Question about a Vanguard study I read (DCA vs LSI)

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
hudson4351
Posts: 307
Joined: Fri Aug 10, 2007 8:30 pm

Question about a Vanguard study I read (DCA vs LSI)

Post by hudson4351 » Fri Nov 14, 2014 3:07 pm

I was reading a paper Vanguard published titled "Dollar-cost averaging just means taking risk later":

https://pressroom.vanguard.com/content/ ... raging.pdf

If you scroll down to Figure 3, it says:
The figures represent rankings of LSI balances subtracted from DCA balances. At the 5th-percentile level, for example, 5% of the resulting amounts were lower than the amount listed and 95% were higher.
If I am reading that correctly, then positive values in Figure 3 should correspond to cases in which DCA outperformed LSI, and negative values should correspond to cases in which LSI outperformed DCA, because, according to my reasoning:

Code: Select all

DCA - LSI > 0 when DCA > LSI
DCA - LSI < 0 when DCA < LSI
However, the paper also says:
The 50th percentile observation is positive (confirming LSI's average outperformance)
Based on the caption they included for Figure 3 and that I included above, a positive value in the table should correspond to cases in which DCA outperformed LSI, because DCA - LSI > 0 when DCA > LSI. Yet this contradicts the second statement I quoted that says "The 50th percentile observation is positive (confirming LSI's average outperformance)".

Is the caption that goes with the table a typo?

User avatar
ogd
Posts: 4854
Joined: Thu Jun 14, 2012 11:43 pm

Re: Question about a Vanguard study I read (DCA vs LSI)

Post by ogd » Fri Nov 14, 2014 3:14 pm

The caption is indeed wrong, the rest of the article makes this clear. The "subtracted from" should read "minus" or "after subtracting". It's possible that an editor changed the language before publishing for stylistic reasons, without realizing that it implied a sign reversal.

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Question about a Vanguard study I read (DCA vs LSI)

Post by Leeraar » Fri Nov 14, 2014 4:24 pm

Thank you for that.
LSI has provided better returns and risk-adjusted returns, on average.
Repeat: LSI has better risk-adjusted returns.

(I had seen that paper, but not noticed the Sharpe Ratio calculation towards the end.)

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

hudson4351
Posts: 307
Joined: Fri Aug 10, 2007 8:30 pm

Re: Question about a Vanguard study I read (DCA vs LSI)

Post by hudson4351 » Sat Nov 15, 2014 12:29 pm

Thanks for confirming what I found. I just sent the message to research@vanguard.com. Maybe they will issue a corrected version of the paper.

Post Reply