Liquidity Providing Models

NikEy · Post by **NikEy** » Fri Apr 24, 2009 12:00 pm

Hi,

I know a few people who are employing trading strategies that are based on liquidity provision (they are not MMs). The idea is as follows:

There exist participants who put on market orders in large chunks (usually at the close, which I guess is due to banks hammering the market at the expense of their clients). Potentially this could lead to an order eating its way through the whole book. By anticipating such orders in certain stocks, the strategy would then aim to buy/sell the asset at a deflated/inflated price by putting on large limit positions on either side of the book with a price at the outer range of the quotes. One then aims to make a profit by offsetting the position the next day using an iceberg or another market impact reducing technique. This of course implies that the market order that triggered it is not based on a corporate event, but rather on a institutional portfolio reallocation or a strategy related execution (such as a reweighting on GSCI).

The sharpe on this strategy ranges between 4 and 8 apparently. Given that I have heard about this strategy from a few independent sources by now I am tempted to believe that there must be some papers circulating about this. Anybody has heard of this before? Anybody could point me in the right direction? Every suggestion is highly appreciated

Regards

Baltazar · Post by **Baltazar** » Fri Apr 24, 2009 1:16 pm

It is not that different than market making with wide spreads.

I think liquidity provision is market making. Sometimes or always, aggressively or more widely but in the end...

Avellaneda has a paper on that i think.

FDAXHunter · Post by **FDAXHunter** » Fri Apr 24, 2009 1:38 pm

The difference between a market maker and a liquidity provider is purely one of obligation. A market maker is obliged to quote (with a maximum spread and a minimum size) either by exchange authorities or effectively through the policy of the institution (you can't say I won't make you a market and expect the customer to call back), otherwise they will lose their status (and business).

In short: Liquidity providers are essentially market makers who can do as they please.

ThomasJ02 · Post by **ThomasJ02** » Fri Apr 24, 2009 3:27 pm

If this strategy was being traded, it should be observable. Look for next-day reversals on stocks that had large movements near the close on the previous day.

sharpend · Post by **sharpend** » Tue Apr 28, 2009 4:30 am

The strategy is more common than lice in kindergarten and old as the hills.

there is a certain amount situational awareness needed for good performance. not sure why you would want a paper.

NikEy · Post by **NikEy** » Tue Apr 28, 2009 8:31 pm

Thanks for the inputs. Well, the strategy sounds simple enough, but to be honest I expected it to be more complicated and thought that a paper would address this more adequately. I agree that this is something one can observe in the market, however, backtesting it on paper seems impossible, since merely interacting with the market in such a manner would change the participants behavior.

Thanks a lot again though

naked · Post by **naked** » Thu Apr 30, 2009 9:50 pm

"If this strategy was being traded, it should be observable. Look for next-day reversals on stocks that had large movements near the close on the previous day."

Would the observation be the other way around?

ie Sellers come in heavy at the close. Liquidity is provided, so shares/contracts are bought.

Next day, you would need to liquidate your long position -- add supply -- and as a result, prices trade lower.

ThomasJ02 · Post by **ThomasJ02** » Thu Apr 30, 2009 10:52 pm

Remember the sellers are also net short though.

Let's say there's an equilibrium price for the asset at E. Before the close, there's a price shock that brings the asset down to E-s (the sellers that come in heavy at the close). The next day, the price ought to recover to become closer to E, since the long-run value of the asset hasn't changed.

Intuitively, you can see this as the sellers covering their short position by buying back from the MMs, although you may actually have something more like fundamental investors buying the asset because it's at a discount.

Another way to think about it is that on average the price must recover, because otherwise the MMs would lose money on all of their transactions (they bought high, sold low) and go out of business.

FDAXHunter · Post by **FDAXHunter** » Fri May 01, 2009 3:30 pm

ThomasJ02: Another way to think about it is that on average the price must recover, because otherwise the MMs would lose money on all of their transactions (they bought high, sold low) and go out of business.

Interesting line of reasoning. Correspondingly then there can't be a credit crisis, because otherwise all the banks would lose money on most of their credit transactions (buy high, recover low) and go out of business.

ThomasJ02 · Post by **ThomasJ02** » Fri May 01, 2009 8:05 pm

But on average, across many years, the banks probably _have_ made money on their credit transactions, even if occasionally they all lose a big chunk. I'm not saying that the MMs won't lose money on any particular day, or even that they won't lose large amounts of money on any particular day. The bid/ask and commissions are also compensation for the inventory risk that they bear.

But saying that banks occasionally lose money is different from saying that MMs lose money whenever there is a price shock, which happen anytime anyone makes a sale or purchase that moves the asset price at all. If MMs lost money anytime the market moved, they would close down about an hour after they opened for business.