Thread
1/

Get a cup of coffee.

In this thread, I'll walk you through 2 key portfolio diversification principles:

(i) Minimizing correlations, and
(ii) Re-balancing intelligently.

You don't need Markowitz's portfolio theory or the Kelly Criterion to understand these concepts.
2/

Imagine we have a stock: ABC Inc. Ticker: $ABC.

The good thing about ABC is: in 4 out of 5 years (ie, with probability 80%), the stock goes UP 30%.

But the *rest* of the time -- ie, with probability 20%, or in 1 out of 5 years -- the stock goes DOWN 50%.
3/

We have no way to predict in advance which years will be good and which will be bad.

So, let's say we just buy and hold ABC stock for a long time -- like 25 years.

The question is: what return are we most likely to get from ABC over these 25 years?
4/

We know ABC stock goes UP 80% of the time and DOWN the other 20%.

So, the most likely scenario is:

During our 25 year holding period, we'll get 20 UP years and 5 DOWN years.

This gets us an annualized return of ~7.39%. Each $1 we put in becomes ~$5.94.

Calculations:
5/

Now let's bring in a second stock: XYZ Inc. Ticker: $XYZ.

Suppose $XYZ exhibits exactly the same statistical characteristics as $ABC.

That is, $XYZ also goes UP 30% in 4 out of 5 years and DOWN 50% in 1 out of 5 years.

But $XYZ and $ABC don't necessarily move *together*.
6/

In some years, *both* $ABC and $XYZ go UP.

In other years, they *both* go DOWN.

And in still other years, they move in *opposite* directions. That is, *one of them* goes UP and the *other* goes DOWN.
7/

The more often these stocks move *together* -- ie, *both* UP or *both* DOWN -- the more *positively correlated* we say they are.

And why do we care?

Because the MORE positively correlated our stocks are, the LESS benefit we'll get from diversification.
8/

Suppose our stocks *both* go UP together X% of the time.

Thus, higher X means higher correlation.

Simple probability math now tells us:

(i) Our stocks will go DOWN together (X - 60)% of the time, and

(ii) They'll move in opposite directions (160 - 2*X)% of the time.
9/

Also, the probabilities in our 2x2 diagram above cannot be negative.

That is, neither (80 - X)% nor (X - 60)% can be negative.

That means X has to lie between 60% and 80%.
10/

At X = 60%, our stocks are as *negatively* correlated as they can be.

And as we raise X from 60% to 80%, their correlation turns strongly *positive*.

The crossover from negative to positive occurs at X = 64%. At that point, there's *zero* correlation between our stocks.
11/

The central idea behind diversification is: everything shouldn't blow up *at the same time*.

In other words, we *don't* want our stocks all moving *together*.

That's why *negative* correlations are "good" and *positive* correlations are "bad".
12/

But to REALLY take advantage of negative correlations, we need another key idea: intelligent re-balancing.

Let's run some numbers to see why.

Suppose X = 60%. That is, $ABC and $XYZ are as negatively correlated as possible.
13/

Now, suppose we build a 50/50 diversified portfolio of $ABC and $XYZ.

We'll hold this portfolio for 25 years.

But we'll never re-balance. We'll just let each individual stock ride.

What will our most likely 25-year return be?
14/

Well, if we never re-balance, each stock is just going to do its thing individually.

Each is likely to give us 20 UP years and 5 DOWN years -- though not in the same order.

This will let us turn $1 into ~$5.94 -- the same ~7.39% annualized return we had before.
15/

Thus, the *negative* correlation between $ABC and $XYZ does NOT help us if we don't re-balance.

Whether we own:

- Just $ABC,
- Or just $XYZ,
- Or a 50/50 "$ABC plus $XYZ" combo,

we're likely to end up at the exact same place -- earning ~7.39% per year.
16/

What happens if we *do* re-balance our portfolio periodically -- say, every year?

For example, if $ABC goes UP and $XYZ goes DOWN during a year, we sell a bit of $ABC at the end of the year and use the money to buy a bit of $XYZ -- so our portfolio is again back at 50/50.
17/

With this re-balancing strategy in place, here are all the outcomes that can possibly happen during any given year -- along with their respective probabilities:
18/

For example, with X = 60% -- our strongest possible negative correlation -- our portfolio:

- Goes UP 30% in 3 out of 5 years, and
- Goes down 10% in 2 out of 5 years.

Over 25 years, our most likely outcome turns $1 into ~$17.85.

That's ~12.22% annualized.

Calculations:
19/

So, when X = 60% (ie, strongly negative correlation), we have:

*Without* re-balancing: $1 --> ~$5.94, and
*With* re-balancing: $1 --> ~$17.85 (>3x as much!),

over the same 25 years.

Thus, negative correlations + re-balancing can be a powerful combination.
20/

If we do this well, our *portfolio* can end up getting us a HIGHER return than any single stock in it!

We just saw an example with 2 stocks. Each got us only ~7.39%. But a 50/50 re-balanced portfolio of them got us ~12.22%.

When I first saw this, I couldn't believe it!
21/

This is the ESSENCE of diversification.

- We minimize correlations, so our portfolio nearly always has *both* risen and fallen stocks, and

- We "cash in" on this gap via re-balancing -- ie, we periodically sell over-valued stocks and put the money into under-valued ones.
22/

Also, negative correlations aren't strictly necessary.

We could use stocks with zero -- or even positive -- correlation.

But the MORE heavily correlated our stocks, the LESS "bang for the buck" we get from re-balancing.

Like so:
23/

As investors, we spend a lot of time analyzing individual stocks.

But we tend to focus less on *portfolio* construction: do our stocks fit well together? Are we re-balancing sensibly? Etc.

I hope this thread shed some light on these topics.

Happy Diwali folks!

/End
Mentions
See All