The Relationship between Liquidity and Volatility as a factor in Assessing Risk

[TheBramble]

I'd like to get some of my thoughts out into open discussion to gain some clarity on this topic as it ain't as straightforward as you might have thought.

I've started looking hard at Volatility and Liquidity, primarily as an additional factor to consider loading into my risk assessment for calculating trading position size rather than as a determinant for trade entry/exit itself.

As usual, I've found it an extremely interesting bag of worms in its own right and will probably lose sight of my initial purpose unless you occasionally whack me over the head to remind me. Feel free.

Rather than dump a whole bunch of stuff into this one post, I'll cobble together those 'bits' I reckon would stimulate interesting conversation and unload them one by one. Anyway, long posts tend to get skipped and the effort of responding to them seems thankless.

First real post in this thread coming up.

 

[TheBramble]

Spread as a Determinant of Volatility

I'll get to the more obvious candidate for discussion "Relationship between Volatility and Liquidity" in a bit. Right now I'd like to focus on Spread.

A few weeks back I added analysis of spread into my magnum opus on trading as important a factor in analysing overall price action and dynamics as is open, close, high, low and volume (there are a few others that I wont bother you with right now).

It's an obvious relationship when you think about is, but I hadn't really tied it into volatility, liquidity or risk per se.

For example, let's take it as a given that "higher spreads are set when price changes are more uncertain". Any disagreement with that statement?

How about "An asset's risk as measured by its volatility is one of the major determinants of its spread". Well OK. But what comes first, the uncertainty of price changes (i.e. expectations of volatility) or actual dynamic volatility itself?

How about "an increase in volatility leads to widening of spreads". OK. Still holds and makes sense, but doesn't address the question above.

Then we also have "The size of bid-ask spread is positively related to the intensity of information flow". Hang on a mo, we're increasingly off down the information flow path (expectations and informed trading [stealth trading]) than pure market dynamics would have us believe is the case.

And finally "there appears to be a dynamic feedback relationship between informed trading and bid-ask spreads. This phenomenon is consistent with microstructure theory that private information induces a dynamic learning (or price discovery) process, through changes in trades and adjustments in the market maker's quotes, which eventually results in prices fully revealing the content of the private information."

{source: http://www.fma.org/Siena/DSS/Proposal_Liquidity_vs_Volatility.pdf.}

So, to summarise my question: What or maybe who, is leading what? Does spread dictate volatility (forget liquidity for now) or volatility dictate spread? Or both? And if either (both) instances, what or who adds what to the process to cause the change?

 

[swededemon]

Volume?

 

[TheBramble]

An increase in trading activity (volume) leads to improvement in liquidity. This is because an actively traded security requires less involvement by a dealer. Their buy and sell orders tend to offset each other. In other words, the dealer's inventory holding costs is low for actively traded stocks. As a consequence, liquidity improves as trading activity increases. But, according to the information asymmetry explanation, an increase in trading activity can lead either to an improvement or a deterioration in liquidity depending on the strategic behaviour of informed traders.

On one hand, a positive relationship is expected between trading activity and liquidity because an increase in liquidity trading enhances market makers' willingness to provide liquidity. However, on the other hand, if an increase trading is associated with an increase in information trading, a negative relationship between trading activity and liquidity is expected.

In essence, trading activity as measured by trading volume seems to improves liquidity, but trading activity as measured by number of transaction deteriorates liquidity.

So where does that leave Volume?

 

[jimbo57]

.

 

[Charlton]

What comes first - chicken or egg ?

I'll get to the more obvious candidate for discussion "Relationship between Volatility and Liquidity" in a bit. Right now I'd like to focus on Spread.

A few weeks back I added analysis of spread into my magnum opus on trading as important a factor in analysing overall price action and dynamics as is open, close, high, low and volume (there are a few others that I wont bother you with right now).

It's an obvious relationship when you think about is, but I hadn't really tied it into volatility, liquidity or risk per se.

For example, let's take it as a given that "higher spreads are set when price changes are more uncertain". Any disagreement with that statement?

How about "An asset's risk as measured by its volatility is one of the major determinants of its spread". Well OK. But what comes first, the uncertainty of price changes (i.e. expectations of volatility) or actual dynamic volatility itself?

How about "an increase in volatility leads to widening of spreads". OK. Still holds and makes sense, but doesn't address the question above.

Then we also have "The size of bid-ask spread is positively related to the intensity of information flow". Hang on a mo, we're increasingly off down the information flow path (expectations and informed trading [stealth trading]) than pure market dynamics would have us believe is the case.

And finally "there appears to be a dynamic feedback relationship between informed trading and bid-ask spreads. This phenomenon is consistent with microstructure theory that private information induces a dynamic learning (or price discovery) process, through changes in trades and adjustments in the market maker's quotes, which eventually results in prices fully revealing the content of the private information."

{source: http://www.fma.org/Siena/DSS/Proposal_Liquidity_vs_Volatility.pdf.}

So, to summarise my question: What or maybe who, is leading what? Does spread dictate volatility (forget liquidity for now) or volatility dictate spread? Or both? And if either (both) instances, what or who adds what to the process to cause the change?

TheBramble

I think that both volatility and spread are primarily symptoms or reflections of the causes, which are uncertainty and power struggles between sellers and buyers.

A seller will attempt to push prices up and, as long as there are buyers willing to purchase at that price, the price will reach that level. Buyers will be looking for a degree of certainty - that is, a certainty of a trend (although of course we know from the other thread you have been visitng that trends do not exist !). If there is a perceived high level of certainty on the upside then the highs will move up and the lows will remain as is, because no seller will be willing to sell at a lower price.

Conversely if there is a high level of certainty to the downside then the lows will move down and the highs remain as is, because no buyer will be willing to pay more.

Volatility as measured by the difference between the highs and lows will eventually become constrained because there will arise a general consensus between buyers and sellers about the trend and thus their power will become balanced.

Where uncertainty exists the same forces will come into play, but no general consensus abounds. Thus at times buyers will have the upper hand and at other times sellers will have the upper hand. The result is that highs and lows will be pushed apart, often with rapid swings during the period in question. No clear trend will exist.

Now spreads are affected likewise. They also represent the relative power between buyers and sellers at each trading point. The more uncertainty, the more disagreement so the wider the spread between bid and ask, as each vies with the other to get their way. There is little consensus.

The factors that give rise to the initial uncertainty will include general economic and political conditions, specific news events and the intentions of the market makers. .

However, interpreting your question as being ot the type "which comes first, the chicken or the egg", once the snowball is rolling up or down the certain trendline or up and down the uncertain big dipper it will gather more and more dirt as it goes.

Thus the price activity will either add to or detract from both the volatility and the spread as the lemmings drop off the cliff.

So in summary neither spread dictates volatility nor volatility dictates spread, but both are the manifestations of the initiating causes and other deeper foces and also the following bandwagon.

Charlton

 

[swededemon]

Volume needs to be processed or absorbed into price. depends how the orders are split and how big they are.

 

[TheBramble]

You're right Jimbo. We can't sensibly discuss this unless we define our terms accurately.

Liquidity

The ability to buy or sell a particular item without causing a significant movement in the price {wikipedia}

The probability that the next trade is executed at a price equal to the last one {wikipedia}

The liquidity of a product can be measured as how often it's bought and sold. For stocks this is known as the volume of trades {wikipedia}

The degree to which an asset or security can be bought or sold in the market without affecting the asset's price. {investopedia}

I can get all I want of the stock, on or off, within 0.25% of the current stock price. {TheBramble}

Volatility

The standard deviation of the change in value of a financial instrument with a specific time horizon {wikipedia}

For a financial instrument whose price follows a Gaussian random walk, or Wiener process, the volatility increases by the square-root of time as time increases. Conceptually, this is because there is an increasing probability that the instrument's price will be farther away from the initial price as time increases. {wikipedia}

A statistical measure of the dispersion of returns for a given security or market index. Volatility can either be measured by using the standard deviation or variance between returns from that same security or market index. Commonly, the higher the volatility, the riskier the security.
{investopedia}

I don't trade Volatility. Not deliberately anyway. What I'm realising is MY idea of Volatility is probably not the 'correct' one and I have all this time being trading precisely that. My previously held idea of what constituted Volatility was a stock that's all over the place, spiking a large number of factors of its open-close range on its high-low range within one or more timeframe bars/candles.

The correct definition, if I'm interpreting the definitions quoted above accurately, is a stock with high volatility simply moves a lot over a small period of time. Yes, I know I really should define 'a lot' and 'small period of time', but I'm not going to. You know what I mean.

It would be good to get others inputs on how they personally define, or perhaps even better, USE these concepts (liquidity and volatility) in their own trading.

 

[jimbo57]

.

 

[Grey1]

Guys , if i was you i would not be worry abou t liquidity at this stage, Just choose 10 core stocks which are liquid enough with low spread ( AAPL , BRCM ,RIMM, SNDK,,, ) even when they become volatile and just trade them .
Donot worry if the spread widens a bit during the volatile situations. Just concentrate on your postion sizing .


Grey1

 

[charliechan]

agreed. i think you are over complicating things a bit bramble. hope you dont suffer from analysis paralysis.

unless you are scalping with an algorithm, i dont think this stuff is of much use.

however, in an attempt to add something, i have noticed spreads widen when the market is quiet.

hope i havent p!ssed on your fire.

 

[TheBramble]

agreed. i think you are over complicating things a bit bramble. hope you dont suffer from analysis paralysis.

unless you are scalping with an algorithm, i dont think this stuff is of much use.

however, in an attempt to add something, i have noticed spreads widen when the market is quiet.

hope i havent p!ssed on your fire.

Not at all mate, appreciate the balanced input.

I'll decide what's of use once I've had the chance to research it more thoroughly. What's of use to me may not be of use to you of course.

Of course spreads widen when the market is quiet! Think it through. That's pretty much what I said in post above.

"An increase in trading activity (volume) leads to improvement in liquidity. This is because an actively traded security requires less involvement by a dealer. Their buy and sell orders tend to offset each other. In other words, the dealer's inventory holding costs is low for actively traded stocks". The converse could also be reasonably expected to hold true. Which is the point you make I think.

No, you carry on p!ssing...

 

[swededemon]

Brambs, i don't want to add further p*ss to the fire, but, you can not play the game, not the game you are watching. You've got to be able to cope with your lot. Nice thread though.

 

[Rhody Trader]

A truly intellectual discussion. I love it!

There's an element to the volume picture which muddles things a bit. That's the difference between transactional and speculative/investment activity. I differentiate the two in terms of transactional being those trades done for hedging and/or business operations purposes which generally are going to be done regardless of price. The latter group is the trading done for the purposes of profit. In stocks the transactional stuff may not be that significant, but in commodities, fixed income, and forex it certainly is.

The importance of that split is in the fact that transactional volume does not necessarily lead to liquidity, depending on the frequency. As was noted, liquidity is at least partially a function of how often trades are done. A market could see very high volume in terms of contracts or shares or whatever, but still be illiquid because those trades are large blocks done infrequently. That is why markets need speculators. They increase trade frequency and thus liquidity.

The flip side of that increased liquidity, though, is the potential for increased volatility because they can be reactionary, which brings in the news element already mentioned.

I define volatility a bit differently than the close to close price change (return) measure outlined earlier. My preference is to think more in terms of the full path prices take along the way, which includes the intra-period highs and lows. My reason for doing so is because volatility doesn't just imply risk, it also implies opportunity. It just depends on which side of the equation you are on.

Here's something to think on in terms of liquidity, though. In exchange-based markets (stocks, futures) where there is a traded price reported for each transaction, that price is generally considered to represent the "value" of that market at that time. But here's the catch - the traded price could be from either the bid or the offer, depending on the underlying action of the trade. This means that even if the bid/offer were to remain 100% stable, the simple fact that trades take place at both prices implies a changing of value.

For example, if the bid/offer for a stock were 100-101 and I were to buy at the offer, the value of the stock, based on the reported last trade would be 101. If someone were then to come through and sell at the bid, the value would drop to 100. If we take the definition of liquidity to mean little to no change in value from transaction to transaction, the exchange traded markets violate that - especially as spreads widen.

That is completely a reporting issue, of course. Were no traded price reported, as in the case of spot forex, then one would never consider value to change from trade to trade if the bid/offer were to remain constant.

Consider this too. Sticking to our definition of liquidity as meaning any given trade creates little to know change in price (value), then bid/offer spreads don't even come in to the equation. If I am equally able to sell $10mm USD/JPY with moving the bid when the spread is 2 pips as I am when it's 10, then we have to define the market as liquid.

Spreads widen in anticipation of illiquidity. They do not necessarily, in and of themselves, imply illiquidity.

 

[TheBramble]

Brambs, i don't want to add further p*ss to the fire, but, you can not play the game, not the game you are watching. You've got to be able to cope with your lot. Nice thread though.

swede/sg/rudeboy/cc - oneliners are a bit tiring. I know you don't want to give too much away about who you are by typing too much, but your old style is beginning to creep back in.

If you genuinely have something to add, spit it out lad otherwise you just come across as trying hard to be deeply cryptic - and not quite making it.

 

[TheBramble]

A truly intellectual discussion.

You know how to hurt a guy.

Great post RT and I do want to address a number of points you make, but don't have the facility right now. Back later.

 

[TheBramble]

There's an element to the volume picture which muddles things a bit. That's the difference between transactional and speculative/investment activity. I differentiate the two in terms of transactional being those trades done for hedging and/or business operations purposes which generally are going to be done regardless of price. The latter group is the trading done for the purposes of profit. In stocks the transactional stuff may not be that significant, but in commodities, fixed income, and forex it certainly is.

The importance of that split is in the fact that transactional volume does not necessarily lead to liquidity, depending on the frequency. As was noted, liquidity is at least partially a function of how often trades are done. A market could see very high volume in terms of contracts or shares or whatever, but still be illiquid because those trades are large blocks done infrequently. That is why markets need speculators. They increase trade frequency and thus liquidity.

The flip side of that increased liquidity, though, is the potential for increased volatility because they can be reactionary, which brings in the news element already mentioned.

The trick then is to have a method for differentiating transactional (informed/business) trade volume from speculative volume.

I define volatility a bit differently than the close to close price change (return) measure outlined earlier. My preference is to think more in terms of the full path prices take along the way, which includes the intra-period highs and lows. My reason for doing so is because volatility doesn't just imply risk, it also implies opportunity. It just depends on which side of the equation you are on.

Couldn't agree more. That's why I wondered in my previous posts if I was the only trader out here who measured or at least considered Volatility to be somewhat different to the textbook definition.

You only have to consider the different perspective in switching timeframes and the impact that will have on standard def Volatility to realise the necessity of considering the High Low as far more useful and indicative or real Volatility. For instance, using the standard definition of Volatility and by way of example, take 12, 5min bars - with the first bar's open at $25.00 and the final bar's close at $25.05. The intervening 10 bars whipsaw around in a High Low range of $15.50 to $45.60. That for me would be one Volatile stock! But the bod using the hourly chart views the Volatility as just 5 cents. Am I correctly interpreting standard definition of Volatility? if I am, it ain't what most traders talk about when they talk about Volatility.

Here's something to think on in terms of liquidity, though. In exchange-based markets (stocks, futures) where there is a traded price reported for each transaction, that price is generally considered to represent the "value" of that market at that time. But here's the catch - the traded price could be from either the bid or the offer, depending on the underlying action of the trade. This means that even if the bid/offer were to remain 100% stable, the simple fact that trades take place at both prices implies a changing of value.

For example, if the bid/offer for a stock were 100-101 and I were to buy at the offer, the value of the stock, based on the reported last trade would be 101. If someone were then to come through and sell at the bid, the value would drop to 100. If we take the definition of liquidity to mean little to no change in value from transaction to transaction, the exchange traded markets violate that - especially as spreads widen.

That is completely a reporting issue, of course. Were no traded price reported, as in the case of spot forex, then one would never consider value to change from trade to trade if the bid/offer were to remain constant.

Not sure I understand. A transaction gets made at one price. That's the value of the stock (instrument) at that time for that single transaction. That's the reality. Where I think I'm going with this though is the relationship of Bid to Offer (i.e. the spread) rather than what's going off on T&S.

Consider this too. Sticking to our definition of liquidity as meaning any given trade creates little to know change in price (value), then bid/offer spreads don't even come in to the equation. If I am equally able to sell $10mm USD/JPY with moving the bid when the spread is 2 pips as I am when it's 10, then we have to define the market as liquid.

Spreads widen in anticipation of illiquidity. They do not necessarily, in and of themselves, imply illiquidity.

Exactly the point I made in post #2. Expectation of liquidity drives liquidity. Much as price drives fundamentals. (I like to slip that in wherever I can ).

A combination of Volume and Ticks to determine the relative balance of information trading/liquidity trading and spread analysis for a view of the MMs' view of expected liquidity.

I am increasingly drawn to an even darker side of dark-siding where even price itself largely becomes an irrelevance in determination of potential future action over your timeframe of choice.

Just Spread, Ticks, Volume and Volatility as defined by the dynamically changing relationship between successive High-Low ranges.

 

[Rhody Trader]

The trick then is to have a method for differentiating transactional (informed/business) trade volume from speculative volume.

In the futures market you can get the Commitment of Traders (COT) data to help in that regard.

Couldn't agree more. That's why I wondered in my previous posts if I was the only trader out here who measured or at least considered Volatility to be somewhat different to the textbook definition.

You only have to consider the different perspective in switching timeframes and the impact that will have on standard def Volatility to realise the necessity of considering the High Low as far more useful and indicative or real Volatility. For instance, using the standard definition of Volatility and by way of example, take 12, 5min bars - with the first bar's open at $25.00 and the final bar's close at $25.05. The intervening 10 bars whipsaw around in a High Low range of $15.50 to $45.60. That for me would be one Volatile stock! But the bod using the hourly chart views the Volatility as just 5 cents. Am I correctly interpreting standard definition of Volatility? if I am, it ain't what most traders talk about when they talk about Volatility.

I can't necessarily say that most traders use close to close change in their definition of volatility. I do know that from an academic perspective, volatility is generally spoken of in terms of variation (standard deviation/variance) in period returns, which obviously are close-to-close changes.

Exactly the point I made in post #2. Expectation of liquidity drives liquidity. Much as price drives fundamentals. (I like to slip that in wherever I can ).

I do think there is a bit of the expectation of liquidity creating it, but that's just part of the equation. A market may attract certain traders because of the expectation of liquidity, but that is only a fraction. Some are there as providers of liquidity (market makers) and others don't really even think about it (commercials, long-term traders perhaps).

A combination of Volume and Ticks to determine the relative balance of information trading/liquidity trading and spread analysis for a view of the MMs' view of expected liquidity.

Here you're getting in to the realm of Market Profile, which I have always found both very interesting and quite useful.

I am increasingly drawn to an even darker side of dark-siding where even price itself largely becomes an irrelevance in determination of potential future action over your timeframe of choice.

Just Spread, Ticks, Volume and Volatility as defined by the dynamically changing relationship between successive High-Low ranges.

I'm right there with you. Shall I bring the torches?

 

[TheBramble]

This is an interesting thread. I was wondering though Bramble, do you actually have (or are looking to have)a spreadsheet or the like, that actually takes these figures, crunches them and then give you some figure as to how much you should expose yourself in a particular instrument?

Lordy! Haven't got to that stage yet, if I ever will at all. Just kicking stuff around to see what I might have been missing.

I do use Excel as my primary scanner and RT alert tool principally because I can customise it quickly, and completely along the lines I want. So yes, not beyond the bounds of possibility that something from this discussion may find its way into my arsenal.

Why? Do you already have it all programmed in???.........

As for "how much you should expose yourself in a particular instrument" I note I haven't really got onto how any of this might be used in assessing risk, which was the titled intent of this thread. I knew it was going to be a long one, and I also suspected the journey would be as interesting as the destination - which is how it's turning out

I guess this would be useful if your trading in a long time frame but from my own experience of short term trading its all better done on the fly as it were. If liquidity drops and volatility rises i always trade smaller. If however volatility goes up but the liquidity stays the same then I probably on balance try to increase my size.

Inherently doing the same myself. I don't deliberately trade illiquidity, so unless I c an see myself comfortable getting in and even more important, getting out on the size I want, without greater than 0.25% instrument price slippage/overhead, I don't go.

Its interesting you speaking of trying to filter speculative money in your analysis. We call this the 'fast money' vs 'real money' (or an old floor saying, the paper). I find watching thre relationship between who is giving the impression of urgency helps. Think about how you react when you know your wrong in the market. You dont (or shouldn't) bother to work a price, you should just get out. Giving the 'edge' up. The skill is trying to see where the greater balance of conviction is, with those giving up the edge versus those working the bid or the offer. I guess for want of a better expression, reading the tape.

When you say 'giving the impression of urgency' do you mean deliberately trying to give the impression or just giving the impression?

Reading the tape. Damn right. Not something that I can with my humble technical abilities easily even think about automating, though I am sure it has been done. I find it taxing enough just 'doing' it, but it's a lot easier and more obvious to me now than it was when I started. And there are enough exceptions always to stop you getting complacent in your personal assessment of your own expertise in this area.

Balance of conviction. Good phrase.

 

[charliechan]

.....

Reading the tape. Damn right. Not something that I can with my humble technical abilities easily even think about automating, though I am sure it has been done. I find it taxing enough just 'doing' it, but it's a lot easier and more obvious to me now than it was when I started. And there are enough exceptions always to stop you getting complacent in your personal assessment of your own expertise in this area.

Balance of conviction. Good phrase.

wouldnt this be quite simple?

add size to count if tick > last tick
minus size to count if tick < last tick
if count > (2*sell) then buy
else stay out

very crude of course. am i embarrassing myself here?

(another bad pint last night and feeling a bit tom & dick)

if sunshine = true then
loop [15]
beer + wine
end loop

if night = true then
bbq
more beer
end

if sunday morning = true then
feel sick as dog
post thoughtless crap until mates pop round for more beer

if monday morning = true then
skip trading cos im even worse for wear
system crash, interrupt error code = 666.

long time since i programmed a computer.......