But that ignores the very reason exchanges exist: To provide a convenient, popular, trustworthy, efficient marketplace. Stocks are more than ledgers, they are an entire economy themselves.
This economy is made useful (and profitable) by Exchanges, Investments Banks, Investors, and the Businesses/Financial Instruments made available by and for those parties.
In addition to having this popular marketplace, made of these (centralized) powers, the separation, and expectation of having those powers in place, provides an environment where all 4 parties' interests can be aligned, and thus thrive.
If you break all this into pieces (decentralization), you lose most of benefits of this system, and swap them for only a literal protocol that takes its place.
Everywhere you look in the world, you will find systems defending my argument. In nature, in finance, in society, in politics, it's the very way of life and the universe itself.
Centralization is easy, decentralization is hard. That's why people flock to centralized solutions. So, although it's a bit of an exaggeration to say that the systems "defend" your argument (they don't, they are just not good enough), I partially agree with you. But we should keep things in perspective and calculate the trade-off's. I remember Nefario making the same argument about a centralized stock exchange.
Companies can keep track of shareholders easily in an automated manner. And accredited investors can run centralized pass-throughs. This is very doable without resorting to future tech. So, let's discuss the downsides.
Convenience: Agreed. Shareholders would have to keep track of issuing companies and relevant escrows. A proper decentralized system that doesn't need an escrow and can harvest a list of issuers will solve this completely and would be even more convenient than having accounts in multiple exchanges. However, in the current atmosphere, it is even less convenient to invest through a centralized exchange, so to me the issue is moot. Will we have risk-free legal exchanges sooner than the proposed decentralized solution? I doubt it.
Popularity: Not sure about this one. Any solution can become popular. If all companies kept their own list of shareholders using the same protocol, there would still be centralized directories and easy to use tools to trade their shares. This isn't as hard a problem as creating a completely decentralized exchange system.
Trust: Disagreed. Is the trust to current exchanges satisfactory? My trust to friedcat is higher than friedcat+exchangeoperator+PToperator. This should be obvious. Although, until we have some system that works without an escrow, there is that.
Efficiency: Agreed. Centralized solutions will always be more efficient. No doubt about that.
Bitcoin actually is a good example.
Bitcoin itself is not as simple to put in this box, as bitcoin is first a protocol, second a commodity, and third a currency. Arguably none of those things are centralized nor decentralized, as they are concepts, not services or products or the like.
We are certainly talking about Bitcoin as a solution to currency. The protocol, the commodity and the currency are all alternatives to centralized systems, so it fits perfectly. It also fits to your above criteria. Central solutions are far more efficient. Transactions are instant, and they can implement any solution to any problem that may arise almost immediately without having to invent a new technology. The protocol is extremely resilient, because you don't need to implement Byzantine fault tolerance. So on and so forth.
Yet, Bitcoin thrives. I have no doubt that people will always prefer a decentralized stock exchange whenever a proper one is developed. Until then, all companies should have a relationship with the shareholders directly. Legal downsides of this is up for discussion, however.
However decentralized bitcoin may be in and of itself, it did not thrive with order. This forum as a communication point, Gox as an exchange point, WOT, Reddit, Havelock, etc, etc, etc. They all have their flaws, but they all serve and continue to serve their purpose.
My argument is simply that a decentralized solution is a unicorn. The moment you think you can grasp it, it will slip away. All of the decentralized concepts ultimately rely on ordered powers to facilitate their usefulness. Maybe they are different areas of centralization, or smaller groups, or higher quantity groups, but any amount of a system being smaller parts, provides more points of failure, more points needing trust, etc.
Without centralized forces, people may have never heard of TAT Investments, nor Neo. We may have never even met to come together for the IPVO. Neo wouldn't have raised enough coin in the "wild" decentralized system, and TATI would not have a reputation as clear as it does currently.
I have already stated the advantages of centralized powers, but don't confuse this with me supporting the concept in its extreme. The answer is always somewhere in between; I simply do not believe true decentralization is possible or desirable. Same as how chaos and anarchy are not practical or desirable. Many people arguing for decentralization are merely describing a
different system (usually with more, smaller parts), not a decentralized one.