Your Results A few thoughts on your results Javier Estrada 4asset - - PDF document

1

• 1. Your Results
• What do the results suggest?
• 2. The Inputs in Portfolio Optimization
• Evidence and implications
• 3. Two Related Issues
• Country diversification v. industry diversification
• Globalization and the market portfolio

Application 4:

Portfolio Optimization

Javier Estrada ADFIN – Winter/2014

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Your Results

• A few thoughts on your results
• 4‐asset optimization
• The weights (portfolios) obviously vary substantially

depending on the goal and restrictions

• Unrestricted portfolios may be ‘weird’ or in many

cases impossible to implement

 Hence you have to specify your restrictions very precisely

• GMM

 Easy to implement (No additional info required)  Very different weights (Very concentrated portfolio)

• 10‐asset optimization
• The optimizer can easily be adapted to any n

 Nowadays optimizers are used mostly across asset classes

• A ‘simpler’ optimizer (and its limitation)

GoXls

2

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Inputs – Some Important Questions

• Consider the standard problem of maximizing

risk‐adjusted returns (Sharpe ratio)

• And consider …
• the inputs of this problem (Ei / Covij / Rf)
• and the following questions
• Do these inputs vary over time?
• If so, are these changes predictable?
• In general, how should these inputs be estimated?

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Inputs – Evidence

• Mean returns
• Fluctuate widely in the short/medium term, around

a rather‐constant long‐term mean

Go

• Volatility
• Fluctuates in the short/medium term, around a

rather‐constant long‐term mean

• Higher in down markets

Go

• Correlations
• Fluctuate in the short/medium term, around a

generally‐increasing trend (Globalization)

• Increasing in volatility
• Higher in down markets
• Hence diversification is there the least when you want it

and need it most

Go

3

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Inputs – Some Further Thoughts

• Two important issues
• Results (portfolios and characteristics) are very

sensitive to changes in the inputs

• Remember GIGO

GoXls

• Can we do better than using historical averages?
• Very (very!) tricky question
• For a short‐term portfolio

 The empirical facts discussed may be useful

• For a medium/long‐term portfolio

 The return of bonds is relatively easy to estimate (The current YTM)

Go

 For the return of stocks, all bets are off (Both history and the RDM may be good starting points)

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Two Related Issues

• Country diversification v. industry diversification
• US domestic/broad (top‐down) portfolio manager
• Concerned with being in the right sectors
• Not concerned with being in the right states (!)
• European broad (top‐down) portfolio manager
• Used to be concerned with being in the right countries

 Hence the ‘country specialist’ was essential

• That is no longer the case

 Countries have become increasingly correlated  The benefits of country diversification have decreased  The current concern is being in the right sectors  Hence the ‘sector specialist’ is now essential

Go

4

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Two Related Issues

• Globalization and the market portfolio
• Segmented markets
• Stocks are affected mainly by local factors

 The local market is the relevant benchmark

• Globalized markets
• Stocks are affected mainly by global factors

 The global market is the relevant benchmark

• Most countries and companies are somewhere in

between the two extremes

• This begs the question about the relevant market

portfolio to be used in the estimation of betas

 Local? Global? A blend?  Perhaps the latter (though it is not entirely clear)

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Appendix

5

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Mean Returns (1900‐2013)

Back Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Volatility (T12M, 1900‐2013)

Back

Volatility in up months: 12.5% Volatility in down months: 13.1%

6

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Correlations – Europe

Source: Freimann (1998). “Economic Integration and Country Allocation in Europe.” FAJ. Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Correlations and Volatility

Source: Solnik, Boucrelle, and Le Fur (1996). “International Market Correlation and Volatility.” FAJ.

7

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Correlations – Up/Down Markets

Source: Erb, Harvey, and Viskanta (1994). “Forecasting International Equity Correlations.” FAJ. Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Correlations – Up/Down Markets

8

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Correlations – Up/Down Markets

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

The Changing Nature of Inputs

Back

9

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

The Expected Return of Bonds

Back Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Country v. Industry Diversification

A recent poll carried out by Merrill Lynch found that only 10% of European fund managers believe that countries matter for their investment decisions – only three years ago the figure was 50%! More than three quarters said it was crucial to be in the right sector.

The Economist (Nov/11/00)

10

Javier Estrada IESE Business School Barcelona Spain ADFIN Winter/2014

Country v. Industry Diversification

Back