Andrew Lo readings

[ravishankar80]

Hi David,

I tried going through the core reading material for “What happened to Quant’s”. I found it quiet complicated. I see in your notes that you have dealt with it in 4 pages. So in there unnecessary verbiage on the part of the authors? I find all of Andrew Lo’s reading intense and time-consuming. I know you always insist on reading core material but in this case finding it tough going? Wanted to understand what is the best way to approach this? In fact I find the whole of investment management lot tougher than any of the other sections.

Ravi

 

[David Harper CFA FRM]

Hi Ravi,

I agree these are time-consuming, difficult, and, yes, I do think they contain some excess material that is not critical for our practical purposes. Here are my thoughts on the Lo readings:

* Did you see/review my episode #13 slides from yesterday, they should give the headlines?

* IMO, the hedge fund replication reading is more important to us (to you) than the "what happened to Quants." Did you see yesterday's newsletter about the key theme? The first thing you should get very comfortable with is the common factor model. Specifically, the concepts of (i) how he extracted factors (by regression), what those mean, and (ii) the idea of creating a linear clone based on common factors. This idea is closely related to Grinold's factor analysis, so it is essential. I believe this is the first stumbling block. But I also believe once you see the linear clones as factor models, you can *mostly* see the rest of this 1st Lo paper as "merely" a catalog of different strategies and their regression results. Here is the acid test of such an understanding, do you understand the table 7 that shows the percentage contribution of each factor and alpha to the strategy's return. I would say your primary goal in this paper is to understand this table (Table 7). Then you really only have the rolling window + liquidity AIMs. For example, what does a high alpha suggest vis a vis a clone? What does a high factor exposure suggest vis a vis a linear clone?

* In regard to the August 2007 paper, I think you only really need a few essentials here (I am of this opinion in part b/c the paper is not conclusive, Lo told a CFA audience while presenting the paper that he isn't after all sure what happened !?). See the AIMs (on my slides). My suggesting here is to start with the limited number of AIMs (i.e., the key themes) and mine the paper for those; i.e., rather than try to process the paper bottom-up, just "go fishing" from the top-down for the key ideas we care about, namely:

1. the unwind hypothesis - if you can state that, you pick up his key ideas: connectedness (betas on the rise), lack of liquidity, and leverage
2. liquidity as measured by autocorrelations (I feel this may be tested, it is Lo's signature finding. I added slide 54 b/c I am betting that "autocorrelation as illiquidity" will find itself on a test question). Especially given events, I expect liquidity to be tested. See my slide 61: Lo finds a significant decrease in liquidity and, here is a relevant theme, he finds liquidity inextricable linked to credit risk.
3. His four conclusions (see slides 64 & 65).

David