2 ms·

No, no I'm not. The rules for markets are man-made. No markets would exist without humans. Humans created markets to serve themselves. As such the rules tha

No, no I'm not. The rules for markets are man-made. No markets would exist without humans. Humans created markets to serve themselves. As such the rules that govern how markets operate can be changed to serve humans. This is very easy to understand, wouldn't you agree? Now comes the question: does HFT serve humans? Some people say yes, others say no. Some argue that they add liquidity and thus earn their fees. Others say no, they don't add liquidity (since they tend not to trade during rough times). Still others say they add negative liquidity: when the markets are plenty liquid they rob market makers of deal flow (thus reducing market makers ability to earn a living) and then when things get tumultuous they shut down first taking the extra liquidity with them and also having the secondary effect of fewer market makers thus reducing liquidity further. Market makers are able to make money because of the uncorrelated order flow that they receive as market makers. But there are HFT firms which are inserting themselves between brokerages and market makers, and paying millions for the privilege. That means they take profits from the market maker and after receiving their cut, provide some of it to the brokerages originating the orders. Is that good or bad? I honestly don't know. I might lean towards it being bad since the HFT firms have no obligation to keep trading. But they might very well reduce spreads. And if they are reducing spreads there is value to that. http://www.businessinsider.com/brokerages-make-millions-selling-orders-to-high-frequency-trading-firms-2014-4 http://www.businessinsider.com/brokerages-make-millions-sell... So let's assume for a moment that they do reduce spreads (good) but they also reduce liquidity during crises (bad). That is the very definition of what is called an engineering tradeoff. What is your design criteria? What are your goals? Engineering tradeoffs are at the heart of why basically everything is the way it is. Why are some cars faster and why do others get better gas mileage? Why are some bridges built with steel and others with concrete and others with wood? Why do some phones have a lot of battery life and other phones have a lot of features? So I've heard plenty about how HFT firms reduce spreads, and many argue that it's a good thing. But they're not an unmitigated good! There are downsides to HFT as well. Which should win out? That's more a matter of opinion than fact.

harryh 13y ago
> As such the rules that govern how markets operate can be changed to serve humans. This is very easy to understand, wouldn't you agree? I agree with this. > Now comes the question: does HFT serve humans? Some people say yes, others say no. I would say that the "others" in this statement are incorrect (except in some isolated and relatively unusual circumstances). Since upthread you said "Technically their arguments are quite good." I had assumed that you agreed with me here, but perhaps not? > but they also reduce liquidity during crises (bad). They don't "reduct liquidity during crises." They might provide less liquidity during a crisis than they do at other times, but that's not the same thing.
- msandford 13y ago
  > I would say that the "others" in this statement are incorrect (except in some isolated and relatively unusual circumstances). Since upthread you said "Technically their arguments are quite good." I had assumed that you agreed with me here, but perhaps not? You can say that they are incorrect, and they can say that they are correct, and neither side will concede to the other. Mostly because they have different goals. When I said "Technically their arguments are quite good" what I mean is "from a technical and/or technicality perspective" If you fill out your tax forms wrong and you pay the government one cent too few you are TECHNICALLY breaking the law and you might owe thousands of dollars in fines and penalties. But most reasonable folks would say that a $0.01 mistake doesn't deserve giant penalties. Similarly if you're doing 51mph in a 50mph zone TECHNICALLY you're breaking the law and depending on jurisdiction you might well have earned a ticket of some sort. But again, most folks would agree it's not reasonable. So the technical arguments of the folks doing HFT are good. On an infinitesimal time scale yes they are 100% right. Technically they (probably) aren't breaking any laws and they are propagating price information quickly. And I'm sure plenty of them have great intentions, I'm definitely not accusing anyone of trying to scam. The problem is that even though they do mean well, and even though they do serve a market function, their presence in the market does make for counter-intuitive outcomes for folks. > They don't "reduct liquidity during crises." They might provide less liquidity during a crisis than they do at other times, but that's not the same thing. Okay so there's two ways to skin this cat. But we have to agree on definitions. If HFT folks want to argue that they "add liquidity" to the market I'll concede that. The amount of liquidity that should be in the market is then the sum of the liquidity provided by market makers (MM) and HFT. Normal Liquidity = MM + HFT Great. But now a crisis hits and prices start gyrating. The HFT folks pull out. Crisis Liquidity = Normal Liquidity - HFT That's removing or reducing liquidity. Now you might say that I'm abusing definitions. Not really. You can't have it both ways. You can't say "HFT adds liquidity[1]" and then in the fine print say "[1] We only have to add liquidity during normal market operations as we define normal" because that's a hell of an asterisk and fine print. Furthermore if HFT folks are doing market making during normal times, that's fine. But if they're making profits, they're doing so at someone's expense right? If we hold the order flow constant and we direct it all at a market maker he's going to make X. Now if we inject an HFT into the same order flow and they make Y, that means he's making X - Y. We can say that because everything else is held constant. So that means that on average the HFT folks are going to reduce the profitability of market makers or the number of market makers or both. Then when the crisis hits there are fewer market makers with less ability to trade and thus, less liquidity. So HFT can reduce crisis liquidity in two ways. First by not providing liquidity during crisis times, and second by reducing the overall number of market makers so that the amount of liquidity that they can provide during a crisis is also reduced. Now again, I'm not necessarily saying that this is bad. But to pretend that HFT is all upside and no downside and that there are no primary or secondary negative effects seems to me to be either ignorant or disingenuous. I can't say if HFT is 10% downside and 90% upside, 50/50, or 90/10. But I can tell you that the split between upside/downside of HFT does depend in some way on how you interact with the market. Different actors will see different cost/benefit ratios. I think I get that you generally think HFT is good. That's OK. I personally don't care all that much since I'm out of equities as of ~3mo ago and I've mostly bought & held. But the folks who think that HFT costs a lot and doesn't add that much aren't really WRONG either. People assign different weights to things.

harryh 13y ago

> As such the rules that govern how markets operate can be changed to serve humans. This is very easy to understand, wouldn't you agree? I agree with this. > Now comes the question: does HFT serve humans? Some people say yes, others say no. I would say that the "others" in this statement are incorrect (except in some isolated and relatively unusual circumstances). Since upthread you said "Technically their arguments are quite good." I had assumed that you agreed with me here, but perhaps not? > but they also reduce liquidity during crises (bad). They don't "reduct liquidity during crises." They might provide less liquidity during a crisis than they do at other times, but that's not the same thing.

msandford 13y ago
> I would say that the "others" in this statement are incorrect (except in some isolated and relatively unusual circumstances). Since upthread you said "Technically their arguments are quite good." I had assumed that you agreed with me here, but perhaps not? You can say that they are incorrect, and they can say that they are correct, and neither side will concede to the other. Mostly because they have different goals. When I said "Technically their arguments are quite good" what I mean is "from a technical and/or technicality perspective" If you fill out your tax forms wrong and you pay the government one cent too few you are TECHNICALLY breaking the law and you might owe thousands of dollars in fines and penalties. But most reasonable folks would say that a $0.01 mistake doesn't deserve giant penalties. Similarly if you're doing 51mph in a 50mph zone TECHNICALLY you're breaking the law and depending on jurisdiction you might well have earned a ticket of some sort. But again, most folks would agree it's not reasonable. So the technical arguments of the folks doing HFT are good. On an infinitesimal time scale yes they are 100% right. Technically they (probably) aren't breaking any laws and they are propagating price information quickly. And I'm sure plenty of them have great intentions, I'm definitely not accusing anyone of trying to scam. The problem is that even though they do mean well, and even though they do serve a market function, their presence in the market does make for counter-intuitive outcomes for folks. > They don't "reduct liquidity during crises." They might provide less liquidity during a crisis than they do at other times, but that's not the same thing. Okay so there's two ways to skin this cat. But we have to agree on definitions. If HFT folks want to argue that they "add liquidity" to the market I'll concede that. The amount of liquidity that should be in the market is then the sum of the liquidity provided by market makers (MM) and HFT. Normal Liquidity = MM + HFT Great. But now a crisis hits and prices start gyrating. The HFT folks pull out. Crisis Liquidity = Normal Liquidity - HFT That's removing or reducing liquidity. Now you might say that I'm abusing definitions. Not really. You can't have it both ways. You can't say "HFT adds liquidity[1]" and then in the fine print say "[1] We only have to add liquidity during normal market operations as we define normal" because that's a hell of an asterisk and fine print. Furthermore if HFT folks are doing market making during normal times, that's fine. But if they're making profits, they're doing so at someone's expense right? If we hold the order flow constant and we direct it all at a market maker he's going to make X. Now if we inject an HFT into the same order flow and they make Y, that means he's making X - Y. We can say that because everything else is held constant. So that means that on average the HFT folks are going to reduce the profitability of market makers or the number of market makers or both. Then when the crisis hits there are fewer market makers with less ability to trade and thus, less liquidity. So HFT can reduce crisis liquidity in two ways. First by not providing liquidity during crisis times, and second by reducing the overall number of market makers so that the amount of liquidity that they can provide during a crisis is also reduced. Now again, I'm not necessarily saying that this is bad. But to pretend that HFT is all upside and no downside and that there are no primary or secondary negative effects seems to me to be either ignorant or disingenuous. I can't say if HFT is 10% downside and 90% upside, 50/50, or 90/10. But I can tell you that the split between upside/downside of HFT does depend in some way on how you interact with the market. Different actors will see different cost/benefit ratios. I think I get that you generally think HFT is good. That's OK. I personally don't care all that much since I'm out of equities as of ~3mo ago and I've mostly bought & held. But the folks who think that HFT costs a lot and doesn't add that much aren't really WRONG either. People assign different weights to things.