Tuesday, October 30, 2012

Value Averaging Doesn’t Work

Andrew Hallam wrote a piece in the Globe and Mail that likened enhancing performance in sports by blood doping to an investing method due to Michael Edelson called “value averaging”. Value averaging is simple enough to understand, and if you use the wrong method of evaluating its results, it seems to boost returns. However, the reality is that it doesn’t boost returns, and it drives up your investing costs.

The idea of value averaging is to keep your portfolio increasing at some target rate, regardless of what happens in the market. For example, if you target a 0.5% return each month, if the market goes up more than 0.5%, you sell some of your portfolio; otherwise, you add more cash to buy more assets. No matter what happens in the market, your portfolio rises steadily.

An immediate problem arises: where do I get this cash to pour into my investments when the market drops? The answer is that you’re supposed to keep a side pot of cash that you either put money into or take money from each month. Over time this pot of cash could either dry up or become quite large depending on how the market moves. To deal with this, the strategy calls for a reset every so often (say 3 years) where you reset the cash level to some fixed percentage of your portfolio’s size.

The selling point of value averaging is that you add money when the market is down and pull money out of the market when it’s up; buy low and sell high. This gives an internal rate of return (IRR) that beats the market return and also beats dollar-cost averaging. So far, what’s not to like?

The problem is that the IRR calculation only takes into account the money actually invested in the market. It ignores the cash that you have to keep on the sidelines. However, this cash is real money that you have to keep around earning low returns.

When you factor in the cash, value averaging gets worse returns in most markets than a simple buy-and-hold strategy. It isn’t hard to see why this is true. At any given time, a buy-and-hold investment is fully invested. However, a value averaging investment is only partially invested. The tendency for markets to go up creates an opportunity cost for the cash on the sidelines. This cost exceeds the boost that comes from a higher IRR.

One remedy for this problem might be to start with no cash on the side and plan to borrow as necessary to run the value averaging strategy. However, there is a gap between the best interest rate you can get on your cash and the lowest interest rate you can get when borrowing. This gap serves as a drag on value averaging returns. Another problem with this remedy is the possibility of becoming highly leveraged if markets drop significantly.

If you are interested in a more technical critique of value averaging, read Simon Hayley’s paper Value Averaging and How Dynamic Strategies Bias the IRR and Modified IRR. He concludes “Value averaging does not boost profits, and will in fact suffer substantial dynamic inefficiency. It also imposes additional direct and indirect costs on investors as a result of its unpredictable cashflows. The strategy thus has very little to recommend it.”

With the promise of lower returns and higher trading costs with value averaging, I’ll happily stick to my buy-and-hold approach with occasional rebalancing.

13 comments:

  1. 121How do the true returns from value averaging differ from a balanced portfolio with a % in fixed income? Seems like having a portion in bonds attempts to achieve the same thing (buy low, sell high) and you benefit from a small return on your cash.

    ReplyDelete
  2. @Anonymous: People typically use bonds to limit portfolio volatility. An investor who doesn't want to exceed some specified level of volatility and wants to use value averaging would have to start with a lower than desired allocation to stocks and then let the strategy play out and possibly have the allocation to stocks rise or fall. This would then give lower expected returns than just using regular portfolio rebalancing with the desired level of volatility.

    ReplyDelete
  3. Could it be, that that it does not ignore the cash that you have to keep on the sidelines?

    Perhaps it assume that you are an average Joe (and not an investment banker) who would diligently invest 10%+ of you income annually anyway. The theory will lead you how to do it and at what time is the more efficient.

    ReplyDelete
  4. @Financial Independence: I think you are talking about dollar-cost averaging where you invest the same amount each month. With value averaging, the amounts invested each month can vary wildly. For example, if you had $200,000 saved, expected 0.5% return each month, and the market dropped 5% one month, you'd have to invest $11,000 of new cash. That's not going to come from the typical person's after-tax and after-expenses income. So, to answer your initial question, yes, there has to be cash (or something very liquid) on the sidelines and value averaging supporters do ignore the opportunity cost of not having it invested.

    ReplyDelete
  5. It seems like the brunt of the arguments against value averaging is that critics assume you are investing outside of a tax sheltered account. Also, when markets do go down and the 'system' needs you to invest a lot more than you might have on hand, isn't that a good time to raise cash by selling a portion of your bond fund? I assume at the bare minimum you would have one equity fund and one bond fund in your account. Please let me know if I have missed something as I am seriously thinking of trying this out inside my RSP with TD e-series index funds (very low MER).

    ReplyDelete
    Replies
    1. @Johnny: I think the answer to your question comes down to why you have a bond fund at all. Owning bonds reduces volatility and increases the chances that you won't panic sell at the bottom of the market. If you're willing to increase your allocation to stocks when the value averaging strategy calls for it, why not just increase your stock allocation now? Historical bond returns are enough lower than stock returns that they offer no increased portfolio return expectation due to diversification into an uncorrelated asset class. Bonds' only virtue is their lower volatility, but you give up this lower volatility if value averaging tells you to sell them.

      Delete
    2. Investing to me seems to be like religion with so many flavours out there. It's hard to argue against diversification though... especially if you were 100% invested in the NasDaq back when it was at it's all time high, or even the TSXventure in the last few years (down 90% I seem to recall). If one were to sell bonds because value averaging called for increased contributions, wouldn't it be fair to say that the allocation had been overweight in bonds at that point anyway, so you're just re-aligning your asset allocation anyway? (Still curious if I am missing anything...) Perhaps I shouldn't of said bond fund, as I would be more disposed to something with a shorter duration, along with Gold and/or commodities/reits/etc.

      Delete
    3. @Johnny: It's true that many people treat investing like a religion: they believe things without seeking any evidence for their beliefs. I agree that it's hard to argue against diversification. However, diversification only improves long-term returns if you diversify into asset classes with similar expected returns. Diversifying from the Nasdaq does not require bonds, gold, real estate, or commodities. Simply owning other US stocks along with foreign stocks gives great diversification. If on top of this diversification you need reduced volatility, then by all means buy some other asset classes. But, it's best to understand why you're doing it. This diversification into other asset classes is unlikely to give better returns over the long run. What it will do is reduce volatility and help you sleep at night. Letting value averaging increase volatility and take away good nights' sleep is counterproductive.

      Delete
  6. Well Mr. James, I think your article at http://www.michaeljamesonmoney.com/2013/03/value-averaging-experiments.html#comment-form
    is starting to convince me that Value Averaging is not the way to go, BUT.... I have been sitting in cash for over a year and wonder if value averaging might be a good way to ease back into the market over a span of say 2 years. You will probably respond with the opportunity cost of being in cash, but opportunity cost appears to be totally random to me.

    Assume for a moment it is June 2008: If I invest it all (50K+) just prior to the crash, I'd had lost about 40% and then I would have to make an 80%+ return just to get back to even. I'm no longer in my 20s (I'm 45), and this is a really big mental hurdle for me.

    ReplyDelete
    Replies
    1. @Johnny: You have to find something you can live with. It sounds like you wouldn't be comfortable with your stock allocation too high. Whatever percentage you choose as a target, there's nothing wrong with easing into it (although 2 years is a long time).

      Delete
    2. Thanks for your insights and especially quick responses. There are quite a few good financial blogs out there but when the writer of the blog doesn't respond for months (if ever) then I tend to come back less.

      Now if I could only get people to visit my blog more often... (cartoons, not finance)

      Delete
  7. I have recently started a live experiment on Value Averaging. I plan to share the progress - check it out at www.investingfunda.com.

    I have made a few adjustments to the vanilla VA strategy.

    ReplyDelete
    Replies
    1. @Investingfunda: VA force you to guess future returns. If you happen to guess well, VA will work for you. If you guess poorly it won't.

      Delete