating on capitalized cost and money factor markups. For a detailed discussion of car price negotiations, see Morton, Zettelmeyer, and Silva-Risso (2001).

Copyright  2008 John Wiley & Sons, Ltd.

lessees receiving higher money factors, but dealers may also attempt to mark up this term for profit. A consumer wishing to lease this vehicle must pay the manufacturer p at the beginning of each of T months. After T months, the consumer lessee can either return the vehicle to the manufacturer, or purchase the vehicle for the contracted residual value r. Under these conditions, the lessor writes the contract where the monthly lease price is: p=

(c − r) + (c + r)f T

For a three-year lease on a $30,000 vehicle with a residual value of $10,000, the consumer must pay monthly installations on two components: 1. the $20,000 depreciation, and 2. three years of interest on the average value of the car during the term of the lease (calculated as the average of the $30,000 capitalized cost and the $10,000 residual value.11 A consumer signing such a contract would pay the depreciation and interest over the length of the lease term, and then typically return the vehicle to the lessor. The consumer also will typically have the option to purchase the vehicle at lease-end for a price equal to the contracted residual value.12 As both c and f increase, so does the monthly payment p. But as r increases, the monthly payment falls. Lessors who forecast residual values that are too high will have attractively low monthly payments, but will also lose money when they sell lease returns at lease-end. Lessors who forecast low residual values may write potentially profitable leases, but may lose customers due to uncompetitive high lease prices. Leases have lower monthly payments than loans,13 which is one reason why leasing expanded rapidly from seven percent in 1990 to 31 percent in 11 For more information on car lease structures, see the Federal Reserve Board’s guide to vehicle leasing at: http://www. federalreserve.gov/pubs/leasing/. 12 Although Johnson and Waldman (2003) suggest that setting a high buyback price can reduce adverse selection, no evidence of this practice yet exists. Instead, nearly all car leases use the residual value r as the buyback price. This could be done for two primary reasons. First, a buyback price below r could qualify the lease as ‘subvented,’ or subsidized, and would require reporting under FASB 11 accounting rules. Buyback prices above r would lower the value of the lease to consumers (by reducing option value) and are likely competed away by creating a perception of overpricing depreciation. Instead, lessors not wanting lease returns will often offer lessees favorable financing or discounts at lease-end in order to ex post lower the buyback price. 13 This is because leases do not build equity, while loans do.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches 1998.14 By 1999 and 2000, however, it was clear that many lessors had severely overestimated the lease-end residual values on most of their leases. Consumers were returning leased cars worth thousands of dollars less than original estimates, and while captive lessors were able to justify many of these losses through increased sales and repeat customers, many independent lessors were forced to acknowledge that they were losing hundreds of millions of dollars on consumer car leases. These widespread losses led to a major shakeout among banks and other independent lessors,15 with overall leasing falling to less than 20 percent of all new vehicle sales by 2003.16 This shakeout, preceded by a similar thinning in the late 1980s (Hobson, 2004), saw independent lessor share of all new cars fall from 18 percent in 1998 to only 10 percent in 2003, after growing dramatically from four percent in 1990. This shakeout, represented in Chart 1, provides the empirical setting for this study. The car leasing shakeout began in 1999, when First Union, a major East Coast lessor, stopped issuing consumer leases. Soon afterward, in 2000, GE Capital announced its exit from the market. GE Capital, with over 150,000 contracts, cited residual losses as a reason for its exit, and was soon followed by major portfolio reductions by Wells Fargo, Chase, Bank One, and Bank of America (Salinas, 2000). As Randall McCathren, president 14

Data from CNW Marketing/Research (http://cnwmr.com/). While many differentiate banks from independent lessors, this study will refer to all lessors not owned by car manufacturers as ‘independent.’ 16 Data from CNW Marketing/Research (http://cnwmr.com/).

329

of Bank Lease Consultants, explained, the banks behaved like lemmings. ‘Everybody sees GE Capital pull out and they say, “what are we doing in this business? If Jack Welch can’t make it, how can we?” So, they all pull out.’ (Salinas, 2000) Other banks exited as well, including National City Bank, Wachovia, Superior Bank, and Union Bank of California. The shakeout was widely attributed to residual value losses, with Bank of America and Bank One charging $257 million and $518 million respectively to their earnings before exiting the industry (Cordle, 2001). While captive lessors also scaled back leasing in response to residual value losses, this reduction was much less than independent lessors. So how could so many lessors perform poorly in forecasting depreciation on a common product with such rich historical price data? The answer lies in the role of the car manufacturers at the core of the model-based ecosystems. Manufacturers initiated much of leasing’s growth in the early 1990s, raising consumer awareness of car leasing from 22 percent of new car buyers in October 1990 to 72 percent in October 1992.17 The automobile leasing industry comprised major captives, pure lessors, and banks, all competing for customers through dealerships. The dominant industry lessors were the manufacturer subsidiaries, which held a 48 percent leasing market share in 2003.18 The captives gained much of their market share through subvention, where their

15

17 18

CNW Marketing/Research. Leasetrak 1995. Data from CNW Marketing/Research.

Chart 1: Independent lessor share of all new car sales

Percentage of new car market

20.0% 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Year

Copyright  2008 John Wiley & Sons, Ltd.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

330

L. Pierce

manufacturer parents subsidized leases to support low-demand models.19 Manufacturers support lowdemand models through multiple levers, including cash rebates and subsidized loan financing, but their ability to directly cut price is highly limited.20 Manufacturers typically offer customers a menu of formal incentives to provide options with different cash flow and tax implications. Under the Financial Accounting Standards Board (FASB) Rule 13, manufacturers should account for all subsidized operating leases and lending in financial statements.21 But lease subsidies can be attractive to managers seeking short-term boosts to sales and earnings, since their true cost can be hidden until lease-end, often four or five years after origination. Despite FASB regulations, it is nearly impossible to distinguish between ‘subsidization’ and ‘overestimation’ due to high depreciation uncertainty, which allows current managers to pass on costs to future managers and shareholders. If lessors had perfect information about the future of the vehicle, then we would expect the total revenue to cover vehicle depreciation. An independent lessor, such as a bank, would seek to price a lease such that it accounts for the risks inherent in the contract, so that the terms written by a lessor reflect the risks borne from the consumer and the specific vehicle. Yet residual value is often highly unpredictable, due to uncertainty about future market conditions, which are affected by general economic conditions, competition, customer preferences, new product pricing, and innovation and new product development. In the automotive industry, manufacturers, financial institutions, and consumers are all highly uncertain about the future value of a newly manufactured car. Automobile lessors take on the risk of obsolescence from consumers through closed-end leases, where the lessee can return the vehicle at lease-end without any responsibility for the residual value agreed upon at the time of the contract. Consumers are not liable for the difference if at the end of the lease term, the vehicle is worth less than the contracted residual value. In fact, consumers hold the option to purchase the vehicle 19 An explanation of the history and practice in the automobile industry can be found in Vertex (1997). 20 See Busse, Silva-Risso, and Zettelmeyer (2006) for an extensive explanation of these multiple subsidization levers and price rigidity. 21 Lending and leasing subsidies are bundled in financial reporting, however, and thus independently unobservable.

Copyright  2008 John Wiley & Sons, Ltd.

should they choose to (typically when the vehicle is worth more than the residual value). Because the lessor is responsible for any discrepancy between residual value and the vehicle’s actual lease-end market value, accuracy in residual value forecasting is critically important to the firm’s financial performance. Lessors who overestimate residual values may incur Type I errors, winning many customers with unprofitable leases. In contrast, lessors that underestimate residual values may incur Type II errors, missing opportunities for profitable leases by pricing too high. Given the negative marginal profit of overestimation, these errors are more severe. The ability of a lessor to accurately predict residual values for a given car is a critical element to leasing profitability. But a firm may also avoid losses by not only accurately forecasting depreciation on the cars it does lease, but also by avoiding those markets with the greatest potential for residual value losses. Performance in car leasing thus may depend both on forecasting capabilities, and the ability of the firm to strategically choose which markets to enter. Recent work by Makadok and Walker (2000) and Durand (2003) suggests that some heterogeneous performance may be due to forecasting competencies based in unique organizational characteristics. Makadok and Walker (2000) find that forecasting ability is an organizational-level capability in the estimation of future interest rates among money market mutual funds. Durand (2003) goes further in identifying some of the processes through which firm forecasting ability differs. Using data on French firms, Durand explores how the organizational illusion of control can lead to positive bias in firm estimates of industry growth. Durand’s findings suggest not only that forecasting is a firm-specific capability, but also that a firm’s internal characteristics, sources of information, and history can bias its forecasts and ultimately affect its performance. This work is also consistent with other research linking estimation capability with organizational context (Bateman and Zeithaml, 1989; McNamara and Bromiley, 1997; Mosakowski, 1998; King and Zeithaml, 2001) and with research documenting individual forecasting biases and errors among investors and financial analysts (Bernard and Thomas, 1990; Abarbanell and Bernard, 1992). Easterwood and Nutt (1999) argue that analysts overreact to good news while underreacting to bad news in adjusting their stock recommendations. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches Mikhail, Walther, and Willis (2003) found evidence that experience reduced analyst forecasting errors, although others introduce a strategic element in forecasts based on internal incentives, suggesting an organizational component to observed forecasting errors (Francis and Philbrick, 1993; Das, Levine, and Sivaramakrishnan, 1998; Lim, 2001).

driving a current model, and a redesign, however slight, will immediately devalue older models. While some redesigns are only cosmetic (e.g., reshaping the body paneling), other design changes include major structural overhauls, including new engines and new platforms. These major redesigns produce great uncertainty regarding the popularity of the new design as well as the devaluation of the old design.

Operationalizing the hypotheses to the car leasing industry

Hypothesis 3a: Independent lessors in markets with major redesigns will suffer larger residual values and thus exit sooner than competitors.

The hypotheses developed in the earlier sections are broadly relevant, but for empirical testing must be redefined in the context of automobile leasing. This section applies Hypotheses 1 and 2 to automotive ecosystems while developing new hypotheses that can be operationalized. This process could be repeated for many industries and markets, as the sources of change, turbulence, and dynamic capabilities vary across business ecosystems. The structure of this hypothesis development is presented below in Figure 2.

Likewise, durability influences used vehicle values as less durable vehicles deteriorate more quickly, reducing their worth on the secondary market. Since durability is often only revealed several years after the contract is written, some lessors will overestimate the durability while others underestimate it. Since those who overestimate durability will tend to win the contract, we should therefore observe those markets with lowdurability cars showing much higher residual value losses. Closed-end leases with purchase options, standard in the automobile industry, only exacerbate this risk for the lessor, as they ensure that only the worst cars within a model will be returned at lease-end. This adverse selection problem in lease returns is a topic of great interest in economics22 and is known to affect auction prices on lease return vehicles.23 The negative effect

Product change Two of the critical components of future vehicle values are new product introductions and durability. Purohit (1992) found that the redesign of models between 1975 and 1985 affected the price of used cars from that model and similar cars from both the same and competing manufacturers. Purohit’s study shows that the magnitude and timing of used car depreciation are strongly dependent on the timing and nature of new model introductions and redesigns. Consumers highly value

22 For a discussion of adverse selection in the automobile market, see Hendel and Lizzeri (1999) and Gilligan (2004). 23 Interviews with several lessors indicated that model-specific auction prices they receive on lease returns are influenced by the model-specific lease return percentage. If customers returned

Core firm product change

Theory

Industry context

H1a

H1b

H3a

Figure 2. Copyright  2008 John Wiley & Sons, Ltd.

331

Core firm penetration

H1c

H2a

H2b

H3b

H4a

H4b

Hypothesis structure Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

332

L. Pierce

of adverse selection on residual value is much stronger for low-durability cars, where more vehicles are lemons. Thus uncertainty about a model’s average durability as well as a specific unit’s durability should lead to an inverse relationship between durability and residual value overestimation, which should ultimately lead to firm exit. Hypothesis 3b: Independent lessors whose portfolios comprise lower durability cars will suffer larger residual value losses and exit sooner than their competitors. Core firm penetration An independent lessor competing in the car lease market must not only battle uncertainty over how market characteristics such as durability and model redesigns affect depreciation. It must also deal with a third major influence in car depreciation: the involvement of manufacturer-owned captive lessors. These captive lessors represent core manufacturers vertically integrating into niche complementary markets. These core firms are heavily involved in the leasing of particular models, writing thousands of identical leases in short periods of time. This behavior causes several problems for the niche independent lessors. First, captive lessors’ size makes their portfolio decisions highly influential on both the primary and the secondary market. Captives’ provide heavy competition in the initial origination of leases, where independent lessors must compete with them for customer contracts. On the secondary market, cars coming off lease must compete with other lease returns, used cars, and new vehicles—all competition that drives down used car prices. When many lease returns simultaneously enter the market, the glut of similar vehicles deflates prices. Second, manufacturers glean other benefits from writing leases, including customer loyalty, marketing information, and customer upgrades. These extra benefits allow captive lessors to charge lower lease prices in order to transfer gains to the parent manufacturer. Any independent lessor winning an auction with bids by captive lessors will be extremely likely to have overestimated the residual value, due to both the only 10 percent of all the lessors’ Ford Escorts, for example, the auctions would assume that these were only the worst cars. A higher lease return percentage would indicate a better average quality among the pool of cars at auction. Copyright  2008 John Wiley & Sons, Ltd.

ex ante low pricing incentives of the captive and the ex post effect of a flooded secondary market at lease-end. Independent lessors can, to some extent, anticipate this effect by observing the number of previously written and in-process leases scheduled to end at certain points in the future. By carefully observing the body of contracts written by themselves and competitors, such lessors can adjust estimates to account for future used car competition. While this future competition will include vehicles currently owned by consumers, whose future sales dates are entirely unknown, it will also include cars scheduled to come off lease at specific times. And while it may be very difficult to track all lease contracts in the market, a lessor may forecast much of this effect by observing the dominant portfolios of the captive lessors. Consequently, niche firms that do not adjust residual values to account for future off-lease competition may suffer unexpected losses and exit the market. Even if they do account for off-lease completion, their cars may still be worth less than manufacturer cars at lease-end. Manufacturers have instituted certified pre-owned programs for their lease returns, which allow them to achieve higher used car market prices than independents. Started during the explosion of leasing in the 1990s, these programs allow manufacturers to raise off-lease prices by five hundred to several thousand dollars by reducing consumer concerns on adverse selection problems through warranties and certification.24 Similar to subsidization, this tie between the new and used market suggests that there are many cars for which independent lessors cannot profitably compete, since certified programs not backed by manufacturers carry little weight with consumers.25 Hypothesis 4a: Independent lessors who lease cars with higher captive lessor penetration will 24

Consumer Reports, as reported in Bell (2007). The economics literature suggests other important roles for manufacturer leasing, such as the ability to control a used good’s competition with new offerings. Waldman (1996) and Hendel and Lizzeri (1999) explain how monopolists can exert market power in the secondary market by scrapping used offlease equipment. Related literature discusses the use of planned obsolescence by monopolists to eliminate the secondary market. While this explanation may be important in a monopoly market, it has little application to the car market, where manufacturers are monopolists in neither the manufacturing nor leasing of vehicles. See Bulow (1986), Kahn (1986), Waldman (1993) for further discussion. 25

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches suffer larger residual value losses and will exit earlier than their competitors. But captive lessors function as more than just another competitor in the market. The effect of captive lessor involvement is magnified when these manufacturer-owned subsidiaries subsidize residual values on specific vehicles. Subsidized leases tend to create large clusters of identical leases within a short period of time. The resulting plethora of identical leases ensures that three or four years later the used market will be flooded with these cars. Furthermore, the high residual values used to initially subsidize the leases will in turn raise the lease-end purchase price, guaranteeing most leased cars will be returned rather than purchased by the lessees. Any lessor writing leases on identical or similar cars will be damaged by these manufacturer incentives, as their own lease returns must compete with the captive’s portfolio at auction. Failing to differentiate from captives puts an independent lessor at risk for residual value losses. Avoiding models with subsidized leases is even more important than simple contract differentiation. Models with subsidized leases will tend to be unpopular and therefore at risk for redesign or cancellation. Any independent lessor holding these models in their portfolio may see this redesign devalue their vehicles. Furthermore, subsidization is a strong signal of market ennui with the model. Models that are unpopular when sold new will likely remain unpopular three to four years later on the used market. Finally, competing with business units that are intentionally pricing below cost is a poor strategy. While captive lessors can afford to write losing leases because of the benefits they provide to their parent manufacturers, independent lessors enjoy few such secondary benefits from leasing. For an independent lessor, the key is to identify subsidized leases, since many times these subsidies are hidden in the residual value. While subsidized loans are easily identifiable by their extremely low interest rates, it is often difficult to differentiate a subsidized residual value from an actual belief about future vehicle values. While the independent may be able to identify lease subsidization by looking for parallel loan and cash incentives, it can also suspect subsidization wherever a vehicle’s sales are underperforming. A lessor that observes a model with underperforming sales could deduce the probability of manufacturer Copyright  2008 John Wiley & Sons, Ltd.

333

subsidization to be high, and that captive residual values on that vehicle are inflated. The firm is then well-advised to either avoid the vehicle altogether, or at least approach leasing these vehicles with very conservative residual values.26 Low market share alone may not identify lease subsidization, as some low-market share vehicles are intentionally niche products, particularly in the sports and luxury segments (e.g., Porsche 911, Jaguar SJ8). Low market share, in combination with heavy captive leasing, however, should indicate the presence of intentionally inflated manufacturer residual values. Hypothesis 4b: Independent lessors that write leases on low market share vehicles with high rates of manufacturer leasing will suffer larger residual value losses and exit earlier than competitors.

THE DATA The primary dataset for this study involves approximately 200,000 individual consumer vehicle lease transactions from the years 1997–2001. These data come from a major supplier of marketing research information and identify detailed vehicle characteristics and the buyrate, monthly payment, capitalized cost, manufacturer rebate, trade-in value, lessor, contract date, and contracted residual value.27 The data are all from vehicle transactions in California, and are biased toward larger dealerships in urban areas. Because the data are based on information collected at the dealership, we cannot observe off-dealership financing. Leases and loans written away from dealerships appear only as cash transactions in the database. Additionally, vehicle model durability data are taken from Consumer Reports online service (http://consumerreports.org) for model years 1997–2002. These data include ex post measures of reliability for 14 vehicle characteristics, written on a five-point scale, and proxy for realized depreciation that may be unobserved ex post of 26 These may achieve the same effect, as conservative residual values will lead to high monthly payments, and will likely sell very few contracts. 27 The residual value reports the predicted residual value in the contract, not what is observed at lease-end, while the buyrate is the interest rate financial institutions offer dealers on the loans and leases for their customers. Dealers will often mark up this rate and offer the consumer a higher interest rate, pocketing the difference.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

334

L. Pierce

the lease contract. While Consumer Reports also provides a predicted depreciation measure, these are based on rough forecasts rather than actual revealed durability. Consumer Reports constructs durability measures from the number of necessary repairs reported by survey respondents. A high reliability rating of five indicates that very few respondents reported problems, while a rating of one represents a high problem report rate.28 Vehicle sales and market share data are taken from the Automotive News Market Data Books, years 1999–2002 (Automotive News, 1999–2002). Using these data, I calculated total segment sales and market share for market segments defined by the transaction data supplier. Vehicle depreciation data are from historical editions of Kelley Blue Book between January, 2000, and January, 2004 (Kelley Blue Book, 2000–2004). Wholesale rather than retail used car prices are used to more accurately reflect amounts lessors would receive at auction for lease return vehicles. These data were compared with the transaction data to estimate the accuracy of residual value forecasts by lessors. Control data on independent lessors are from several sources. Loan portfolio data came from the lease transaction supplier. Firm specialization data in consumer car financing came from OneSource (http://www.onesource.com), a business content aggregation service. New model introduction and redesign data are from Intellichoice.com. These data identify when and how an automobile manufacturer redesigned or introduced a new model. These data included dummy variables identifying whether or not the model change involved a new platform, which is considered a major update. The individual contract 28

While these data are the standard deterioration measures in the durable goods literature and the industry standard for vehicle reliability measurement, they are not without criticism. The principal complaint with this measurement is that the type of consumers reporting vehicle problems is not uniform across all brands. Owners of one brand may treat their vehicles more harshly than owners of other brands, or may have different standards in identifying a problem with their car. While these are certainly valid concerns in measuring actual product quality, they pose less of a problem in evaluating the durability effect on used car values. If a car model on average deteriorates significantly, this deterioration will affect the model’s usedmarket value, regardless of whether this deterioration was based on initial quality or average abuse level. Discrepancies across brands in survey reporting may be a concern in evaluating brand durability, but should be mitigated in this study by the use of manufacturer dummy variables. Since this study will address within-firm variation in product durability, firm-level bias in the durability measurements should be controlled. Copyright  2008 John Wiley & Sons, Ltd.

data identify the date on which the first vehicle of a new model or model update was sold. Thus for each of approximately 200,000 vehicles, the data include the contract terms and lessor, the realized durability of the model year, and the nature of the next redesign. This study uses only the approximately 70,000 contracts written by independent lessors. These data represent leases originated at dealerships. Since they do not include leases arranged directly with banks, they may not constitute an entirely representative sample of independent leases. Given the 45 percent national market share of captives, and the 60 percent market share in the California sample, one can infer that a significant number of independent leases are missing from the data.29 The difference between independent leases originated at dealerships and those arranged directly with the lessor is unobserved here, although one might expect differences in lessee populations and prices similar to studies of car lending (Cohen, 2006; Charles, Hurst, and Stephens, 2008). Where leases are originated at banks, competition for customers is lower, as the bank need not competitively bid for the customer as at dealerships. This should allow higher lease prices for the banks, and less financial pressure and risk on average. The durability and redesign decisions of the manufacturer will still affect the residual value in the same way, however, and so their implications for market choice and residual value forecasting remain important.30

IDENTIFYING HAZARDS FROM CORE FIRMS I first attempt to identify how the actions of the core manufacturer in an ecosystem increased hazards in certain markets, leading to residual value losses and exit in the niche market of leasing. To do this, I empirically examine the shakeout that occurred between the years 1998–2002 among 29 My best estimate is that approximately 40 percent of independent leases are missing here. 30 Furthermore, customers who acquire leases away from dealerships also present less default risk, since they are typically observed in person by a bank officer and may have long-term financial relationships with the institution. This is consistent with loan data from Charles et al. (2008). If these customers are also more responsible in maintaining the car, this could reduce the risk of durability, because used car auctions would be willing to pay higher prices for bank-originated cars.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches independent car lessors. I identify the market characteristics that increased residual value losses and accelerated firm exit from model-specific markets. I first use an ordinary least squares (OLS) regression model with fixed effects to identify how these market characteristics determine the magnitude of the residual value error. To support these results, I then use a time-variant Cox proportional hazard model with random effects to identify the effects on firm exit. Measuring residual value error I measure residual value forecasting performance by directly comparing the lessor’s ex ante residual value forecast in the contract data with the actual blue book car value at lease-end. I compare these two values by calculating the percentage error from the original value of the car. I define


r − v
the absolute value error as AV E =
c
and v the straight error as SE = r − c , where r is the contracted residual value, v is the lease-end market value, and c is the capitalized cost, or original value of the new car. The first definition, the absolute value of the percentage difference between the forecasted and the realized values, assumes equal value to the firm of setting residual values too high and too low. Under the absolute value measure, overestimating and underestimating by equivalent amounts are treated as equal forecasting performance, which may generally be the case. But it does not account for firms strategically setting conservative residual values, and thus may only partially reflect forecasting performance. Consequently, the second performance measure is the straight percentage error, where positive values reflect an overestimation and negative values reflect either an underestimation or conservative contracting. And because we might believe that winning a lease while underestimating the residual value is a good thing, the straight error seems to be the more appropriate measure.31 OLS models The determinants of vehicle depreciation from manufacturer decisions and product characteristics provide hypotheses on how core product design, 31 While winning a lease with a low residual value may be highly profitable, it may also represent that the lessor is losing in many other markets due to overpricing.

Copyright  2008 John Wiley & Sons, Ltd.

335

pricing, and entry strategies might lead to niche market firm losses and exit. These hypotheses suggest that any firm lacking understanding of vehicle durability variance may have overestimated residual values on the worst vehicles, predicting them to have average deterioration. The hypotheses also suggest that many lessors failed to account for design obsolescence, and overestimated residual values on vehicles with pending major redesigns. The largest pitfall, however, may be failure to avoid leasing cars with high captive lessor involvement. And while failure to differentiate may in itself lead to firm exit, this effect should be exaggerated on models where subsidization is likely. In order to test the hypotheses on how core manufacturer decisions affect lessor forecasting performance, I use an OLS model with measures of residual value accuracy as the dependent variables. This equation, which includes the explanatory and control variables presented in Table 1, predicts the forecasting accuracy for an individual lease contract written by an independent lessor. I also include dummy variables for year, month, manufacturer, and vehicle segment to control for unobserved heterogeneity. To control for unobserved lessor characteristics, I alternatively include lessor fixed effects αi . For any given contract, this model is: Errori = αi + λXi + βKi + γ Fi + φPi + ωDi + εi Explanatory variables The X matrix represents the variables meant to test the hypotheses. These include vehicle durability, new product introduction, and captive subsidization measures. New product introduction variables include the Intellichoice major redesign dummy and dummies for the first and last year in a current design. Captive subsidization is measured by interacting model-specific captive leasing penetration with one of two measures of low demand: vehicle segment market share and dealer profitability. Vehicle market share is defined at the vehicle segment level, as defined by J.D Power and Associates. Low dealer profit indicates low demand relative to supply, suggesting the potential need for manufacturer lease and loan subsidization.32 32

See Busse et al. (2006) for an explanation of this price rigidity. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

336

L. Pierce

Table 1. OLS variable descriptions Dependent variables Straight RV error pct. Abs. value RV error ct. Market controls Price (log) Bidder count Segment Herfindahl Dealer credit risk First year dummy Last year dummy Lessor controls Portfolio size (log) Loan credit risk Specialist Explanatory variables Major redesign dummy Durability Market share Captive penetration Dealer profit

The residual value forecasting error as a percentage of the new vehicle price The absolute value of the residual value error The new vehicle price (logged) The number of firms leasing this car The Herfindahl market concentration in this vehicle segment The average interest rate on loans written at the originating dealership 1 if the first year of a redesign. 0 otherwise 1 if the last year before a redesign. 0 otherwise The number of contracts in the lessor’s loan portfolio (logged) The average interest rate on the lessor’s loan portfolio 1 if the firm’s exclusive business is consumer automobile finance. 0 otherwise 1 if a major platform design occurred during the period. 0 otherwise The Consumer Reports durability score (1 to 5) The average model family market share in its vehicle segment The ratio of captive leases to total production The average dealer profit margin for this model

Control variables The K matrix represents car variables, including the logged vehicle invoice price, manufacturer and segment dummies, and vehicle durability. The F matrix consists of firm-level control variables, including the loan portfolio size, loan credit risk, and car industry specialization. Loan portfolio size is meant to measure past success and current size in the related car loan market. Several studies suggest that size and success in previous ventures will affect forecasting performance. Durand (2003) found that firm size was positively related to both the directional and absolute value errors in firm forecasts of industry growth rates. Durand’s finding is inconsistent with results from Makadok and Walker (2000), who find a self-reinforcing positive feedback loop in money market mutual funds (MMMF). Their results, consistent with evolutionary growth models (Nelson and Winter, 1982; Arthur, 1989), show that larger size causes a subsequent improvement in MMMFs’ Treasury Bill forecasting accuracy. Loan credit risk is the average interest rate on loans in those portfolios reflecting inherent credit risk. Firms with high loan interest rates are subprime lenders operating in markets where understanding vehicle depreciation is already critically important. Subprime lending requires knowledge on depreciation because its payment structures typically involve long periods where consumers are Copyright  2008 John Wiley & Sons, Ltd.

‘upside-down,’ or owe more money than the car is worth.33 High depreciation lengthens this period, during which the lender is guaranteed to lose money on any default. Similar to home lending, where borrowers can be upside-down even without depreciation, the expected length of this period can be incorporated into loan terms.34 This subprime lending produces prior related experience, which has been shown to improve survival (Mitchell, 1989, Klepper; 2002, Thompson, 2005). Car industry specialization is a dummy variable reflecting whether or not lessors’ exclusive business is in consumer car financing, or whether they have a

33 While car lenders and lessors often ask for down payments to reduce this risk, down payments (called capital reduction in leases) are often infeasible requirements for subprime loans and leases, since these customers are typically cash constrained (Mannering, Winston, and Starkey, 2002). Nonlinear payment structures produce similar problems in addition to irritating the dealers who originate leases, because a standardized monthly payment structure simplifies advertising and negotiation. 34 In home lending, upside-down borrowers typically result from down payments not large enough to account for the transaction cost of selling (e.g., real estate agent fees and closing costs) or fluctuations in the housing market. Home lenders typically account for this risk by requiring a second, higher interest loan for any amount over 80 percent of the home price, or by requiring the borrower to pay for private mortgage insurance (PMI). Car lenders instead typically account for this additional risk through a single higher interest rate loan. Gap insurance, used in both car leases and loans, serves a similar role to PMI, but only applies to vehicle loss from theft or collision while the car is worth less than the amount owed.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches more diversified business line. Models of multitasking in financial services (Holmstrom and Milgrom 1991; Kanatas and Qi 1998; Puri 1999; Ross 2007) suggest agency issues may disadvantage diversified banks compared to specialized independent lessors, consistent with findings on incentives and monitoring in diversified banks (Acharya, Hasan, and Saunders, 2006; Winton 1999). The P matrix consists of market variables, including model segment competition (Herfindahl index), a consumer credit proxy, and the number of lessors for that model. The Herfindahl index for model market share is calculated as the sum of the squared market shares of all cars in a given car segment. The consumer credit proxy is the average interest rate offered on (non-lease) loans at the dealership of lease origination, and serves to control for any differences between lessors in the credit rating of their lessees. The number of lessors for a model is used to proxy the number of bidders for a given customer in order to control for model-specific leasing competition and a possible winner’s curse.35 The D matrix consists of month 35 Since dealership-based leases operate in an auction format, the potential exists for a serious winner’s curse. The winner’s curse, a phenomenon popularized by Richard Thaler’s (1994) book of the same title, results in auction bidders paying an overvalued price for an item. In auctions with many bidders, the winner will tend to be the one with the highest estimated asset value, and will

Table 2.

337

and year time dummies. The unit of analysis is an individual vehicle lease contract, of which I use 70,337. OLS model summary statistics The summary (descriptive) statistics for the OLS model are listed in Table 2. For the 70,337 lease contracts, the average straight error is 10 percent of new vehicle price. The average vehicle price is $24,080 with the logged value mean at 10.05. Eighty-four percent of redesigns are major platform changes and only 16 percent are minor. Twenty-one percent of the vehicles were leased in their first design year, while only 13 percent were leased in the year before the redesign. The average car model had nearly 26 successful lease contract bidders. The average vehicle durability is 4.27 out of 5. The average market share was eight percent, with an average Herfindahl index of 8.86 reflecting manufacturer concentration in certain car segments likely have overestimated as well. Evidence shows that banks are susceptible to a winner’s curse in consumer credit evaluation (Shaffer, 1998), although it should be noted that if banks monitor other bids and update beliefs on the residual value, this effect may disappear. While it would be ideal to have data on losing bids for each contract, these data are unavailable. Consequently, I assume that the relative number of lessors in a given model market approximately reflect the relative number of bidders on a given contract.

OLS descriptive statistics Observations

Mean

Standard deviation

Min

Max

70 337 70 337

10.07 11.57

9.52 7.62

−58.65 0.00

71.51 71.51

70 337 70 337 70 337 70 337 70 337 70 337

10.05 25.79 7.97 8.86 0.21 0.13

0.28 7.86 4.29 1.19 0.40 0.34

8.84 3.00 3.59 1.64 0 0

11.44 39.00 51.01 20.31 1 1

70 337 70 337 70 337

10.27 9.01 0.03

1.74 1.30 0.15

1.38 6.10 0

12.67 21.00 1

70 337 70 337 70 337 70 337 70 337

0.84 4.27 8.05 0.10 7.09

0.36 0.52 6.55 0.06 2.03

0 2.64 0.48 0.00 1.97

1 5.00 57.11 0.45 19.16

Dependent variables Straight RV error pct. Abs. value RV error pct. Market controls Price (log) Bidder count Segment Herfindahl Dealer credit risk First year dummy Last year dummy Lessor controls Portfolio size (log) Loan credit risk Specialist Explanatory variables Major redesign dummy Durability Market share Captive penetration Dealer profit Copyright  2008 John Wiley & Sons, Ltd.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Copyright  2008 John Wiley & Sons, Ltd.

1.00 −0.01 1.00 −0.01 0.04 1.00 −0.01 −0.06 0.03 1.00 −0.01 0.08 −0.21 −0.25 1.00 0.01 −0.16 −0.18 0.09 −0.06 1.00 −0.03 0.03 −0.08 0.06 −0.02 −0.02 1.00 −0.20 0.27 0.03 0.08 −0.02 −0.01 −0.06 1.00 −0.01 −0.03 0.01 −0.13 0.00 −0.07 0.01 −0.07 1.00 −0.20 −0.02 0.02 0.00 −0.05 −0.19 0.01 0.02 0.04

10 9 8 7 6 5 4 3 2 1 OLS correlations Table 3.

The results from the OLS model, presented in Table 4, identify the effects of core manufacturers on the financial performance of niche market lessors. Models 1–4 use the straight residual value error as the dependent variable, while Models 5–8 use the absolute value of the residual value error. Within these subgroups, the first model uses only control variables to form a baseline. The second model adds explanatory variables to observe their overall predictive value. The third model replaces lessor controls with lessor fixed effects. The fourth model uses dealer profit as an alternative proxy for subsidization risk. All standard errors are robust and correct for error clustering at the lessor level. The results from the OLS model support the hypothesized influence of core manufacturer choices on niche market lessor performance. The major redesign dummy is positive and statistically significant in all models, consistent with Hypothesis 3a. Disruptive product change in specific car markets appears to increase uncertainty and reduce forecasting accuracy. Major redesigns on average result in overestimations of about three percent, or $600 on a $20,000 car. Also supporting this hypothesis are the coefficients for the first and last year dummies, where the uncertainty of the new redesign impact is largest. Market share positively influences residual value losses. Models that are very popular when new may crowd the used market in later years, particularly when popularity has waned. In models using the market share variable (2 and 3), captive lessor involvement increases the residual value error, consistent with Hypothesis 4a. This effect suggests that as captive lessor involvement increases in a market, residual value errors among independent lessors magnify. This coefficient could reflect

11

OLS model results

1.00 0.78 1.00 0.02 0.05 1.00 0.22 0.23 −0.08 1.00 0.14 0.15 −0.23 −0.01 1.00 −0.02 0.02 −0.18 0.12 0.04 1.00 0.03 −0.01 −0.05 −0.02 0.03 0.01 −0.01 0.01 0.03 −0.08 −0.08 0.00 0.00 0.00 −0.05 −0.01 −0.04 −0.06 −0.03 −0.02 −0.04 0.03 0.03 0.01 0.04 0.04 0.02 −0.01 0.01 0.04 0.11 0.09 0.01 0.11 −0.16 0.04 −0.21 −0.23 −0.16 −0.03 −0.03 0.08 0.19 0.18 −0.18 0.39 0.55 −0.02 0.05 0.02 0.46 −0.13 −0.33 −0.16 0.05 0.13 0.09 −0.05 0.27 0.00

12

13

14

15

16

(e.g., luxury, mid-sized). The average dealer profit margin was 7.1 percent. For firm characteristics, the average interest rate written on the firm’s loans was nine percent, with variation from low-risk portfolios of six percent to subprime portfolios of 21 percent. While 20 percent of the sample firm population are car industry specialists, these lessors wrote only three percent of all leases, reflecting their significantly smaller size. The average logged loan portfolio size among these firms was 10.3. Correlations of the variables are listed in Table 3.

1.00

L. Pierce

1. Straight RV error pct. 2. Abs. value RV error pct. 3. Price (log) 4. Bidder count 5. Segment Herfindahl 6. Dealer credit risk 7. First year dummy 8. Last year dummy 9. Portfolio size (log) 10. Loan credit risk 11. Specialist 12. Major redesign dummy 13. Durability 14. Market share 15. Captive penetration 16. Dealer profit

338

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

OLS models

Copyright  2008 John Wiley & Sons, Ltd. (0.732)∗∗ (0.031)∗∗ (0.312)∗∗ (0.122)∗ (0.293)∗ (0.348)∗∗

0.3432 70 337

Y Y N

0.283 (0.127)∗ −0.412 (0.195)∗ 1.680 (0.787)∗

3.276 0.105 1.935 −0.293 .717 1.521

(1) Straight

(0.788)∗∗ (0.038) (0.332)∗∗ (0.135) (0.227)∗∗ (0.117)∗∗

Y Y N

(0.38)∗∗ (0.238) (0.049)∗∗ (2.759)∗∗ (0.221)∗∗

0.3746 70 337

3.398 0.176 0.372 11.468 −1.125

0.194 (0.132) −0.469 (0.175)∗∗ 2.784 (1.12)∗

2.678 −0.027 2.360 −0.114 1.200 2.270

(2) Straight

Robust standard errors clustered at the lessor level in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1%

R2 Observations

Market controls Price (log) Bidder count Segment Herfindahl Dealer credit risk First year dummy Last year dummy Lessor controls Portfolio size (log) Loan credit risk Specialist Explanatory variables Major redesign dummy Durability Market share Captive penetration MarkShare∗ CapPen Dealer profit DealProf∗ CapPen Month/year effects Manufact./segment effects Lessor fixed effects

Dependent var: RV error

Table 4.

Y Y Y

(0.338)∗∗ (0.219) (0.045)∗∗ (3.043)∗∗ (0.182)∗∗

(0.728)∗∗ (0.037) (0.336)∗∗ (0.119) (0.214)∗∗ (0.094)∗∗

0.3804 70 337

3.401 −0.250 0.378 11.525 −1.235

2.809 −0.034 2.364 −0.141 1.227 2.248

(3) Straight

(8.802)∗ (0.101)∗∗ (1.219)∗∗ Y Y Y

21.302 −0.665 −3.988

0.3827 70 337

(0.335)∗∗ (0.172)∗∗

(0.729)∗∗ (0.025)∗ (0.356)∗∗ (0.126) (0.226)∗∗ (0.173)∗∗

2.905 0.662

3.514 0.053 2.445 −0.129 1.651 2.523

(4) Straight

(0.509)∗∗ (0.016)∗∗ (0.205)∗∗ (0.054)∗ (0.161)+ (0.296)∗∗

0.2816 70 337

Y Y N

0.182 (0.100)+ −0.394 (0.148)∗∗ 1.608 (1.042)

3.236 0.081 1.581 0.114 .286 1.217

(5) Absolute

(0.677)∗∗ (0.015)∗ (0.274)∗∗ (0.055)∗∗ (0.216)∗ (0.286)∗∗

0.3403 70 337

Y Y N

2.614 (0.247)∗∗ 0.492 (0.278)+ 0.139 (0.041)∗∗ 7.103 (3.233)∗ −0.488 (0.239)∗

0.143 (0.104) −0.413 (−0.148)∗∗ 2.147 (1.184)+

3.079 0.033 1.613 0.209 .556 1.405

(6) Absolute

(0.645)∗∗ (0.013)∗ (0.284)∗∗ (0.033)∗∗ (0.198)∗∗ (0.289)∗∗

0.3549 70 337

Y Y Y

2.582 (0.203)∗∗ 0.451 (0.270)+ 0.139 (0.042)∗∗ 6.969 (3.461)∗ −0.537 (0.229)∗

2.991 0.028 1.642 0.209 .551 1.405

(7) Absolute

(4.939)+

(0.236)∗∗ (0.268)∗∗

(0.566)∗∗ (0.008)∗∗ (0.294)∗∗ (0.034)∗∗ (0.187)∗∗ (0.274)∗∗

0.3539 70 337

0.203 (0.066)∗∗ −0.560 (0.781) Y Y Y

8.729

2.581 0.837

2.898 0.074 1.697 0.189 .579 1.476

(8) Absolute

Big Losses in Ecosystem Niches 339

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

340

L. Pierce

increased competition on both the leasing and secondary markets. Furthermore, as market share goes down, this effect decreases, supporting Hypothesis 4b. Models including dealer profit (4 and 8) also support Hypotheses 4a and 4b. High dealer profit reduces residual value losses, particularly when captive penetration is high. Captive penetration again increases both straight and absolute residual value losses. The results on durability are mixed. While durability increases the absolute residual value error in Models 5–8, it appears to inconsistently affect the straight residual value error. Given its negative correlation with both measures in Table 3, low-durability cars clearly have higher residual value losses, but this correlation may be spuriously related to multicolinearity in the model. Models run without vehicle segment dummies produce consistently positive durability coefficients that are statistically significant at the five percent level. This, combined with raw correlations, suggests that most of the durability variation is captured at the segment level. Independent lessors in lowdurability vehicle segments suffer larger losses, so long as they cannot identify this risk. Given this colinearity, the results seem to provide only weak support for Hypothesis 3b. Among the market controls, the segment Herfindahl index has a consistently positive effect on residual value errors. This suggests that higher levels of core firm competition across ecosystems produce lower profits in complementary niche markets. Bidder count also consistently raises residual value losses, suggesting competition within the complementary niche market lowers profits. The coefficient for loan portfolio size is positive in all the models, although only marginally significant, suggesting that firms with past success and size in related activities may suffer from overconfidence and a positive bias in forecasting, consistent with findings from Durand (2003). These results do not truly conflict with Makadok and Walker (2000), because size in this market does not reflect experience specific to this niche market. Similarly, the negative coefficients on loan credit risk suggest that related experience in subprime lending improves residual value performance. Finally, car financing specialists appear to have outperformed diversified banks, supporting the argument that diversification in financial services may hinder individual firm functions such as forecasting. The Copyright  2008 John Wiley & Sons, Ltd.

results from these firm-level controls strongly suggest value in future work examining the firm-level characteristics that improved residual value forecasting. Identifying niche firm exit While the OLS models identify residual value forecasting performance, they do not identify firm exit, a likely result of overestimation. In order to identify firm exit, a semi-parametric Cox proportional hazard model predicts the duration of a specific lessor in a model-specific market.36 This specification is particularly appropriate for continuous time observations, because no two observations are likely to have the same event time. These data identify the day of each market entry and exit, so this problem is unlikely. The hazard of exiting a market is the product of the exponential of a linear function of independent variables and the undefined baseline hazard function. The market exit hazard is a function of an unspecified baseline hazard and the exponential of a matrix of explanatory variables. For lessor i in market j , the Cox model is: hij (t) = λ0 exp{βX} While the Cox model is appropriate in several ways, it suffers from two flaws in this setting. First, since the data are right truncated at January, 2002, I treat all failures within 300 days before this date as right-hand censored.37 Second, and more severe, since these data are left truncated it is impossible to identify the true entrance into a market. Many of the firms thus were at risk before the data begin, making the hazard rate estimates inaccurate. One can typically correct for this if the age at the left truncation point is known, but these data are unavailable for this study. Consequently, I eliminate all firm/market pairs where entry occurs within the first 300 days of the data. This correction reduces the number of observations, but ensures accurate identification of the true ‘birth’ date for each observation. Unfortunately, it also limits interpretations of the Cox model results. These results reflect only the group of firm/market 36 An example of a model-specific market would be the lease market for Honda Accords. 37 This model was repeated using 200 and 100 days as the censor point, both with similar results.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches pairs entering late in the market, and any extrapolation to the broader population of lessors is problematic.

341

Table 5 presents the Cox model variable definitions, with Table 6 presenting summary statistics for the explanatory variables. There are 837 observations, with each observation representing a lessor/market pair and 600 resulting in failures. Correlations are presented in Table 7. The results for the Cox model are presented in Table 8. The analysis uses models both with and without lessor random effects to account for a given lessor’s exit decisions being correlated.38

Columns 1 and 2 serve as a baseline with vehicle price as a control. Columns 3 and 4 show the magnitude of the average straight residual value error to increase the hazard of exit, consistent with residual value error driving niche market shakeouts. Columns 5 and 6 replace residual value error with the explanatory variables shown in the OLS specification to predict it. Similar to the OLS models, Columns 7 and 8 use dealer profit as an alternative to market share. The captive penetration interaction variables are omitted due to the difficulties of interpreting interaction terms in nonlinear models. Durability shows no statistically significant effect on hazard rates, which is consistent with the weak findings in the OLS model. Major redesigns

38 Lessor random effects are from a gamma-distributed sharedfrailty model. Variance tests show the shared-frailty assumption

to be appropriate at the one percent level. The use of fixed effects in Cox models is not feasible.)

Results from Cox models

Table 5.

Cox model variable definitions

Price (log) Residual value error Major redesign dummy Captive model penetration Durability Dealer profit Market share

Table 6.

The new vehicle price (logged) The percentage difference between the contracted and realized residual value 1 if a major platform design occurred during the period. 0 otherwise The ratio of captive leases to total production The Consumer Reports durability score (1 to 5) The average percentage dealer profit margin on this vehicle make The average model family market share in its vehicle segment

Cox descriptive statistics Observations

Mean

Standard Deviation

Min

Max

837 837 837 837 837 837 837 837

1168.53 10.11 0.15 4.24 0.72 12.38 5.88 7.32

394.23 0.35 0.25 0.51 0.45 9.23 7.16 2.63

80 8.95 −0.68 2.93 0 0.32 0.11 2.06

1823 11.27 1.48 5 1 71.96 51.71 23.32

Time in market Price (log) Residual value error Durability Major redesign dummy Captive penetration Market share Dealer profit

Table 7.

Cox correlations

Variable 1. 2. 3. 4. 5. 6. 7.

Price (log) Residual value error percentage Durability Major redesign dummy Market share Captive penetration Dealer profit percentage

Copyright  2008 John Wiley & Sons, Ltd.

1

2

3

4

5

6

7

1.00 −0.03 −0.04 −0.17 −0.11 0.49 0.22

1.00 −0.30 0.10 0.16 −0.01 −0.13

1.00 −0.19 0.03 −0.12 −0.07

1.00 −0.08 −0.08 −0.23

1.00 −0.25 0.01

1.00 0.09

1.00

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Copyright  2008 John Wiley & Sons, Ltd.

837 600 −4690.74

Y N

0.449 (0.066)∗∗

Y Y 1.247 (0.240)∗∗ 837 600 −4456.97

0.659 (0.098)∗∗

(2) Time to exit

837 600 −4644.18

Y N

0.481 (0.072)∗∗ 2.855 (0.300)∗∗

(3) Time to exit

Y Y 1.225 (0.237)∗∗ 837 600 −4425.43

0.698 (0.105)∗ 2.851 (0.431)∗∗

(4) Time to exit

837 600 −3606.04

Y N

0.961 (0.108) 1.338 (0.133)∗∗ 3.376 (2.115)+ 0.981 (0.008)∗

0.505 (0.092)∗∗

(5) Time to exit

Y Y 1.171 (0.234)∗∗ 837 600 −3456.89

1.008 (0.119) 1.389 (0.147)∗∗ 3.335 (2.108)+ 1.001 (0.009)

0.725 (0.134)+

(6) Time to exit

837 600 −3605.22

0.949 (0.019)∗∗ Y N

0.936 (0.106) 1.234 (0.126)∗ 2.163 (1.118)

0.572 (0.111)∗∗

(7) Time to exit

Lessor random effects are from a gamma-distributed shared-frailty model. The variance theta significantly different from zero represents a shared frailty. + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1%

Number of subjects Number of failures Log likelihood

Car segment effects Lessor random effects Theta

Dealer profit

Market share

Captive penetration

Major redesign dummy

Durability

Residual value error

Price (log)

(1) Time to exit

Table 8. Cox proportional hazards models

0.913 (0.019)∗∗ Y Y 1.233 (0.243)∗∗ 837 600 −3445.88

0.994 (0.117) 1.203 (0.131)+ 2.248 (1.183)

0.982 (0.191)

(8) Time to exit

342 L. Pierce

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches continue to increase hazard rates in all models. Similarly, captive penetration increases hazard rates, consistent with the OLS model. The results on market share are conflicting, with no effect under the statistically appropriate random-effects model. The substitution of dealer profit for market share in Columns 7 and 8 shows that low-profit vehicles yield higher risk of exit to lessors. Overall, these results support the OLS specification, although the findings are statistically less significant on average.

DISCUSSION AND CONCLUSION This study considers how the strategic actions of core firms generate shakeouts in the complementary niche markets of a business ecosystem. These core firms’ dynamic product design and niche market entries create turbulent ecosystems that generate financial losses and exit for independent niche market firms. This strategic change is dangerous for niche market firms, particularly when the health of complementors is of little importance to the core firm. The hazards of core firm actions are severely amplified in high-technology markets, but they apply in other industries as well. Hazardous activities can include pricing, product design, marketing, diversification, and regulatory influence. Niche firms must constantly monitor these activities to formulate which strategies are optimal in response, whether competing in or exiting the market. In order to test these propositions on survival and performance, I examine car leasing, a complementory niche market in an automotive ecosystem. Using unique data of nearly 200,000 individual lease transactions, I find evidence that niche complementor performance and failure was a function of the activities of the core manufacturers. Changes in product redesign, product durability, manufacturer entry, and subsidization all raised the probability of niche lessor exit and directly resulted in measurable financial losses. These findings support my hypotheses on how core firm actions can drive complementor losses and shakeouts. This study also raises questions about the heterogeneity of firm performance in hazardous markets. While many of the complementary niche markets in automotive leasing had poor performance and high exit rates, these results were not uniform Copyright  2008 John Wiley & Sons, Ltd.

343

across firms. This suggests that some complementary firms may have possessed the knowledge or related experience necessary to adapt to this hazardous ecosystem. In markets where hazards stem from constantly changing core firm activities, performance is likely determined by adaptive response to these actions. Firm survival in the face of strategic change by core firms involves anticipating, exploiting, and expeditiously reacting to the activities of these core firms—capabilities that are truly dynamic. These dynamic capabilities may be developed in a number of ways, but the literature suggests several potential sources. First, this performance may be based in superior dynamic forecasting capabilities. Scholars have long recognized the importance of forecasting in firm profitability (Barney, 1986; Nelson and Winter 1982; Teece et al. 1997), and there is evidence from accounting (Bernard and Thomas, 1990), finance (Lim, 2001) and management (Barnes, 1984; Hambrick, 1997; Makadok and Walker, 2000; Durand, 2003) that forecasting ability may be heterogeneous at both the individual and organizational level. Second, related experience of the niche market firm likely plays an important role in predicting survival and performance. Experienced firms can use past successes and failures to ‘develop new products and processes, and design and implement viable business models.’ (Teece, 2007: 1320). Effects of pre-founding experience have been observed in preexisting firms diversifying into new markets, where prior related experience and knowledge improves probability of entry and, ultimately, survival (Mitchell, 1989; Klepper; 2002; Phillips, 2002; Thompson, 2005; Roberts, Klepper, and Hayward, 2007). While some of this tacit knowledge may be technical, firms likely benefit from market knowledge on the dynamics of competition. These findings contribute to the strategy and management literature in several ways. First, they develop the understanding of how competition in business ecosystems evolves. The existence of a core firm in an ecosystem provides dangerous implications for firms considering entering niche complementary markets. Second, they show how the structure of an ecosystem or market can influence predictions on shakeouts. Third, they suggest the importance of dynamic capabilities in business ecosystems, even when that ecosystem is not primarily driven by technology considerations. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

344

L. Pierce

Finally, they provide a clear industry example of these theories, and use a remarkable dataset to test developed hypotheses on niche market shakeouts. The data in this study are valuable in their potential for isolating pure measures of financial performance across hundreds of niche markets and numerous complementary firms. Large and otherwise successful financial institutions did remarkably poorly in this market. One interesting implication of these results is that managers cannot assume that success in past activities, even when seemingly related, will ensure profitability in new endeavors. Even an expansion within car financing from lending to leasing may require completely new sets of market knowledge. The leasing niche is significantly more turbulent than lending, as it is driven by residual value forecasts extremely susceptible to manufacturer activities. This market was evolving during this period with manufacturer penetration and subsidization never before observed, and new innovations like certified pre-owned programs changing manufacturer strategies. Given this period of disequilibrium, it is perhaps not surprising that the same institutions currently collapsing under the weight of the subprime mortgage crisis performed so poorly when entering a new market. These findings have implications beyond the automotive industry to the diversification of firms into new markets. This research underscores that the replication of past successes in new markets is not a trivial activity. There are limits to the expansion of the firm that must be acknowledged in any diversification strategy. Further work is required to better understand the sources of heterogeneous performance in these ecosystems. While these data are extremely rich at the market and contract level, and particularly suited for measuring performance, they lack internal firm detail on the lessors in the sample. While field interviews and coefficients on firm-level controls suggest that experience and specialization play a role in performance, more detailed firmlevel data will be important in cleanly identifying the sources of these dynamic capabilities.

ACKNOWLEDGEMENTS My appreciation to Garrick Blalock, Meghan Busse, Aaron Chatterji, Bronwyn Hall, Steven Copyright  2008 John Wiley & Sons, Ltd.

Klepper, Robert Lowe, Jeff Macher, David Mowery, Patricia Murphy, Jackson Nickerson, John Pierce, Dan Snow, Jason Snyder, David Teece, Mike Toffel, Todd Zenger, and Florian Zettelmeyer for their feedback and support throughout the revision of this paper. I also thank seminar participants from Carnegie Mellon, Harvard, UC Berkeley, and UCLA. I am particularly indebted to Editor Richard A. Bettis and two anonymous reviewers whose work produced a substantially better paper. The Institute of Management, Innovation, and Organization and the Intel Corporation supported much of this work. All mistakes are my own.

REFERENCES Abarbanell JS, Bernard VL. 1992. Tests of analysts’ overreaction/underreaction to earnings information as an explanation for anomalous stock price behavior. Journal of Finance 47(3): 1181–1207. Acharya VV, Hasan I, Saunders A. 2006. Should banks be diversified? Evidence from individual bank loan portfolios. Journal of Business 79(3): 1355–1412. Agarwal R, Echambadi R, Franco AM, Sarkar MB. 2004. Knowledge transfer through inheritance: spinout generation, development, and survival. Academy of Management Journal 47(4): 501–522. Aghion P, Bolton P. 1987. Contracts as a barrier to entry. American Economic Review 77(3): 388–401. Albornoz F, Yoguel G. 2004. Competitiveness and production networks: the case of the Argentine automotive sector. Industrial and Corporate Change 13(4): 619–641. Alchian AA, Demsetz H. 1972. Production, information costs, and economic organization. American Economic Review 62(5): 777–795. Amburgey TL, Rao H. 1996. Organizational ecology: past, present, and future directions. Academy of Management Journal 39(5): 1265–1286. Arthur WB. 1989. Competing technologies, increasing returns, and lock-in by historical events. Economic Journal 99(394): 116–131. Automotive News. 1999–2002. Automotive News Market Data Book. Automotive News (Crain Communications, Inc.): Detroit, MI. Barnes JH Jr. 1984. Cognitive biases and their impact on strategic planning. Strategic Management Journal 5(2): 129–137. Barney JB. 1986. Strategic factor markets: expectations, luck, and business strategy. Management Science 32(10): 1231–1241. Bateman TS, Zeithaml CP. 1989. The psychological context of strategic decisions: a test of relevance to practitioners. Strategic Management Journal 10(6): 587–592. Bell C. 2007. Are certified pre-owned cars worth the extra money? Bankrate.com: http://www.bankrate.com/ Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches brm/news/auto/20070319 certified preowned car a1. asp?prodtype=auto (19 March 2007). Bernard VL, Thomas JK. 1990. Evidence that stock prices do not fully reflect the implications of current earnings for future earnings. Journal of Accounting and Economics 13(4): 305–340. Brandenburger A, Nalebuff B. 1996. Co-Opetition: 1. A Revolutionary Mindset that Combines Competition and Cooperation in the Marketplace; 2. The Game Changing Strategy that is Changing the Game of Business. Currency Doubleday: New York. Bulow JI. 1986. An economic theory of planned obsolescence. Quarterly Journal of Economics 101(4): 729–749. Busse M, Silva-Risso J, Zettelmeyer F. 2006. $1,000 cash back: the pass-through of auto manufacturer promotions. American Economic Review 96(4): 1253–1270. Carroll GR. 1988. Organizational ecology in theoretical perspective. In Ecological Models of Organizations, Carroll GR (ed). Ballinger: Cambridge, MA; 1–6. Carroll GR, Hannan MT. 2000. The Demography of Corporations and Industries. Princeton University Press: Princeton, NJ. Charles K, Hurst E, Stephens M. 2008. Rates for vehicle loans: race and loan source. American Economic Review 98(2): 315–320. Cohen M. 2006. Imperfect competition in auto lending: subjective markup, racial disparity, and class action litigation. Vanderbilt Law and Economics Research Paper No. 07–01. Available at SSRN: http://ssrn.com/abstract=951827. Collis DJ. 1994. Research note: how valuable are organizational capabilities? Strategic Management Journal , Winter Special Issue 15: 143–152. Cordle IP. 2001. Auto leases throw banks off course. South Coast Today. 3 January. Available at: http://archive.southcoasttoday.com/daily/01-01/ 01-03-01/a03tr125.htm. Das S, Levine CB, Sivaramakrishnan K. 1998. Earnings predictability and bias in analysts’ earnings forecasts. Accounting Review 73(2): 277–294. Dasgupta P, Stiglitz J. 1988. Learning-by-doing, marketstructure and industrial and trade policies. Oxford Economic Papers-New Series 40(2): 246–268. DiMaggio J, Powell WW. 1983. The iron cage revisited: institutional isomorphism and collective rationality. American Sociological Review 48: 147–160. Dobrev SD, Kim T, Carroll GR. 2002. The evolution of organizational niches: U.S. automobile manufacturers, 1885–1981. Administrative Science Quarterly 47(2): 233–264. Dobrev SD, Kim T, Carroll GR. 2003. Shifting gears, shifting niches: organizational inertia and change in the evolution of the U.S. automobile industry, 1885–1981. Organization Science 14(3): 264–282. Durand R. 2003. Predicting a firm’s forecasting ability: the roles of organizational illusion of control and organizational attention. Strategic Management Journal 24(9): 821–838. Copyright  2008 John Wiley & Sons, Ltd.

345

Easterwood JC, Nutt SR. 1999. Inefficiency in analysts’ earnings forecasts: systematic misreaction or systematic optimism? Journal of Finance 54(5): 1777–1797. Economides N, Salop SC. 1992. Competition and integration among complements, and network market structure. Journal of Industrial Economics 40(1): 105–123. Eisenhardt KM, Martin JA. 2000. Dynamic capabilities: what are they? Strategic Management Journal , October–November Special Issue 21: 1105–1121. Farrell J, Saloner G. 1986. Installed base and compatibility: innovation, product preannouncements, and predation. American Economic Review 76(5): 940–955. Francis J, Philbrick D. 1993. Analysts’ decisions as products of a multitask environment. Journal of Accounting Research 31(2): 216–230. Franco AM, Filson D. 2000. Knowledge diffusion through employee mobility. Research Department Staff Report 272, Federal Reserve Bank of Minneapolis. Available at SSRN: http://ssrn.com/abstract= 1223445. Freeman JH, Hannan MT. 1983. Niche width and the dynamics of organizational populations. American Journal of Sociology 88(6): 1116–1145. Freeman JH, Lomi A. 1994. Resource partitioning and foundings of banking cooperatives in Italy. In Evolutionary Dynamics of Organizations, Baum J, Singh J (eds). Oxford University Press: New York; 269–293. Garud R, Kumaraswamy A. 1993. Changing competitive dynamics in network industries: an exploration of Sun Microsystems’ open systems strategy. Strategic Management Journal 14(5): 351–369. Gilligan TW. 2004. Lemons and leases in the used business aircraft market. Journal of Political Economy 112(5): 1157–1180. Gort M, Klepper S. 1982. Time paths in the diffusion of product innovations. Economic Journal 92(367): 630–653. Greve HR. 2000. Market niche entry decision: competition, learning, and strategy in Tokyo banking, 1893–1936. Academy of Management Journal 43(5): 816–836. Hambrick DC. 1997. Corporate coherence and the TOP management team. Strategy and Leadership 25(5): 24–29. Handfield RB, Ragatz GL, Petersen KJ, Monczka RM. 1999. Involving suppliers in new product development. California Management Review 42(1): 59–82. Hannan MT, Carroll GR. 1992. Dynamics of Organizational Populations: Density, Legitimation, and Competiton. Oxford University Press: New York. Hannan MT, Freeman J. 1977. The population ecology of organizations. American Journal of Sociology 82: 929–964. Hannan MT, Freeman J. 1989. Organizational Ecology. Harvard University Press: Cambridge, MA. Haunschild PR. 1994. How much is that company worth? Interorganizational relationships, uncertainty, and acquisition premiums. Administrative Science Quarterly 39: 391–411. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

346

L. Pierce

Heeb R. 2003. Innovation and vertical integration in complementary markets. Journal of Economics & Management Strategy 12(3): 387–417. Helfat CE, Finkelstein S, Mitchell W, Peteraf M, Singh H, Teece D, Winter S. 2007. Dynamic Capabilities: Understanding Strategic Change in Organizations. Wiley-Blackwell: Malden, MA. Helfat CE, Lieberman MB. 2002. The birth of capabilities: market entry and the importance of pre-history. Industrial and Corporate Change 11(4): 725–760. Helfat CE, Peteraf MA. 2003. The dynamic resourcebased view: capability lifecycles. Strategic Management Journal, October Special Issue 24: 997–1010. Hendel I, Lizzeri A. 1999. Adverse selection in durable goods markets. American Economic Review 89(5): 1097–1115. Henderson RM, Clark KB. 1990. Architectural innovation: the reconfiguration of existing product technologies and the failure of established firms. Administrative Science Quarterly 35(1): 9–30. Hobson P. 2004. The many faces of leasing. Auto Remarketing 1 June. Holmstrom B, Milgrom P. 1991. Multitask principalagent analyses: incentive contracts, asset ownership, and job design. Journal of Law, Economics, & Organization 7: 24–52. Iansiti M, Levien R. 2004a. The Keystone Advantage: What the New Dynamics of Business Ecosystems Mean for Strategy, Innovation, and Sustainability. Havard Business School Press: Boston, MA. Iansiti M, Levien R. 2004b. Strategy as ecology. Harvard Business Review 82(3): 66–76. Johnson JP, Waldman M. 2003. Leasing, lemons, and buybacks. Rand Journal of Economics 34(2): 247–265. Kahn C. 1986. The durable goods monopolist and consistency with increasing costs. Econometrica 54(2): 275–294. Kanatas G, Qi J. 1998. Underwriting by commercial banks: incentive conflicts, scope economies, and project quality. Journal of Money, Credit & Banking 30(1): 119–133. Katz ML, Shapiro C. 1985. Network externalities, competition, and compatibility. American Economic Review 75(3): 424–440. Kelley Blue Book. 2000–2004. Kelley Blue Book Used Car Guide. Kelley Blue Book Co.: Irvine, CA. King AW, Zeithaml CP. 2001. Competencies and firm performance: examining the causal ambiguity paradox. Strategic Management Journal 22(1): 75–99. Klepper S. 2002. The capabilities of new firms and the evolution of the US automobile industry. Industrial and Corporate Change 11(4): 645–666. Klepper S, Graddy E. 1990. The evolution of new industries and the determinants of market structure. Rand Journal of Economics 21(1): 27–44. Klepper S, Simons K. 2000. The making of an oligopoly. Journal of Political Economy 108(4): 728–760. Klepper S, Sleeper S. 2005. Entry by spinoffs. Management Science 51(8): 1291–1306. Lim T. 2001. Rationality and analysts’ forecast bias. Journal of Finance 56(1): 369–385. Copyright  2008 John Wiley & Sons, Ltd.

Makadok R, Walker G. 2000. Identifying a distinctive competence: forecasting ability in the money fund industry. Strategic Management Journal 21(8): 853–864. Mannering F, Winston C, Starkey W. 2002. An exploratory analysis of automobile leasing by US households. Journal of Urban Economics 52(1): 154–176. McNamara G, Bromiley P. 1997. Decision making in an organizational setting: cognitive and organizational influences on risk assessment in commercial lending. Academy of Management Journal 40(5): 1063–1088. Metcalfe JS, Gibbons M. 1989. Technology, variety and organisation: a systematic perspective on the competitive process. In Research on Technological Innovation, Management and Policy (Vol. 4), Rosenbloom R, Burgelman R (eds). JAI Press: Greenwich, CT; 153–193. Mikhail MB, Walther BR, Willis RH. 2003. The effect of experience on security analyst underreaction. Journal of Accounting & Economics 35(1): 101–116. Mitchell W. 1989. Whether and when: probability and timing of incumbents’ entry into emerging industrial subfields. Administrative Science Quarterly 34(2): 208–230. Moore JF. 1993. Predators and prey: a new ecology of competition. Harvard Business Review 71(3): 75–86. Morton FS, Zettelmeyer F, Silva-Risso J. 2001. Internet car retailing. Journal of Industrial Economics 49(4): 501–519. Mosakowski E. 1998. Managerial prescriptions under the resource-based view of strategy: the example of motivational techniques. Strategic Management Journal 19(12): 1169–1182. Nelson RR, Winter SG. 1978. Forces generating and limiting concentration under Schumpeterian competition. Bell Journal of Economics 9(2): 524–548. Nelson RR, Winter SG. 1982. An Evolutionary Theory of Economic Change. Belknap Press: Cambridge, MA. Nickerson JA, Zenger TR. 2004. A knowledge-based theory of the firm: the problem-solving perspective. Organization Science 15(6): 617–632. Phillips DJ. 2002. A geneological approach to organizational life chances: the parent-progeny transfer among Silicon Valley law firms, 1946–1996. Administrative Science Quarterly 47(3): 474–506. Puri M. 1999. Commercial banks as underwriters: implications for the going public process. Journal of Financial Economics 54(2): 133–163. Purohit D. 1992. Exploring the relationship between the markets for new and used durable goods: the case of automobiles. Marketing Science 11(2): 154–167. Rauschenberg B. 2001. 2001 mid-year used vehicle report & forecast. Auto Remarketing 1 July. Available at: http://www.autoremarketing.com/ar/news/story. html?id=869. Roberts PW, Klepper S, Hayward S. 2007. Founder backgrounds and the evolution of firm size and scope. Working paper, Goizueta Business School, Emory University, Atlanta, GA. Available at: http://www.goizueta.emory.edu/faculty/PeterRoberts/ documents/WP-RelatedPFEandSS.pdf. Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj

Big Losses in Ecosystem Niches Romanelli E. 1989. Environments and strategies of organization start-up: effects on early survival. Administrative Science Quarterly 34(3): 369–387. Rosenbloom RS, Christensen CM. 1994. Technological discontinuities, organizational capabilities, and strategic commitments. Industrial and Corporate Change 3(3): 655–685. Ross DG. 2007. On bankers and their incentives under universal banking. Working paper, Stern School of Business, New York University, New York. Available at: http://finance.eller.arizona.edu/documents/se minars/2006-7/DRoss.BankerCompUB8 JMF01-07. pdf. Salinas P. 2000. Dealers profit from rising lease losses. Used Car News 7 August. Shaffer S. 1998. The winner’s curse in banking. Journal of Financial Intermediation 7(4): 359–392. Sorenson O. 2000. Letting the market work for you: an evolutionary perspective on product strategy. Strategic Management Journal 21(5): 577–592. Sorenson O, Audia PG. 2000. The social structure of entrepreneurial activity: geographic concentration of footwear production in the United States, 1940–1989. American Journal of Sociology 106(2): 424–461. Sutton J. 1998. Technology and Market Structure. MIT Press: Cambridge, MA. Teece DJ. 1986. Profiting from technological innovation: implications for integration, collaboration, licensing and public policy. Research Policy 15: 285–305. Teece DJ. 2007. Explicating dynamic capabilities: the nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal 28(13): 1319–1350. Teece DJ, Pisano G, Shuen A. 1997. Dynamic capabilities and strategic management. Strategic Management Journal 18(7): 509–533. Thaler RH. 1994. The Winner’s Curse: Paradoxes and Anomalies of Economic Life. Princeton University Press: Princeton, NJ.

Copyright  2008 John Wiley & Sons, Ltd.

347

Thompson P. 2005. Selection and firm survival: evidence from the shipbuilding industry, 1825–1914. Review of Economics and Statistics 87(1): 26–36. Tirole J. 1988. The Theory of Industrial Organization. MIT Press: Cambridge, MA. Tripsas M. 1997. Unraveling the process of creative destruction: complementary assets and incumbent survival in the typesetter industry. Strategic Management Journal Summer Special Issue 18: 119–142. Tushman ML, Anderson P. 1986. Technological discontinuities and organizational environments. Administrative Science Quarterly 31: 439–465. Vertex Consultants. 1997. Research on vehicle lease data in Canada and the United States, prepared for the Canadian Banks Association by Vertex Consultants Inc., October 1997. Available following DesRosiers Automotive Consultants Inc. report at http://www.fin.gc.ca/taskforce/research/pdf/rr7 e.pdf. Waldman M. 1993. A new perspective on planned obsolescence. Quarterly Journal of Economics 108(1): 273–283. Waldman M. 1996. Planned obsolescence and the R&D decision. Rand Journal of Economics 27(3): 583–595. Williamson OE. 1985. The Economic Institutions of Capitalism. Free Press: New York. Winter SG. 2000. The satisficing principle in capability learning. Strategic Management Journal October–November Special Issue 21: 981–996. Winter SG. 2003. Understanding dynamic capabilities. Strategic Management Journal October Special Issue 24: 991–995. Winton A. 1999. Don’t put all your eggs in one basket: diversification and specialization in lending. Working paper, Finance Department, University of Minnesota, Minneapolis, MN. Available at: http://fic.wharton.upenn.edu/fic/papers/00/0016.pdf.

Strat. Mgmt. J., 30: 323–347 (2009) DOI: 10.1002/smj