FICC market-making desks in large investment banks operate in principle similarly than 20 years ago. However, the wider banking environment around them has changed significantly in recent years.
From my perspective, these shifts have fundamentally altered the relationship between quant and trading roles, changing how they collaborate and create value.
Back then, Rates & FX trading was scattered across products and regions. Capital was loose and leverage plentiful, desks hired quants directly to build models and pricing spreadsheets, booking and risk management systems were sourced from third party vendors and were notoriously not fit for the purpose.
Most of the problems were concentrated aroud multi-curve bootstrapping, OIS discounting, emerging market currencies etc...
In the past, scattered tooling and scattered market data were plaguing banks with internal arbitrage. Inability to calculate capital charges was misallocating regulatory capital. Inability to see aggregate risks across businessess was blindfolding senior management.
Not to mention the massive work duplication on desk level and operational risks associated with the “spreadsheet pricing”.
To make things worse, before global financial crisis, banks showed almost unlimited appetite to sell unnecessarily complex derivatives to clients. The more complex product, the fatter the commission. Balance sheet usage and capital requirements were treated only as an afterthought, if at all.
Afer the global financial crisis in 2009, few things changed:
Regulatory and operational costs became the new battleground. It quickly became evident that the old operating model could not survive in a world where regulatory risk charges mattered as much as the desk P&L. Today, many profitable trades simply won’t happen, because they consume too much regulatory capital.
It became clear that many complex derivatives served no meaningful purpose in the real economy, and their continued existence could not be justified. As the market sobered, these structures faded into obscurity. Majority aspiring quants today don’t realize that advanced stochastic calculus survives mostly as an interview filter rather than a practical tool in day-to-day finance.
It’s much more profitable to build scalable and cost-efficient franchise serving institutional and corporate clients (real-money), rather than trying to outsmart hedge funds and prop trading firms, or sell complexity no one needs. With the electronification of trading, the consolidation of technology stacks, and front-to-back automation, most of the remaining value lies in improving operational efficiency.
The fact that e.g. Citadel Securities, Jane Street, et. al generates significantly more profit per employee than most of the “bulge bracket” banks’ CIB divisions is driven to large extent by their efficient modus operandi, not by any magical model able to predict tomorrow’s market.
In this environment, traditional investment banks now face fierce competition from far more agile market makers. What began as a revolution in ETF market making by afforementioned challenger firms is now steadily spreading into the broader bond and FICC markets as well.
This is the world I walked into almost two decades ago, and it’s shaped everything I’ve worked on since. Over the last decade I helped to deliver strategic analytics and help aligning front office processes to this new reality.
Most of the wall street quants are coming from physics, math or chemistry backgrounds. Many of them studied stochastic calculus and most of them are mistaken that they will deal with “complex math formulas”, most of the time. Critical thinking and judgement which comes from math education is here to stay, but most of the mathematical wizardry will be taken away by powerful AI.
In fact, the truth is that big portion of complexity they will ever deal with comes from the lack of focus - an unavoidable byproduct which is created each time when multiple people work together, often under budgetary or timeline contraints. The larger the bank, the harder to keep this focus. Moreover, contrary to cracking the complex math, AI will not help much here.
On the flip side, this is not necessarily a bad thing. The most successful people I met didn’t chase their academic passions; instead, they focused on making themselves useful, and they learned to love what makes them valuable.
From my own experience, a typical example of what most of the market-making desks are battling with is the failure to deliver the following:
Pre-trade price which is displayed to a trader and quoted to the counterparty comes from the same methodology as the one used for official end-of-day snap, regardless of which desk and in which geographical region is making that price.
Almost every bank is affected by something like this to lesser or larger degree. The solution is conceptually simple but very difficult to achieve in practice. In past few years, I have been participating in the following initiatives:
Streamlining sales-trader workflow by limiting usage of custom analytics, custom market data and on-board desks to bank-wide strategic analytics.
Designing robust systems which can cope with round-the-clock operations in major financial centres (London, New York, APAC), with multiple P&L snaps per day.
Unifying concepts and deliver consistent market conventions across Rates, FX and credit markets.
Building quant libraries and interfaces which can be integrated into heterogeneous environments and simplify the life of technology teams.
Last but not least, most investment banks developed their FICC businesses from Western financial centres. This is often reflected in the core design of their trading and risk systems. While G10 will remain on the top of the game for foreseable future, the emergence of CNY as a challenger to USD hegemony and rapid evolution of FICC emerging markets elsewhere is fact not to be ignored.
⚠️ DISCLAIMER: Opinions are my own and not those of my employer.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.