There are several views of what constitutes "decentralized". Often, people will confuse three independent design factors:
To me, the practical question is, "How many unique individuals do you have to coerce or seduce to gain control of the block chain?"
With Bitcoin, over half the hash power is vested in just 5 entities who sign over half the blocks and therefore control the chain.
With BitShares, no signer can sign more than 1% of the blocks, so it takes 51 of them to control the chain, no matter what. These individuals must have well-known reputations to get elected. Reputations they can lose if they misbehave.
The holders of bitcoins have no say in this process - its up to those who provide hashing, i.e. the coin's "employees". Market forces drive them to support a very small number of "most profitable" signers.
The holders of BitShares have all the say in the process. The signers with the top 101 most BitShareholder approval take random turns signing the blocks.
So, by this metric BitShares is much more decentralized than Bitcoin.
But our position is that both are "decentralized enough" because it is the transparent public ledger that matters. You can detect bad behavior and theoretically eliminate bad actors (if your chain has a mechanism to do so).
With POW coins, you can't remove hashing power from its owner, no matter how badly they behave. Theoretically, a bad mining pool operator can have its hashing power removed from it by those using the pool. However, in practice, this does not seem to happen. For example, I have seen the most profitable pool in a POW chain continue to be favored even when it was not including any transactions at all in its blocks.
With BitShares, not only can BitShareholder's detect it,
they can do something about it by removing approval from a bad actor. Since there is no economic incentive for continuing to support a bad actor (like there is in mining pools), they don't last long. Even while waiting for shareholders to wake up and fire them, they can't sign more than 1% of the blocks.
There is also the cost of decentralization to be considered. Bitcoin spends hundreds of millions of dollars per year to spread its control across just 5 pool operators. BitShares spends about 1% of that to spread control across 101 delegates chosen by all BitShareholders.
So, to answer your OP question: BitShares is "decentralized enough" to be safe while remaining efficient enough to earn a profit for its BitSharesholders.
Now the point of comparison becomes speed of operation (e.g. 10 second confirmations) and all the things a particular chain can do (like run a decentralized exchange and issue stable smart currencies).
That ought to be enough to merit some attention, no?