TL;DR: There are two types of quants: “buy side” quant and “sell side” quant. A buy side quant might work in a hedge fund coming up with directly applicable money making strategies (using maths/stats/coding). A sell side quant would work in an investment bank crunching numbers for clients. Financial engineering degrees typically prep you to be a sell side quant, not a buy side quant.
Square brackets in the text below refer to sources (links to sources are at the very bottom of this post).
What is a quant?
When someone says quant online, they are probably referring to a fancy maths/stats/coding-heavy role in algorithmic/systematic trading, not someone doing pricing [QN1]. As we will see, this is basically a “buy side” quant. There are really only two types of “quants”: buy side quants working in hedge funds, and sell side quants working in investment banks [QN5].
Sell side quants
Before the 2008 financial crisis, sell side quants reported to Sales and Trading (S&T) and owned a share of profit/loss (PnL) [QN2, Q2] but after the crisis their roles changed. To understand what changed, we need to understand that a sell side firm makes money by being a liquidity provider and/or taking a cut in the form of commissions: they create, price, hedge and arrange the sale/purchase of tailor made financial products for their clients, or they buy/sell products held by the firm and make money from the bid/ask spread [B, C, D].
Sell side firms traditionally employed quants in large swathes for pricing and portfolio modelling to price and hedge the positions they took when servicing clients [WSO1]. The modelling/pricing task for a sell side quant is predetermined by the products a client is buying/selling [WSO1]. As such, a front office sell side quant is what comes closest to a buy side quant.
Nowadays though, sell side quants are in the Risk department of a firm (doing pricing, risk management, hedging, etc); they only support the S&T department indirectly and do not own PnL [Q1]. “Sell side quants more often use traditional mathematical finance - stochastic calculus, differential equations, etc.” [R], so this is where an MFE degree would be helpful. Traditional mathematical finance topics are [B, C, QN5]:
- Derivatives pricing
- Partial differential equations (e.g. Black-Scholes equation)
- Numerical methods
- Ito (stochastic) calculus
- Others (risk management, model validation, CVA, XVA, HPC, etc)
There has been a trend of people trying to switch over from the sell side to the buy side for many reasons: fewer working hours, more interesting work, salary ceiling is higher, etc. However, switching over is not easy because skills such as derivatives pricing may be irrelevant to the buy side firm [Q4]. If a sell side quant with pricing knowledge is trying to move to the buy side, the value of their knowledge depends on the function within the firm they have applied to [QN4].
Lastly, a year or two of being in a sell side quant role is going to give everything that could ever come in useful in a buy side quant role, technically speaking, and one should try to move roles soon thereafter [WSO2].
Buy side quants
Buy side firms are not as concerned about pricing because they are not making structured products [WSO1], and the traditional definition of quant (pricing and structuring of derivatives) is mostly outdated on the buy side [Q1].
In contrast to the sell side, who are liquidity providers, buy side firms can be thought of as liquidity seekers [D] who want to buy/sell various products (e.g. trading in open exchanges or over-the-counter) and grow retail, institutional and/or house accounts by taking calculated gambles on the markets [2, 3].
The value in the buy side lies in coming up with money making strategies that work [WSO1]. Buy side quants develop, refine and manage these money making strategies [Q1] to generate alpha (returns above market growth, [D]) using whatever means available to them, usually maths/stats/coding. These means can be almost anything: stats, time series and machine learning, etc [R, 2, QN5].
Note that while topics like numerical methods and
C++ are often said to be core skills for quants, this only true for those working on pricing libraries (who are generally sell side quants), and buy side quants would code in
Python, not
C++ [QN3].
Sources:
