Value Averaging, IRR Bias/Misconception: Have we been misled?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
AndroAsc
Posts: 1127
Joined: Sat Nov 21, 2009 7:39 am

Value Averaging, IRR Bias/Misconception: Have we been misled?

Post by AndroAsc » Tue Jan 24, 2017 2:11 pm

I've been doing Value Averaging since I started doing investments over a decade ago.

I like VA for two reasons vs conventional DCA + rebalancing:
1) I can save towards a final "retirement number" which I can control.
2) I like the idea of methodically "buying low and selling high", because it should increase overall returns.

To the best of my knowledge, VA has a downside of having too much "cash drag" especially if your projected returns are too far off from the actual portfolio returns. This means potentially lower performance at retirement vs DCA + rebalancing, but I am okay with that.

More recently, I came across an article published in 2013:
https://papers.ssrn.com/sol3/papers.cfm ... id=1606347

The main claim is that IRR is not a "fair" metric to make the claim that VA can outperform other mechanical investment strategies. It is also true that almost all prior VA research has used IRR to demonstrate its superiority, but the point here is higher IRR does not equate to more money at retirement.

This to me is quite concerning, as it strikes at the foundation of reason (2). If higher IRR is not going to translate to more money at retirement, then that's a big concern for me.

For those who do Value Averaging. Are you aware of this? What are your thoughts?

To keep discussion on point, please refrain on comments on "I don't do VA, I do DCA, and I don't believe in VA". I would like to focus this discussion on whether we have been misled by earlier VA papers that used IRR as a proxy measurement of the superiority of their technique. And if yes, how much of a concern is it going to be?

AndroAsc
Posts: 1127
Joined: Sat Nov 21, 2009 7:39 am

Re: Value Averaging, IRR Bias/Misconception: Have we been misled?

Post by AndroAsc » Sun Oct 14, 2018 2:41 pm

No discussion for over a year, I thought I would bump this out. I'm really curious to understand if it's true a higher IRR will not translate into a higher end value. Do we have any Bogleheads doing VA here?

MotoTrojan
Posts: 2469
Joined: Wed Feb 01, 2017 8:39 pm

Re: Value Averaging, IRR Bias/Misconception: Have we been misled?

Post by MotoTrojan » Sun Oct 14, 2018 5:10 pm

Hadn’t heard about this. Enjoyed reading the top level description.

If you’re at $100K and inherit $2M, do you reset your target or sit on 95% cash for years?

User avatar
raven15
Posts: 344
Joined: Sun Nov 30, 2014 8:01 pm

Re: Value Averaging, IRR Bias/Misconception: Have we been misled?

Post by raven15 » Sun Oct 28, 2018 2:23 am

AndroAsc wrote:
Tue Jan 24, 2017 2:11 pm
I've been doing Value Averaging since I started doing investments over a decade ago.

I like VA for two reasons vs conventional DCA + rebalancing:
1) I can save towards a final "retirement number" which I can control.
2) I like the idea of methodically "buying low and selling high", because it should increase overall returns.

To the best of my knowledge, VA has a downside of having too much "cash drag" especially if your projected returns are too far off from the actual portfolio returns. This means potentially lower performance at retirement vs DCA + rebalancing, but I am okay with that.

More recently, I came across an article published in 2013:
https://papers.ssrn.com/sol3/papers.cfm ... id=1606347

The main claim is that IRR is not a "fair" metric to make the claim that VA can outperform other mechanical investment strategies. It is also true that almost all prior VA research has used IRR to demonstrate its superiority, but the point here is higher IRR does not equate to more money at retirement.

This to me is quite concerning, as it strikes at the foundation of reason (2). If higher IRR is not going to translate to more money at retirement, then that's a big concern for me.

For those who do Value Averaging. Are you aware of this? What are your thoughts?

To keep discussion on point, please refrain on comments on "I don't do VA, I do DCA, and I don't believe in VA". I would like to focus this discussion on whether we have been misled by earlier VA papers that used IRR as a proxy measurement of the superiority of their technique. And if yes, how much of a concern is it going to be?
I saw this go through a couple weeks ago and remembered that I would reply (since nobody else has), but didn't have time right then. And I am not doing anything useful or entertaining now. Yes, value averaging will absolutely be expected to give you a possibly higher IRR but most likely lower ending $ value. It is roughly between dollar cost averaging and lump sum investing in expected results. If you look at dollar cost averaging, there are many periods where it would have returned a higher IRR, and yet a lower ending amount than the same inflation adjusted lump sum. You can backtest that in portfolio visualizer. If it was that easy to beat the market by keeping money out of it, everybody would do it. :wink: Sorry, it isn't that easy.

My understanding of value averaging (I actually read the book :shock:) is that it is intended to give the investor the greatest amount of cash for spending or to feel safe, relative to the goal of saving for a fixed amount at a given time. It is not expected to beat the real result of simply throwing money into the stock market whenever you are able over many decades. That is not its purpose. It is more about improving risk adjusted returns for intermediate term goals. It is probably best in the 2-20 year range. Shorter than that and you would seek to be safer, longer and you would want more risk.

In theory I will do this to save for a house. So far I am still building an equity position. I have no fixed goal, any money I don't need to meet stock growth will go into an ever larger pile of bonds. At least I'll have more equity exposure in down markets and less in up markets, I like that I won't be sacrificing too much in expected return even as my portfolio becomes progressively safer. But, I do not see an advantage for an objective more than say 20 years out, and even that is more of a theory because who wants to religiously follow an arcane rule every quarter for decades for at best marginally better results? Less work to just use all stocks or a fixed asset allocation.
It's Time. Adding Interest.

Post Reply