11. Theory of Active Portfolio Management

Question 1

Active portfolio management

Answer 1

Seeking alpha

Question 2

Pure CAPM theory

Answer 2

• alpha = 0 for all shares: all stocks are on SML (stock is over-/underpriced if not)
• If not on SML, stock will move back to SML (prices goes up/down)
• In reality there are a lot of stocks not on SML (f.e. small stocks)
• Taking position on CML = passive investment.

Question 3

Problems with CML in practice

Answer 3

• Not 10, but 5000 stocks (var covar-matrix is large)
• Some stocks have very little data (young firms)
• Negative variances are possible
• No room for strategy (preferences)
• alpha ≠ 0 (value, size, momentum)
• CAPM doesn’t hold empirically

Question 4

How to make it more realistic?

Answer 4

• Improve SR (risk-return tradeoff)
• Allow alpha to differ from zero
• Allow investors to partially manage their portfolios actively (hold an actively managed part next to their passive investment in market portfolio and risk free asset)
• Use investor knowledge/ managerial insights to actively select stocks
• Drawback: only two types of risk: market and idiosyncratic (no industries, no correlations between 2 stocks that are actively managed etc.)

Question 5

Treynor & Black

Answer 5

Investors only care about risk and return (maximize return and minimize risk)

• Allow α to differ from zero
• Add high alpha socks in actively managed portfolio, not entire wealth in highest alpha

Question 6

Treynor & Black

Allow α to differ from zero

Basic idea

Answer 6

Allow for more idiosyncratic risk in order to get α -> market/passive + actively managed part (exposed to idiosyncratic risk)

Question 7

Treynor & Black

Active part

Answer 7

Select stocks with high alpha and low idiosyncratic volatility

1. Relative attractiveness of each share within active part (lies in alpha and exposure to idiosyncratic risk)
2. Position in active portfolio
3. Fraction allocated to active part by making beta correction
4. Final weights
5. Data of optimal risky portfolio

Question 8

Black-Litterman

Answer 8

Traynor-Black is more at stock selection level (which stocks to pick), little room for tuning

Black-Litterman is about asset allocation -> tune confidence

Question 9

Difference BL and TB

Answer 9

1. BL ad TB are complements

2. Models are identical with respect to optimization process

Question 10

Black-Litterman model

Answer 10

1. BL combine data with managerial opinion
2. Not restricted to asset returns, can integrate relative opinions (B>S), no need to specify values about bonds and stocks
3. Validity rests upon way in which confidence about views is developed
4. Use for asset allocation
5. Confidence is manually set
6. (-) Cooking the results: if too much weight to manager
7. (-) Inconsistency

Question 11

Treynor-Black model

Answer 11

1. Not applied in the field, because it results in “wild” portfolio weights
2. Extreme weights are consequence of failing to adjust alpha value reflect forecast precision
3. Use for security analysis with proper adjustment of alpha forecast
4. Confidence is estimated

Question 12

M^2

Answer 12

Return gain for same level of risk

• When right: M2 increases with confidence
• When wrong: M2 decreases with confidence

Question 13

Treynor & Black

Add high alpha socks in actively managed portfolio, not entire wealth in highest alpha because:

Answer 13

• Limit amount of idiosyncratic risk in actively managed part
• Rank stocks in vector on alpha
• Be aware that alphas are estimates/outcome of regression, might be lower in reality