2015 Summer projects- SAS credit risk project summary

SAS Credit Risk Project

Designed & led by Financial Math Alumnus- Jonathan Leonardelli

Objective:
For students to apply credit risk concepts while developing SAS programming skills

Goals:
1. Become Base SAS certified
2. Have an understanding of CCAR and Basel II calculations
3. Learn how to model PD / LGD / EAD
4. Use equations to calculate Expected Loss (EL), RWA (Risk Weighted Assets), and capital ratios

By Aisha Barnes & Preethi Kankanala- The purpose of the summer SAS case study was to develop an understanding of the different steps that are involved in calculating the loss portion of CCAR (Comprehensive Credit Analysis & Review). CCAR is a regulatory framework that ensures Bank Holding Companies (BHCs) have enough capital under the worst scenarios. This is tested under various stress-testing scenarios.

In our case study, we analyzed a portfolio of different products and estimated the capital that is required to hold the portfolio under three different scenarios. For this, we have estimated the historic Probability of Default (PD)*, Loss Given Default (LGD)* and Exposure at Default (EAD) and forecasted the future values using a variety of techniques (e.g., regression models) in SAS. Then we used these values to estimate risk weighted assets and capital.

This exercise helps BHCs ensure that they have enough capital if there is any change in the economic conditions. If the capital plan does not pass regulatory review, then the company has to change it to ensure there is adequate regulatory capital.

IMG_6600

"Throughout the project, Financial Math Alumnus and Board Member, Jonathan Leonardelli, directed and mentored our team. We gained knowledge and enhanced our technical and business skills under his guidance. The project provided us hands-on experience on estimating the credit metrics and how to apply them with real world problems."- Preethi Kankanala, December 2015 Graduate

"This summer I experienced real application of how I will use my Financial Mathematics degree. I learned how to program in SAS and plan to gain certification. I used SAS to find the amount of capital a bank reserves to meet the Basel II requirements. I feel confident in having these skills."- Aisha Barnes, December 2015 Graduate

*PD (Probability of Default) = likelihood that a loan will default in the future
*LGD (Loss Given Default) = amount a bank will lose if a customer defaults on their loans

(All the project groups presented their summaries at the end of the summer session. More info to come soon from the other groups).
IMG_8194

Join our new workshop- “Introduction to Financial Risk”

NC State's Financial Math program has partnered with 2004 Alumnus, Jonathan Leonardelli, to create a new workshop series "Introduction to Financial Risk" for all NC State students and faculty. The workshop is also opened to the public.

Students and faculty in Mathematics, Statistics, Economics, Finance, Operations Research, MBA and other related programs are welcome to join!

Here is what you will learn:

Risk Workshop Overview

Presenter information:

Jonathan Leonardelli, FRM, MFM

Jonathan Leonardelli, Risk Consultant at the Financial Risk Group, specializes in credit and market risk management. Over the course of his career he developed a diverse knowledge of retail banking risk as well as the technical skills needed to integrate risk assessment processes into a company’s business and technology infrastructure.

Jonathan’s career started with positions in the credit risk groups at Wells Fargo (Wachovia) and BB&T. In these positions, he developed expertise in acquisitions and portfolio risk management.  In his current position, Jonathan develops and implements processes that provide quantitative risk assessment and reporting capabilities for clients that include banks, hedge funds, and asset management companies.

Jonathan is an experienced presenter and author.  He is a certified SAS® Risk Dimensions Instructor. His papers in financial risk management covered topics such as the Dodd-Frank Act and its implications for risk professionals, as well as techniques for handling missing data. He has also authored a Webinar for the Insurance & Finance SAS® Users Group (IFSUG) regarding loss estimation using roll rate matrices.

Jonathan holds an Masters of Financial Mathematics from North Carolina State University and is a member of the Global Association of Risk Professionals (GARP).

FRM designation since 2010

SAS Certified Advanced Programmer for SAS 9

***************************************************

Those interested- please contact Leslie Bowman, Director of Career Services- leslie_bowman@ncsu.edu 

Workshop begins Friday, September 5th 2014- registration deadline, August 28th

New Series- “Meet our Financial Math Alumni”- Coffee Chat with Jonathan Leonardelli

"Meet our Financial Math Alumni" is an up-close interview series with select Financial Math alumni to learn more about their career, experience and knowledge after receiving their Masters in Financial Math degree from NC State University. Alumni are important part of our program for many reasons. They provide support, vision and strategy to ensure the success of our program. They are role models and mentors for current students. They strengthen the reputation of the Financial Math program. They provide job leads and recruitment activity for students. Our alumni are intelligent and awesome! Thank you to those who participant in this series"- Leslie Bowman, Director of Career Services

*****

Meet Jonathan Leonardelli, FRM, Risk Consultant (Graduated 2004)

Image

Interview conducted and summarized by Yi Chao and Xiaohong Chen, Financial Math Interns, May 2015 Graduates

Jonathan is currently a Risk Consultant at Financial Risk Group. He is also one of the first students to graduate from NCSU’s Financial Mathematics program. We were honored to have the opportunity to interview him.

Part I: Education & Job Background

1) Interviewer: Why did you decide to get a Masters in Financial Math at NC State?

Jonathan: I had just moved here and was interested in changing careers. I wanted to find a program that combined mathematics and finance. As it happened, NC State was in the process of creating the Financial Mathematics program. Although the start of the program was still a couple years off, that time allowed me to get the necessary mathematics background I needed. Entering the program produced the exact result that I wanted: it gave the mathematics I need to do interesting work in the banking industry.

2) Interviewer: How did the program prepare you for your job?

Jonathan: The program really gave me the depth of math that I needed to work in the field of risk. It also taught me how to apply rigorous logic to a problem to help find a solution. Beyond this, though, I learned that life was filled with randomness. This randomness, as a result, causes the quantification of some metrics to be difficult.

Image

Part II: Analytic techniques

3) Interviewer: "Big Data" is a hot specialization in the field. Do you see this as long term trend or something that might pass as a fad?

Jonathan: It is definitely not a fad. In this day and age many actions we take, especially when using a piece of technology, is probably captured and stored in some database. Now, think about all those action and all the data that comes along with it. What is a company going to do with this data? They are going to improve their sales, improve their risk practice…the list goes on. Having skills to work with big data, to be able to find the relevant information and then integrate it into a model, is very useful.

4) Interviewer: The trend of “Big Data” implies that historical data can shed some lights on future prediction. However, this contradicts with “efficient market theory” to some degree. What are your thoughts about this?

Jonathan: I think concerns should always be involved when using historical data because the tacit assumption is that the future is going to behave like the past. That being said, there are ways to mitigate these concerns. For example, when we calibrate models one of the first things we do is test it with a different period of data to see if the model is robust. Sometimes we might use the model on data representative of a stressed scenario (i.e., a scenario that is uncommon but still possible) and see how the model performs. If it performs badly, we try to assess why that is. Are the parameter estimates wrong? Are different variables needed?

5) Interviewer: Does your company use stochastic models to predict interest rate? What kind of models are used?

Jonathan: To be honest, in my current job the main stochastic (i.e., diffusion based) models I have used are CIR (Cox-Ingersoll-Ross) and GBM (Geometric Brownian Motion) with the occasional jump-diffusion model thrown into the mix. Most of the models I have worked with recently are linear regression, logistic regression, and Markov chains.

6) Interviewer: In your area of specialization, what is your favorite method or model and why? Do you believe it is perfect?

Jonathan: Markov chains and their resulting transition matrices. This comes from the years when I worked in the banking industry. At a glance, the transition matrix tells you the behavior of different segments of accounts. Depending on how the states of the Markov chain are defined, the transition matrix can tell you: 1) The probability of going to default, 2) the probability of paying off, 3) the probability of curing, 4) the probability of moving across multiple states over a given time, etc.

The transition matrix is a great summary tool. Of course, it is not perfect. Always keep that in mind when you build a model. Even though the model looks pretty and deals well with the data – now – it is not perfect. It is an easy move from complacency, when the model is performing well during good times, to anxiety when the model is performing poorly during a financial crisis.

Image

Part III: Risk Management

7) Interviewer: How do the regulation policies enacted after crisis affect the behavior of your company?

Jonathan: The regulations have not impacted the behavior of our company. However as a risk consulting company, we have seen more requests from financial institutions asking us to help them comply with the regulations.

8) Interviewer: The goal of risk management is to achieve a balance between returns and risks. Thus, with a lot of capital and human resource spent, risk management may, to some extent, reduce a company’s profits. Driven by the motivation of maximizing the profits, will the companies pay enough attention for risk management?

Jonathan: I may be biased, given my chosen career path, but I think with the recent financial crisis still fresh in our memories as well as all the regulations that were created as a result of it, businesses will continue to pay enough attention to risk management. And, I think, it will be that way for a while.

Part IV: Suggestions & Advice

9) Interviewer: Any tips for those interested in getting into the field?

Jonathan: First, be comfortable with the idea of randomness. Randomness is uncertainty and uncertainty is risk. Second, companies do not only seek candidates that are good at math and programming. They also seek those candidates who can clearly present their ideas in writing and presentations. Third, get perspective from areas outside of mathematics. I strongly suggest taking business classes because it gives you a different view of a business. A company is multifaceted and does not solely revolve around models.

10) Interviewer: What courses do you recommend?

Jonathan: The courses in the Financial Math program are very good. Among them, I definitely think the probability course is the most important one. That course takes you deep into the world of randomness and, hopefully, makes you comfortable with it. If you have an option of taking a course on risk or financial regulation that would be good. I took econometrics as an elective and, more often than not, draw upon the math tools I learned from that class more than others. Another very good course I took, and have used frequently, is time series analysis.

*****

"It is a great experience to interview Jonathan. He is humorous and smart. We learned a lot when we talked with him. Thanks Jonathan for participating in our interview! We are sure that your answers will shed some light for those interested in Financial Mathematics!"- Yi Chao and Xiaohong Chen

 

Financial Math Alumni Panel Discussion on Big Data, High Frequency Trading and more…

Image

(Above left to right- Jeff Scroggs, Jonathan Leonardelli, Jared Bogacki, Ryan Wesslen, Albert Hopping, Emmanuel Sanchez)

Jeff Scroggs, Director of the Financial Mathematics Program, conducted a panel of industry experts on trends in financial mathematics and quantitative risk.  Three of the practitioners,  Jared Bogacki (BB&T),  Jonathan Leonardelli (Financial Risk Group), and Emmanuel Sanchez (Allianz), were from the class of 2004 – the first class to graduate with a Masters of Financial Mathematics.  Two of the panelists,  Albert Hopping (SAS Institute, class of 2007) and Ryan Wesslen (Bank of America, class of 2009), were more recent graduates.

The panel offered the audience an opportunity to see what role quants play in optimal business practices.

All the panelists agreed that their Masters of Financial Mathematics from NC State opened opportunities for career advancement, ranging from a entry-level quant positions to promotions to lead quant.

Image

Jared Bogacki started with the topic of 'Big Data'.  'Big Data' is the use of data sets so large and complex that it has become difficult to process them using traditional tools or data processing applications.  It is currently a hot specialization for quants, and will likely remain a sector of the job market that is hungry for well-qualified people. 'Big Data' is a broad term that covers several topics including analytics used to glean market sentiment as well as some aspects of high frequency trading.

High-Frequency Trading (HFT) is algorithmic trading that uses algorithms to rapidly trade securities. The methods involve proprietary trading strategies carried out by computers to move in and out of positions in fractions of a second.  Of course, there are many different approaches to HFT that range from geometric observations (Golden Cross) to taking advantage of arbitrage opportunities across markets (e.g. New York vs London).  Albert Hopping was asked, “Do you think HFT is good or bad for markets?”  He pointed out that HFT is really a response to the way electronic trading in markets such as the New York Stock Exchange and Chicago Mercantile Exchange function are regulated.  HFT reduces friction by providing liquidity, but it can also cause flash crashes that force markets to temporarily halt trading.  There was consensus that such trading is impossible to control – regulations always lag advances in technology.

Image

Effective deployment of quantitative risk management is a challenge for all businesses, large and small. Jonathan Leonardelli led this discussion.  There are many aspects to risk management, ranging from data mining for parameter estimation to the creation of dashboards in the context of Enterprise Risk Management. The push to use more quantitative risk measures can come from inside the business or from outside.  For example, regulations like the Dodd-Frank act require more transparency from banks and reliable quantitative measures for stress testing.

Image

Emmanuel Sanchez was asked to lead the discussion on how weather-related risks can be controlled. Climate change has brought the spotlight on some of the impacts of short-term and long-term weather. Catastrophe bonds can provide protection against large-impact short-time events such as hurricanes and floods; whereas, weather insurance provides coverage based on measures such as annual rainfall and heating degree days (when it is cold enough to need a furnace).  The availability of these securities and derivatives allows the sharing of risk inherent in sectors like farming and homeowners insurance.

Ryan Wesslen was asked to share his favorite/best model in the area of consumer credit risk and/or counterpart credit risk.  Of course, quants do not share the best model to predict recent trends, but many share their wisdom in hindsight after a particular model offers no significant competitive advantage.  No model is likely to produce a clear crystal ball with excellent valuations, and all models provide some level of insight. For example an individual’s credit scores, like many holistic indicators, are good at the extreme high and low ends. But scores in the middle aren’t good predictors of default risk.

Thank you Jared, Albert, Jonathan, Emmanuel and Ryan!

Keep a lookout for an upcoming new blog series- interviews with a featured alumnus to learn even more about his or her experience and expertise.