'intelligent' asset allocation

Asset allocation, risk, diversification and rebalancing. Pros/cons of hiring a financial advisor. Seeking advice on your portfolio?
Post Reply
User avatar
ClosetIndexer
Silver Ring
Silver Ring
Posts: 259
Joined: 27 Feb 2012 02:20
Location: Vancouver, BC
Contact:

'intelligent' asset allocation

Post by ClosetIndexer » 27 Feb 2012 02:46

Hi FWF!

Longtime lurker, and Norbert's probably saved me thousands with his gambit, but I didn't have a good enough reason to post until today. (Closest I came was when I had the exact same thought as Icarus in this post (viewtopic.php?p=238019&sid=0b5545c57be7 ... ce#p238019), but then found... that post.)

Anyway, there's something I've been internally debating for a while, and I think I've finally found a satisfactory way of explaining it to myself. I, like many/most people on this forum, am of the belief that indexing is the most intelligent strategy for retail and likely institutional investors. It's been shown time and again that on average, market timers fail to add alpha. Even stock picking is so difficult as to be essentially impossible (on average) for a retail investor. And even if the occasional Peter Lynch or Warren Buffet exists, fees and impact costs will eliminate their contribution too. So, passive indexing is the only thing that makes sense. And yet, many 'indexers' can't help but make occasional bets on the market.

Even William Bernstein, after discussing re-balancing on P. 161 of 4 Pillars, writes "Ideally, when prices fall dramatically, you should go even further and actually increase your percentage equity allocation, which would require buying yet more stocks." Unsurprisingly, Benjamin Graham gave similar advice to 'enterprising' investors. Both warn that this technique is only for those who are willing to put in the time and effort to understand the logic behind their decisions, but if done intelligently, it should be sound. For example, contrast these two situations just involving two asset classes:

Situation 1: S&P 500 P/E is 11, long bond interest rate is 3%
Situation 2: S&P P/E 30, long bond interest 8%

All else being equal, does it really make sense to have the same equity/bond split in these two situations? Obviously the long term expectation of the stock market in the first case is better. (In fact, with the dividend yield we can put a numeric estimation on it using the Gordon equation.) And the long term expectation of long bonds are given by their interest rate. So many people, even if trying to be 100% passive, can't help but shift their weighting more to stocks in case 1 and bonds in case 2. (Note that I'm not talking about 'tactical asset allocation', which is defined as shifting allocations to take advantage of short-term inefficiencies, but rather taking into account long-term expectations, and varying allocations if those long-term expectations change.)

But of course, you still have to look at that and think "Ah! Market timing!" How can we reconcile that market timing is doomed to failure barring short-term luck, but shifting your allocation based on market conditions appears to have merit? I've finally found an explanation that at least to me seems right. I'm sure this is nothing new, but I figured I'd share it in case anyone else finds my particular phrasing of the ideas useful. Or disagrees with me in a compelling way!

SO, the classical market timing concept is basically to invest in stocks when they're expected to go up, then sell out (or even short the market) when they're expected to go down, for some definition of 'expected to go up/down'. As I see it, there are two problems with this. First, it's true that when stock prices are inflated (and therefore yields are reduced), the long-term expectation of the market is lowered. (According to the Gordon equation, the long-term expectation of the market can be estimated by adding its dividend yield to its expected dividend growth.) However, even at inflated prices, the market's long-term expectation is lower, but is still positive! Therefore, any technique that involves cashing out, or worse, shorting, will fail unless it gets lucky. That is fundamentally different from shifting allocation from an asset class with a low long-term expectation to another with a higher expectation. (Of course, there are logical reasons to accept reduced expectation, like diversification, but that's another discussion.)

The other issue is that the majority of market timers tend to focus on short-term predictions of market direction. Like most of you, I've accepted that predicting short term market fluctuations is impossible. Certainly on average it must be, by definition. In the long-term though, it is possible to make a logical estimate. Based on that, it is theoretically logical to set an asset allocation, then eventually, if the long-term expectation changes, to shift that allocation. And considering how that would happen, it makes sense. Say you find yourself in my 'Situation 1' above. So you decide to shift your weighting slightly toward equities. Then some years later, you re-evaluate and find that the interest rate has risen, so it makes sense to sell some of your stocks and shift to more bonds. If interest rates rose, that means bond prices decreased, so your bond underweight paid off. Similarly, say you re-evaluate and find that the P/E of the market is no longer so low, so again, you shift some money out of stocks and into bonds. But that means the price of the stock market has risen relative to earnings, so your over-weighting of stocks likely paid off. Finally, if you re-evaluate and see the same picture, no problem, just re-balance normally. Basically, you set your allocation based on long-term expectations, and only change it if those long-term expectations change. But if they do change, by definition, your allocation has done its job. The difference from market timing is that you don't know when or if that will happen, but fortunately, you also don't care.

That about sums it up. All that said, I firmly believe that a fixed asset allocation along with dollar cost or value averaging is probably the best approach for almost everyone, and is likely the approach I will continue to take. Aside from being easier, it removes any possibility of psychological factors playing a roll. But I do think it's theoretically possible to do somewhat better by taking into account long-term market expectations when setting (and occasionally modifying) asset allocations - as long as those expectations are arrived at logically.

What do you think?
Last edited by Peculiar_Investor on 07 Feb 2014 07:04, edited 1 time in total.
Reason: replace old domain name with www.financialwisdomforum.org to reflect new domain name effective 19-Jan-2014

User avatar
newguy
Gold Ring
Gold Ring
Posts: 8073
Joined: 10 May 2009 18:24
Location: Montreal

Re: 'intelligent' asset allocation

Post by newguy » 27 Feb 2012 10:13

I thought about something like that,
viewtopic.php?f=29&t=111659

There's some links to follow as well. I still think it's a good idea but I'm not sure that it makes more money than just regular rebalancing.

newguy
Last edited by Peculiar_Investor on 07 Feb 2014 07:04, edited 1 time in total.
Reason: replace old domain name with www.financialwisdomforum.org to reflect new domain name effective 19-Jan-2014

User avatar
ClosetIndexer
Silver Ring
Silver Ring
Posts: 259
Joined: 27 Feb 2012 02:20
Location: Vancouver, BC
Contact:

Re: 'intelligent' asset allocation

Post by ClosetIndexer » 28 Feb 2012 02:14

Thanks for the link. That does seem to be along the same lines. The link in that thread no longer works, but I assume this was the article (or similar)? http://www.passionsaving.com/how-to-of-investing.html

I think I agree with a lot of the approach (ie "He lowers his stock allocation at times of high valuations, putting the money into super-safe asset classes like Treasury Inflation-Protected Securities (TIPS), IBonds, or certificates of deposit (CDs)."), but not necessarily with the rational behind it ("He thereby limits his losses during times when stock prices are headed downward, leaving him with more to invest in stocks when prices return to reasonable levels." (emphasis mine)). That's the part I maintain no one can know, and is the difference in mindset between a market timer and a true value investor. I would instead say something like "He thereby reduces his exposure to stocks when their long-term expectation is reduced compared to other asset classes." Just like the stock-picking form of value investing, I do think it's important not to expect it to always pay off with a quick mean-reversion. Yes, usually a bear market and a drop in stock prices follows after the P/E of the market hits some level X, but there's no guarantee it will always do so. Graham discussed this in the Intelligent Investor, iirc. He basically said you can control the price you're willing to buy something at, but not when (or even if) the market will eventually offer a higher one. In the meantime though, you can just be satisfied that you bought at a reasonable price and therefore are presumably earning a good dividend yield, and have reduced your downside exposure (at least compared to having bought at a higher price).

On the other hand, I guess as long as you arrive at the same strategy, your exact reasoning behind it don't matter so much. Unless it causes you to bail out when things don't go as planned I suppose!

Edit: Regardless, I definitely agree with you here
newguy wrote:I like the idea of value investing, but not the work. If you combine this with the inconsistency of screeners and the time needed to look at each company individually, I wouldn't bother. I'd like to come up with some formula for valuation of the entire market and then adjust allocations based on that.
and I think I've basically reached the same point as idOp
idOp wrote:Then, if you can come up with some predicted returns, the question is, how are you going to base your allocation on that? This seems like a question it would be easy to overkill, and any answer is going to be based on arbitrary premises anyway. So maybe it would be best to keep it simple and just set ranges for equity/bonds/cash that you can live with (risk and return) and adjust your position in there somehow based on the differences between projected and average returns of the classes. "Somehow" meaning try to do something very simple that adds up to 100% but puts a little more weight on what's expected to do best. :)

As you can see from my rambling, even putting this much into practice was too overwhelming for me. :)

User avatar
newguy
Gold Ring
Gold Ring
Posts: 8073
Joined: 10 May 2009 18:24
Location: Montreal

Re: 'intelligent' asset allocation

Post by newguy » 28 Feb 2012 08:03

I also meant this link to a bit of a study posted here.

viewtopic.php?f=4&t=106998&start=0

newguy
Last edited by Peculiar_Investor on 07 Feb 2014 07:04, edited 1 time in total.
Reason: replace old domain name with www.financialwisdomforum.org to reflect new domain name effective 19-Jan-2014

George$
Gold Ring
Gold Ring
Posts: 2612
Joined: 18 Feb 2005 20:46
Location: Toronto

Re: 'intelligent' asset allocation

Post by George$ » 28 Feb 2012 08:23

ClosetIndexer wrote:.....
The other issue is that the majority of market timers tend to focus on short-term predictions of market direction. Like most of you, I've accepted that predicting short term market fluctuations is impossible. Certainly on average it must be, by definition. In the long-term though, it is possible to make a logical estimate.
I'm impressed with the effort that went into your thoughtful post - and I will re-read it more carefully later (I'm rushed at the moment)

I agree that the short term has so much market noise that I think it unlikely to exploit it in any way.

But even in the long-term I don't think the market has to follow a script from the past. Look what has happened to the Japanese market over decades and counting.

My second thought was - you don't mention inflation. This is a major issue - in my mind the real return is what matters - not the nominal returns. It seems to me you may be focused only on the latter.

Possible major inflation (double digit maybe?) is what scares me. And I'm not alone. Look at ReaL Return Bond yields. In 2000 they were briefly at 5.0%. Today they are one tenth that, at about 0.5% real yield. This says to me that there are many others who are seriously scared about the likelihood of major future inflation.

Just my two cents. :)
“The search for truth is more precious than its possession.” Albert Einstein

User avatar
ClosetIndexer
Silver Ring
Silver Ring
Posts: 259
Joined: 27 Feb 2012 02:20
Location: Vancouver, BC
Contact:

Re: 'intelligent' asset allocation

Post by ClosetIndexer » 29 Feb 2012 04:56

newguy wrote:I also meant this link to a bit of a study posted here.

viewtopic.php?f=4&t=106998&start=0

newguy
Wow... had no idea what a hotly debated subject this was! Some good thoughts there, if you can filter out the... irrational exuberance.
George$ wrote:I'm impressed with the effort that went into your thoughtful post - and I will re-read it more carefully later (I'm rushed at the moment)

I agree that the short term has so much market noise that I think it unlikely to exploit it in any way.

But even in the long-term I don't think the market has to follow a script from the past. Look what has happened to the Japanese market over decades and counting.

My second thought was - you don't mention inflation. This is a major issue - in my mind the real return is what matters - not the nominal returns. It seems to me you may be focused only on the latter.

Possible major inflation (double digit maybe?) is what scares me. And I'm not alone. Look at ReaL Return Bond yields. In 2000 they were briefly at 5.0%. Today they are one tenth that, at about 0.5% real yield. This says to me that there are many others who are seriously scared about the likelihood of major future inflation.

Just my two cents. :)
Thanks! And yes, I agree that you certainly can't just project the past forward to make a long-term prediction. That's why the Gordon Equation seems like a logical method. Also implicitly deals with inflation. On further inspection though it's likely too uncertain to base useful decisions on, aside from perhaps your initial allocation. http://raddr-pages.com/research/gordon.htm. Also of course inflation could affect your alternate asset classes.

Even if we decide we can't make long-term predictions of the stock-market with sufficient accuracy to inform allocation shifts, it seems to me you should be able to do so with bonds. Especially at times like this when interest rates literally can't fall much further. All else being equal, wouldn't it make sense to have somewhat less of your allocation in fixed income (and therefore more in equities) when interest rates are near zero? (Of course, all else is never equal!)
Last edited by Peculiar_Investor on 07 Feb 2014 07:04, edited 1 time in total.
Reason: replace old domain name with www.financialwisdomforum.org to reflect new domain name effective 19-Jan-2014

like_to_retire
Gold Ring
Gold Ring
Posts: 4664
Joined: 27 Feb 2005 07:14
Location: Canada

Re: 'intelligent' asset allocation

Post by like_to_retire » 29 Feb 2012 07:52

All else being equal, wouldn't it make sense to have somewhat less of your allocation in fixed income (and therefore more in equities) when interest rates are near zero? (Of course, all else is never equal!)
In the short term perhaps, but the present economic situation would indicate that interest rates will rise over the long term (more people will leave the market for fixed income's higher rates, while interest rate sensitive companies suffer). Your course of action in the quote appears more like market timing in the short term than any long term allocation call.

ltr

flywaysuzy
Gold Ring
Gold Ring
Posts: 1221
Joined: 23 Feb 2005 10:04

Re: 'intelligent' asset allocation

Post by flywaysuzy » 29 Feb 2012 09:07

'Interest rates can't fall much further'- this is the talk of people only armed with whole numbers-there are plenty of folks out there in poor countries actually paying to keep their money in bank accounts. I imagine people paying bank fees here in Canada are in the same boat (perhaps called the Costa Integera?) As global instability becomes the norm, I could see banks charging money to keep large amounts of cash here as well...
suzy

User avatar
ClosetIndexer
Silver Ring
Silver Ring
Posts: 259
Joined: 27 Feb 2012 02:20
Location: Vancouver, BC
Contact:

Re: 'intelligent' asset allocation

Post by ClosetIndexer » 29 Feb 2012 16:51

like_to_retire wrote:
All else being equal, wouldn't it make sense to have somewhat less of your allocation in fixed income (and therefore more in equities) when interest rates are near zero? (Of course, all else is never equal!)
In the short term perhaps, but the present economic situation would indicate that interest rates will rise over the long term (more people will leave the market for fixed income's higher rates, while interest rate sensitive companies suffer). Your course of action in the quote appears more like market timing in the short term than any long term allocation call.

ltr
Hmm.. not sure about this for a couple reasons. I would argue that making decisions on the known fact of where interest rates are at now is a more general/long-term decision than basing things on where you expect them to go, which no one knows for sure. ie rates could stay flat for a long time from here, or they could rise, or they could even half again, but regardless we 'know' that they'll be lower on average over the long term than if they were starting out at a higher level.

While I do agree that it seems likely in the long term they will rise somewhat from this point, if we are in for long-term rising rates, that would almost certainly be harder on bonds than equities, assuming you're properly diversified. In the future if/when rates are higher, you're free to re-examine your allocation (which if the logic in my original post holds, would be a smart thing to do).

Post Reply