Impact of Open-Source Community on Cryptocurrency Market Price: An Empirical Investigation

ABSTRACT

Although the prices of cryptocurrencies remained volatile for the past decade, the factors that impact the price dynamics of the new type of investment instrument have not been fully identified yet. In this study, we recognize the dual nature of cryptocurrencies, that is, being a software program and a financial instrument, and examine the impact of software advancement on the price dynamics of cryptocurrencies. The open-source software (OSS) platform functionality enables social behaviors that we use as signals. Using data from the largest OSS platform, we establish the connection between open-source activities and the price movement of cryptocurrency. In particular, as project popularity (forks and watches) and users’ feedback (issues) increase, the market price increases by 4.3 percent, 2.4 percent, and 4.4 percent per annum, respectively. On the contrary, the number of code corrections (pull requests) is negatively related to prices leading to a 5 percent annual price decrease. Our results suggest that OSS contributions create a perfect selection mechanism, where higher quality projects receive more developers’ attention and user feedback, whereas lower quality projects do not, thus creating the separating equilibrium.

KEYWORDS:

• Cryptocurrency
• open source software
• blockchain
• signaling theory
• OSS
• online platforms

Supplementary Information

Supplemental data for this article can be accessed online at https://doi.org/10.1080/07421222.2023.2267322

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

^1. Total amount estimated by authors based on [97].

^2. #51 most visited website in Finance/Other Finance category in the US (https://www.similarweb.com/, accessed June 14, 2023).

^3. #15 most visited website in Finance/ Investing category in the US (https://www.similarweb.com/, accessed June 14, 2023).

^4. Calculated by authors using CoinMarketCap data on August 31, 2018. Based on the data sample, 1,274 projects had their GitHub link posted on their page out of 1,934 cryptocurrencies listed.

^5. https://docs.github.com/en/get-started/quickstart/set-up-git (accessed June 14, 2023).

^6. Open-source Unix-like operating system (https://en.wikipedia.org/wiki/Linux, accessed June 14, 2023).

^7. See www.coinmarketcap.com (accessed June 14, 2023).

^8. See www.github.com (accessed June 14, 2023).

^9. Note that we refer to forking as the activity at GitHub. Social coding involves contributing to someone else’s project by creating a personal copy of someone’s code, and a fork is used to make changes without modifying the original code. Hence, forking on GitHub is merely a convenient instrument for developers to coordinate their work. Nevertheless, sometimes forked cryptocurrency projects can be turned into new cryptocurrencies (hard forks). For this to happen, software changes made via forking code from GitHub repositories require consensus from the blockchain network. Therefore, we distinguish between GitHub workflow forks and cryptocurrency hard forks. While forking in the OSS platform is a usual component of the workflow, forking GitHub does not necessarily entail forking crypto.

^10. Google’s serverless, highly scalable enterprise data warehouse. https://cloud.google.com/bigquery/ (accessed June 14, 2023)

^11. Figure 2 is constructed by the authors using the GitHub Archive data for the respective period.

^12. Daily closing price is reported as volume-weighted average of market pair prices for the cryptocurrency, denominated in USD, at 11:59 pm UTC of the corresponding date.

^13. Typically, the maximum lag length is selected based on the comparison of statistics that minimizes the trade-off between the model fit and the number of parameters from models with different lag lengths. We follow the Andrews and Lu [4] approach and select the optimal lag for the model selection based on J-criterion (similar to information criteria like BIC and HQIC). The third-order panel VAR was chosen. Selection procedure results are reported in Online Supplemental Table 1.5.

^14. The stability condition implies that the panel VAR is invertible and has an infinite-order vector moving-average representation, providing a known interpretation of estimated impulse-response functions [3, 96]. We use the pvarstable function to find the characteristic roots and check if they are equal to or greater than 1.

^15. We estimate the IRFs up to a 10-month lag. The impulse responses are derived from a Cholesky decomposition to orthogonalize the shocks, i.e., make them mutually uncorrelated.

Additional information

Notes on contributors

Mariia Petryk

Mariia Petryk ([email protected]) is an assistant professor of information systems and operations management in the School of Business, George Mason University. She received her PhD in Information Systems and Operations Management from the Warrington School of Business at the University of Florida. Dr. Petryk’s research interests span across information systems (IS), finance, management, and organization science, including emerging technology management (blockchain, artificial intelligence), digital and mobile platforms, and social networks. Her work appeared in Information Systems Research and at premier IS conferences, including Conference on Information Systems and Technology and INFORMS.

Liangfei Qiu

Liangfei Qiu ([email protected]) is the PricewaterhouseCoopers Associate Professor and UF Research Foundation Professor at Warrington College of Business, University of Florida. He also serves as the PhD coordinator for the Department of Information Systems and Operations Management. Dr. Qiu’s research focuses on social technology (social networks, social media, and prediction markets), platform technology (sharing/gig economy, e-commerce platforms, and healthcare analytics), and telecommunications technology. His research has appeared in premier academic journals, including Information Systems Research, Journal of Management Information Systems, MIS Quarterly (MISQ), Production and Operations Management (POM,), and others. He is an Associate Editor of MISQ and other journals, and a Senior Editor of POM.

Praveen Pathak

Praveen Pathak ([email protected] is the Robert B. Carter Professor in the Information Systems and Operations Management Department in the Warrington College of Business at the University of Florida. He received his PhD from the Ross School of Business at the University of Michigan. Dr. Pathak’s research interests include machine learning and its applications, blockchain technology, healthcare IT, and Web mining. His research has been published in various journals including Management Science, Information Systems Research, Journal of Management Information Systems, Journal of Operations Management, Decision Support Systems, and Journal of the Association for Information Systems.