QuantCandy's Blog

Lots of great interview questions and tips on quantlife here

I have been asked this more than once, it seems to be a perennial favorite!  In one sentence, tell me what an outer join is?  Now how about an inner join?  When would you use an outer join vs. an inner join?

You are rolling a fair six-sided die.  You will receive a reward equal to the roll (that is, if a 3 is rolled you get $3 etc.).  You can accept the result of the first roll and stop, or you can reject it and roll again, taking that roll as your win.  What is the fair value of this game?  What if you roll a third time, what is the pattern?  Can you use the answer from the first two rolls to answer this?

 You have a friend who has two children.  You do not have any other information regarding their ages or gender.  You go to his house and ring the doorbell.  A little boy answers (you can assume that this is his son.)  What is the probability that the other child is a girl?

Is it possible to have a bond with convexity of zero?

How would you hedge default risk?

Here are some interview questions:

Time Series
What is a stationary time series?
How do you handle non-constant volatility?
What is a logistic regression?
What is a moving average process?

See notes here and here

Credit Derivatives
I was on the subway the other day reading my textbook on Credit Derivatives and a random stranger asked “What is a Credit Derivative?” I thought this would make a great interview question, as I had about 60 seconds to try to come up with an answer before my stop.

Why do companies have their debt rated, but not their equity?

In the Credit Derivatives course, the lecturer brought up lots of very practical facts about ratings agencies and how they work.  I am attending this class in person, and also taking another class by Professor Cont online.  So far the courses work very well together.  One is highly theoretical while the other is very practical.  In the first lecture, the professor mentioned that there is a great article in the Wall Street Journal detailing the structure of a CDO. It’s really interesting because he also went over the reason that most ratings agencies have about 20 ratings categories. S&P, for example, has AAA, AA, A, BBB, BB, B, CCC, CC, C, D and NR with “notches” of + and -. Too many categories, and how would you really differentiate between them; too few, and they would be too undifferentiated to provide much useful information. He also discussed the NR (Not Rated) category and reasons for it. Is the fact that a company decided not to continue with a rating derogatory or not? It depends on the reason for this. If a company knows it is facing bankruptcy, for example, it might not want to continue to pay for ratings, so a rating of NR would have a negative connotation in this case. However M&A activity may lead to NR ratings as well.

And here is an article on the anatomy of a ratings downgrade.

During the class, he asked many questions that I thought would make fantastic interview questions! They include:

1) How would you generate a monte carlo analysis of correlated normal random variables?
2) What is the shape of an equi-probable distribution (say, p = 0.001) of a bivariate normal distribution?
3) What is the Matlab command to generate a bivariate normal cdf?
4) What is the distribution of the variance of a binomial distribution?
5) Explain why an A rated security has a credit spread term structure that may be downward sloping and why a B rated security may have one that is upward sloping.
6) Why do we need ratings agencies for the debt of a corporation, but not for the equity?
7) What happens to the equity price of a firm when there is a ratings downgrade? (Answer: it depends on the reason for the downgrade.)

Here is some discussion on a hedge fund manager’s speculation on reasons people didn’t anticipate the subprime crisis.  And the WSJ profiles John Paulson, a hedge fund manager who did see this coming.