2.7
Risk-Neutral Valuation Principle
In
both the riskless hedge and the synthetic option approaches, you could calculate
the value of an option without referring to any of the following:
1) |
the expected return from the stock |
|
2) |
the expected return from the option |
|
3) |
the risk preferences of investors |
|
4) |
the probability that the stock moves up or down |
|
It
is sufficient to know the current stock price, and the risk-free
interest rate, and the fact that there are only two possible future
stock values.
This
observation leads to an elegant valuation argument.
Since the option value is independent of these factors, the
option price is the same whether you are in a risk-averse world or in a
risk-neutral world. But in
a risk-neutral world, we know how to value an asset: You simply discount
the expected future value by the risk-free interest rate.
The
question that arises, then, is whether we can transform our world into a
risk-neutral world. The key
thing to keep in mind is that in such a world, the rate of return on
every asset must equal the risk-free interest rate.
In particular, the return on the stock (and on any option) must
equal the risk-free return.
Therefore,
we first need to discover whether the binomial stock model is consistent
with some risk-neutral world. In
particular, can we choose the probabilities so that the expected return
on the stock equals the risk-free interest rate?
If
we can find such probabilities, called "risk-neutral
probabilities," then we have a parallel world in which all the
prices are the same as in the original world.
It turns out that this is a very general principle, and that in
fact, such probabilities always exist as long as there is no arbitrage.
Deriving the Risk-Neutral Probabilities
The simplest method for deriving the risk neutral probabilities is to solve for the probability of an up-tick that equates the expected present value of the stocks terminal values to the current stocks price:
By rearranging and canceling the stock price S reveals that:
Alternatively you can see why such a probability is implied from a closer inspection of the synthetic option approach. In this approach, we used a stock and the risk-free asset to replicate the payoffs from a call. We then solved for the end-of-period value as:
and
to
get
and
By
substituting these values, we can rewrite this as follows:
where
Since
u > r > d, observe that both p and (1-p) are
between 0 and 1. Therefore
they can be interpreted as probabilities.
We
can now provide a risk-neutral interpretation for the value of the call
option:
The
numerator of the right hand side is the expected end-of-period value
where the probability used to evaluate the expectation is
p. The
denominator is the risk-free interest rate, and therefore C equals the
expected future value discounted by the risk-free rate.
Thus
in this world the expected return on the option when evaluated relative
to p
is the risk-free rate of return. To
complete the derivation, we must demonstrate that this probability has
the property that the expected return for the stock is also equal to the
risk-free rate of return. This
is because in the risk-neutral world, every asset must earn the
risk-free return.
Consider:
By
canceling out common terms and dividing both sides by r,
we find that the stock also has an expected return equal to the
risk-free rate:
For
this reason, p is
called the "risk-neutral probability."
You
should remember that the risk-neutral valuation principle does not imply
that the true expected return on the option is equal to the risk-free
rate of interest. Indeed
this will not be the case for a capital market with risk-averse
investors. This is because
the actual expected return on the option is evaluated with respect to
the true probabilities, not the risk-neutral probabilities.
Probabilities and Returns
Recall
that with respect to the risk-neutral probabilities the option value is
and
the stock value is
In
a risk-averse market, the "true" probability will be some p
> p,
and thus the "true" expected returns for the stock and option
(respectively rs
and rc)
are greater than r.
It
is not always the case that we can determine the option price
independently of risk preferences; we discuss this in the next topic, Reconciliation
of Risk-Neutral Valuation with CAPM.
(C) Copyright 1999, OS
Financial Trading System