ORIA (Orthogonal Risk Integrated Alpha) Explanation Part 2 - Joshua Kyan Aalampour — Transcript

Full transcript

1. 0:00Hello everybody, this is part two of
2. 0:01explaining Oria, which is my original
3. 0:03quantitative architecture. It stands for
4. 0:05orthogonal risk integrated alpha. In
5. 0:07part one, we last left off with the
6. 0:09orthogonalization step. So okay, now
7. 0:11we're done with that, the allocator
8. 0:13takes over. So we start from a
9. 0:14post-close market state vector. So I'll
10. 0:17go ahead and list some example
11. 0:18conditions here. We might have some
12. 0:19things like volatility, dispersion,
13. 0:21correlation, breadth, trend, slippage,
14. 0:23etc. You know, stuff like that. And then
15. 0:25for each strategy, we want to form a net
16. 0:27utility like quantity. So how are we
17. 0:29going to do this? We're going to
18. 0:30penalize uncertainty and expected
19. 0:32implementation cost.
20. 0:34>> [music]
21. 0:34>> So it's going to look something like
22. 0:35this. And so what does this mean? So
23. 0:37this over here would be like our
24. 0:39estimated edge. And then this over here,
25. 0:42this is
26. 0:43like our uncertainty
27. 0:45>> [music]
28. 0:45>> penalty. And then this over here would
29. 0:47be like our expected trading cost.
30. 0:50[music] So we have the estimated edge
31. 0:52minus the uncertainty penalty minus
32. 0:54expected friction. That's basically what
33. 0:56we're dealing with here. And then we
34. 0:58have the allocator do its thing. So over
35. 1:00here we have the sleeve utility. This is
36. 1:02the overlap penalty. And then this is
37. 1:04the turnover penalty. So the goal of
38. 1:06this is that if one strategy looks
39. 1:08attractive, but it overlaps in some way
40. 1:10with the others or for example, if
41. 1:12reallocating into it might cause too
42. 1:14much churn, then that strategy gets
43. 1:16penalized. Okay, there's another
44. 1:18scenario we have to take into account,
45. 1:19which is, you know, what if there isn't
46. 1:21enough clean opportunity for any of the
47. 1:24strategies right at a given time. You
48. 1:26know, that is possible.
49. 1:27And the answer is pretty simple. It is
50. 1:29just [music]
51. 1:30cold hard cash. Now the allocator also
52. 1:32sets the total portfolio risk. So this
53. 1:35right here is the predicted risk of the
54. 1:37sleeve blend.
55. 1:38>> [music]
56. 1:38>> And then this part accounts for
57. 1:39volatility targeting. And this is our
58. 1:41stress penalty. So essentially what we
59. 1:43have so far is a stack that has properly
60. 1:45calculated sleeve budgets, cash, and a
61. 1:48total risk scalar. Okay, I think that's
62. 1:51where I'm going to end part two. Cuz
63. 1:53again, I don't want to make these too
64. 1:54long. But just to recap, we talked about
65. 1:57the allocator in this part.
66. 1:59The next part I want to talk about the
67. 2:00risk box and execution governor. So
68. 2:03I hope to see you there.