what do you think of Liquidnet
I think it's really, really neat, and meets a need which was once available in the markets and has since been whittled away over the years. That need is the ability for institutions - buyside entities - to trade amongst themselves without sellside participation. Called the "fourth market," other originally "buyside institution-only" systems have strayed from this mandate by permitting broker-dealers (dealing and proprietary) and hedge funds (usually momentum-oriented) access. The decision to broaden the types and number of firms who had access was a great boon to liquidity, but didn't permit buysiders to honestly, and fearlessly, post bids or offers for 100K+ size without the impact that comes from information leakage.
I think that eVWAP probably belongs on an active sellside trader's desk, and that the LiquidNet P2P interface belongs on the buyside trader's desk; this is to say, they both meet a need.
It's like this - again, in my humble opinion: where there is a bona fide, unique routing and/or execution proposition within an ECN, crossing system, or other alternative trading system, the sellside - and perhaps (but less likely and less advisably) - the buyside should want connectivity to it.
it seems as if the absence of a feature requiring those who transmit (or respond to) expressions of interest in trading a large block to actually trade [is a weakness]
I think that it is an Achilles Heel, but that there are ways to keep the system from being gamed. First, to set very clear guidelines on penalties for peeking and then backing away. Maybe a "one strike and you're out [of that issue via the ATS]" rule barring access to looks for that stock for, say...2 weeks.
There was a similar ATS that sought to do almost the same thing but with individual intermediaries that didn't, as I recall, workout - largely due to the buyside reluctance (this was four or five years ago, so attitudes might have shifted a bit) to working with anyone but their usual sellside contacts (again, relationships). I'd guess that a second approach to addressing those who don't "play fair" in a system like LiquidNet could be to add some superficial human facilitation.
They're some smart folks, and I'm sure they'll come up with more ingenious solutions than these.
Do you think it's true that there is, in effect, an honor code among traders that when a large block order from a buy side desk hits the market in a private auction amongst the top sell side traders, the party on the the buy side desk is supposed to let the sell side traders bidding on the block know what the true "total size" of the order is? I can't imagine the counterparty to that block being real happy if buys 1,000,000 shares without knowing that there are 4,000,000 more coming down the pipe.
You're right. If a block transaction is being negotiated, as I previously mentioned, one of the first sellside questions is, "Is this it, or will there be more behind it?" If you're asking if buysiders are generally honest when asked this, the answer is a resounding yes. Absolutely. They don't want to do anything that can ruin their sellside relationships' performance and, therefore, make them a less important customer.
And this is where the decision to take a loss once in awhile comes in. A great buyside client who's done lots of business with you over the years, has always been honest in circumstances like the above - hell, has even let you, the sellsider, off the hook once or twice for a screw-up over the years - sure: you'll take a hit for him once in awhile. That's something the Buffoonigans just don't get.
If a buyside entity were to misrepresent the shares they had coming to market, it would (a) - as you astutely pointed out - make for some furious counterparties who (again, how this business works) won't be there for you the next time you want help placing stock; (b) probably result in overall mishandling of the order because of the inability to move the whole thing at once and the information leaks risked in the meantime; and (c) will reduce the trust inherent, or that should be inherent, in the buyside/sellside relationship.
When sellsiders are bidding on a large block, they get two major pieces of data with which to formulate their bids. The first is the number of shares. The second is a factlist about the stock including its' beta, listed or NASDAQ, common or preferred, etc. The major bit of information withheld are the name(s) of the stock(s). It is sometimes easy to figure out, of course, knowing what issues has been active, who the major holders are (from Form ADV filings, and the like), what issues' blocks might be looking for new homes, etc.
And in all cases, extremely sophisticated software applications, some utilizing artificial intelligence like neural networks or fuzzy logic, can scan the data given, comb the universe of stocks, and spit out both (a) the likely issue(s) bidded for, and (b) what, given the profile of the stock, an adequate bid might be. As always, individual, human trader savvy is the ultimately decisive factor for a block bid.
LPS5 |
