Why Bitcoin will Fail to Scale

While Bitcoin has garnered enormous attention with its promises of inexpensive, fast and trust-less payments, it falls way short of the scale provided by banks. Bitcoin throughput is limited because its ledger accepts a block of fixed maximum capacity that can accommodate approximately 2200 transactions every 10 minutes. The Bitcoin community is actively considering technology upgrades to increase the block capacity. Intuitively, one would expect that increasing the block capacity would solve this scaling problem. However, in this paper, we show that increasing the block capacity would be futile for scaling Bitcoin.

We analyze the strategic interactions of miners, who are heterogeneous in their computational power, and users, who are heterogeneous in the value of their transactions, using a game-theoretic model. We show that a relaxation of throughput congestion can facilitate large miners to tacitly collude – artificially reducing the actualized throughput via the strategic partial filling of blocks to receive higher transaction fees. This strategic partial filling is sustained if the computing power of the smallest colluding miner is larger than a threshold. In addition, Bitcoin can only serve 50% of the demand if the colluding group power is above a threshold. We provide empirical evidence of such strategic partial filling of blocks by large miners of Bitcoin. 

We show that a technological intervention, such as banning large miners, can eliminate collusion. However, this also makes the system less secure. A strategic adversary faces a trade-off between earning money as a miner and carrying out a double-spend attack to steal a large amount of money in one shot. Such an attack becomes preferable as revenues from mining decrease in the absence of collusion. Therefore, surprisingly, collusion indirectly protects the Bitcoin system by increasing miner revenue. Overall, we show that there is an economic limit to the scalability of Bitcoin. On the one hand, increasing Bitcoin’s capacity invites collusion; on the other hand, if the collusion is suppressed, the system becomes insecure, which drives away the demand regardless. Our analysis raises antitrust concerns with respect to Bitcoin and suggests at least some of the anticipated competitive gains from decentralization in Bitcoin may be difficult to realize.

Workshop in Information Systems and Economics. San Francisco, CA, Dec 2018.

Workshop on the Economics of Information Security. Cambridge, MA, Jun 2019.

Marketing Science Conference. Rome, Italy, June 2019.

Bitcoin Disruption in Payments – Winners and Losers


Bitcoin is a cryptocurrency which allows financial transactions between users, independent of financial intermediaries such as a payment processing providers like Visa/Master Card. Bitcoin validates transactions through a distributed consensus mechanism which circumvents the need for a financial intermediary; this allows it to have potentially low transaction fees. Using a game theory model involving Bitcoin, traditional channels of verification, and users, we study how Bitcoin would disrupt the payments processing industry. There are a few subtle but important differences between the value propositions offered by Bitcoin and traditional channels. For example, post-transaction services (e.g., reversal of an unintended or fraudulent transaction), which are readily available on traditional channels are not present on Bitcoin. We show that owing to this difference, the presence of Bitcoin leads to a filtering effect - the proportion of transactions needing post-transactions services increases on the traditional channel in the presence of Bitcoin. Since it is costlier to process such transactions, the transaction fee charged by the traditional channel increases. 

We also find that Bitcoin pricing mechanism makes is more suitable for high value transactions as users compete fiercely to use its limited throughput. Developers have been investing significantly to increase capacity and achieve a solution that allows micro payments. We show limited throughput is not a technological shortcoming but a key property that keeps the platform stable. Overall, the transactions of small sizes, which are not suitable for Bitcoin and need to use Visa/Master Card, have to pay a higher transaction fee on the traditional channel in equilibrium. Therefore, paradoxically, even though Bitcoin increases competition in the payments processing industry, the users who need to process transactions through the traditional channels end up paying higher transaction fees.

Conference on Information Systems and Technology. Phoenix, AZ, Nov 2018. (Nomination for Best Conference Paper)

Marketing Science Conference. Rome, Italy, June 2019.

How to invest in Cryptocurrencies - A model of value capture in decentralized application ecosystem

The scale of investments raised by Initial Coin Offerings (ICO) has generated a debate on how to value Crypto coins and what ICO’s to invest in. We first present a contrast against traditional monopoly platform to highlight the potentially revolutionary benefits of decentralized Proof of Work (PoW) crypto platforms. A monopoly investor profits from collection of transaction fee revenue (Operational Value) in every period while a PoW investor extracts a one time Turnover Value by sale of tokens. Unlike a monopoly, PoW’s Turnover Value driven revenue mechanism does not incentivize investment in platform quality. PoW’s potential to expand user surplus by banishing monopoly control over transaction fee is diluted by a low investment in platform quality. Proof of Stake (PoS) which promises to eliminate PoW’s wasteful puzzle solving turns out to instead raise the fees for users. From an investors point of view PoS allows monetization of both Operational Value and Turnover Value, thus a preferable alternative in most cases. PoS platforms, as a middle ground between PoW and monopolies, are best bet to disrupt existing monopoly intermediaries.

We further assess partition of value between decentralized protocols (Bitcoin, Ethereum, EoS) and applications (Augur, CryptoKitties, uPort) on top of them. Our model supports the Fat Protocol hypothesis i.e. larger value capture in protocols. We independently show three contributing factors - first mover advantage in setting platform design by protocol investor, Store of Value characteristic of protocol tokens and direct network effects. We also present empirical evidence in the form of two prevalent phenomenon - Ecosystem Funds and Protocol Forking. Investors looking for the Amazon’s and Google’s of the decentralized era must look to high quality Proof of Stake protocols with a vision to either create or subsidize creations of applications.