February 23rd, 2018, 5:02 pm
Capital structure: Without a central entity that can vet and legally obligate counterparties to behave, the need for escrowed collateral goes up and the availability of margin goes down. If one has $X in cash for trading, the number or size of positions might be much smaller in a DLT market than in centrally managed market. Moreover, high-quality traders will prefer centrally managed markets where they get favorable terms based on their identities and credit worthiness in a known legal jurisdiction. That means that DLT markets will have a larger portion of low-quality, under-capitalized traders. It would quite be an Akerloff lemon situation, but there will be some assortment by participant quality.
No doubt "settlement finality" can be improved although non-POW seems prone to other risks of fraud, flooding, gaming, etc. And it would seem to be very hard if not impossible to have both a distributed system and hard real-time performance guarantees.
Capacity bounds and transaction cost issues seem a bit more intractable to me. If you can find a trustworthy central marketplace provider thats running 1 copy of the ledger on 1 CPU, they are always going to be faster and more energy efficient than some amorphous network of N copies of the ledger running on N CPUs with all the added overhead of network synchronization between them. BitCoin's initial promise of disintermediating traditional financial services firms and offering lower transaction costs was based on BitCoin not offering the same services such as dispute resolution and repudiation of transactions. Otherwise BitCoin would always be more expensive.
Obviously, the key word is "trustworthy" and that's no small hurdle. There will be competition between centralized versus decentralized system with centralized touting speed + consistent performance and decentralized touting trust assurance + access to a broader array of untrustworthy counterparties.