Monitoring Risk While Pursuing High Returns: The Importance of Being Quantitative
“A quantitative model provides a very different insight than a rating,” said Rajan Singenellore, Global Business Manager of Risk & Valuations at Bloomberg. Models are based on objective inputs and usually publicly available information, whereas ratings are based on subjective factors and inside information. Singenellore was speaking about quantitative models as the second part of a two-part webinar presentation “Monitoring Risk While Pursuing High Returns” on March 7, 2013 organized by the Global Association of Risk Professionals (GARP). “A quantitative model lets you understand risk drivers and sensitivities,” he said, such as the effects of input changes on default risk […]
Managing Model Risk: Part 3. Collective Hubris
There are more things in heaven and earth, Horatio, Than are dreamt of in your philosophy. – Wm. Shakespeare, Hamlet, Act 1, Scene 5. “We are building the language in which to discuss model risk,” said Boris Deychman, Director of Model, Market and Operational Risk Management at RBS Citizens Financial Group. He drew an analogy with the world of wine experts, who have developed specific vocabulary to talk about aroma and taste. “They don’t say just: this tastes like wine.” Deychman was the third and final speaker of a panel invited by the Global Association of Risk Professionals to discuss […]
Managing Model Risk: Part 2. The Impact of Basel III
“More consideration must be made for regulatory capital changes that focus on small but potentially devastating risks or risk factors,” said Peter Went, VP, Banking Risk Management Programs at GARP. He was speaking about Basel III and its impact on modelling regulatory capital at a webinar on January 29, 2013, organized by the Global Association of Risk Professionals to discuss model risk management. Went mentioned three developments in the Basel III that directly impact capital models and model risk in regulatory capital determination. First, Basel III introduces a risk-invariant leverage ratio of 3 percent in recognition that models may be […]
Managing Model Risk: Part 1. “Models are Always Wrong”
“Models are always wrong,” said Joe Pimbley, Principal at Maxwell Consulting, via webinar on January 29, 2013. He was the first of a panel of three speakers invited by the Global Association of Risk Professionals to discuss model risk, what it is, and how to assess and validate it after the financial crisis. Models are always wrong, Pimbley clarified, because they are only simplifications. “Wrongness” he defined as some type of error or omission that materially impacts the results as understood by the user. A model can be wrong because the meaning of the result differs from that understood by […]
MATLAB for Excel Users: “Discoverability”
When looking for good data analysis tools, many financial professionals turn first to MS Excel which displays numeric data, contains advanced functions, and can be programmed with Visual Basic. A growing number of professionals are turning to MATLAB which has strengths complementary to ordinary spreadsheets. At a webinar on January 15, 2013, Adam Filion, application engineer at MathWorks, showed approximately three hundred audience members how easily features of MATLAB software adapt themselves to rapid analysis of large datasets. In Excel functions, the input is specified as a cell location and the math is hidden. In MATLAB, the interface has a […]
Analytical Tools to Gain Insight and Speed Up Numerical Analysis
“Symbolic computing can be a practical part of the solution to your problem,” said Deepak Ramaswamy, technical marketing manager at Mathworks. On January 8, 2013 he showed about three hundred participants via webinar how they could switch between analytical tools in a Notebook app to numerical analysis of the same problem in the MATLAB interface. He stepped his way through one “classic” and two “real-life” problems: the damped oscillator model; fuel consumption of a rocket-powered car; and kinematics of a double-jointed robot arm. The three problems can be written as systems of Ordinary Differential Equations (ODEs) and this was the […]
Mining Microeconomics Using MATLAB
Modelling the economics of an iron ore mine “is a complex task that can be made more reliable,” said David Willingham to a webinar audience on December 5, 2012. Willingham, an application engineer at Mathworks, was demonstrating how a typical mine’s economics could be modelled using MATLAB and then embedded within an Excel spreadsheet. Developing a mine involves significant capital expenditures and long time frames. Willingham aimed to take the audience through a good model that would take into account the microeconomics of a particular mining company, integrated with the macroeconomic environment, such as interest rates and iron ore prices. […]
Contingent Capital: The Case for COERCs. Part 2.
“Many people have misgivings about [contingent convertible bonds] because they just don’t know how to value them,” said George Pennacchi, Professor of Finance at University of Illinois. He was the second speaker at the November 29, 2012 GARP webinar on the subject of Call Option Enhanced Reverse Convertible (COERC) bonds. Click here to go to Part 1. Pennacchi, along with Theo Vermaelen and Christian Wolff, co-authored a recent paper proposing a new type of cocobond. [Contingent convertible bonds, or “cocobonds,” are bonds that convert into equity when the market value of capital falls below a trigger level.] “The paper provides a […]
Develop & Deploy Financial Models
“Focus on modelling, not programming,” urged Ameya Deoras, senior applications engineer at MathWorks. He was speaking during a webinar on December 3, 2012 about the use of MATLAB in the construction of financial models. Deoras’ talk covered four examples to varying depth during the hour. The first example, the calculation of the efficient frontier for large-cap stocks, allowed him to show the easy data importation from an ODBC-compliant database. Each step of the way he showed how the input could be visualized with a single click. If the data exist in a relational database (think tables and fields such as […]
Tree Bagger & Tree Booster: MatLab for Data-Driven Fitting
Let’s say you want to create a predictive model without assuming an analytical form to the model. How would you go about it? On August 14, 2012, Richard Willey, Technical Marketing Manager at MathWorks, demonstrated via webinar how input data could be fit using machine-learning approaches. The emphasis here is data-driven, as opposed to model-driven, fitting. “A limitation of regression techniques is that the user must specify a functional form,” said Willey, and the choice of that model is usually based on the domain model. Typically the data points are fit with high-order polynomials or Fourier series. Or, the user might run the data […]