Investigation of 1998 'Russian Contagion' for the Hungary-Germany Pair of Interest Rates in a Reduced-Form Model

In this paper I use a reduced-form, tree-based model to fit the term structure for the AAA-rated (German) and risky (Hungarian; spread 120-140 bp) bonds in the 2-month period immediately preceding Aug. 1998 Russian default. The term structure was optimized by the stochastic, derivatives-free algorithm with accuracy typically better than 1% for risk-free and 1.5% for risky rates for all maturities. This study was motivated by the desire to observe contagion, because the default was such a momentous event for emerging-market finance, far exceeding the significance of Russian economy for the world markets in 1998. Both Germany and Hungary at the time were heavily exposed to the Russian market, Germany, in particular, as Russia's largest creditor. Moreover, the dynamics of Hungarian interest rates was closely following German rates (the coefficient of correlation ≥ 0.8). Despite an obvious strain on Russian finances, which was obvious to German holders of Russian debt well before August, we do not observe any particular structure in the pre-default hazard function which allows us to speak of contagion. Shortened version of this paper under the title "Contagion hypothesis and the risk of the sovereign bond portfolios" has been published in the book "Financial Risks: New Developments in Structured Finance & Credit Derivatives", edited by C. Gourieroux and M. Jeanblanc, Economica--Elsevier, 2009