Download presentation

Presentation is loading. Please wait.

1
**LTCM’s Analysis of Risk Management**

February 28, 2002 Frank Burke Larry Kissko Gurkan Salk Heather King

2
**Agenda LTCM Background Swap Spread Trading Strategy Project Analysis**

Comparison/measurement of LTCM’s Risk Assessment Discussion on return and spread distribution, calculated implied std deviation Estimate of LTCM’s Value-At-Risk Proxy Tests Take-aways

3
LTCM Background August 21, 1998, fund lost $550m mostly from swaps spreads and equity volatility bets. LTCM believed this event would occur 1 in every 800 trillion years (or an 8.3 std dev move). Swap spreads shot up from 60 bps to 80 bps intraday vs. an average daily move of 2 bps LTCM’s swap position represented 2.4% of global swap market in December 1997 Leverage ratios varied from 28:1 to a high of 55:1 in late 1998

4
**LTCM Trading Strategy We focused on of one of LTCM’s biggest trades:**

Swap Spread Relative Value Trade Swap spread – difference between the fixed rate on a fixed-for-floating swap and the yield on a coupon-bearing Treasury bond of comparable maturity Speculative strategy that spread would converge to its historical mean Long swap/short the treasuries (in 1998) Crisis: Aug 21, spreads spiked 21 bps intra-day

5
**Swap Spread Frequency: “the bet”**

6
**Project Analysis Parametric VAR – assumes normal distribution**

Historical VAR – based on actual data distribution Proxy search – difficult to find a strong correlation BAA- 10 year treasury AAA- 10 year treasury MBS - 10 year treasury Forecasted daily variance Value At Risk – defined as the expected maximum loss over a target horizon within a given confidence interval

7
**Swap Returns Distribution (thru 7/98)**

8
**.16% = 4 observations over 10 year period**

Analytic Results Risk analysis LTCM Satchmo Return distribution Normal Curve Non-normal: w/Kurtosis & fat tails 99.7% confidence interval [- 6.07%, %] from the mean return 0.01% [ %, %] from the mean return 0.01% Implied Daily Std. deviation 2.03% 3.46% Value at Risk (VAR) - estimated $60M $95.2M Probability of Aug 21 event 10-13 Or % .16% = 4 observations over 10 year period

9
**Value at Risk (VAR) Principal measure of risk at LTCM**

LTCM parametric VAR measure Capital (assume $1b) x daily std dev of returns (.02) x std dev of required confidence interval (3 = 99.85% 1-tail) $1.0b x 2% x 3 = $60,000,000 Our historical VAR measure $ 1.0b x % = $95,238,000

10
Take-Away Thoughts VAR not necessarily suspect – correct inputs are critical Cannot blindly apply normal distribution Dig into your data If data is not complete consider: Developing a risk proxy Assuming fatter tails in distribution (Student’s T curve)

11
Appendix - charts August 21, 2002

12
Appendix - charts

13
Appendix - charts

14
Appendix - charts

15
Appendix - charts

16
**References Jorion, P., 2000 Risk Management Lessons from LTCM**

Kolman, Joe, 1999, “LTCM Speaks”, Derivatives Strategy (April) p.12-17 Lewis, Michael, 1999, “How the Egg-Heads Cracked” New York Times Magazine, January 24, p 24-77 Anonymous, 1998, “Too Clever By Half”, The Economist Magazine, November 14 Whaley, Robert, 2001, “Derivatives” Class Presentation Scholes, Myron, 2000, “Crisis and Risk Management- The Near Crash of 1998”, AEA Papers and Proceedings Vol 90 No. 2, May. Bloomberg – Swap spread data

Similar presentations

OK

Options, Futures, and Other Derivatives, 5th edition © 2002 by John C. Hull 16.1 Value at Risk Chapter 16.

Options, Futures, and Other Derivatives, 5th edition © 2002 by John C. Hull 16.1 Value at Risk Chapter 16.

© 2018 SlidePlayer.com Inc.

All rights reserved.

Ads by Google

Ppt on slum areas in india Ppt on bionics Ppt on lunar and solar eclipse Free ppt on mobile number portability pdf Ppt on product specification meaning Circulatory system anatomy and physiology ppt on cells Ppt on employee provident fund act 1952 presidential election Ppt on series and parallel circuits diagram Ppt on transportation in india Ppt on political parties and electoral process