Mike Horia Teodorescu - Faculty & Research - Harvard Business School
Photo of Mike Horia Teodorescu

Mike Horia Teodorescu


Doctoral Student

I am currently in transition to an Assistant Professor position at a different institution after the HBS graduation; please use my alumni email hmteodor [at] post.harvard.edu to reach me in the interim this summer. My research interests are below. Thank you for visiting.

Mike Teodorescu studies innovation, machine learning, entrepreneurship, and international business. Mike's research interests include startup strategy, innovation within firms and across borders, applications of machine learning to business strategy, and MNC strategy. Mike's teaching interests include the applications of Big Data to management problems, business analytics, technology strategy, and intellectual property policy. He is an award winning teacher and has teaches at both undergraduate and graduate (masters of science and MBA) levels. In his teaching interests Mike wishes to make the concepts and fundamental methods of big data and machine learning accessible to everyone, as analysis of large datasets is becoming essential for firm strategy in ever growing fields.

Mike is currently a research affiliate of the U.S. Patent and Trademark Office (within the Office of the Chief Economist) where he studies the effects of uncertainty on startup outcomes (job market paper research) as well as serves as a machine learning/big data resource. Throughout all three of his doctoral papers, Mike combines econometric methods with new approaches from Machine Learning, Natural Language Processing, and broadly computer science methodologies. He hopes to act as a bridge between the fields of economics and computer science through his academic work.  

Prior to joining HBS, Mike has worked as a software engineer at Microsoft on several major software products. His research interests prior to HBS included artificial intelligence and applications of computer algorithms to the sciences. He is the author of several patents in technology fields (US Pat. 7,822,181 B2; US Pat. 8,143,607 B2; RO Pat. 123562) and co-author of several software related patent applications. Mike has several publications in engineering and computer science, primarily clustered around the field of artificial intelligence, and secondarily applications of AI to medicine. Mike has served on boards of several startups, and has been the first in business to support, fund, and create a successful business plan for the award winning 'operating room in a backpack' startup out of MIT's D-Lab - SurgiBox. Mike serves on the Board of the Summer Science Program, a leading scientific education nonprofit established in 1959 as a collaboration of universities (MIT, Caltech, Purdue, NM Tech, and Harvey Mudd). Prior to Harvard, Mike was the first worldwide to win the international NASA Space Settlement Design Contest twice, and has published some of his findings from that experience in aerospace engineering outlets. Mike also serves as a Harvard College Alumni Interviewer and lends his business experience through volunteering in his home community. He holds several awards for inventions as well.

Mike's work has most recently been featured in an article on the future of machine learning in business, featured in McKinsey Analytics, IOT Solutions, HBS Working Knowledge  for an article on methods in machine learning, and other outlets. Mike's engineering work has appeared prominently on Romanian news outlets, including TV stations ProTV, TVR1, Antena1, and several local and national Romanian newspapers.

Mike received a bachelor’s degree cum laude in Computer Science with High Honors in field from Harvard College in 2011 and wrote a thesis on swarm intelligence. He is an IEEE member, AOM member, SMS member, and INFORMS member. Mike holds a student license from the FAA and works on his piloting skills in his free time (flying four seat airplanes). Mike also loves biking and traveling to learn about cultures in various parts of the world. Mike holds US citizenship.

Working Papers
  1. The Need for Speed: Effects of Uncertainty Reduction in Patenting

    Mike Horia Teodorescu

    Patents are essential in commerce to establish property rights for ideas and to give equal protection to firms that develop new technologies. Young firms especially depend on the protection of intellectual property to bring a product from concept to market. However, the market for technology ideas has been recognized as an inefficient market in the management and economics literatures. While information asymmetry and expropriation risks have been studied extensively, the question of the effects of pre-patent grant uncertainty on firm outcomes remains open. This paper introduces a novel analysis based on internal US Patent and Trademark Office databases, exploiting an exogenous shock to startup firms from a previously unstudied executive action involving reduction of patent pendency (time from application to patent decision) for green technology patents. The aim of the paper is to determine whether reduced patent pendency improves firm outcomes for startups and to explore its implications. The findings are that treated startups (with accelerated patenting) have increased sales (by 30%), greatly increased venture funding (50%) and increased employment (over 25%). The paper also introduces a novel method for constructing a control group using a classification algorithm rooted in natural language processing, which can be used in conjunction with traditional econometric approaches such as difference-in-differences analysis beyond the topic of this paper.

    Keywords: patents; Startups; natural language processing; Machine learning; Patents; Business Startups; Risk and Uncertainty; Outcome or Result; Green Technology Industry;

    Citation:

    Teodorescu, Mike Horia. "The Need for Speed: Effects of Uncertainty Reduction in Patenting." Working Paper, September 2017. (Job Market Paper.)  View Details
  2. Machine Learning Methods for Strategy Research

    Mike Horia Teodorescu

    Numerous applications of machine learning have gained acceptance in the field of strategy and management research only during the last few years. Established uses span such diverse problems as strategic foreign investments, strategic resource allocation, systemic risk analysis, and customer relationship management. This survey article covers natural language processing methods focused on text analytics and machine learning methods with their applications to management research and strategic practice. The methods are presented accessibly, with directly applicable examples, supplemented by a rich set of references crossing multiple subfields of management science. The intended audience is the strategy and management researcher with an interest in understanding the concepts, the recently established applications, and the trends of machine learning for strategy research.

    Keywords: Machine learning; natural language processing; classification; Decision Trees; strategic decisions; Strategy; Research; Information Technology;

    Citation:

    Teodorescu, Mike Horia. "Machine Learning Methods for Strategy Research." Harvard Business School Working Paper, No. 18-011, August 2017. (Revised October 2017.)  View Details
  3. Knowledge Flows within Multinationals—Estimating Relative Influence of Headquarters and Host Context Using a Gravity Model

    Prithwiraj Choudhury, Mike Horia Teodorescu and Tarun Khanna

    From the perspective of a multinational subsidiary, we employ the classic gravity equation in economics to model and compare knowledge flows to the subsidiary from the MNC headquarters and from the host country context. We also generalize traditional economics gravity models to include new distance measures and a similarity measure novel to the management literature and test our theoretical predictions using a hand collected dataset of U.S. patents filed by the top 25 patenting multinational firms headquartered in the US. We find that as the size of the subsidiary grows, the host country’s influence onto knowledge flows into the subsidiary grows faster than the influence of the headquarters. We also present some evidence that immigration policy positively affects knowledge flows.

    Keywords: multinationals; knowledge flows; Cosine Similarity; immigration; gravity model; Multinational Firms and Management; Knowledge Dissemination; Business Headquarters; Immigration;

    Citation:

    Choudhury, Prithwiraj, Mike Horia Teodorescu, and Tarun Khanna. "Knowledge Flows within Multinationals—Estimating Relative Influence of Headquarters and Host Context Using a Gravity Model." Working Paper, July 2017.  View Details
  4. The Attenuating Effect of Banking Relationships on Credit Market Disruption

    Stefan Dimitriadis and Mike Horia Teodorescu

    This article examines how the relationship between banks and corporations moderates the effect of credit market disruptions. The 2008-09 financial crisis led to a dramatic restriction in the supply of credit to corporations via the syndicated loan market (Chodorow-Reich 2014; Ivashina and Scharfstein 2010a). We examine whether this shock was moderated by the strength of pre-crisis relationships between corporations and banks. By accounting for the strength of lending relationships in Chodorow-Reich’s (2014) baseline model of lender-borrower relationships we find partial evidence that relationships attenuate credit shocks. Specifically, we find that corporations that were more closely tied to banks suffered weaker credit shocks during the financial crisis. Otherwise stated, the strength of the corporation-bank relationship attenuated the effect of the credit shock. This result contributes to the literature on the effects of credit disruptions on firms and the literature on lending relationships between firms and banks. Furthermore, through careful analysis of the Chodorow-Reich model, we were able to determine that an instrument used in the literature may not, in fact, be appropriate for small sized banks.

    Keywords: Banks and Banking; Relationships; Financial Markets; System Shocks; Banking Industry; United States;

    Citation:

    Dimitriadis, Stefan, and Mike Horia Teodorescu. "The Attenuating Effect of Banking Relationships on Credit Market Disruption." Working Paper, July 2016.  View Details
Cases and Teaching Materials
Refereed Articles - Engineering
  1. A Collective Biological Processing Algorithm for EKG Signals

    Mike Horia Teodorescu

    We establish and explore an analogy between hunting by packs of agents and signal processing. We present a version of adaptive ‘Hunting Swarm’ algorithm (HSA), apply it to EKG signals, and investigate the influence of the model parameters on the filtering of stationary and nonstationary noise. We show that results obtained with the HSA filter may outperform results obtained with several other filters.

    Keywords: artificial intelligence; Technological Innovation; Health Care and Treatment;

    Citation:

    Teodorescu, Mike Horia. "A Collective Biological Processing Algorithm for EKG Signals." Proceedings of the International Conference on Bio-inspired Systems and Signal Processing 4th (2011): 413–420. (IEEE BIOSIGNALS 2011.)  View Details
  2. Effects of Pseudo-Lensing and Pseudo-Dispersion in Curved Radiation Shields and Collimators: Effects on Measurements

    Mike Horia Teodorescu

    Citation:

    Teodorescu, Mike Horia. "Effects of Pseudo-Lensing and Pseudo-Dispersion in Curved Radiation Shields and Collimators: Effects on Measurements." In Sensors for Harsh Environments III. Vol. 6757, edited by Hai Xiao and Anbo Wang. Proceedings of SPIE—the International Society for Optical Engineering. Bellingham, WA: SPIE, 2007.  View Details