Notes from Session 5 of the NYU Corporate Finance Class.

Starts at 2:30

Lets define Danger.

CAPM: Capital Asset Pricing Model

Mean-Variance Framework

Measure of actual returns around an expected return

 A one-year government bond, zero risk. Exact return.

A ten-year government bond, price may change with interest rates.

Risk-free has to be used in the context of a time horizon.

If you buy Con Ed stock, you expect 7% return. In a year, what were your returns? 2%? Maybe 10% wider possible returns, more risk.

Risk takes place over time risk does not exist in the past, since we already know what happened.

AD: All our data comes from the past, but all of our worries are in the future.

The higher uncertainty an investment faces, the riskier it is.

A single stocks price history does not tell a whole story: volatility, risk, etc are all observed in context. Is your stock riskier than the rest of its sector?

Do you live in a mean-variance world?

Really interesting thought experiment at 10:40

Demonstrates that the premise that CAPM is built on, the mean variance assumption, is not necessarily reflected in the real world.

In a mean variance world, casinos would be empty. Lotteries would be spectacular failures.

Would your answer change, if one potential investment had a much bigger, but improbable, downside?

When you use the CAPM, you’re going to over/under estimate risk due to missed variables

AD: “Stocks that move a lot are more risky than stocks that don’t.”

Think about all of the potential risk you’re exposed to as a stockholder

@15:55 discussion of Disney risk stories

Disneyland Hong Kong starts making money good news, but only involves Disney. Project risk.

John Carter loses $200 Million again, only bad news for Disney. Project Risk

Universal Studios opened a new park in Singapore. Risk affects Disney, but comes from competition. Competitive Risk.

New FCC Rule affects how TV is governed. This affects all broadcast companies. Sector Risk.

Dollar rises against other currencies: mixed blessing for Disney. Exchange Rate Risk.

Bernanke testimony affects markets. Hits everyone. Market Risk.

Spectrum ranges from one firm to the full market of publicly traded stocks.

AD: “Risk that affects a few firms cuts both ways. It will get averaged out, due to diversification.”

Diversification works, not because of Finance, but because of common sense.

What won’t diversify out? Market Risk factors, like Ben Bernanke.

Risk is diversifiable.

Median # of stocks in a US investor? 3-5.

It’s not whether individual investors are diversified, but the marginal investors.

The marginal investor has to own enough shares to have an impact

The marginal investor has to be able to trade those shares

Look at risk through the eyes of the marginal investor

That is who sets stock prices.

In Corporate Finance, we assume our Marginal Investor is diversified (why?)

If we accept this assumption, our belief that few-firm risk averages out holds.

How to identify the Marginal Investor in your firm?

See 28:30 for chart

Look at some investor bases

Disney, Deutsche Bank, Aracruz and Tata

Besides Jobs, every name for Disney is an institutional investor.

In the case of Disney, we feel safe, since institutional investors are almost certainly well diversified.

Top investors in DB, again, institutional and thus diversified.

Aracruz: Top investors of voting stock are held again, by diversified institutions.

Tata: 29% insiders, but note that the other Tata companies “count” as insiders. Don’t trade.

Tata: Who’s the marginal investor? Probably institutions, right?

AD: It turns out to be Foreign funds, in fact. See 36:00

Defined risk as variance

Defined risk tha can and cannot be diversified away

The only risk to care about is macro risk that cannot be diversified away.

In the CAPM, you go further than assuming diversified investors.

Why do investors (individual or institutional) stop diversifying?

Why not buy all the stocks?

The benefit of each new stock has a decreasing utility

Adding each new stock has a cost fees, etc

We tend to stop as benefits get smaller and costs add up.

Must be more reasons, the cost of adding stocks are smaller and smaller.

If you own the whole market, you simply match the market. Totally average.

We want to beat the market. So do institutional investors.

Two reasons we stop: costs piling up, and desire to beat the market.

“The Market Portfolio”

This is what the CAPM assumes we’d end up with

A perfectly diversified portfolio, holding every traded fund in the world.

Individual investors will adjust according to risk aversion

Percentage of risk-free (T-Bills) and risky (Market Portfolio)

Living in a CAPM world, we’d hold a combination of only these two assets.

AD: “I know this is unreasonable, but this is what our assumptions present.”

What about new assets to the market? How do we measure the risk of that asset?

How much does the asset move with the market? (Called covariance)

This is what a beta measures, covariance.

We divide an asset’s covariance by the variance of the actual market.

Beta is carrying a lot of baggage mean variance world, CAPM world, etc

Expected Return = Riskfree Rate + Beta * (Expected Return on the Market Portfolio Riskfree Rate)

AD: “It’s built on blocks… you can agree or disagree with each block.”

3 Criticisms of the CAPM:

Unreasonable assumptions (AD: I’ll take useful simplicity over useless realism any day)

Parameters cannot be estimated precisely (Definition of Market Index, Firm may change during estimation)

The model doesn’t work well ( should be linear relationship btw returns and betas, not true)

Other variables (size, price/book value) seem to explain returns better. Beta only explains

8% of returns.

Friedman: “It takes a model to beat a model.”

AD: “Before rejecting the CAPM, find an alternative that works better.”

Alternatives to the CAPM:

APM: Arbitrage Pricing Model

Introduced a number of different betas relative to different factors. Created by using a computer to model 30 years of stock returns.

Why didn’t it catch on?

The factors identified are not human-readable. Factor One, Factor Two, etc.

Even though there may be five factors, without useful names, no one would accept the model.

Next Step: Identify the factors in Multi Factor Modeling

After 2 years, each factor had a macroeconomic term.

Problem: factors were assigned looking backwards, but turn out to be poor forecasters.

Not largely adopted.

Next: Proxy Modeling

Looked over same 30 years of data to find common features of successful stocks.

Found correlations and assigned risk to successful stocks.

Proxy models are add ons to other models

You’re simply letting something stand in for risk.

Plus, none of these models do better than the CAPM in forecasting, which is really the key of the whole exercise.

Inputs for the CAPM

Expected Return = Riskfree Rate + Beta * (Expected Return on the Market Portfolio Riskfree Rate)

Easiest number to get = Riskfree Rate

To be RIskfree:

The actual return is equal to the expected return.

Can be no default risk, implying it must be government issue.

There can be no uncertainty about reinvestment rates.

Zero Coupon Ten Year Treasury Bond

AD: “I’ve just created a nightmare for you. ”

A way out: approximate weighted duration assets.

Find a default free ten year bond, use it as risk free rate

Not perfect but it will do.

Let’s find the Riskfree rate in US Dollars right now.

Find the 10 year T Bond rate. Right about 2%

It can and will change. Use object oriented spreadsheeting to protect your equations.

Riskfree rate for the Euro?

Riskfree rate in Euros for a company in Ireland

About a dozen companies that issue 10-year bonds in Euros

Adjust local currency government borrowing rate for default risk by checking Moody’s currency rating. If Indian bonds are available at 7% and the default spread is 3%, your riskfree rate is 4%