[rahimnathwani]

Jack is not upset because he couldn't buy the shares at the price he wanted. He is upset because someone was offering shares at a specific price, and Jack was willing to pay that price, but the order was not executed. The reason the order was not executed is not because someone else accepted the offer before him, or because Jill cancelled before he tried to accept. It was because Jill was able to see his acceptance in transit and cancel part of her offer as a result of this new information. This isn't how markets are supposed to work. If Jill had posted a single offer for half the number of shares that Jack wanted, it would be different.

Someone shouldn't need to use Thor or some other delaying mechanism to accept open offers. The latency between different exchanges (which was exploited in this example) does nothing to increase market efficiency.

You're right that I shouldn't care about this as a practical matter, as the impact on me is very small, but that doesn't mean it's right.

I really appreciate the clarity in your comments on this thread.

[throwaway283719]

Here is our point of disagreement, then - I think that this is exactly how markets are supposed to work (in the presence of multiple exchanges).

The job of a market maker is to supply liquidity at a price/risk tradeoff that is reasonable to them, subject to the information available to them. If there are multiple exchanges, and someone trades with them on one exchange, then the set of information available to them has changed (specifically, their knowledge of the supply/demand balance for a particular stock has changed). It's only natural that they will want to change their prices in response.

Now, we could change legislation to either (a) go back to having a single exchange or (b) restricting the ability of market makers to move their quotes on one exchange if they trade on another. But that won't necessarily result in a better deal for non-market makers, because instead of quoting 20,000 shares split across 4 exchanges, the market maker now quotes 5,000 shares on 1 exchange.

The benefit is that all market participants have a more accurate idea of the true liquidity available in the market. The disadvantage is that you have removed the element of competition between exchanges, so the exchange is no longer incentivized to offer low fees and keep improving its service.

[rahimnathwani]

OK. So it's because there are multiple markets. We want multiple markets because competition, but multiple markets mean latency which means HFT using info from one exchange as a signal in another exchange.

If I could wave a magic wand then we would have one exchange which was run as a public service, by some beneficent person with no profit motive. I don't have a magic wand :(

I agree that the current situation is the result of the market structure (multiple competing markets) but I don't agree that this is how markets are supposed to work. The reason I say this is that people with no knowledge of market microstructure (e.g. ordinary people or people with undergraduate degrees in economics) would not expect this kind of 'arbitrage' to be possible. I hesitate to call it front-running because this is a term better reserved for instances where a client relationship and non-public information exists.

I take your point that the situation we have now (multiple exchanges without specialists) may be better than we had before (a single exchange with specialists) but I still don't think it's _fair_.

I wonder what would happen if I could wave my magic wand and have multiple exchanges with no specialists and zero latency...

[harryh]

If you could wave that magic wand big players with proprietary information (like hedge funds) would win a bit and everyone else (which includes you by the way) would lose because of slower price discovery.

[_delirium]

Do you have opinions about a third proposal that's occasionally floated, (c) discretize the market's timeline to something smallish but not miniscule? For example, the market maker can move their quotes however often they want, but changes take effect on the next tick, which happens every (say) 1 second. So therefore you can't trade on advance knowledge in the sub-second range, and market makers can't gain a trading advantage vis-a-vis investors solely by having a slightly lower latency connection to the exchange.

It's possible there's some massive downside to that, but afaict the advantages of liquidity that market-makers provide mostly accrue at larger timescales. So it's not clear the millisecond-shaving game is really improving markets (though it provides interesting challenges for technologists).

[kasey_junk]

How do you pick the timeframe? Seconds are still too fast for humans. Minutes would be too fast for people who are not professional traders, hours would be too fast for people who can't be near a computer all day.

There is always a locality advantage in the market, this has been true as long as there have been markets, and it will be true forever. Why do we as market participants care?

The other problem with your scenario is that you make market making more risky. The riskier it is, the higher the profits must be. This means that the market makers must keep the bid/ask spread higher (their means of making a profit). This cascades to all of us in the form of higher execution costs.

[_delirium]

> The other problem with your scenario is that you make market making more risky.

The problem with the current scenario is that it makes market making more expensive, as it requires a lot of technological investment into the microsecond arms race. This means the market makers must pull in more revenue from their trading to cover these expenses, before they even get to thinking about making a profit. This cascades to all of us in the form of higher execution costs. The huge amount of money being spent on HFT infrastructure, software development, etc. is ultimately being paid by market participants. It's worth considering if this is an arms race worth funding to the max, or if 99% of the benefits could be had much more cheaply just by putting a floor on execution latency, thereby rendering this whole millisecond-shaving industry unnecessary.

At the very least, I'd be interested in seeing rigorous models that show a benefit to, say, markets that can trade at 1-microsecond granularity vs. 1-millisecond vs. 1-second.

[hft_throwaway]

The claim that market makers pass costs on to end users is only true if they have pricing power. In reality, on-exchange liquidity provision is basically the kind of perfect competition that only exists in economics textbooks. Market makers are selling a commodity product (you don't care or control who you trade stocks with) in a market where buyers are purely sensitive to price (tightest market always wins and is enforced by exchange matching rules).

So what actually ends up happening in a market with multiple competitive market makers? To make money, a market maker needs to trade a lot of volume. The only way to trade a lot of volume is to put up the most aggressive (worse for the market maker, better for end users) prices at any time. Market makers can only do this by charging a smaller spread than their competitors. They can only charge a smaller spread by either reducing their margins or getting smarter at deciding when to be in or out of the market, usually a combination of both. The end result is extremely tight markets that react to information very quickly (i.e. cheap to trade and very efficient).

Competition keeps markets honest. If you had one very fast guy, he would clean up, but when you have a dozen guys who are roughly equally fast, they all compete one another down to barely making profit above their cost of doing business. Only the most efficient can survive. If anything, we want more HFT by removing barriers to entry rather than creating a lot of regulations that would ironically help incumbents by killing off weaker competitors.

[kasey_junk]

Except thats demonstrably not what has happened. Market making has gotten cheaper, not more expensive. Spreads have tightened, not gotten wider. Fees have gone down not up. Literally every cost to trading has been reduced.

The single biggest cost to any market maker is their market risk. Latency is exceedingly cheap in comparison. Any increase in latency raises market risk thereby raising their biggest cost.

[throwaway283719]

I basically agree with what you're saying, but I think there's another point to make - market making is simply not as profitable as it once was.

I've seen estimates that the entire high frequency trading industry made $1 billion profit in 2013, down from $5 billion in 2009. [0]

In contrast, JPMorgan made $6 billion profit in the last quarter, and that was reported as "not particularly impressive"! [1]

[0] http://www.reuters.com/article/2014/04/06/us-dark-markets-an...

[1] http://dealbook.nytimes.com/2014/07/15/jpmorgan-earnings-dec...

[harryh]

Moving to a 1 second tic wouldn't save money on infrastructure because HFTers would still have an incentive to wait for the last possible moment before the tic to update their orders to make sure that they had access to all possible information when making their decisions.