Financial Econometrics and Empirical Market Microstructure

Japan Case Study

We’ll use a recent Japan stress test case study to illustrate how this would work. As a consequence of massive quantitative easing, Japanese equities and bonds entered an exceptional bull market from 2012 to early 2013. By April 4, 2013 Japan was entering bubble territory with JGB yields reaching a historical low at 35 bp and with equities up 80 % over the last year. JGB downside (price) outliers then started escalating as Gold crashed on April 12 and 15. Then on May 23, after the Fed announced potential tapering of QE and after a lower than expected China PPP announcement, Japanese equities dropped by 7.32 % in 1 day. This was a classic early warning signal—after a long ebullient run tied to low interest rates, risk finally returned. This analysis was noted in our PRMIA Emerging Stress Scenarios community on NextThought. com, and we contemplated the potential repercussions of a Japan meltdown on global markets.

The visualization (Fig. 20) shows how a 10 % daily drop in the Japanese Stock Index (EWJ) would likely affect other asset classes, using partial correlation analysis. As opposed to using current data, we applied a historical stressed period during the GFC (a month period starting 4 May, 2008).

Statistically significant partial correlations are shown as links between the nodes (ETFs). Link widths denote the strength of dependence. Node sizes scale with the predicted 1-day return on the day of the stress event. Node color denotes positive (green) or negative (red) returns.

This analysis shows three layers of relatively weak connection between Japan and other asset classes. Japan only shows a moderate primary link to EAFE (logical, given its 21 % weighting in the index). So it was not surprising that as Japan stocks continued to slide by over 10 %, EAFA dropped by just under 5 %, while broader markets hardly reacted.