Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

There is still something that nags me.

I've heard a bunch of explanation about liquidity and how HFT allows for large orders to be fulfilled, but it seems like this is quite the opposite.

What purpose does this serve? Is society as a whole better off when Jill is able to make this .05 per share more? I wouldn't frame the debate as 'god given rights' and 'competitive advantage'. What I really want to know is why a society where trades and quotes happen on a millisecond scale is better off than one where they happen on a second scale.

It's an honest question. Someone please convince me.



What's great is that we don't have to engage in thought experiments about what would happen if we didn't have trading activity with fast computers on a millisecond scale - we can just look back to any time before the 1990s, when most market making was done by humans, on human time scales.

Before 2001 the minimum tick size on any exchange was 1/16th of a dollar ($0.0625) and before 1997 it was 1/8th ($0.125), so the absolute minimum you would pay for a round trip (buying a stock and later selling it) was that much. Frequently, the bid-offer spread would be many ticks wide, so you could easily be paying $0.25 or $0.50 for each round trip.

The current minimum tick size is $0.01, and there are many stocks which trade at that level. Even if you suffer $0.05 of slippage on a round trip, you're still better off than you would have been under the old regime.

In the old regime, instead of high frequency traders, you had floor brokers who would work orders. Fortunately, floor brokers were paragons of virtue and morality, who would certainly never front run their clients orders, and would take any trade even if it worked to their disadvantage (NB in case you don't get it - this is sarcasm. In the 1987 crash, most brokers wouldn't even pick up their fucking phone because too many people were trying to sell stock, and the brokers didn't want to buy).

I honestly find it hard to believe that some people think that was better than what we have today.

---

Edit: The other thing I don't get is why ordinary investors (by which I mean anyone with less than $100m to invest) care about this. For a small investor, you are actually getting an even better deal because your order for 1000 shares or whatever can get filled instantaneously, in one chunk, for a great price! It's only when you're trying to buy hundreds of thousands of shares in a few minutes that you end up suffering price slippage.

The standard response is that ordinary investors have their money invested in mutual funds and pensions, who are large investors. But in that case you are already paying 0.5-2% per year to your fund manager, and why do you give a shit if they lose 10 basis points (0.1%) in price slippage because the market is more efficient than it used to be?

In fact, why is my pension fund manager trading so fucking much anyway? I don't have a pension because I think the fund manager is some genius stock picker, I have it because it's tax efficient and my employer contributes to it. Just buy the S&P500 and sit on it.


I see a false dichotomy here. The fact that floor brokers were dishonest doesn't imply we need to have continuous computerized trades. We could still have discrete steps and computers filling orders, right?


Right, and maybe one day the market will work that way. But

(a) what we have at the moment is still a lot better than what we had before - incremental progress!

(b) it's not at all obvious (to me) that discrete time steps would be better than what we have now. Market makers would be taking more risk, so that would quote in smaller size and at wider spreads, which could make trading more expensive for everybody.


Thanks for the perspective - very helpful.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: