Why Model Development is Broken in Banking

Model development in the banking industry has been broken but bankers don't realize it because they aren't great model developers. First off great talent is very rare. It takes good recruiters to find talent, good managers to train talent, and a great person who is driven to continue learning after school. Second you need a management team that understands data determines the model structure and not tradition or a manager's opinion.

The reason model development is broken is due to poor talent and poor management. Employees lack the skills and drive to eat, sleep, and breath risk management and statistics while banks are trying to pass regulations with as little resources as possible. The result is the constant use of one model structure, models with short life spans, and a continual rotation of talent. All three of these results are problems within themselves and costs banks millions of dollars as they need more modelers to redevelop models, hire new employees, and train employees.

In general I feel banks are falling behind in basic theory and in new methods such as machine learning. If you have poor talent there is no hope in building great models, attracting great talent, and reducing the costs associated with model development.

So how do we fix this? Create a culture for quants and not MBAs. Banks need to start understanding that the perfect company culture for business minded employees is very different than the perfect company culture for quants. Quants need a job where they are constantly challenged with complex problems, a high amount of focus on learning, a company that supports new technology while removing tech security barriers, a culture that promotes the brightest and hardest working, and dare I say a relax environment (even from a guy who chooses to wear a suit every day I'm in the office). The difference between business and quant workers is the complexity. Business people like challenges however their challenges are typically qualitative and driven by Excel. When quants hear Excel we cringe as Excel is highly inefficient for large complex problems. Business people love fancy charts with great formatting while quants just want to see the results regardless of the colors and font size. A great example is when banks promote learning by giving us Lynda.com. Lynda is great for learning Excel, PowerPoint, and how to give a presentation however these are useless for quants. We want to know how and why to implement a random forest over neural network. Lynda helps business people however it doesn't help quants. If you want to promote learning then give us Coursera and time to learn with it.

I get a bit off track but the point is the best quants aren't going into banks, they are going into hedge funds, tech companies, and data science. These other industries offer a mix of better pay, flexible benefits, challenging problems to solve, and managers with similar interests and educations. So why I am still in risk management? I love banking and so do most quants however there needs to be change so that the best talent wants to be in banking and not fancy hedge funds and frilly tech firms.

What do you think? Am I right or am I crazy?

If you want more content like this or want to see me change the banking industry follow me on LinkedIn or better yet SUBSCRIBE to my YouTube channel.