The Cult of the Equity

<!--EZCODE BOLD START--> Contradiction Number 1 <!--EZCODE BOLD END-->
This Cult of the Equity is based entirely on an analysis of past performance.

Yet past performance is otherwise decried in the "book" as a pointer to choice of fund manager or fund or future performance.


Reference is made to 94% of 20 year periods since 1871. In fact there have only been seven such independent periods ( overlapping periods do not count as they are simply repeats of the same data)

Hi Sea Pigeon

I am working on a revision of these chapters and so I have cut these comments from an earlier topic and reposted them in your name.

"The past performance of investment managers is no guide to the future performance of investment managers. "What this means is that just because the Focus 15 has outperformed over the past year, doesn't mean that it will do it again next year. The fact that New Ireland have been the best performer over 15 years, doesn't say anything about their future comparative performance.

The irrelevance of past performance does not apply to everything. I expect that this coming winter will be colder than the past summer. This sort of past performance does persist.

You and your equine friends have made very valid arguments that some stockmarkets may be overvalued and I have attempted to address these issues. But I see no relevance to the past performance of investment managers argument.

<!--EZCODE BOLD START--> How many 20 year periods?<!--EZCODE BOLD END-->

I handn't noticed this point at all until today and I am still trying to figure it out.

Let's assume for the moment that long term past performance of stockmarkets is a reasonable guide to the future.

Cash has beaten equities in 6% of 20 year periods over the past 130 years. Am I correct in saying that if you had picked a starting period at random over the past 130 years, that you would have had a 94% chance of doing better in equities on a 20 year holding period ?

If I am not correct, how would I estimate the chance of equities beating cash over the next 20 years ?

Brendan

As you know, Oh Leader, Sea Pigeon is no longer with us but I will attempt in my humble way to step into his hoofs.

The point about past performance was banal and your refutation is sound. However, we are still talking about predicting human behaviour and development which doesn't quite have the same scientific regularity as the planetary motions or the seasons.:smokin

<!--EZCODE BOLD START--> How many 20 year periods have there been?<!--EZCODE BOLD END-->

Yes, if you pick a point in time at random over the last 130 years there is a 94% chance that Equities will have outperformed Cash. But how statistically significant is that observation? How many such <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> periods have there been?

You could have picked a starting point from any of 110 years. Or from 1,320 months. Or from 40,000 days. An <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> sample of 40,000 would be truly statistically significant.

(<!--EZCODE ITALIC START--> Aside: This sort of "backtesting" is sometimes very misleadingly used to analyse say the past returns on the EuroStoxx 50, which has only been around for a few years but can be represented as having had many starting months. <!--EZCODE BOLD START--> The Book<!--EZCODE BOLD END--> might carry wealth warnings about these backtesting demonstrations.<!--EZCODE ITALIC END-->)

Let my try and explain.

How <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> is the period from, say, <!--EZCODE ITALIC START--> 1/1/1940 to 1/1/1960 <!--EZCODE ITALIC END--> compared to the period <!--EZCODE ITALIC START--> 1/2/1940 to 1/2/1960 <!--EZCODE ITALIC END-->. If the first period has done well the second has almost certainly done equally well but this is not really giving us any extra statistical information.

In a similar way the perod from 1940 to 1960 is really very closely linked to the period from 1941 to 1961. So even if you restrict your sampling to picking a start year you do not have an <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> sample of 110 choices which a reader of <!--EZCODE BOLD START--> The Book<!--EZCODE BOLD END--> might glean from the sweeping statement that <!--EZCODE ITALIC START--> "over the last 130 years etc."<!--EZCODE ITALIC END-->

What was Sea Pigeon getting at?

He was pointing out that our experience of 20 year periods in the modern economic environment is really very limited and not a reliable statistical guide as to what the next 20 years might hold. I might even suggest as a sort of separate point that the world of 1870 is so patently different from today's as to be totally irrelevant. We might be down to at most 2 or 3 <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> 20 year periods worthy of comparison with our current condition.

Your estimate of 94% is as good as any of the outrageously sophisticated and pretentious stochastic models that I have seen.:rolleyes

What is at question is how significant or reliable is your estimate as a guide to the future. The statement in <!--EZCODE BOLD START--> The Book<!--EZCODE BOLD END--> suggests that ONE HUNDRED AND THIRTY years' experience back up your assessment whereas really one has only had a handful of <!--EZCODE BOLD START--> independent<!--EZCODE BOLD END--> 20 year periods in modern human socio economic history.

My suggestion - tone down the significance of your statistical observation.|I

What I would like to know today is what is the liklihood that equities will outperform cash over the next 20 years ?

No one can answer this with certainty. We can look at recent history and we can look at long term history. We could also look at current valuations to see how they look compared to historical valuations.

If recent history or if long term history is any guide, it looks good for equities.

I don't understand why the observations need to be independent. I appreciate that a good year will show up in 20 different samples, but so too will the bad years.

While you say that it is as good as any other, do you think it is good enough ?

Brendan

Oh dear, us horses are very poor communicators.|I

Let's see if an example will show what I am neighing on about.

Imagine a product which promises the growth in ISEQ over 5 years. I might want to show how this product would have done in the past to give a flavour of its potential. Two such possible statements might be:

(a) ISEQ has never fallen in value over a 5 year period <!--EZCODE ITALIC START--> (let's assume that's true, I'm not sure if it is)<!--EZCODE ITALIC END-->.

(b) If this product had been available then on none of the 3,000 days since ISEQ commenced would it have shown a loss.

Statement (b) is grossly misleading (from a statistical viewpoint). There have really only been about three distinct 5 year periods in ISEQ's history. To suggest that the product has shone through on thousands of occasions is multiple counting the same experience.

I know you haven't gone quite that far but it still reads as if there is a vast history of 20 year periods from which you gleaned your 94% statistic.

What I would say, if I was a disciple of the <!--EZCODE BOLD START--> <!--EZCODE ITALIC START--> CoE<!--EZCODE ITALIC END--><!--EZCODE BOLD END-->, as you obviously are, would be:

<!--EZCODE ITALIC START--> If history is anything to go by, the experience of the last century would suggest that equities have a 94% chance of being on top over the next 20 years. But there are those who point out that a century is a very short space of time for drawing inferences as to what will happen in the next 20 years.<!--EZCODE ITALIC END--> :smokin

Hang on a minute !

<!--EZCODE QUOTE START--><blockquote>Quote:<hr> A century is a very short space of time to draw inferences...<hr></blockquote><!--EZCODE QUOTE END-->

<!--EZCODE QUOTE START--><blockquote>Quote:<hr> I might even suggest as a sort of separate point that the world of 1870 is so patently different from today's as to be totally irrelevant.<hr></blockquote><!--EZCODE QUOTE END-->

Are you saying that 100 years is too short, while 130 years is too long ?

Or maybe,we should draw absolutely no inferences from the past at all ? The recent past is too short and the distant past is irrelevant !

Brendan

I have just had an unexpected visit from a fine backgammon player and statistician and have put your argument to him.

He suggests extending the sample by looking at the 14 independent 10 year periods and the 28 independent 5 year periods. While that is a very fine suggestion, I don't have the data to do this report. Over to you Sir Ivor ?

Brendan

Re: Not long enough, Too long

<!--EZCODE QUOTE START--><blockquote>Quote:<hr> <!--EZCODE BOLD START--> <!--EZCODE ITALIC START--> "Are you saying that 100 years is too short, while 130 years is too long ?"<!--EZCODE ITALIC END--><!--EZCODE BOLD END--><hr></blockquote><!--EZCODE QUOTE END-->

That is precisely what I am saying and there is really no contradiction here.:\

I suggest you need about 1,000 years of similar socio- economic conditions to make <!--EZCODE BOLD START--> confident<!--EZCODE BOLD END--> statistical inferences about the next 20. We haven't got 1,000 years of similar socio economic conditions, in fact we scarcely have 40 such years, certainly the conditions of 1870 are totally unrecognisable from today's conditions.|I

Your Backgammon friend seems to appreciate the point and if his initials are SC I would certainly respect his views.

Sorry, but I don't have that sort of data to hand. In any case, I suggest there is simply not enough data to be statistically confident about the next 20 years, but there is no problem in your statement that the last century or so has come out 94 to 6 in favour of the equity. It's just we shouldn't take that too seriously as a predictor of the next 20 years. Ask the Japanese!:smokin

My basic problem with <!--EZCODE BOLD START--> The Book<!--EZCODE BOLD END--> is that your endorsement of the <!--EZCODE BOLD START--> <!--EZCODE ITALIC START--> CoE<!--EZCODE ITALIC END--><!--EZCODE BOLD END--> is purely based on an empirical analysis of past experience. There is no attempt to actually fundamentally argue that equities should outperform.

Certainly I agree that the many quasi economic arguments are faulty for they miss the point that this is a market and if it was an economic certainty that equities should outperform, the markets would build that into the price. Have they already done so with P/Es of 20 and more?

One possible market based argument for equities is that the market will always demand a premium for volatility/uncertainty. Thus if you ever read a stockbroker report they will invariably forecast future earnings and then discount these back at a discount rate maybe 3% ahead of the risk free rate.

If all market participants are either explicitly or implicitly doing these same sums then there might be a case that over time, provided the market's forecasts are on average correct, the market will deliver 3% extra return over the risk free rate.

I'm not convinced. :smokin

Hi Sir Ivor

Your raise very valid questions. I have argued that the stockmarket has been the best performing investment in all economies over the long term. Therefore, every long term investor should be 100% invested in the stockmarket. I have also suggested that they should not attempt to time the market.

You have accepted the historical fact, but you ask why has it done so and why should it continue ?

I got up early this morning and reread <!--EZCODE ITALIC START--> Stocks for the Long Run<!--EZCODE ITALIC END--> by Jeremy Siegel to see could I find an answer. This is what he says:
<!--EZCODE QUOTE START--><blockquote>Quote:<hr> The reasons for the persistence and long-term stability of stock returns are not well understood. Certainly the returns on stocks are dependent on economic growth, productivity and the return to risk taking…

Political or economic crises can throw stocks off their long-term path, but the resilience of the market system enables them to regain their long-term trend. Perhaps that is why stock returns transcend the radical, political, economic, and social changes that have impacted the world over the past two centuries.

The superior returns to equity over the past two centuries might be explained by the growing dominance of nations committed to free-market economics... The robustness of world equity prices in recent years might reflect the emergence of the golden age of capitalism… Yet even if capitalism declines, it is unclear which assets, if any, will retain value. <hr></blockquote><!--EZCODE QUOTE END-->

So here was I on my own struggling on my own to answer your probing questions, when Professor Siegel can just say "The reasons …are not well understood".

<!--EZCODE BOLD START--> Investor returns from market peaks<!--EZCODE BOLD END-->

<!--EZCODE QUOTE START--><blockquote>Quote:<hr> Many investors, although convinced of the long-term superiority of equities, believe that they should not invest in stocks when stock prices appear at a peak. But this is not true for the long term investor. <hr></blockquote><!--EZCODE QUOTE END-->

Siegel looked at 30 year holding periods from the major market peaks of this century. From the 1929 peak, for a $100 investment, the total real return on stocks was $565 compared to $141 for bonds and $79 for cash.

From the January 1966 peak, the total real returns were $447, £205 and $145 for stocks, bonds and bills.

The average 30 year return after the 6 stockmarket peaks ( 1901, 1906, 1915, 1929, 1937 and 1966) was $510 in stocks, $177 in bonds and $125 in bills.

<!--EZCODE BOLD START--> Where does all this leave me ? <!--EZCODE BOLD END-->

I have two concerns. I would like the book to be correct. I don't want to either overstate or understate the case for equities. But I also want to know what to do with my own investments.

I have decided that, apart from my home, I will stay fully invested in equities. In fact, because I have a mortgage, I have effectively borrowed to invest in equities.

I have not yet decided how to rewrite the bits in the book. I think I will leave the overall advice stand and include any well written CounterViews. Sir Ivor, I would be delighted if you would write a CounterView which I could include.

Brendan

Hi Oh Leader,

Your reference to your personal situation caused me to consider my own. I have my mortgage paid and I have a modest Nest Egg equal to about 3 times my gross salary.

I have that Nest Egg 75% in cash and always have had. The equities I do have are the odd Capetbag and shares from my employer which they keep shoving at me.:eek

So you can see that despite my equine features there is a strong ursine streak in me. This year has been my year, but unfortunately my strategy has not been the correct one in general. All this chat has me wondering what to do, is now a buying opportunity? If I sit in cash, will I regret it in a couple of years' time? I am very confused.|I (<!--EZCODE ITALIC START--> Though I will be certainly investing my SSIA in equities whenever I get round to taking one out. It's investing a lump sum in them that gives me the jitters, one is so vulnerable to the timing.<!--EZCODE ITALIC END-->)

<!--EZCODE ITALIC START--> Siegel's<!--EZCODE ITALIC END--> stuff was fascinating. However, he his still groping for an economic justification of stockmarket outperformance. This is entirely valid looking back, but looking forward, my argument is that if there is still a compeling economic case for a return on equity this will be built into the prices of that equity in the secondhand market.

The only thing that will convince me is if I can believe that investors will always demand a risk premium for investment in equity simply to compensate for the volatility/uncertainty, for which indeed there may be some evidence.

I'd say at this stage we are boring all other AAM contributors (<!--EZCODE ITALIC START--> Mith<!--EZCODE ITALIC END--> excepted, of course) to death.:O

I must decline your kind offer to allow me to reflect the equine/ursine view.:smokin

CofE

Hi All,

Delighted to see this debate continuing (and learning from it all the time).

I was struck by a quote from John Bogle, the indexing guru and founder of Vanguard, which was included in one of the Jim Grant articles posted in another thread recently. Bogle wrote:

<!--EZCODE QUOTE START--><blockquote>Quote:<hr> "The future is not only unknown but unknowable. Yet with the acceptance of Modern Portfolio Theory, the ease of massaging data with the computer .... investors seem to have developed growing confidence that they can forecast future returns in the stockmarket. If you fall into that category, I send you this categorical warning: the stockmarket is not an actuarial table.

When everyone assumes, at least implicitly, that the market is an actuarial table, that the past is inevitably prologue, and that common stocks, held over an extended period, will always provide higher returns than bonds - and at lower risk - then stocks will inevitably be priced to reflect that certainty. <hr></blockquote><!--EZCODE QUOTE END-->

And that seems to me to be exactly what has happened in the recent past. If I had a € for every time someone told me, the US market will go up until 2020, because that's when the average baby boomer will retire and cash in his 401k plan, then ...

So where does that leave us. I'm somewhere between Brendan and Sir Ivor. I accept that stocks probably will give the best long-term returns - for broadly the same risk premium reason as Sir Ivor advances - so I have just over half my assets in stocks. But I worry that (a) I may be wrong, and that (b) the stockmarket is a capricious mistress, so I have just under half my assets elsewhere. Even in boring old assets like cash and bonds, which have been sneered at elsewhere on this board. I try to add to my stock portfolio when I think markets are in low ground (like two weeks ago) and not too when I think they're in high ground. But I know I don't <!--EZCODE BOLD START--> know <!--EZCODE BOLD END--> whether they're in high or low territory. And I try not to make investments that are too lumpy, again for the reason of timing advanced by Sir Ivor. Brendan, I seem to recall your disparaging pound-cost averaging elsewhere on the board, but it does have that major advantage over sporadic lump-sum investment.

Re: CoE

Do the honourable thing and register, Dynamo. The chances of getting a complex posting like your last one right first time are remote and a distraction anyway. The power to edit makes life real easy. You have to give an e-mail address but the rest of the input can be rubbish. I don't know how compromising to one's privacy/anonymity giving out an e-mail is. :\

I sense from your last contribution that you are somewhat nearer to me than to our Leader on this equity thing.

But that <!--EZCODE BOLD START--> Pound Cost Averaging<!--EZCODE BOLD END--> hogwash.:eek That really is a loada rubbish. If you really want to drag me into that cul de sac, my contributions would be more properly made in the <!--EZCODE ITALIC START--> Letting Off Steam<!--EZCODE ITALIC END--> forum.:rolleyes