150 Quant Questions Flashcards
Put options with strikes 30 and 20 on the same underlying asset with the same maturity are trading for $6 and $4 respectively. Can you find an arbitrage?
The value of a put option with strike zero is 0. So we have P(30) = 6, P(20) = 4, and P(0) = 0. These three points are on the line P = (1/5)K. But this contradicts the fact that put options are strictly convex functions of strike price, so there is an arbitrage opportunity. The out with strike 20 is overpriced, and you can buy low, sell high. Buy ⅔ put options with strike 30 and sell 1 put option with strike 20. At maturity T, the value of the portfolio is V(T) = 2max(30 - S(T),0) - 3max(20 - S(T),0)
The number 2^29 has 9 digits, all different. Without computing 2^29, find the missing digit.
For any positive integer n, the sum of the digits is D(n). n - D(n) is divisible by 9. D(2^29) = (sum from 0 to 9 of j) - x = 45 - x. Break 2^29 into factors, with one being as close to 45 as possible: 2^5 * (2^6)^4 = 2^29. Then you have 9 | (45 - x) - 2^5 = 13 - x. x must make it such that 9 is divisible, so x must be 4.
Let W_t be a Wiener process, and let X_t = the integral from 0 to t of W_τ dτ . What is the distribution of X_t? Is X_t a martingale?
dX = Wdt = Wdt + 0dW. Therefore, X is a diffusion process with only drift part W, and thus not a martingale. Using integration by parts:
X_t = tW_t - int(0,t)τ dW_τ
= tint(0,t)dW_τ - int(0.t)τ dW_τ
= int(0,t)(t - τ)dW_τ
~ N(0, int(0,t)(t-τ)^2dτ )
= N(0,t^3/3)
X_t is a normal random variable of mean 0 and variance t^3/3
Alice and Bob stand at opposite ends of a straight line segment. Bob sends 50 ants towards Alice, one after another. Alice sends 20 ants towards Bob. All the ants travel along a straight line. Whenever two ants collide, they simply bounce back and start traveling in the opposite direction. How many ants reach Bob and how many ants reach Alice?
Bob gets 20 ants, Alice gets 50. The number of collisions is 20 * 50 = 1000 collisions.
Find all values of ρ such that
(1 0.6 -0.3
0.6 1 ρ
-0.3 ρ 1)
is a correlation matrix.
The det(M) >= 0. Find det(M) and solve.
How many independent random variables uniformly distributed on [0,1] should you generate to ensure that there is at least one between 0.7 and 0.72 with probability 95%?
N is the smallest number of random variables such that P(at least one rv in [0.70,0.72]) >= 0.95. The probability that a rv uniformly distributed on [0,1] is not in [0.70,0.72] is 0.98. So the probability that none of the N independent variables are in [0.70, 0.72] is 0.98^N. So P(at least one rv in [0.70,0.72]) >= 0.95 = 1 - 0.98^N >= 0.95. Solve for N.
Show that the probability density function of the standard normal integrates to 1
The probability density function of the standard normal is (1/sqrt(2π))e^(-t^2/2)dt. Show that the integral of this from -∞ to ∞ = 1
You travel exactly 1 mile south, 1 mile east, and 1 mile north. You find yourself at the starting point. If you are not at the North Pole, where can you be?
Anywhere that has a latitude circumference of 1near the South Pole. But any mile north of such a latitude with circumference 1/k when you loop the globe k times will put you back where you started.
Solve the Ornstein-Uhlenbeck SDE
The OU SDE is
dr_t = Λ(θ - r_t)dt + σ dW_t. It is used in the Vasicek model for interest rates.
Write a function that computes the n-th Fibonacci number
def fibonacci_iterative(n):
if n <= 0:
return 0
elif n == 1:
return 1
else:
# Initialize the first two Fibonacci numbers
a, b = 0, 1
# Iterate from 2 to n
for _ in range(2, n + 1):
# Compute the next Fibonacci number
a, b = b, a + b
return b
What is the value of i^i (where i is sqrt of -1)?
Recall i = cos(π/2) + isin(π/2) = e^(iπ/2)
i^i = (e^(iπ/2))^i = e^(-π/2)
Which number is larger, π ^e or e^π ?
e^π
Show that
(e^x + e^y)/2 >= e^((x+y)/2)
Let a = e^x and b = e^y. solve for the case is >= 0
Solve x^6 = 64
The six unit roots of z^6 = 1 are:
z_k = exp(2kπi/6) = exp(kπi/3) = cos(kπ/x) + isin(kπ/3)
therefore x_k = 2cos(kπ/3) + 2isin(kπ/3)
What is the derivative of x^x
By chain rule, (x^x)’ - (x^x) * (lnx + 1)
Calculate sqrt(2 + sqrt(2 + sqrt 2 + ….)
2
Find x such that x^x^x^x^x… = 2
sqrt(2)
Which of the following series converge (first two sums from 1 to ∞, second sum from 2 to ∞):
Σ 1/k ; Σ 1/k^2 ; Σ 1/(klnk)
Since all the terms in the middle one are positive, partial sums are uniformly bounded and thus it is a convergent series. It equals π^2/6
Compute integrals:
xln(x) dx and xe^xdx
Use integration by parts: int(udv) = uv - int(vdu) where u is ln(x) and dv is x
Compute integral (x^n)ln(x)dx
Use integration by parts: int(udv) = uv - int(vdu) where u is ln(x) and dv is x^n
Solve the ODE:
y’’ - 4y’ + 4y = 1
y(x) = c1e^(2x) + c2xe^(2x) + ¼
Find f(x) such that
f’(x) = f(x)(1 - f(x))
Use y and integrate:
f(x) = Ce^x/(1 + Ce^x)
Define the Black-Scholes PDE
It estimates the theoretical value of derivatives based on other investment instruments, taking the impact of time and other risk factors into account. It is still regarded as one of the best ways to price an options contract.
The prices of three put options with strikes 40, 50, 70 but otherwise identical, are $10, $20, and $30 respectively. Is there an arbitrage opportunity present? If yes, how can you make a riskless profit?
When plotting, draw a line between the first and last points to see if there is a violation of the convexity criterion. The point at P(50) = 20 is above that line, violating the convexity criterion, so an arbitrage opportunity is available. Buy 3 of the overpriced option, and sell 2 of the $10, and 1 of the $30 options. This will generate a positive cash flow of $10. The value at maturity will be:
V(T) = 2max(40 - S(T),0) + max(70 - S(T),0) - 3max(50 - S(T),0)
The price of stock is $50. In three months, it will either be $47 or $52, with 50% probability. How much would you pay for an at-the-money put?
Real world probabilities do not play any role in valuing an option in a (one period) binomial tree model. Thus, the 50% is just there to throw you off. The value of the option is the discounted expected value of the payoff of the option in the risk-neutral probability measure. Since the interest rates are zero, this can be written as
P(0) = p(RN,up)P(up) + p(RN,down)P(down)
The up and down factors are u = 52/50 and d = 47/50. Thus the risk-neutral probabilities of up and down are
p(RN,up) = (1-d)/(u-d) and
p(RN,down) = (u-1)/(u-d)
The ATM put pays $3 if the stock goes down to $47, so P(down) = 3 and expires worthless if the stock goes up to $52, so P(up) = 0. Putting hte numbers in gives 1.20
The price of stock is $50. In three months, it will either be $60 or $40, with 50% probability. The value of a three months at-the-money put on this stock is $4. Does the value of the three months ATM put increase or decrease, and by how much, if the probability of the stock going up to $60 were 75% and the probability of the stock going down to $40 were 25%?
The probabilities don’t play a role, so the value would still be $4.
What is risk-neutral pricing?
Risk-neutral pricing, or valuation, refers to valuing derivative securities as discounted expected values of their payoffs at maturity, under the assumption that the underlying asset has lognormal distribution with a drift equal to the risk-free rate. It cannot be used for path dependent options such as American options, barrier options, and Asian options (from European perspective).
Describe briefly how you arrive at the Black-Scholes formula
There are several methods. 1) risk-neutral pricing 2) the Black-Scholes PDE solution 3) Binomial tree model pricing. there are 12 different ways to derive it that can be found in Wilmott.
How much should a three months at-the-money put on an asset with a spot price $40 and volatility 30% be worth?
P(ATM) ~ 0.5σS_0sqrt(T)
where S_0 is the spot price and σ is the volatility. T is time in years. Plug and chug.
If the price of a stock doubles in one day, by how much will the value of a call option on this stock change?
You can approximate the value of a deep in the money call on a non-dividend paying asset by using the put-call parity, as C ~ S - Ke^(-rT) where K and T are the strike and maturity of the option, and r is the constant risk free rate. Thus, if the spot prince S doubles, the call option will be even deeper in the money and the value will be approximately 2S - Ke^(-rT). i.e. the value of deep in the month call options will double if the price of the stock doubles.
If the option is around at the money, the percentage generated by doubling the stock price is 1 oom larger since the option will become deep in the money.
If the option is deep out of the money, it trades for fractions of cents. The doubling of the spot price would result in changing of the price of the option by several oom
What are the smallest and larges values that Delta can take?
-1 and 1
What is the Delta of an at-the-money call? What is the Delta of an at-the-money put?
0.5. -0.5
What is the Put-Call parity? How do you prove it?
Show that the time value of a European call option is highest at-the-money
What is implied volatility? What is a volatility smile? How about a volatility skew?
What is the Gamma of an option? Why is it preferable to have a small Gamma? Why is the Gamma of plain vanilla options positive?
When are a European call and a European put worth the same? What is the intuition behind this result?
What is the two year volatility of an asset with 30% six months volatility?
How do you value an interest rate swap?
By how much will the price of a ten year zero coupon bond change if the yield increases by ten basis points?
A five year bond with 3.5 years duration is worth 102. What is the value of the bond if the yield decreases by fifty basis points?
What is a forward contract? What is a forward price?
What is the forward price for treasury futures contracts What is the forward price for commodities future contracts?
What is a Eurodollar futures contract?
What are the most important differences between forward contracts and future contracts?
What is the ten-day 99% VaR of a portfolio with a five-day 98% VaR of $10 million?
How would you compute π using Monte Carlo simulations? What is the standard deviation of this method?
What methods do you know for generating independent samples of the standard normal distribution?
How do you generate Brownian motion stock path using random numbers from a normal distribution?
How do you generate a sample of the standard normal distribution from 12 independent samples of uniform distribution on [0,1]
What is the rate of convergence for Monte Carlo methods?
What variance reduction techniques do you know?