Do not put all eggs in one basket!
This post explains two fundamental approaches of diversification: Naive and Markowitz. Diversification means risk reducing by spreading capital into different investments. One basket could be one sector, country or investment type.

Why should investors hold whole portfolios?

By holding stocks, investors get rewarded for two fundamental different parts. First, interests or returns which represents no-risk. Risk-free returns compensates investors for deferred consumption. Additionally, a risk premium is given depending on the risks involving with the concerning company. This part can be sub-divided info the specific and systematic share. At the end, investors are awarded for the non diversifiable systematic risk only. Contrary systematic risk, specific risk can be reduced by diversification.

Why are risks important?

All investors try to reduce risks as much as possible and simultaneously maximize returns. The so-called risk-return ratio should be high. Risks are widely seen in volatility of market assessments (stock prices). So, to get a lower risk respectively volatility, investors should eliminate specific risks as much as possible.

How do they do that?

Holding entire capital in one stock, goes in hand with the maximum specific risk possible. There are two extremes which can be found for reducing. First, the oldest diversification method existing since ancient greek, also aimed at Eternal Yield: Naive diversification. Here, investors hold as much different investments (e. g. stocks, land, raw materials) as possible, all equally weighted. Some publications show that by 15 random stocks 80 % of specific risks can be reduced.

Drawbacks:

1. Many investment simultaneously needed.
2. Regular balancing necessary, in order to get equal positions at all times.
3. Correlated companies reduce risks only partially.

Advantages:

1. Robust strategy. Equal and clear loss limitations per position.
2. Easy to implement.
3. Fast risk reduction possible with small stock amounts, if  stock picking follows right interplay (e. g. measured by correlation).

What about the other strategy?

On the other hand, there is the more recent, nobel price awarded approach via Markowitz. Investors estimate mathematically future stochastic measures like variances and covariance between companies and their expected stocks prices. Therefrom, relative investing weights of all considered stocks can be deduced. Stochastic measures are calculated by past seen ones or through market knowledge.

Drawbacks:

1. Forward looking expectations. Past numbers cannot be used for future developments.
2. Accurate estimation necessary. Strategy stands and falls with ability to estimate future interplay.
3. Strategy can be randomly complex.

Advantages:

1. With exact numbers, returns above market average with 0 % specific risks possible.
2. Individually adaptable to well-known active investment strategies (e.g. turtle or momentum trading).
3. Incorporation of insider knowledge or other valuable information possible.

Recent publications show that sometimes naive and sometimes Markowitz perform better. Furthermore, many combinations can be found in practice, but they all are based on these fundamental two. So, there is still a controversial discussion, which one is more practicable, especially for private investors.

Which one is better?

That cannot be answered clearly. It depends on the investor’s point of view about the market as a whole, his trust in estimation future stochastic properties, investing strategy and target time frame. But, diversification must be done always – otherwise you lose money.

[J.P. Morgan – Risks and Rewards of a Concentrated Stock Position (pdf)]