Fractal Indonesia

When I lived in New York, every other night I'd have a drink with our head of financial engineering. Beer in the winter, G&Ts in the summer.

In any case once he was telling me how in Japan even ordering coffee was tricky, because the Japanese didn't want to cause offence by not knowing English well, and he being somewhat familiar of Japanese culture, wanted to be polite as possible and not cause undue embarrassment. A veritable minefield.

Never missing a chance to talk about Indonesia, I told him about the time when I asked a waitress for a coffee in my terrible Indonesian and she literally started rolling on the floor laughing.

Ironically, I suspect Indonesia's as hard as any country to really get to know well, hundreds of cultures make it fractal-like; but you can have a lot of fun while trying!

In any case, I was recently staring at the Nikkei, unable to make heads nor tails of it when I turned to the JKSE for light relief and instantly found that using a Hurst Exponent strategy is very profitable.

The Hurst exponent is analogous to autocorrelation, but uses techniques derived from fractal geometry.

The strategy predicts upturns and downturns for the next day, weights those predictions and goes long or short accordingly.

The Sharpe Trajectories remove superficial differences between the strategies and focus solely on the efficiency of holding the JKSE or applying the Hurst Exponent to the JKSE.

Investing in the JKSE over almost 20 years returns a nice 11% pa. However the Sharpe is just 0.42. Volatility is a killer.

Or in Sharpe Trajectory terms - it's inefficient.

Applying the Hurst Exponent strategy boosts efficiency.

The Sharpe becomes 2.2.

Max Wait to profitability halves from 4 years to 2.

The large negative skew of -0.4 (lots of nasty surprises) is transformed to +1.7 (no nasty surprises!).

You can see that during 2008 the JKSE drops like a rock whereas Hurst cashes in on the chaos.

This is further underlined when there was rape and pillage on the streets of Jakarta in the late '90s - the hell which gave birth to the current democracy and peace. The JKSE plummets but Hurst stays steady.

Two provisos. I didn't take into account FX rates (we are taking the view of a domestic Indonesian investor) and I didn't apply a risk free rate. Couldn't find BI rates (sure to be large!) and felt US rates are not applicable.

Nevertheless, good to show the opportunities available in more exotic markets.

Code for the Lazy Backtesting IDE is here.