The Portfolio MYTH-----Do you need to change your thinking?

[Value Hunter]

Not everybody uses stop loses. I do not use a stop loss (I have different sell criteria based on fundamentals).

 

[minwa]

So ? You still manage it as you just said.

Take away the stop in portfolio A & B and put on a fundamental sell signal - the bottom line risk is the same. 1 is 1 scandal away from massive drawdown the other have a few "lives".

The fact that you can't guarantee nothing major will ever happen to the one stock in a major negative way makes A higher risk. Your research will lower the probability, not eliminate it.

 

[skc]

SKC you misunderstood me I meant top 5 positions by weighting in the portfolio. In other words the stocks they (fund managers) invest the most money in (percentage allocation), on average perform the best (compared to other stocks in their own portfolio) because they tend to invest the largest sums of money in their best ideas.

I see. That makes slightly more sense. I can't tell, from a quick glance of the paper, how they determined top positions by weighting. I know the 5 biggest positions in my portfolio have performed much stronger than the average... but that's because they were the best performers. They are the bigger positions after their stellar runs, and ongoing validation of business performance means that I increased position size along the way. I certainly can't say that they are my high conviction picks from day one however.

 

[howardbandy]

What a lot of good discussion.

Most agree that it is better to favor holding a small number of issues.
... It is hard to monitor a larger number.
... The top performers do better than the average.

I am a little confused by DeepState's scenario (in post 107). Is this correct:

Two separate models. One for stock A, the other for stock B. Resulting in two systems -- A and B.

Each system makes a prediction every day. The prediction is the direction for one day holding. A correct prediction -- a winning trade -- returns 1.0%. An incorrect position returns -1.0%. Each is 65% accurate.

The systems and their results are independent -- correlation between them is 0.0.

We need a statement of risk tolerance in order to normalize risk. I use:
We are trading a $100,000 account, forecasting two years. I want to hold the probability of a drawdown greater than 20% to a chance of 5% or less.

---------

Compare two trading management alternatives:

1. Trade A (or B) alone. Each trade will be made using whatever portion of the account is determned by safe-f.

2. Trade both A and B. Profit and loss of both trades are accumulated into a single balance, which is split between the two for the next trade. Each system is trading half the composite account, at the safe-f of each.

This much is straight forward and can be analyzed.

----------

I am confused about the 0.3 correlation mentioned. How does it come into play?

----------

There is statement that risk can be approximated by standard deviation. That fits the frequentist / modern portfolio theory approach.

I prefer the alternative Bayesian approach, where we let the data guide us. That is, form many equally likely trade sequences, compute the final equity and maximum drawdown of each, form the distribution of drawdown, adjust position size to normalize the risk, compute the distribution of final equity, estimate profit potential at the 25th percentile.

Best, Howard

 

[Value Hunter]

Minwa another one of the problems with simplistically looking at the number of stocks I'm your portfolio is that it does not tell you much about diversification by itself. For example if you own a company like Nestle or 3M or Unilever you have exposure to a huge number of product lines, countries and potentially currencies. However somebody that buys a Greek stock market index gets exposed mostly to the problems of the Greek economy. If you buy an Australian index fund you get mainly exposure to banks and resources and the Australian dollar. If you buy Berkshire Hathaway you get exposure to well managed companies across a huge range of industries.

Would you agree that owning shares in Washington H.Soul Pattinson gives you more diversification than owning Telstra and TPG telecom?

 

[DeepState]

http://www.lazardnet.com/us/docs/sp6/3048/LessIsMore-ACaseForConcentrated_LazardInvestmentFocus.pdf

The above is a link to a study which shows that on average a fund managers top 5 picks outperform their portfolio as a whole.

The Lazard paper which, in turn, refers to the Yeung paper, is showing that cross sectional momentum works. This result has been published for decades.

Unless someone is really interested, please just accept that the analytical process which backs the Yeung base paper produces a strong lean towards momentum even if fund managers are coin flippers. This assertion is confirmed by their own analysis [P-value of alpha on Carhart 4 regression is far from significant but loads on MOM with P = significant across a whole range of tables in https://www.uts.edu.au/sites/default/files/wp18.pdf].

Naturally Lazard has something to sell and the paper wants to create some attention at a conference. The paper is actually a thumbs up for momentum, not concentrated portfolios on the basis of generating superior returns from focused stock picking insight.

Predictably, the other paper mentioned is Petajisto which argues that very concentrated portfolios (measured by 'Active Share') outperform. Let's consider this, Abbott and Costello both own stocks A and C in equal measure. The benchmark is 50/50 to each. They are index managers with zero active share. No outperformance. According to Petajisto, when they trade such that Abbott holds 100% in A and Costello holds 100% in C, their active share shoots to 50% and both are somehow supposed to outperform the index because their active share goes up. I kid you not. Anyway, AQR is now at war with Petajisto on this matter. It's not even worth buying popcorn.

These are routine arguments trying to pull money away from index and quant money. They are head fakes.

 

[DeepState]

Hi Howard

What a lot of good discussion.

I am a little confused by DeepState's scenario (in post 107). Is this correct:

Two separate models. One for stock A, the other for stock B. Resulting in two systems -- A and B.

Each system makes a prediction every day. The prediction is the direction for one day holding. A correct prediction -- a winning trade -- returns 1.0%. An incorrect position returns -1.0%. Each is 65% accurate.

The systems and their results are independent -- correlation between them is 0.0.

All good. I made an error in my example by saying that the portfolio is long only. We can relax this given it is a binary process. All the figures actually relate to a full long-short implementation. The point would still stand if you wanted to restrict to long only or, indeed, short only or a mix.

We need a statement of risk tolerance in order to normalize risk. I use:
We are trading a $100,000 account, forecasting two years. I want to hold the probability of a drawdown greater than 20% to a chance of 5% or less.

You can use whatever works for you. Ultimately is will come down to the distributions I have pre-specified and are not subject to uncertainty. The safe-f, for whatever risk tolerance you want to specify is identical for both stocks individually and also for their combination. The diversification effects will hold.

---------

Compare two trading management alternatives:

1. Trade A (or B) alone. Each trade will be made using whatever portion of the account is determned by safe-f.

2. Trade both A and B. Profit and loss of both trades are accumulated into a single balance, which is split between the two for the next trade. Each system is trading half the composite account, at the safe-f of each.

This much is straight forward and can be analyzed.

No. This is where some confusion seems to arise. In option 2, you are not trading the safe-f allocation of each. You are trading the safe-f allocation of the portfolio consisting of a 50/50 mix of strategies A and B. For whatever definition of risk tolerance you reasonably want, the nominal allocation to this combination of stocks will be greater than the nominal allocation for a single stock alone for any given safe-f. In this example, the nominal allocation to the portfolio of stocks will be about 1.4x the allocation to either A or B alone (option 1). This 1.4x comes about due to portfolio diversification gains. That figure isn't even taking in to account return benefits, only risk [yet more in your favour].

----------

I am confused about the 0.3 correlation mentioned. How does it come into play?

Don't worry about it. It is the equivalent of saying 65% hit rate for a binary outcome.

----------

There is statement that risk can be approximated by standard deviation. That fits the frequentist / modern portfolio theory approach.

I prefer the alternative Bayesian approach, where we let the data guide us. That is, form many equally likely trade sequences, compute the final equity and maximum drawdown of each, form the distribution of drawdown, adjust position size to normalize the risk, compute the distribution of final equity, estimate profit potential at the 25th percentile.

I am making this easy for you so that you do not need a Bayesian estimator, although you can weaken the outcomes if you want by applying it. I have fully specified the system and revealed its full workings for this exercise, tilting everything heavily in your favour. Your process, Bayesian or other, requires a 75% hit rate to justify a 1-stock portfolio.

If, in the wild, you have to apply Bayesian estimates and let the data speak, it implies that there is estimation error in play of the type which makes the level of insight required to justify a 1-stock portfolio even more out of reach.

 

[minwa]

Minwa another one of the problems with simplistically looking at the number of stocks I'm your portfolio is that it does not tell you much about diversification by itself. For example if you own a company like Nestle or 3M or Unilever you have exposure to a huge number of product lines, countries and potentially currencies. However somebody that buys a Greek stock market index gets exposed mostly to the problems of the Greek economy. If you buy an Australian index fund you get mainly exposure to banks and resources and the Australian dollar. If you buy Berkshire Hathaway you get exposure to well managed companies across a huge range of industries.

Would you agree that owning shares in Washington H.Soul Pattinson gives you more diversification than owning Telstra and TPG telecom?

Being a diversified company does not eliminate the possibility of a big move against you.

For a 2 stock portfolio, one would not have them both in the same sector so don't see how your example applies. You're using examples of how a multistock portfolio should not be constructed to compare against the one stock portfolio.

 

[Value Hunter]

Minwa, my point was that looking at the number of stocks owned in isolation to other factors does not tell you about the level of diversification or risk.

Deep state I will concede the point to you on the Lazard paper, but perhaps you would care to address my other posts/arguments?

 

[Value Hunter]

Minwa owning an index fund (or 20 individual stocks, etc) does not eliminate the possibility of a big move against you as would have occurred in 2008.

1. Deep state my essential issue is against the middle ground which is the worst of both worlds. If an investor lacks skill and wants to own index funds that is a prudent decision. If an investor has skill and wants to own 5 individual stocks that is all well and good also. What I disagree with (unless its a mechanical approach) is a fund manager or individual that owns a diversified portfolio with no more than say10% in any single holding and owns say 20-30 stocks. This person should either throw in the towel and buy index funds or trim their portfolio to a maximum of say 4-8 positions. Now I understand some fund managers have too much money too manage and due to size constraints must diversify, but for small funds and individuals I think concentration is the answer.

 

[DeepState]

Minwa owning an index fund (or 20 individual stocks, etc) does not eliminate the possibility of a big move against you as would have occurred in 2008.

1. Deep state my essential issue is against the middle ground which is the worst of both worlds. If an investor lacks skill and wants to own index funds that is a prudent decision. If an investor has skill and wants to own 5 individual stocks that is all well and good also. What I disagree with (unless its a mechanical approach) is a fund manager or individual that owns a diversified portfolio with no more than say10% in any single holding and owns say 20-30 stocks. This person should either throw in the towel and buy index funds or trim their portfolio to a maximum of say 4-8 positions. Now I understand some fund managers have too much money too manage and due to size constraints must diversify, but for small funds and individuals I think concentration is the answer.

Hi VH

I think an industrial conglomerate with dispersed industries inside the banner has inherent diversification in it. However, Minwa has a point that there are still elements of being part of the same conglomerate which reduces the diversity and exposes the lot to event risk or otherwise diminishes risk relative to stand alone holdings in each of the underlying businesses. This arises due to commonality of management and, sometimes, financing.

I am not particularly averse to the view that you have outlined regarding 20-30 stocks and use of index funds. It is quite reasonable and would find support from many quarters. I think it comes down to a view on loss of focus that comes with holding an increasing number of positions. Each situation is unique and this might lead to different conclusions between fund managers and individuals depending on their skills, abilities and approach.

Stock numbers per se are just the most obvious way of talking about diversification, but it is a proxy for some concept of effective diversification. What is wanted is maximum reward for risk and an appropriate deployment of risk. To that extent, Howard's approach is taking broad steps in the right direction although I clearly diverge in how that it actually done.

Although the attention scope of an individual might lead them to invest in only 2 stocks in a truly optimal sense (using everything I have outlined throughout), it might be better for that person to admit that their situation isn't ideal and to outsource if those assets are material to them.

 

[Value Hunter]

Deep state as I stated before in my own stock portfolio (stocks make up around 30% of my total gross assets but I am highly leveraged) one stock is over 95% of the stock allocation. Due to portfolio leverage(no margin loans though) if that stock goes to zero 80% of my net assets (my life savings and a six figure sum which is meaningful to me) will be wiped out. I mangage risk not by diversifying but by:
A) Having a long term horizon and being agnostic of drawdowns (unless the value of the business is permanently impaired)
B) Having strong knowledge of the company
C) Choosing a company with a strong balance sheet, strong earnings track record, good management and a bright future (my own subjective assessment) with diversified earnings streams
D) monitoring the company. If the quality of the business ever detiorates I will likely sell.

I still maintain that somebody holding a portfolio of six speccy stocks to me is taking more risk than someone who has 100% invested in a company like Kraft or Unilever or Nestle or 3M.

Diversification is one tool in the toolbox that can be used to manage risk. Other factors such as the ones I stated above are other tools as are things like stop losses (which I do not use). There are many risk management tools with diversification just being one such tool which is useful in some cases but not the right tool in others.

 

[Value Hunter]

Another point I would like to raise is looking at number of stocks or even the number of industries is just one way to look at risk exposure. You have other ways such as looking at exposure to risk/event factors (e.g. how many stocks in your portfolio are vulnerable to a China slowdown?)

Another method is to look at story types as Peter Lynch did (Cyclicals, Asset Plays, Turnarounds, Slow Growers, Stalwarts and Fast Growers)

 

[Value Hunter]

Deepstate I think you need to define what you consider risk. I agree with Warren Buffet that risk is the probability of a permanent loss of capital (as opposed to say beta, drawdown, etc). In my methodology risk is to a large extent qualitative and not measurable in a strict numerical sense and rather can be described in a broad manner. Given that is the case how would you define or measure "maximum reward for risk"?

The way I look at is adequacy rather than maximisation. Is the reward adequate for the risk I am taking? Due to having imperfect information and dealing with things which are subjective and difficult to measure I do not think I can be more precise than that. I think maximising is an unrealistic goal.

 

[DeepState]

Deep state as I stated before in my own stock portfolio (stocks make up around 30% of my total gross assets but I am highly leveraged) one stock is over 95% of the stock allocation. Due to portfolio leverage(no margin loans though) if that stock goes to zero 80% of my net assets (my life savings and a six figure sum which is meaningful to me) will be wiped out. I mangage risk not by diversifying but by:
A) Having a long term horizon and being agnostic of drawdowns (unless the value of the business is permanently impaired)
B) Having strong knowledge of the company
C) Choosing a company with a strong balance sheet, strong earnings track record, good management and a bright future (my own subjective assessment) with diversified earnings streams
D) monitoring the company. If the quality of the business ever detiorates I will likely sell.

I still maintain that somebody holding a portfolio of six speccy stocks to me is taking more risk than someone who has 100% invested in a company like Kraft or Unilever or Nestle or 3M.

Diversification is one tool in the toolbox that can be used to manage risk. Other factors such as the ones I stated above are other tools. There are many risk management tools with diversification just being one such tool which is useful in some cases but not the right tool in others.

VH, do what you feel is right. When I debate, it's mostly on the basis of assertions which have been made and the course of logic pursued. Nothing in what I have said makes it wrong for your to do what you are doing per se. It is unusual and uncommon. But I cannot say that it is wrong...for you.

I will suggest that the protections in terms of fundamentals are not as strong as might be hoped. This statement comes from many years in the 1990s when funds management was growing, but not the monolith it is today. Quant had barely begun to show itself in a meaningful way.

Manager after manager would espouse risk management in the way you have articulated. They have almost all been swept away with time.

Focus and strong knowledge doesn't improve predictability very much beyond a point. I can know everything there is to know about a (frictionless) billiard table and the balls on it. Not allowing for the planetary motion of Mars will change the outcomes of the billiard ball by enough that my forecasts of motion will be wildly off within a very short space of time. Billiards is nothing compared to investment.

 

[craft]

Deep state as I stated before in my own stock portfolio (stocks make up around 30% of my total gross assets but I am highly leveraged) one stock is over 95% of the stock allocation. Due to portfolio leverage(no margin loans though) if that stock goes to zero 80% of my net assets (my life savings and a six figure sum which is meaningful to me) will be wiped out. I mangage risk not by diversifying but by:
A) Having a long term horizon and being agnostic of drawdowns (unless the value of the business is permanently impaired)
B) Having strong knowledge of the company
C) Choosing a company with a strong balance sheet, strong earnings track record, good management and a bright future (my own subjective assessment) with diversified earnings streams
D) monitoring the company. If the quality of the business ever detiorates I will likely sell.

I still maintain that somebody holding a portfolio of six speccy stocks to me is taking more risk than someone who has 100% invested in a company like Kraft or Unilever or Nestle or 3M.

Diversification is one tool in the toolbox that can be used to manage risk. Other factors such as the ones I stated above are other tools as are things like stop losses (which I do not use). There are many risk management tools with diversification just being one such tool which is useful in some cases but not the right tool in others.

Is the stock you refer too as representing over 80% of your net worth CCP?

 

[Value Hunter]

Deep state can you provide evidence to support your assertion that fundamentals do not provide strong protection? I would argue that using fund managers as an example is a suboptimal example because of the conflicts of interest and agency costs. The job of the funds managment companies is to maximise profit for themselves not risk adjusted returns for investors. The job of a portfolio manager is to keep his job not to maximise the risk adjusted return to investors.

 

[Value Hunter]

Yes Craft it is, the stock representing 25-30% of my gross assets (and a much higher chunk of my net assets is Credit Corp Group (CCP). And I am happy with it. I know you think it is a risky company but I disagree with your assessment.

 

[Value Hunter]

Some people put most of their life savings in the home they own, some people put their life savings in the small business they own. In either scenario most people will not bat an eyelid. Suddenly if somebody wants to put a huge chunk of their life savings in a single stock people think they are nuts.

 

[DeepState]

Deepstate I think you need to define what you consider risk. I agree with Warren Buffet that risk is the probability of a permanent loss of capital (as opposed to say beta, drawdown, etc). In my methodology risk is to a large extent qualitative and not measurable in a strict numerical sense and rather can be described in a broad manner. Given that is the case how would you define or measure "maximum reward for risk"?

The way I look at is adequacy rather than maximisation. Is the reward adequate for the risk I am taking? Due to having imperfect information and dealing with things which are subjective and difficult to measure I do not think I can be more precise than that. I think maximising is an unrealistic goal.

Risk, for me, is not getting what you want from your investments. This can happen due to permanent capital impairment or the market not being rational (seeing things as you want them to - which may arise from our own delusion) in the period that you care about. Or screwing up. This is often correlated to measures like MDD and standard deviation.

Although we want maximum reward for risk, all that you can go for is maximum expected-reward-for-risk. The chances of that particular outcome being the actual best with the benefit of hindsight are remote.