|
|
|
|
EMMA Program was down 0.9% in August and up 4.8% YTD.
Another dull month with subdued volatility. Investors' focus this month was on NAFTA and emerging market woes as a whole. While North American trade tension eased, Turkey, India, Argentina, Indonesia, Brazil, South Africa have all seen their currencies plunge to multi-year/historic low. US dollar also gained on euro, pound, aussie, and consolidated against the yen. The strength of the dollar and the weakness of developing economies (many of them commodity exporters) are putting a lot of pressure on commodity prices. Market still does not believe the Fed's rate hike schedule but what's keeping the yield down seems to be policy divergence from Europe and Japan.
We noted last month that the US stockmarket is acting like a safe haven amidst the storm brewing all around the globe. In our new HFM-CTA Intelligence article we argued that volatility is poise to return soon when reluctant buyers stop hedging. However, we are wary of the Republicans' willingness to keep the #MAGA party going. Besides complaining about monetary tightening, the Republicans' only hope to goose the Trump economy ahead of the midterm election is fiscal stimulation. Another tax cut is likely coming.
Switching subject and delving deeper into our methodology based on the three-legged stool: (F)undamentals, (S)entiment, and (T)echnical. Their interplay can be explained by the concept of mathematical function x -> f(x).
Most of the time FST interact and form complex feedback loops. When an event , say a Trump tweet, mine strike, or inflation headline happens, we don't (just) attempt to get a precise output on the superficial functions. Rather we update our beliefs and we aim to be accurate on the direction of travel, trade sizes etc by having the right algorithm.
Naïve investors typically place bets based solely on observing x, or formulating f(x), or blindly following price/trend signals generated by g(f(x)). Successful trading goes deeper. Function h needs to consider timing, macro factors, structural issues e.g. whether trend will continue, shifting correlations, and specific portfolio construction and risk management considerations. Trading is not about so much about accurately predicting future events; it is about modeling scenarios (providing structure) and continually adjusting weights of parameters as well as posterior probabilities and expectations. With enough data, machine learning methods utilizing neural network (weight learning) or Bayesian network (probabilistic learning) are indeed suitable for this type of problems. Rules-based induction learning or analogical reasoning, not so much. This is our ongoing research subject. Stay tuned.
Given the recent choppiness due to low volatility and low-ish correlations, chasing returns requires high leverage and frequent position switching/tuning. We scaled back our positions and adopted a wait-n-see approach.
I look forward to seeing many of you at the CTA Expo in Chicago next week.
|
|
Lastly, EMMA program is currently open to new investors. Please let us know if you'd like more information.
Regards,
Jeff Lee, CTA
|
|
|
|
|
|
|
|
