AyG
Crazy Mango Extraordinaire
Posts: 5,871 
Likes: 4,555
|
Post by AyG on Feb 7, 2018 11:55:43 GMT 7
As I've previously written, looking at very short term (e.g. daily) movements in prices trying to see correlation (or better, inverse correlation) is usually meaningless. I did wonder, however, whether looking at the sharp movements of the last week might teach us anything.
I considered four categories of instrument (equities, government bonds, corporate bonds, and government index linked bonds), and looked at the one week price movements for representative iShares ETFs. First looking at the UK, traditional wisdom tells us:
- Equities and Gilts should typically be negatively correlated
- Corporate bonds should move further than Gilts
- Index link bonds should rise on an expectation of rising inflation
Well, what do we see?
FTSE (ISF) -5.78%
Gilts (IGLT) -0.42%
Corporate Bonds (SLXX) -1.17%
Index Linked Bonds (INXG) -1.44%
- No evidence of negative correlation, though the bonds fell less than the equities, so have helped to preserve wealth
- Corporate bonds moved further than gilts, as expected
- Index linked bonds fell, rather than rose, which is odd given that one of the stated causes of recent price movements has been the expectation of a rise in inflation
Turning to the US (and accounting in USD)
S&P 500 (CSPX) -5.03%
US Treasuries (CBU0) +1.34%
US Corporate Bonds (LQDE) +0.59%
US TIPS (IDTP) +1.23%
- Here bonds and equities have been negatively correlated, as expected
- Corporates have moved less than treasuries, which is slightly odd.
- TIPS have risen, as would be expected.
In short, the US markets have acted fairly “classically”, whilst the UK markets have reacted differently.
I'm not sure what this all means, if anything perhaps
It is a tale
Told by an idiot, full of sound and fury,
Signifying nothing.
|
|
chiangmai
Crazy Mango Extraordinaire
Posts: 6,481
Likes: 5,570
|
Post by chiangmai on Feb 7, 2018 16:32:14 GMT 7
Ah surely every piece of data tells us something useful, even if it's only that we need more data or that we're looking at the wrong data!  As I said earlier I haven't looked very closely at the damage done, I merely glanced at a few numbers to see if there was anything glaring. But probably tomorrow or the day after I will look more closely, the purpose in doing so is as you have written, to see if there are lessons that I can learn, bearing in mind that we are both in different classes and study years. Off the top of my head however I would say that my quick peak aligns with what you've reported, the global equity funds I hold were down the most, my index-linked Gilts we're second and that surprised me also and my bond funds were down the least, in some cases minimally. I will report back some detail of my portfolio's performance in due course, I have an expectation though that my five start fund managers for whom I am paying a fair whack, will have minimised my losses on the downside, it'll be interesting to see if that's the case. |
|
|
|
Post by Fletchsmile on Feb 7, 2018 17:57:29 GMT 7
A key thing to remember with correlations is that they vary over 1) time and 2) different circumstances
What we've seen in particular in the last decade or so are circumstances that have been unusual compared to many years that people are familiar with, and circumstances have changed.
Flood the world with liquidity, and in an excess of liquidity its no surprise that asset classes become more correlated than they have in other times and we see this now classic risk-on / risk off talk
So I'd disagree that equities should be negatively correlated with bonds as a blanket statement. We need to look at the circumstances.
There's been a fair bit written about inflation linked bonds in the last few years too. Inflation linked bonds are not just driven by inflation and expectations about it though, but also by interest rates. This can get overlooked. The 2 factors play interest rates and inflation rates can play off to an extent in determining total return.
These are incidentally factors among others why I'd disagree when people say conventional bonds are a waste of time. They have their place as do inflation linked bonds.
So for me knowing that correlations vary over time and in different circumstances it doesn't come as a surprise to see things not always behaving as we first expect, particular over the short term.
A common problem with economic theories is that they pick one or two factors and assume these levers alone are impacting the economic universe. The whole economic ecosystem is much more complex. Certain factors may dominate from time to time and be more influential than others, but there's really just a lot of factors at play. One at the moment you heard a lot recently is "Is the Phillips curve broken" for example. I don't think it's broken, just it's not a complete picture and the whole ecosystem is much more complicated than those simple factors.
|
|
mistermember
Crazy Mango
Not all that is gold glitters.
Posts: 821
Likes: 376
|
Post by mistermember on Feb 7, 2018 18:00:33 GMT 7
Correlation before marriage is a sin.
|
|
|
|
Post by Fletchsmile on Feb 7, 2018 18:04:30 GMT 7
Here's an article from a while back that struck me on the issue of index linked bonds. The idea that people have done well out of them by getting it wrong. Banking on higher inflation. However, as there hasn't been any, rates have been cut and yields go down, so bond prices go up  ========================================================================= Will UK inflation-linked bonds be the choice of the wise? www.ft.com/content/48231068-581c-11e7-80b6-9bfa4c1f83d2May need to sign in so here's a few excerpts worth a read as it then gives a few reasons of why they may be a good choice to buy |
|
chiangmai
Crazy Mango Extraordinaire
Posts: 6,481
Likes: 5,570
|
Post by chiangmai on Feb 8, 2018 6:53:56 GMT 7
A key thing to remember with correlations is that they vary over 1) time and 2) different circumstances What we've seen in particular in the last decade or so are circumstances that have been unusual compared to many years that people are familiar with, and circumstances have changed. Flood the world with liquidity, and in an excess of liquidity its no surprise that asset classes become more correlated than they have in other times and we see this now classic risk-on / risk off talk So I'd disagree that equities should be negatively correlated with bonds as a blanket statement. We need to look at the circumstances. There's been a fair bit written about inflation linked bonds in the last few years too. Inflation linked bonds are not just driven by inflation and expectations about it though, but also by interest rates. This can get overlooked. The 2 factors play interest rates and inflation rates can play off to an extent in determining total return. These are incidentally factors among others why I'd disagree when people say conventional bonds are a waste of time. They have their place as do inflation linked bonds. So for me knowing that correlations vary over time and in different circumstances it doesn't come as a surprise to see things not always behaving as we first expect, particular over the short term. A common problem with economic theories is that they pick one or two factors and assume these levers alone are impacting the economic universe. The whole economic ecosystem is much more complex. Certain factors may dominate from time to time and be more influential than others, but there's really just a lot of factors at play. One at the moment you heard a lot recently is "Is the Phillips curve broken" for example. I don't think it's broken, just it's not a complete picture and the whole ecosystem is much more complicated than those simple factors. For my benefit and clarity: when you write in para. 1 that they vary over time, I presume you mean that circumstances change over time rather than time (or duration) itself is a factor? That being the case it would support the complexity issue rather than trying to suggest one or two simple rules govern everything. |
|
|
|
Post by Fletchsmile on Feb 8, 2018 10:56:06 GMT 7
For my benefit and clarity: when you write in para. 1 that they vary over time, I presume you mean that circumstances change over time rather than time (or duration) itself is a factor? That being the case it would support the complexity issue rather than trying to suggest one or two simple rules govern everything. I just meant that when circumstances change correlations often change. For the time aspect, I meant that at and over different periods of time, even though circumstances may appear to be similar, correlations may change. As to exactly why, I guess someone one need to look more. Could be just randomness, could be changes in circumstances that we maybe can't see easily, could be other factors as time and circumstances are unlikely to be the only two factors, although two of the biggies |
|
|
|
Post by Fletchsmile on Feb 8, 2018 11:03:32 GMT 7
Over the years I've had a fair bit of experience with correlations, from academic theory, general reading, my own investing, experiences at work etc. Couldn't think of any particular articles or texts to recommend, so did a quick google. This came at the top of the google list. Only skim read quickly, but seems a decent article and relevant in the context of this thread, including some of AyGs observations and CM's monitoring elsewhere of your own portfolio. Probably not a perfect article, but some good take aways from it www.vanguard.com/pdf/s130.pdf |
|
