BUFN742 Financial Engineering Term B Fall Semester 2022
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
BUFN742 Financial Engineering
Term B Fall Semester 2022
Course Description:
Financial engineering has become one of the most popular areas of applied finance in the last twenty years or more, bringing together financial mathematics, financial programming, and asset pricing theory and mechanics to develop and value innovative financial instruments. Examples of such instruments include collateralized mortgage obligations (CMOs) and REMIC securities, collateralized debt obligations (CDOs), credit-linked notes (CLNs), and credit default swaps (CDS), collateralized loan obligations (CLOs) are just a few of the more noteworthy structured finance instruments that transformed markets. Today, financial engineering continues to play a major role in financial markets as new structures are being developed by Freddie Mac and Fannie Mae, for example, to transfer trillions of dollars of credit risk to private investors such as reinsurance companies, hedge funds, banks and asset management firms. The tools required to develop these new securities include extraction, sampling and manipulation of large loan level datasets for complex statistical analysis, simulation of cash flows affected by house price and interest rate stochastic processes affecting twin embedded options in a standard mortgage loan; namely borrower default (put option) and prepayment (call option). The valuation of these embedded options in a fixed-income-like mortgage presents one of the most challenging exercises in financial engineering, particularly relating to the allocation of losses and prepayments to various investor tranches of a credit risk transfer (CRT) security instrument.
Students will work in a team-based setting to price a representative CRT transaction recently issued by Freddie Mac (STACRs or ACIS) and Fannie Mae (CAS), leveraging techniques described above, among others. The 35 million loan level database of credit performance of the GSEs will serve as the basis for sampling historical performance from which to estimate and validate survival models of mortgage default and prepayment used to value the embedded options. The teams will then estimate and validate stochastic interest and house price models. These activities will allow students to empirically gain hands on experience with the issues, data and analysis associated with building complex term structure models and understanding asset pricing dynamics in a realistic setting. Students will also learn how to efficiently “link” the stochastic processes with the default and prepayment option behavioral equations in order to generate a lognormally-shaped credit loss distribution from a simulation of 500 paths of interest rates and house price changes over the life of a pool of mortgages. Cash flows from the pool will then be tranched according to rules for each CRT instrument including clean-up call provisions and discounted in order to establish a price for each tranche. These results will then be calibrated against external ratings loss subordination levels to effectively mark the loss distribution to market. The course is built to mimic the type of team-based environment you would experience at a company that has hired you to conduct complex modeling. As such, in addition to building strong financial engineering skills you will also hone your ability to work in teams under tight deadlines, communicate technical findings to senior audiences (your presentations will be a sort of mock board presentation to myself and your classmates), and develop your skills at multitasking in an environment with imperfect information. As a result, students will have one of the most interesting and value-added experiences in their academic program which will provide skills that are in high demand by major financial institutions of all types.
Learning Objectives:
· Ability to manipulate and assess large loan level databases for applied financial analytics
· Empirical estimation of embedded options in financial instruments
· Apply industry standard model validation techniques to econometric models
· Empirical modeling of asset price dynamics via stochastic processes
· Empirical modeling of term structure dynamics via stochastic processes
· Application of simulation techniques to financial instruments
· Generation of credit loss distributions for risky financial instruments
· Calibration of credit loss distribution using external ratings
· Pricing and valuation of complex structured credit instruments
· Interpret technical findings for nontechnical audiences
· Ability to develop and adhere to strict project plans for quantitative analysis
· Demonstrate ability to use SAS, R and Python for analysis
· Ability to conduct complex analytics in a team-based setting
Meeting Dates/Times
Course Website: https://myelms.umd.edu/
Location: VMH Rm. 1505
Tuesday and Thursday 12:00-1:50pm
Professor
Clifford Rossi
Office: 4465 Van Munching Hall
TA
Shrushti Mahesh Shah
Office Hours & COVID-19
By appointment on Zoom
Note also that due to social distancing requirements, I ask that you reach out to me during class with questions on lecture topics or follow up via email or request a Zoom meeting to discuss.
Team Reminders
I will send out an announcement before our course begins with team assignments. I am making these assignments on a random basis. These assignments will be final. Select one person from your team to be the point of contact with me and the GA during the course. For questions that arise during the course please have this individual contact me or the GA directly to be efficient in answering questions. Also, when emailing for questions please make sure all team members are copied on the email – unless it is of a personal nature. Each team must manage its project. This is a critical skill to learn as you will find in industry it is essential not just to build models but to ensure they are deployed in time.
COURSE AGENDA
Week 1 Course Overview, Data Development
Lectures (Tuesday 10/25 and Thursday 10/27)
· Project overview and objectives
· Walk through course Google drive
· Mortgage cash flows and embedded options (default and prepayment)
· Mechanics of credit risk transfer and examples (STACR, ACIS and CAS)
· Valuation model structure and key components
· Project data requirements and structure
Live FinEng Lab – Thursday 10/27 12-12:50
End of Week 1 Deliverables Due Tuesday 11/1
· Form project teams
· Create Sample Course Loan Level Credit Performance Data for Estimation of Survival Models – Representative of Recent Originations
Week 2 Data Analysis and Variable Transformation
Lectures (Tuesday 11/1 and Thursday 11/3) MAY NEED TO SCHEDULE 11/3 CLASS DUE TO FDA ADVISORY COMMITTEE MEETING
· Theoretical model considerations
o Borrower attributes
o Loan Attributes
o Property Attributes
o Macroeconomic Factors
· Specification Issues
o Linear vs nonlinear variables
§ Piecewise linear regression and spline techniques
o Interpretation of odds and hazard ratios
Live FinEng Lab – Thursday 11/3 12-12:50
End of Week 2 Deliverables Due Tuesday 11/8
· Create transformed variables for current LTV and relative median UPB
· Complete Univariate and Bivariate Analysis of Key Variables
Week 3 Survival Analysis of Embedded Options
Lectures (Tuesday 11/8 and Thursday 11/10)
· Modeling embedded default and prepayment options using survival techniques
o Proportional hazards models, assumptions and techniques
o Model structure and hazard math concepts
FinEng Lab – Thursday 11/10 12-12:50
End of Week 3 Deliverables Due Tuesday 11/15
· Complete survival model specifications of default and prepayment (SAS)
Week 4 Model Validation Techniques
Lectures (Tuesday 11/15 and Thursday 11/17)
· Review of model validation techniques and procedures
· Building holdout samples for model validations (pros and cons)
· Goodness-of-fit Diagnostics – Overall and by key cohorts
o KS
o AIC
o Error rate analysis
Live FinEng Lab – Thursday 11/17 12-12:50
End of Week 4 Deliverables Due Tuesday 11/22 (NO CLASS on 11/24)
· Provide model diagnostics on default and prepayment (SAS)
Week 5 Stochastic Process Analysis
Lectures (Tuesday 11/22 and Tuesday 11/29
· Stochastic process basics
· House price dynamics and estimation methods
· Term structure model dynamics and estimation methods
· Data sources for stochastic modeling
Live FinEng Lab – Tuesday 11/29 12-12:50
End of Week 5 Deliverables Due Thursday 12/1
· Conduct specification of term structure model (R)
· Conduct specification of house price model (R)
Week 6 Simulation Methods
Lectures (Thursday 12/1 and Tuesday 12/6)
· Mechanics of simulation of default and prepayment models leveraging terms structure and house price stochastic models
· Understanding the lognormal loss distribution and assumptions regarding loss severity
· Market-based calibration of loss distribution using external ratings loss subordination levels
Live FinEng Lab – Tuesday 12/6 12-12:50
End of Week 6 Deliverables Due Thursday 12/8
· Build simulation code and generate log normal credit loss distribution (Python)
· Assess reasonableness of loss distribution and calibrate to ratings loss subordination of recent CRT transaction
Week 7 Pricing Credit Structured Notes
Lectures (Thursday 12/8 and Tuesday 12/13)
· Anatomy of a GSE credit risk transfer security
· Discounted cash flow analysis and risk tranching of CRT
· Price and yield analysis and interpretation
Live FinEng Lab – Tuesday 12/13 12-12:50
End of Week 7 Deliverables Due Friday 12/16
· Generate CRT tranche prices and yields (Python)
· Final presentation and all code and results posted to your team Google drive by 12/16
· FINAL EXAM – Date TBD
Dedicated Course Google Drive
Once enrolled in the course, I will grant each student access to a Google Drive that has been set up for the teams to use. We will walk through its contents on our first class to acquaint everyone with it. The drive is set up with weekly folders containing subfolders of materials; Code (example software of the type you will be using – i.e., SAS, R and Python), Data (both GSE loan level and macroeconomic data) and Output (example of what you will be producing from your own analysis).
Team Google Drives
Each team will create their own Google Drive providing access only to the team and your professor. You will create a Weekly folder (Week 1, Week 2, etc.) that will be used for you to post your weekly deliverables for review and assessment by your professor. Please do this by the first week of class.
Course Dedicated Server
I have arranged with Smith IT to set up a dedicated server to perform your team analysis.
Some caveats that are extremely important before instructions on how to access the resource:
· This is a shared server resource, please be aware that other students may be working on their project when you are working on it as well
· There is a 5Tb storage device which is attached as the E: drive, labelled as Save Work Here. This is where you should save any data files or work. There are two subfolders, Private work and Shared work, you can create a folder with your name in Private work and save your data there, only you can see it, or if you need to share work with others you can create folders in the Shared work folder.
· There is limited storage space on the C: drive, if you save your work or files there then the drive may become full and the server will stop working. Please use the E: drive
· Software installed: SAS 9.4, R, Python and Office 2016
· As a shared server, do not attempt to change settings or software or install anything else, other people rely on this resource. If you need other items please request them.
· Server maintenance will be performed on a monthly basis as we do with all servers, server will be rebooted so save your work and log off prior to then.
· If you are finished working on the server for a period of time, please sign out (logoff) from the server, do not simply disconnect, your session will continue to run using system resources someone else can put to better use than maintaining an empty session. Disconnected sessions are monitored constantly and if abused then automatic timeout and logouts will need to be implemented.
· If you need technical assistance contact the Office of Smith IT Service Desk on 301 405 2269 or via [email protected], for project or classwork assistance please contact the TA and/or professor.
The resource has been published in vSmith and should show up shortly. Login to vsmith.umd.edu as you normally would and look for the ELP Desktop and launch that resource. The server will open in a new window (assuming you have installed the vSmith client previously, see it.rhsmith.umd.edu/vsmith for getting started information).
Each student will have their assigned server available at vsmith.umd.edu. Once logged in they can search for ELP in the search bar at the top orclick on catalog and scroll down to ELP Desktop. If this is their firsttime using vSmith they can find a short how to video here:
https://www.youtube.com/watch?v=yJw7NJ8W5ho
Reading Materials
Survival Analysis Using SAS, A Practical Guide, Paul D. Allison
Federal Housing Finance Agency, Model Risk Management Guidance, Number: AB 2013-07, 11/20/2013
2017 Federal Housing Administration Actuary Report
Freddie Mac Credit Risk Transfer Investor Presentation
Fannie Mae CRT Investor Presentation
Various research studies and articles on mortgage default modeling, simulation and valuation to be assigned at the time of class
Resources
Programming Basics
SAS – SAS contains an enormous online help library on every procedure (PROC) statement used as well as data and modeling tutorials and examples. Online help is available directly through the software or at the SAS website – one of the many useful videos available for starting to program with SAS is below.
R – As an open sources software application, it like Python has an extensive following with extensive resources available online to learn R programming and techniques. The R-Project site in particular https://www.r-project.org/ provides a variety of materials to R users. A basic starting primer on learning R can be downloaded at the following link.
https://cran.r-project.org/doc/manuals/r-release/R-intro.pdf
Likewise, there are many web-based tutorials as well on learning R.
Python – As in the case for R, Python has a large user following and a useful site for learning more about Python is found at https://www.python.org/. For those not exposed to Python in their coursework, a handy reference for learning how to program in Python is found at the following link: https://www.python.org/about/gettingstarted/
Student Background
BUFN 610 Financial Management is a prerequisite for taking the course. However, ideally, MQF and MFin students will have had courses in Financial Mathematics, Financial Programming and Econometrics prior to attending this course. MFin students may take Financial Engineering, however, it is highly desirable that they have attained comparable mathematical and econometric knowledge prior to taking Financial Engineering. Other desirable courses include Fixed-Income, and Fixed-Income Derivatives.
Grading
Grades will be assigned on the performance of your team. This is in keeping with industry practices that emphasize the importance of effective team relationships in achieving important strategic and business objectives. Also in accordance with industry practice of 360-degree feedback on individual performance, in addition to a faculty grade (weighted 60%), each team member will provide a final grade (anonymously – the average of team member grades 15%) for each team member other than themselves. Final grades will be allocated as follows:
Percent of Final Grade
· Team Collective Grade
o 15% from your teammate anonymous survey at the end of the course
o Final Project - faculty portion of the grade
§ 45% of the grade will be based on the substance of your work, data, modeling, and application to the credit risk transfer security. That includes all developed code, data, and output from deliverables through the term.
§ 15% of the grade will be based on the quality, clarity, comprehensiveness, and accuracy of the final presentation.
· Individual (i.e., not team-based) Final Exam – 25%
The final examination is an individual exam that will be provided in class for a duration of 2 hours covering materials discussed in class.
ACADEMIC INTEGRITY
It is essential that all students rigorously follow the University’s policies on academic integrity.
The University's Code of Academic Integrity is designed to ensure that the principles of academic honesty and integrity are upheld. All students are expected to adhere to this Code. The Smith School does not tolerate academic dishonesty. All acts of academic dishonesty will be dealt with in accordance with the provisions of this code. Please visit the following website for more information on the University's Code of Academic Integrity:
http://www.inform.umd.edu/CampusInfo/Departments/JPO/AcInteg/code_acinteg2a.html
On your final presentation and any other course materials provided, you and your team will be asked to write out and sign the following pledge. "I pledge on my honor that I have not given or received any unauthorized assistance on this exam/assignment." In terms of acceptable modes of learning, please take note that you and your team may not use any other course materials including code, data and output without express permission from your instructor. You may only work with your individual teams with respect to project deliverables. Violations will be strictly enforced consistent with university policy.
Special Needs
Any student with special needs should bring this to the attention of the instructor as soon as possible, but not later than the second week of class. This includes any prior arrangements regarding examinations. Make-up examinations will only be allowed with proper advance notice to the instructor and only in family emergency situations or illness. Schedule conflicts with other courses or non-academic meetings will not be grounds for rescheduling exam dates.
2022-12-15