Stocks and bonds move in opposite directions and improve results? Really?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

I find the concept that money's got to go somewhere a good foundation for designing a portfolio which is always likely to have uncorrelated assets

e.g. SCV, Emerging Markets, Gold and LT Treasuries - money's probably flowing from one into another at any time, like an aggressive Harry Browne portfolio - but the diversification ameliorates a lot of the risk (so little point chasing middle-of-the-road returns) .. the endowment principle (SCV and gold standing in for private equity and hedge funds)

Three times the return of a three-fund portfolio with half the max. drawdown

Image
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

There are lots of other places money can flow to. Real estate, farmland, livestock*, etc. You also have to take into account that sometimes, there just isn't as much money in the economy as it used to be. E.g. great depression.


*Okay, I will be REALLY surprised if livestock becomes an investment "thingy", but it can happen.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

lee1026 wrote:There are lots of other places money can flow to. Real estate, farmland, livestock*, etc. You also have to take into account that sometimes, there just isn't as much money in the economy as it used to be. E.g. great depression.


*Okay, I will be REALLY surprised if livestock becomes an investment "thingy", but it can happen.
One argument for holding the global market portfolio - there really aren't many other places it can go (perhaps Beanie Babies?)

(PS - Hillary Clinton .. 20,000% in one year trading livestock futures)
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

Why can't there simply be less money in the world after the crash? The money doesn't have to go anywhere. It can just disappear.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

lee1026 wrote:Why can't there simply be less money in the world after the crash? The money doesn't have to go anywhere. It can just disappear.
I'm sure someone with a better grasp of economics can explain it better (or correct me) - but in a sense as soon as you invest in stocks, the money's stopped being money, and become an estimation of value

And that estimation comes from buying and selling - and when stocks crash, it means there are more sellers than buyers, and those sellers will need to move their money somewhere else

You probably wouldn't want $10 million in a bank account (because it could disappear if those bank stocks take a turn for the worse) - so you might want to invest in short-dated government bonds ... But then you're still not safe if your economy/currency collapses - so you might choose gold or bonds in a safer region ... So if there's mass selling going on in one part of the market, it's probably a good bet there's mass buying somewhere else ... And when you look at the global market portfolio (as difficult as it is to model) - or approximations, like the Permanent Portfolio - it's a fairly smooth ride
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
edge
Posts: 3485
Joined: Mon Feb 19, 2007 7:44 pm
Location: NY

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by edge »

There can definitely be less net value after a crash. This is especially obvious in a crash of illiquid assets where the 'fleeing' action does not obfuscate the raw losses - e.g. housing crash. In a more liquid market value is still destroyed; some escapes to other assets.

Re: random variables. You cannot model variables that have strong inter-dependencies using independent random variables....because well...they aren't independent.
Last edited by edge on Fri Dec 11, 2015 1:38 pm, edited 1 time in total.
azanon
Posts: 2947
Joined: Mon Nov 07, 2011 10:34 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by azanon »

You don't need different asset classes that are negatively correlated. All you need is uncorrelated assets. If two asset classes are significantly below 1.0 (FULL) correlation, then there is a potential beneficial effect, in terms of risk-adjusted returm, to owning both of them, w/rebalancing, especially if said asset classes independently have a real return greater than cash. And, of course, that's easy to accomplish given that cash's real return is quite negative.
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

And that estimation comes from buying and selling - and when stocks crash, it means there are more sellers than buyers, and those sellers will need to move their money somewhere else
I don't think that is accurate - the amount that sellers sell and the amount that buyers buy needs to be exactly the same. So the sellers needs to move their money somewhere else, but the buyers also had to sell something to get the money to buy it in the first place.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

lee1026 wrote:
And that estimation comes from buying and selling - and when stocks crash, it means there are more sellers than buyers, and those sellers will need to move their money somewhere else
I don't think that is accurate - the amount that sellers sell and the amount that buyers buy needs to be exactly the same. So the sellers needs to move their money somewhere else, but the buyers also had to sell something to get the money to buy it in the first place.
The number of stocks, bonds and gold don't change, but when there are more sellers than buyers, the value of (say) stocks goes down ... And when there are more buyers than sellers, the value of (say) gold goes up

So perhaps it's better to forget the term "money", and just focus on "value"
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

So perhaps it's better to forget the term "money", and just focus on "value"
I don't think there is any rule that the value of all assets needs to remain constant?
The number of stocks, bonds and gold don't change, but when there are more sellers than buyers, the value of (say) stocks goes down ... And when there are more buyers than sellers, the value of (say) gold goes up
Err.... what does it mean to have more buyers than sellers? In dollar weighted terms, the number of buyers and sellers must always match.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

lee1026 wrote:
So perhaps it's better to forget the term "money", and just focus on "value"
I don't think there is any rule that the value of all assets needs to remain constant?
The number of stocks, bonds and gold don't change, but when there are more sellers than buyers, the value of (say) stocks goes down ... And when there are more buyers than sellers, the value of (say) gold goes up
Err.... what does it mean to have more buyers than sellers? In dollar weighted terms, the number of buyers and sellers must always match.
If 100 people are looking to buy a house in Malibu, and supply only extends to 50, then price goes up until we reach an equilibrium

So value changes based on buyers vs sellers ... It's not necessarily a constant, but money/value does have to go somewhere
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
User avatar
nisiprius
Advisory Board
Posts: 41993
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by nisiprius »

Maynard F. Speer beat me to it.
lee1026 wrote:...Err.... what does it mean to have more buyers than sellers? In dollar weighted terms, the number of buyers and sellers must always match...
I've been brooding about this, and I think it's misleading. It is sort of like Zeno's paradox of the arrow: at any instant in time an arrow must alway be at one specific place, so at any instant of time it is at rest, therefore it can't possibly move. Although differential calculus makes my head hurt if I think about it too hard, I'm willing to believe that there is something about infinitesimals at the margin. In some sorta-kinda way, at any instant of time, the number of buyers and sellers must match, but the urge to buy and the urge to sell doesn't necessarily match. If you want to buy and I want to sell, we may not want to sell at the market price, and the relative power of our urges will determine whether the price of our trade is a trifle higher or lower than the market.

If the price of a stock is $100 and there are "more buyers than sellers," that means there is a larger number of people who want to buy at $100 than there are people who want to sell at $100. The buyers keep buying at $100 until everyone willing to sell at $100 has sold. Now there are people who want to buy at $100 but nobody willing to sell at the price. Well, if there is a crowd of people who "wanted to buy at $100" and can't, it's really unlikely that $100 is truly a hard limit for the whole crowd. Somewhere, there is someone willing to pay some sufficiently tiny increment over $100. $100.00000001, perhaps. $100.001, perhaps. $100.01, perhaps. Did I hear $100.01? Sold, to the lady in the green dress for $100.01.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

If 100 people are looking to buy a house in Malibu, and supply only extends to 50, then price goes up until we reach an equilibrium
Agreed. The number of sellers and buyers depends on the price, and any differences in the number of sellers and buyers in a liquid market is mediated by a change in price. That is, if the stock market fell today, it means that there is more willing sellers than buyers at yesterday's price. But the number of sellers and buyers at this instant at the instant price still matches.

More importantly for the money needs to go somewhere argument, however, is that the number of dollars that the sellers got and the number of dollars that the buyers paid needs to match, and the dollars that the buyers used to buy from needs to come from somewhere.
money/value does have to go somewhere
That is the part that I am objecting to, and will keep objecting to. Let's construct a rather simplistic market: There is $100 billion dollars in stocks and $100 billion in bonds. There are no other assets in this economy*. Something spooks the stock market and it makes a lot of shareholders want to sell out. Not a lot of people want to buy because the "something" that spooks the stock market scares the bondholders as well. Stocks fall by 50% before the buyers and sellers match, and the 50% of the stock market changes hands in this epic day.

So we have former shareholders who got $25 billion that day. You are correct that they need to put it somewhere, and in this market, they have put it in the bond market. But that $25 billion have to have came from the former bond holders. So on the net, you will see $25 billion flowing into the bond market from the shareholders, and $25 billion flowing out of the bond market into the stock market. You wouldn't expect bonds to go up in this case for purely the reason that the old shareholders are buying bonds.

*Let's pretend that all real estate is owned by REITs or something.
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

Frans wrote:It is often said that stocks and bonds have a tendency to move in opposite directions and that this behavior noticeably improves returns and lowers risk of combinations of stocks and bonds. When stocks underperform, the rationale goes, bonds have the tendency to compensate for this by doing better, resulting in better overall returns and less variation. Equally, when stocks do better, bonds have the tendency to underperform, with similar results. Really?
Yeah, this is why my portfolio is both the S&P500 and the inverse S&P500. teehee.

I think your digging in the data and others have thoroughly discussed this. But, folks saying negative correlation is desirable is a HUGE pet peeve of mine.If one likes to short markets, then okay negative correlation works in this scenario. But, if one is a buy and hold investor, anything that is negatively correlated with core assets is a real drag (such as cash).

Okay, nothing unique to add here, just a voice in the choir.
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

nisiprius wrote:Maynard F. Speer beat me to it.
lee1026 wrote:...Err.... what does it mean to have more buyers than sellers? In dollar weighted terms, the number of buyers and sellers must always match...
I've been brooding about this, and I think it's misleading. It is sort of like Zeno's paradox of the arrow: at any instant in time an arrow must alway be at one specific place, so at any instant of time it is at rest, therefore it can't possibly move. Although differential calculus makes my head hurt if I think about it too hard, I'm willing to believe that there is something about infinitesimals at the margin. In some sorta-kinda way, at any instant of time, the number of buyers and sellers must match, but the urge to buy and the urge to sell doesn't necessarily match. If you want to buy and I want to sell, we may not want to sell at the market price, and the relative power of our urges will determine whether the price of our trade is a trifle higher or lower than the market.

If the price of a stock is $100 and there are "more buyers than sellers," that means there is a larger number of people who want to buy at $100 than there are people who want to sell at $100. The buyers keep buying at $100 until everyone willing to sell at $100 has sold. Now there are people who want to buy at $100 but nobody willing to sell at the price. Well, if there is a crowd of people who "wanted to buy at $100" and can't, it's really unlikely that $100 is truly a hard limit for the whole crowd. Somewhere, there is someone willing to pay some sufficiently tiny increment over $100. $100.00000001, perhaps. $100.001, perhaps. $100.01, perhaps. Did I hear $100.01? Sold, to the lady in the green dress for $100.01.
Thanks for the reminder. Some folks forget Macroeconomics 101.
lee1026
Posts: 395
Joined: Sat Jun 21, 2014 7:47 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by lee1026 »

I think your digging in the data and others have thoroughly discussed this. But, folks saying negative correlation is desirable is a HUGE pet peeve of mine.If one likes to short markets, then okay negative correlation works in this scenario. But, if one is a buy and hold investor, anything that is negatively correlated with core assets is a real drag (such as cash).
If there a magical fund offered such that annual returns are always guaranteed to be the inverse of S&P 500 + 5%, this fund would have negative average returns. But I would still buy as much of this fund as you would possible let me since it cancels out the risk from my stock investments but doesn't cancel out all of the return. If I want more risk, I would simply lever up.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

lee1026 wrote:So we have former shareholders who got $25 billion that day. You are correct that they need to put it somewhere, and in this market, they have put it in the bond market. But that $25 billion have to have came from the former bond holders. So on the net, you will see $25 billion flowing into the bond market from the shareholders, and $25 billion flowing out of the bond market into the stock market. You wouldn't expect bonds to go up in this case for purely the reason that the old shareholders are buying bonds.

*Let's pretend that all real estate is owned by REITs or something.
I'll have to think about your example a little more to make sure I'm not misunderstanding you - but it seems to me if both are perceived as risky, you wouldn't expect value to shift .. More likely we'd all wind up holding 50:50 - stock and bond-heavy investors in need of diversification equally happy to buy and sell at current prices ... As you say, which is what one would expect

JZinCO wrote:I think your digging in the data and others have thoroughly discussed this. But, folks saying negative correlation is desirable is a HUGE pet peeve of mine.If one likes to short markets, then okay negative correlation works in this scenario. But, if one is a buy and hold investor, anything that is negatively correlated with core assets is a real drag (such as cash).

Okay, nothing unique to add here, just a voice in the choir.
See I find low or negative correlations - ideally with equity-like returns - key to good portfolio construction .. The one 'free lunch' you don't have to be a genius to capture

Sequence-of-returns is a genuine risk .. Lose 50% and you need to make 100% back to break even; lose 80% and you need 400% to get back .. So taking a large hit early on in your investing career can be much worse than taking one later on .. Anything to smooth the sequence of returns removes risk, and means you're more likely to get a fair return based on the level of risk you're exposed to

Take gold .. It's returned 3% less than the S&P 500 over period 1972-today, but because it's a flight-to-safety asset, it's often had useful negative correlations when markets need them .. So here, a 25% allocation to gold (despite the lower return) improves both risk-adjusted and absolute returns over long periods by smoothing out the sequence of returns

Image
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

Maynard F. Speer wrote:
JZinCO wrote:I think your digging in the data and others have thoroughly discussed this. But, folks saying negative correlation is desirable is a HUGE pet peeve of mine.If one likes to short markets, then okay negative correlation works in this scenario. But, if one is a buy and hold investor, anything that is negatively correlated with core assets is a real drag (such as cash).

Okay, nothing unique to add here, just a voice in the choir.
See I find low or negative correlations - ideally with equity-like returns - key to good portfolio construction .. The one 'free lunch' you don't have to be a genius to capture

Sequence-of-returns is a genuine risk .. Lose 50% and you need to make 100% back to break even; lose 80% and you need 400% to get back .. So taking a large hit early on in your investing career can be much worse than taking one later on .. Anything to smooth the sequence of returns removes risk, and means you're more likely to get a fair return based on the level of risk you're exposed to

Take gold .. It's returned 3% less than the S&P 500 over period 1972-today, but because it's a flight-to-safety asset, it's often had useful negative correlations when markets need them .. So here, a 25% allocation to gold (despite the lower return) improves both risk-adjusted and absolute returns over long periods by smoothing out the sequence of returns

Image
Sorry, you're arguing a red herring in two ways.
First, You made a brilliant argument for diversification. Diversification is holding assets with the lowest correlation possible. I have no qualms with this. This is my strategy as a passive, buy and hold investor. Your statements suggest I argued against building portfolios with low correlations. I don't and did not say that and therefore you aren't refuting my statement.
Second, you said, negative correlations are good in certain market conditions. This is also true. This is called shorting. It is not a buy and hold tactic and therefore you aren't refuting my statement.

Try running a 50/50 portfolio of SH and SPY (almost a perfect negative correlation, minus tracking error). You'll end up with a near 0% CAGR. I repeat: If a buy and hold investor holds an asset with a negative correlation against core assets, that asset is a drag. It is a mathematical truth.
---
I'll rephrase my pet peeve: conflating no correlation and negative correlation. You just demonstrated this.
I think this confusion stems from the fact that most people only get a taste of statistics in undergrad in some poorly taught, rudimentary 101 class. Most people confuse low with negative. The mathematical absolute lowest correlation is 0. The absolute highest is -1 or 1.
User avatar
nisiprius
Advisory Board
Posts: 41993
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by nisiprius »

JZinco, I don't think you have it quite right. If the only way you can get negative correlation is to construct it synthetically, with derivatives and short positions, then you are correct, because the same mechanism that cancels out risk is canceling out return at the same time. (Portfoliovisualizer is showing me that you're right about a 50/50 mix of SH and SPY which is interesting... I would have expected the combination to have a distinctly negative return due to the effects of daily rebalancing and costs).

If, however, it were the case that you had two real assets that both had robust positive real return and robust negative correlations, it would be magic. You'd cancel out risk without canceling out return. This is one of the cases people like to fantasize about. In theory, an asset with zero return would improve a portfolio if it had a nice strong persistent reliable robust negative correlation with, say, stocks.

Where Frans comes in is that there are many casual statements that seem to imply that stocks and bonds actually are such a pair. I don't think they are.

It's my belief that negative correlation is magic, like a magician pulling a rabbit from a hat. The only way you can pull a rabbit from a hat is if you first put the rabbit into the hat. The only way you can get negative correlation is by using the derivatives and short positions that also give you negative return, so, if that's true, you are right.

The factual question is: can you, in fact, find situations where there is good reason to expect continuing negative correlation between non-synthetic, real assets, or do you only find them when you've put the rabbit into the hat yourself?
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

JZinCO wrote: Sorry, you're arguing a red herring in two ways.
First, You made a brilliant argument for diversification. Diversification is holding assets with the lowest correlation possible. I have no qualms with this. This is my strategy as a passive, buy and hold investor. Your statements suggest I argued against building portfolios with low correlations. I don't and did not say that and therefore you aren't refuting my statement.
Second, you said, negative correlations are good in certain market conditions. This is also true. This is called shorting. It is not a buy and hold tactic and therefore you aren't refuting my statement.

Try running a 50/50 portfolio of SH and SPY (almost a perfect negative correlation, minus tracking error). You'll end up with a near 0% CAGR. I repeat: If a buy and hold investor holds an asset with a negative correlation against core assets, that asset is a drag. It is a mathematical truth.
---
I'll rephrase my pet peeve: conflating no correlation and negative correlation. You just demonstrated this.
I think this confusion stems from the fact that most people only get a taste of statistics in undergrad in some poorly taught, rudimentary 101 class. Most people confuse low with negative. The mathematical absolute lowest correlation is 0. The absolute highest is -1 or 1.
I think you're taking the term "negative correlation" in a very literal sense ... It's unlikely we'd be talking about perfectly inverted correlations in a buy-and-hold context

Your maths isn't quite right though ... We measure correlation coefficients linearly and in the context of timescales - so one could have a negative correlation of -0.6 over 12 months, but a positive correlation over 3 years
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

nisiprius wrote: If, however, it were the case that you had two real assets that both had robust positive real return and robust negative correlations, it would be magic. You'd cancel out risk without canceling out return. This is one of the cases people like to fantasize about. In theory, an asset with zero return would improve a portfolio if it had a nice strong persistent reliable robust negative correlation with, say, stocks.

Where Frans comes in is that there are many casual statements that seem to imply that stocks and bonds actually are such a pair. I don't think they are.

It's my belief that negative correlation is magic, like a magician pulling a rabbit from a hat. The only way you can pull a rabbit from a hat is if you first put the rabbit into the hat. The only way you can get negative correlation is by using the derivatives and short positions that also give you negative return, so, if that's true, you are right.

The factual question is: can you, in fact, find situations where there is good reason to expect continuing negative correlation between non-synthetic, real assets, or do you only find them when you've put the rabbit into the hat yourself?
Canceling out returns. Exactly. That is what negatively correlated assets do. when one asset goes up, the other goes down. The closer negative correlation is to 1, the closer you are to zero sum returns. In regards to your last question, there are very few actual assets that are negatively correlated in the long term.In fact this matrix suggests only low negative correlation in some instances at best.http://www.dividendtree.net/risk/asset- ... ification/
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

Maynard F. Speer wrote: Your maths isn't quite right though ... We measure correlation coefficients linearly and in the context of timescales - so one could have a negative correlation of -0.6 over 12 months, but a positive correlation over 3 years
Correlation is correlation. It is always calculated the same way. sometimes the data is truncated and always the future values are censored. Just like annual returns, just like sharpe ratio, just like CAGR.
I have repeated said, I am speaking of the case for the buy and hold, non market timer. Therefore, your discussion about 12 mo vs 3 yr is moot. It also does not matter what the magnitude of negative correlation. With any value between -.01 to -1, the asset would be a drag. It were a correlation of -1, it would wholly reduce your return to 0%. If were a correlation of -0.01, it would reduce your return by 1 percent.
The key in MPT is to diversity with noncorrelated assets, such that the conditions which might reduce Asset A's returns, do not influence B's returns. Therefore neither asset drags on one another but they move independently.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

No .. This is the case only with a perfect negative correlation

We can still have negatively correlated assets - such as oil and bonds .. However with anything other than a perfect negative correlation, it's important to understand the period we're looking at .. Short-to-medium-term negative correlations are generally desirable

Image

Image

http://www.assetcorrelation.com/majors
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

Maynard F. Speer wrote:No .. This is the case only with a perfect negative correlation

We can still have negatively correlated assets - such as oil and bonds .. However with anything other than a perfect negative correlation, it's important to understand the period we're looking at .. Short-to-medium-term negative correlations are generally desirable
I can't say anything I haven't said already.
You're arguing against an argument that I've never made.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

JZinCO wrote:
Maynard F. Speer wrote:No .. This is the case only with a perfect negative correlation

We can still have negatively correlated assets - such as oil and bonds .. However with anything other than a perfect negative correlation, it's important to understand the period we're looking at .. Short-to-medium-term negative correlations are generally desirable
I can't say anything I haven't said already.
You're arguing against an argument that I've never made.
I think it was actually you picking up on an argument that was never made

I meant to address this: "I have repeated said, I am speaking of the case for the buy and hold, non market timer. Therefore, your discussion about 12 mo vs 3 yr is moot. It also does not matter what the magnitude of negative correlation"

This is not 'market timing' .. What we're looking to do is reduce short-term losses of capital (and especially black swans that could lead to long-term loss of capital) .. You've heard examples of what would happen if we removed the 10 worst days from a decade of stock returns? Or how much you'd need to break even from a 50% loss? .. Well a negative short-term correlation minimises the impact of these events across a portfolio; while a positive long-term correlation means you're not "shorting" .. Simple rebalancing takes care of the rest
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Rodc »

JZinCO wrote:
nisiprius wrote: If, however, it were the case that you had two real assets that both had robust positive real return and robust negative correlations, it would be magic. You'd cancel out risk without canceling out return. This is one of the cases people like to fantasize about. In theory, an asset with zero return would improve a portfolio if it had a nice strong persistent reliable robust negative correlation with, say, stocks.

Where Frans comes in is that there are many casual statements that seem to imply that stocks and bonds actually are such a pair. I don't think they are.

It's my belief that negative correlation is magic, like a magician pulling a rabbit from a hat. The only way you can pull a rabbit from a hat is if you first put the rabbit into the hat. The only way you can get negative correlation is by using the derivatives and short positions that also give you negative return, so, if that's true, you are right.

The factual question is: can you, in fact, find situations where there is good reason to expect continuing negative correlation between non-synthetic, real assets, or do you only find them when you've put the rabbit into the hat yourself?
Canceling out returns. Exactly. That is what negatively correlated assets do. when one asset goes up, the other goes down. The closer negative correlation is to 1, the closer you are to zero sum returns. In regards to your last question, there are very few actual assets that are negatively correlated in the long term.In fact this matrix suggests only low negative correlation in some instances at best.http://www.dividendtree.net/risk/asset- ... ification/
Blue highlighted. Unless I misunderstand what you are saying this is wrong.

Correlations are computed from differences from the mean of each time series. That is is the ups and downs from the means that go up and down together that give rise to positive correlation, or go in opposite directions that give rise to negative correlations.

For example try this out in a spreadsheet.

Make a set of "noise returns" = rand().

Make a set of returns A = 0.10 + noise (that is a mean return of 10% but with ups and downs)

Then make a set of returns B = 0.10 - noise (use the same column of random numbers, not a new set of random numbers)

Now you compute the correlation and you find a perfect -1.

But the sum of the two is the magic portfolio that returns 10% every year with zero volatility. The (mean) returns do not cancel, the noise cancels.

If you want play with A and B where you again use the same fundamental noise but make A =0.10+0.6667*noise and B = 0.05+.3333*noise

So B has a smaller return and a smaller "risk", again you get correlation = -1. But if you want to get the noise to cancel you need to use 1/3 A and 2/3 B - This is more like stocks and bonds (though with a perfect corr= -1 rather than the more realistic corr ~= 0).

Next if you want to continue playing see what happens when you use independent sets of noise.

So that is the math. I doubt in the real world you can do much better than zero correlation such as stocks and treasuries.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

Rodc wrote:
JZinCO wrote:
nisiprius wrote: If, however, it were the case that you had two real assets that both had robust positive real return and robust negative correlations, it would be magic. You'd cancel out risk without canceling out return. This is one of the cases people like to fantasize about. In theory, an asset with zero return would improve a portfolio if it had a nice strong persistent reliable robust negative correlation with, say, stocks.

Where Frans comes in is that there are many casual statements that seem to imply that stocks and bonds actually are such a pair. I don't think they are.

It's my belief that negative correlation is magic, like a magician pulling a rabbit from a hat. The only way you can pull a rabbit from a hat is if you first put the rabbit into the hat. The only way you can get negative correlation is by using the derivatives and short positions that also give you negative return, so, if that's true, you are right.

The factual question is: can you, in fact, find situations where there is good reason to expect continuing negative correlation between non-synthetic, real assets, or do you only find them when you've put the rabbit into the hat yourself?
Canceling out returns. Exactly. That is what negatively correlated assets do. when one asset goes up, the other goes down. The closer negative correlation is to 1, the closer you are to zero sum returns. In regards to your last question, there are very few actual assets that are negatively correlated in the long term.In fact this matrix suggests only low negative correlation in some instances at best.http://www.dividendtree.net/risk/asset- ... ification/
Blue highlighted. Unless I misunderstand what you are saying this is wrong.

Correlations are computed from differences from the mean of each time series. That is is the ups and downs from the means that go up and down together that give rise to positive correlation, or go in opposite directions that give rise to negative correlations.

For example try this out in a spreadsheet.

Make a set of "noise returns" = rand().

Make a set of returns A = 0.10 + noise (that is a mean return of 10% but with ups and downs)

Then make a set of returns B = 0.10 - noise (use the same column of random numbers, not a new set of random numbers)

Now you compute the correlation and you find a perfect -1.

But the sum of the two is the magic portfolio that returns 10% every year with zero volatility. The (mean) returns do not cancel, the noise cancels.

If you want play with A and B where you again use the same fundamental noise but make A =0.10+0.6667*noise and B = 0.05+.3333*noise

So B has a smaller return and a smaller "risk", again you get correlation = -1. But if you want to get the noise to cancel you need to use 1/3 A and 2/3 B - This is more like stocks and bonds (though with a perfect corr= -1 rather than the more realistic corr ~= 0).

Next if you want to continue playing see what happens when you use independent sets of noise.

So that is the math. I doubt in the real world you can do much better than zero correlation such as stocks and treasuries.
Rodc, you just described something interesting. Interesting, but different. You described perfectly the case where two highly positively correlated assets have perfectly negative correlation of rate of change in returns.

In this case, a return rate change in asset A is completely canceled out by a return rate change in asset B. The rate of change in return between the assets is perfectly negatively correlated. You preserve the growth rate at 10%. Because correlation is measured as a relationship not in the rate of change in return , but in change in value, these assets are highly positively correlated. Because we always know that the average return of Asset A is 10 + avg(rand)= 10.5 (using excels rand function, where rand draws from a uniform pdf with min 0 and max 1), and
Asset B is 10 - avg(rand)=9.5.
you see, the prices of these assets are positively correlated regardless of the noise. One just moves up higher than the other.

Instead, let's show the market use of correlation (price relationships between assets). Okay, so we're no longer discussing rate changes in returns but just prices.
Let's say hold two assets and they are each valued at $10. Now, do your spreadsheet game, where you add rand() to asset A and subtract that value from asset B.





What's your return? Did the return of one asset cancel the other return out? We preserved value but we locked in a 9% growth.
now what if the future value of Asset A was 10 + rand(), but the future value of Asset B was 10 +0.1*rand(). It dragged only 10% of returns correct?

If you want a brief review on correlation and why non-correlation is most desirable, check out Investopedia's video.
http://www.investopedia.com/terms/n/neg ... lation.asp
Last edited by JZinCO on Mon Dec 14, 2015 11:13 pm, edited 1 time in total.
JZinCO
Posts: 168
Joined: Fri Mar 20, 2015 6:32 pm

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by JZinCO »

NEWS BULLETIN
I was describing correlation of prices. you were describing correlation of returns. We are talking about different things and not being clear, or I was stubborn :)
This page describes correlation brilliantly (http://financialexamhelp123.com/covaria ... rrelation/). I was right, you were right.
But, in a way you and Maynard were more right because correlation of returns is the metric that is more 'standard'. Websites devoted to market fundamentals 101 are totally sloppy with regards to discussing correlation. For example, Investopedia describes it as correlation of prices; boglesheads' blog describes it as correlation of returns. This very important but often neglected modifier is the root of confusion.

So, this might go back to OP.
When asset prices are negatively correlated, variability in principle is reduced.
When returns are negatively correlated, variability in returns is reduced.

From the website (long):
A third error that occurs in the context of finance is that people tend to be very casual (i.e., sloppy) in their language... This is particularly important because the casual language can allow you to confuse, for example, correlation of returns with correlation of prices. If two assets have a strong negative correlation of returns, they can have a negative correlation of prices, zero correlation of prices, or a positive correlation of prices. The same is true if they have a strong positive correlation of returns: the correlation of prices can be negative, zero, or positive. We’ve seen examples above where two securities have perfect negative correlation of returns and strong, positive correlation of prices
Ahhhh, it feels great when I'm viewing concepts narrowly and I'm dug in only to learn of my ignorance. What a great learning opportunity.
Thanks for the discussion guys and/or gals.
User avatar
Maynard F. Speer
Posts: 2139
Joined: Wed Mar 18, 2015 10:31 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Maynard F. Speer »

JZinCO wrote:Ahhhh, it feels great when I'm viewing concepts narrowly and I'm dug in only to learn of my ignorance. What a great learning opportunity.
Thanks for the discussion guys and/or gals.
That's what I come here for :beer
"Economics is a method rather than a doctrine, an apparatus of the mind, a technique of thinking, which helps its possessor to draw correct conclusions." - John Maynard Keynes
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Stocks and bonds move in opposite directions and improve results? Really?

Post by Rodc »

JZinCO wrote:NEWS BULLETIN
I was describing correlation of prices. you were describing correlation of returns. We are talking about different things and not being clear, or I was stubborn :)
This page describes correlation brilliantly (http://financialexamhelp123.com/covaria ... rrelation/). I was right, you were right.
But, in a way you and Maynard were more right because correlation of returns is the metric that is more 'standard'. Websites devoted to market fundamentals 101 are totally sloppy with regards to discussing correlation. For example, Investopedia describes it as correlation of prices; boglesheads' blog describes it as correlation of returns. This very important but often neglected modifier is the root of confusion.

So, this might go back to OP.
When asset prices are negatively correlated, variability in principle is reduced.
When returns are negatively correlated, variability in returns is reduced.

From the website (long):
A third error that occurs in the context of finance is that people tend to be very casual (i.e., sloppy) in their language... This is particularly important because the casual language can allow you to confuse, for example, correlation of returns with correlation of prices. If two assets have a strong negative correlation of returns, they can have a negative correlation of prices, zero correlation of prices, or a positive correlation of prices. The same is true if they have a strong positive correlation of returns: the correlation of prices can be negative, zero, or positive. We’ve seen examples above where two securities have perfect negative correlation of returns and strong, positive correlation of prices
Ahhhh, it feels great when I'm viewing concepts narrowly and I'm dug in only to learn of my ignorance. What a great learning opportunity.
Thanks for the discussion guys and/or gals.
Every article on correlations in investing I have ever read is in regards to returns - that is what drives Modern Portfolio Theory (well to be clear: covariance in the returns) and what matters as far as building a portfolio - you pick (with extremely rare exception) assets with positive mean return and with low correlation in returns (to get a nice positive mean return and low variance).

Can you explain why one would care about correlation of price rather than correlation of returns?
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Post Reply