John

1,626 reputation

312

bio	website
	location
	age
visits	member for	1 year
	seen	2 days ago
stats	profile views	153

	1,626 reputation	bio	website		visits	member for	1 year
312 badges		location			seen	2 days ago

summary answers questions tags badges favorites bounties reputation activity

146 Comments

suggestions reviews revisions comments badges posts accepts all

Sep 28	comment	Comparing MVO with Resampled Efficient Frontier I really like that last point.
Sep 28	comment	Comparing MVO with Resampled Efficient Frontier If you want to just compare the frontiers, why not just plot both of them?
Sep 27	comment	S&P 500 P/E percentile What do you mean Bloomberg accounted for it, but you didn't. If you select the S&P500 composition today and then get their P/E on 5/27/2007 and calculate 5 year return of them, then some of the stocks that are in the S&P500 today may not have existed five years ago (like a spin-off). Alternately, if you took the S&P500 as it existed on 5/27/2007, sorted by top percentile and calculated the return, then companies like Lehman may no longer exist and not have five year returns.
Sep 27	comment	Fastest solver possible for portfolio optimization I take it you mean you want to do a backtest that looks over the past n years using 30 different strategies. You might want to look to parallel processing. That tends to work well in this sort of situation.
Sep 27	comment	S&P 500 P/E percentile Does Bloomberg account for the changing composition of the S&P500?
Sep 26	comment	Average beta of index consitutents w.r.t. the index is 0.60 Just saying they should be consistent, but I'm not 100% sure it would be a big impact.
Sep 26	comment	Average beta of index consitutents w.r.t. the index is 0.60 Just saying you should be consistent. Total return vs. total return or price versus price.
Sep 26	comment	Average beta of index consitutents w.r.t. the index is 0.60 I'm not sure it would make a big difference (since the total return series should be highly correlated with the price series), but are you sure the index is total return if you're using the total return of the individual stocks? Outside of that I'd have to actually look at the data and code to have an idea. No other idea.
Sep 25	comment	Average beta of index consitutents w.r.t. the index is 0.60 How about using the market capitalization weights you have and calculating the returns on this portfolio. Then plot and compare and calculate the beta. My guess is that you have some kind of data problem, like the index being in a different currency or something.
Sep 22	comment	Average beta of index consitutents w.r.t. the index is 0.60 You might try calculating the weighted version as well. Another alternative, if you don't have market caps, is to calculate the return on the equally weighted portfolio of stocks you do have and then calculate the betas with respect to those returns.
Sep 21	comment	Is inverted Japanese style curve persistent when negative rates are real / market - observed? To be honest, I still don't understand what you're asking. I doubt you'll see a significant yield curve inversion when the short-term yield is kept at 0%. I just pulled on the Japanese curve on BB and saw inversion at the end of 1990 when rates were high, but nothing recent.
Sep 21	comment	Is inverted Japanese style curve persistent when negative rates are real / market - observed? The reason no one has answered this question is that it contains too many extraneous details. You need to simplify it.
Sep 20	comment	Discrete returns versus log returns of assets If you take normally distributed log returns and convert them to arithmetic, then they will become log normal. That's what I mean by estimating distributions easier. Also, it is easier to project log returns to the appropriate horizon due to time aggregation. As for invariance, see: symmys.com/node/85
Sep 20	comment	Discrete returns versus log returns of assets I'm not sure there needs to be a "study." You seem well aware of the reasoning. Arithmetic returns allow for easier cross-sectional aggregation and log returns allow for easier time-aggregation. The reason people use log returns is that (for equities) they are approximately invariant and are easier to work with in estimating distributions. However, proper procedure is to convert the log returns to arithmetic returns for the purposes of portfolio optimization and risk management.
Sep 19	comment	Unsystematic and systematic risk of a portfolio symmys.com/node/196
Sep 18	comment	Markowitz mean-variance optimization as “error maximization” @QuantGuy That's gated for me, but I think I recall similar complaints. Personally, I don't use the resampling technique.
Sep 18	comment	Markowitz mean-variance optimization as “error maximization” Well imagine that there is one element in the covariance matrix, so the inverse is one over the variance of the asset. If the standard deviation is 20%, the inverse of the variance is 25. This is why small changes in the means matter.
Sep 17	comment	Markowitz mean-variance optimization as “error maximization” I think of Michaud's resampling as an application of Bayesian techniques. Under Black-Litterman/Entropy Pooling framework, if you take views on every asset and perform unconstrained optimization with equal confidence in the views, then the optimal portfolio is an average of the individual portfolios if you had full confidence in them. You could come up with arbitrary views by sampling from $\mu_{r}\sim N\left(\mu,\frac{\Sigma}{T}\right)$ where $T$ is the number of observations. More observations, less estimation error.
Sep 10	comment	Can Hurst exponent be used to characterize nonlinear dependence in time series? The Hurst exponent can be used to distinguish between an AR(1) process and an ARIFMA process. Dependence is more often used contemporaneously. For instance, a t copula could capture the non-linear dependence between two time series. Presumably you could use some sort of t autocopula to do something similar, but I'm skeptical to what benefits you could get from doing that.
Sep 7	comment	Does Ito/Malliavin calculus have any applications helpful for direction based trading? When you say direction-based trading, do you mean like longer-term and not high frequency trading? Ito calculus (not familiar with the other) is very important for options pricing. If the strategy you're looking at uses options, then it would be very important to know, but it depends a lot on what you're doing.