Role of Latecomers’ Imitation in Overtaking Industry Leaders under Schumpeterian Competition

Role of Latecomers’ Imitation in Overtaking Industry Leaders under Schumpeterian Competition [Link to SSRN]


Since Schumpeter emphasized the importance of innovation for latecomers to leapfrog and overtake industry leaders, strategy research has viewed innovation as the primary impetus for leadership change. On the other hand, imitation in this literature has been portrayed as a strategy by which a latecomer can narrow the gap with, but not surpass, leaders. Thus far, little research has examined whether imitation can help latecomers with disadvantageous positions catch up with and eventually overtake leaders. By employing a computational model, we find that latecomers who pursue innovation alone are unlikely to overtake leaders, whereas latecomers who mix innovation and imitation are more likely to do so. When the latecomer allocates all its R&D budget to innovative R&D, it is likely to run out of capital quickly and be unable to continue to invest in R&D. A mix of innovation and imitation better maintains a steady cash stream, helping the latecomer continue to invest in R&D and thereby leaving room for leapfrogging opportunities.

Keywords: Schumpeterian competition, imitation, innovation, technological leadership change, difficulty of innovation

Does Toolkit Licensing Facilitate the Survival of Latecomers in the Presence of Industry Cyclicality?: Evidence from DRAM Industry 2005-2014

Toolkit Sharing and Industry Cyclicality: Evidence from DRAM Industry 2005-2014


Does toolkit license from incumbents facilitate the survival and growth of the latecomers? We argue that if industry cyclicality exists, toolkit licensing is detrimental to the survival of latecomers as well as low-performing incumbents. The DRAM industry, which is known as a cyclical industry, offers an unusual setting to test these competing hypotheses. During the period 2005-2014, Taiwanese companies experimented with a new business model, the DRAM foundry without an R&D unit. Since all these companies could not license new chip designs during the 2007-2009 downturn of the silicon cycle, we examine how an exogenous reduction of R\&D units during this period influenced firm-level performance. Using all the DRAM manufacturers’ profitability and technology data from iSuppli, DRAMExchange, and SEMI databases, we find that during a downturn, the void of the R&D units led to dramatically low profitability (-84%), which came from low systemic innovation which bridges product innovation and process innovation. The results highlight that the importance of product innovation may not decrease in a fast-changing technological environment.

Keyword: Schumpeterian competition, toolkit licensing, industry cyclicality, DRAM industry

Software Development Kits and Product Innovation

Software Development Kits and Product Innovation:
Modularity and Ecosystem Perspectives


We observe a growing number of software development kits (SDKs) are externally available. We explore how using externally available SDKs affects the product innovation process. We characterize (1) product development by using externally available SDKs as solving nearly modular problems and (2) product development by building inhouse SDKs as solving an integral problem (i.e., relatively non-modular). We explore the video game industry, in which software development kits are called game engines, and game developers explore theme innovations (i.e., non-technological dimension) as well as technological innovations. Findings suggest that on average, using commercial game engines facilitates module-level innovations but less likely to introduce system-level innovations which require a cross-module coordination. Also, the number of commercial game engine users is an important predictor of product innovation. If when the number of users is not sufficiently large, its weakness in system-level innovations exacerbated, and the strength in module innovation is also weaker, showing that harnessing the power of positive feedback (i.e., ecosystem effect) between the number of users and the quality of engines matters.
Keywords: Software Development Kits, Product Innovation, Modularity, Ecosystem

Technological Opportunity and Technological Leadership Change

Technological opportunity, technological leadership change, and the latecomer’ R&D resource allocation between innovation and imitation


This study examines when and how latecomers can surpass incumbents in technological capabilities with a focus on the role of technological opportunity. There is a disagreement in theoretical prediction and evidence on whether technological opportunity is conducive to the change of technological leadership. To reconcile this disagreement, we build a computational model on Schumpeterian competition in which incumbents and latecomers compete with innovation and imitation R&D. First, results suggest that technological opportunity indeed has the two opposing effects (i.e., positive and negative) on technological leadership change, and these effects create an inverted-U relationship. When there are few opportunities, leadership change is unlikely to happen because latecomers will hardly come up with a breakthrough. Also, abundant opportunities may not be conducive to leadership change either because incumbents move forward faster than latecomers. We further examine why these opposing effects exist by exploring latecomers’ R&D allocation between innovation and imitation. Results highlight that imitation R&D is a necessary condition for latecomers to leverage technological opportunity (i.e., enabling the positive effect of technological opportunity) and overcome their disadvantages under technologically munificent environment (i.e., mitigating the negative effect of technological opportunity).

Keyword: Schumpeterian competition, Technological opportunity, Technological leadership change, Imitation, Innovation

Two Faces of Decomposability in Search: Evidence from the Recorded Music Industry 1995-2015

Two Faces of Decomposability in Search:
Evidence from the Recorded Music Industry 1995-2015


We propose that decomposability may generate a trade-off across different stages of search. We compare (1) decomposed search, the process of searching by producing a decomposed module, and (2) integrated search, the process of searching by producing a full-scale product. In the variation generation stage, decomposability can allow firms to experiment with more alternatives at the same time than an integrated search. However, in the selection and retention stages, a decomposed search may be more vulnerable to imperfect evaluation than an integrated search. It may increase the chance of missing out on promising alternatives after the first evaluation because the low cost of a decomposed search makes firms less committed to each alternative. We test our theory with a unique empirical setting, the recorded music industry, where singles (i.e., decomposed products) and albums (i.e., integrated products) have coexisted since the early twentieth century. In the variation generation stage, single-producing firms experiment with 35.22% more new artists than album-only-producing firms. In the selection and retention stage, single-producing firms are 69.57% more likely to neglect top-tier artists who failed in their first releases because single-producing firms have a higher performance target (i.e., lower commitment) than album-only-producing firms.

Keywords: Decomposability, Evolutionary Perspective on Search, Behavioral Theory of the Firm, Alternative Evaluation



Music Databases and APIs


Here is the link to the list of online music database in Wikipedia.

Also, below is a summary of music databases I have worked on. I added some notes.

MusicBrainz Database : Free and data dumps are available. (PostgreSQL), a good starting point

The Echonest API : Rosetta Stone project -> mapping multiple music APIs’ ids -> helping integrate multiple databases and APIs.

Spotify Web API : popularity data and user listening history, will integrate the Echonest API soon.

Discogs: contains label information.( -> Data dumps are available.)

Quantone: (Fee) Mapping ids among spotify, discogs, and musicbrainz ids.

7Digital API: Music audio samples are available.

LastFM API: airplay data on the internet radio

Nielsen Soundscan: Not free. Traditionally used for music sales. (info from Wikipedia)

Billboard Chart Weekly Top 100: Terminate API services in 2013, chart rankings are based on Nielsen Soundcan (some scraped data from

Million Song database: data on music features

Musixmatch: data on lyrics


Dynamics of Imitation versus Innovation in Technological Leadership Change: Latecomers’ Catch-up Strategies in Diverse Technological Regimes


We examine the role of latecomers’ optimal resource allocation between innovation and imitation in latecomers’ catch-up under diverse technological regimes. Building on Nelson and Winter (1982), we develop computational models of technological leadership change. The results suggest that one-sided dependency upon either imitation or innovation deters technological leadership change. At an early stage with low-level technologies, latecomers should focus on imitation; then, as the technological gap decreases, they should allocate more R&D resource to innovation. We also examine the role of several variables, such as appropriability, cumulativeness, and cycle time of technologies (CTT), as related to technological regimes. The simulation results show that while low appropriability tends to increase the probability of technological leadership change, it makes imitation a more effective strategy compared to innovation; in addition, while a higher level of cumulativeness tends to reduce the probability of leadership change, it makes imitation a more valuable option because innovation becomes more difficult for latecomers. We also find an inverted U-shaped relationship between the CTT and the probability of technological leadership change. When the CTT is short, it makes sense for latecomers to allocate more resources to imitation, especially when their technology level is initially low.

Number of Pages in PDF File: 47

Keywords: latecomers, technological leadership change, imitation, innovation, technological regime, catch-up

JEL Classification: O31, O32

[Link to the paper on SSRN]

Growth Logics: Market vs. Technological Relatedness in Product Entry

Growth Logics: Market vs. Technological Relatedness and the Direction of Organizational Growth

Sungyong Chang,            J.P. Eggers, and Daniel Keum
Columbia Business School         NYU Stern School of Business

[ SSRN Working Paper Version]


How do technological and market knowledge drive organizational expansion decisions? Extending research on knowledge, resources, and organizational growth, we suggest that the relatedness of the firm’s market resources and knowledge to a potential market for entry will better predict entry behavior than technological relatedness. Using cross-industry data with granular product entry information covering over 2,681 product segments, we find robust evidence that market relatedness is an independent and significant predictor of product market entry, while technological relatedness does not predict entry. Due to the valuable but non-tradable nature of market resources, firms must enter new markets to capture that value, while firms can capture the value of technological resources without entry. These findings contribute to ongoing discussions about the directions of organizational growth within the Resource Based View of the firm and the role that markets for technology are playing in reshaping organizational boundaries.


Key words: firm knowledge; product market entry; diversification; resource-based view; markets for technology


Figure 1. Histogram of Technological Relatedness
between Firms and Product Categories



Global Diversification Discount and Its Discontent

Global Diversification Discount and Its Discontents:
A Bit of Self-selection Makes a World of Difference

Sungyong Chang, Bruce Kogut, and Jae-Suk Yang

Columbia Business School, Columbia University

Forthcoming, Strategic Management Journal

[SSRN Accepted Paper Version]


The documented discount on globally diversified firms is often cited, but a correlation is not per se evidence that global diversification destroys firm value. Firms choose to globally diversify based on their firm attributes, some of which may be unobservable. Given these exogenous firm attributes, the decision to diversify globally is endogenous and self-selected. Using the same specifications save for the Heckman selection instrument, our results contradict past research that did not address endogeneity. We posit that the global premium should reflect the value of multinational operating flexibility. We use the 2008-2009 financial crisis as creating exogenous variation to permit a test for the positive change in firm valuation due to global diversification. During and after the 2008-2009 financial crisis, the premium associated with global diversification became larger and more significant than before the 2008-2009 financial crisis. The churn of subsidiaries entering and exiting countries increased during the crisis, pointing to the value of an operating flexibility to restructure the geography of the multinational network. In all, the results contradict past findings and finds evidence that operating flexibility is more valued during times of high volatility, thus generating the diversification premium.

Keywords: Global diversification; Self-selection; Operating flexibility; Financial crisis

Figure 3.  Country-level Valuation Effect of Global Diversification

Panel A. With Controlling Self-selection


Panel B. With Controlling Self-selection


Note: Color shows t-values. (Blue-Premium; Red-Discount, Grey-No data)

Roles of Giant Cluster in Innovation


Roles of Giant Clusters in Knowledge Diffusion and Recombination

Sungyong Chang,

Columbia Business School, Columbia University

Jeho Lee and Jaeyong Song

 Graduate School of Business, Seoul National University

[SSRN Working Paper Version]


      In this paper, the inventor collaboration networks in the semiconductor division of Samsung are examined during the period from 1982 to 2006. The data analysis demonstrates that in the beginning, the collaboration networks were made up of unconnected small clusters corresponding to subunits. In the 1990s and onwards, a giant cluster made up of bridges connecting previously unrelated small- and medium-sized clusters emerged. However, hubs or extremely well-connected individuals are absent. To identify the role of the giant cluster in facilitating innovation, we develop computational models. Two roles are identified: knowledge integrator and knowledge reservoir. The giant cluster acts as a knowledge integrator by facilitating knowledge diffusion between previously unconnected clusters. As the cluster grows and more ties are formed, knowledge diffusion increases monotonically. However, in the case of knowledge recombination, performance does not increase monotonically with additional bridges. With too many bridges or hubs, knowledge recombination performance declines because giant clusters cannot keep diverse knowledge for future use—i.e., it is unable to play a role of knowledge reservoir. We find that only giant clusters with a modest proportion of bridges can act as a knowledge reservoir and facilitate innovation by better preserving diverse knowledge.

Key words: Network; Knowledge; Evolution; Innovation