Monday, March 25, 2013

Value Averaging Nonsense

Andrew Hallam is promoting value averaging again. This is an investing strategy that involves choosing a pre-determined rate of growth for your portfolio and then either making up the difference when markets disappoint or taking money out of the market when returns exceed your expectations. It doesn’t work.

Several months ago I explained in detail why value averaging doesn’t work. A quick summary: you have to have a pile of cash on the sidelines available to pour into the market if needed. Either that or you have to borrow deeply if the cash isn’t available.

Value averaging proponents calculate their strategy’s returns using the “internal rate of return” (IRR). The IRR can be a good way to measure returns, but in this case it means that we ignore the opportunity cost of idle cash and ignore the interest costs on borrowed money.

Hallam gives some results over a period cherry-picked to make value averaging look good: the 5 years from February 2008 to January 2013. Note that 2008 was the year the market crash began. Let’s say that you had $500,000 invested in U.S. stock index ETF VTI on 2008 Feb. 1 and began value averaging with a 9% per year return target and making adjustments each month.

By the beginning of March 2009, you would have borrowed a total of just over 363,000! It’s ridiculous to think that investors would have had this much cash on the sidelines or would have had the nerve to borrow this much to buy stocks while they were crashing. It’s equally ridiculous not to count the opportunity cost of cash on the sidelines or borrowing costs.

Here is Hallam’s attempt to acknowledge these concerns:
“Critics have wondered where the cash for aggressive purchases is supposed to come from when markets keep falling. Such a drop could force a depletion of the money market cash reserves if investors are continuing to transfer assets from cash to stocks. In this case, Bruce Ramsey suggests if there’s not enough capital in the money market account, then you either do nothing or invest whatever cash is available.”
Are the return figures quoted by Hallam based on Ramsey’s suggestion to do nothing when you’re out of cash or are they based on borrowing hundreds of thousands of dollars? If they are based on the borrowing, does he factor in interest costs? Does he seriously think that investors should leverage themselves in this way?

Value averaging is a bad idea that sounds appealing until you really crunch through some numbers with real-life scenarios.

13 comments:

  1. Indeed. At the end of 2008 I thought the crash was a historic opportunity to invest more. I borrowed about a year's worth of expected savings and invested at that point -- adding about 4-5% leverage. I didn't like being leveraged even that much and covered by May 2009. I'm one of the most risk-tolerant investors I personally know, so I can't imagine a lot of people sticking with a plan that calls for 1:1 leverage in the midst of the chaos of a market crash.

    ReplyDelete
    Replies
    1. @Potato: I didn't even really use leverage. I normally have dribs and drabs of cash in various trading accounts and I wait until they build up to $3000 to $5000 before investing. During the crash I was more diligent about getting smaller sums into the market. So, I was just a little closer to 100% invested.

      Delete
  2. So, I checked your earlier post on VA, you did no math but based your opinion on somebody's work. How do you know they are right? First, I don't like how M.E. Edleson compared DCA to VA using IRR. Nor do I like the Chart he used, for his comparison. However using simple logic, VA should beat DCA most times when you compare total out of pocket money to total profit. What you should really compare is DCA,VA and Robert Lichello's Synchrovest program from his book superpower investing,1974. Synchrovest is the only periodic investing program I know that keeps track of your average cost per share to determine when to invest extra money in the market.

    ReplyDelete
    Replies
    1. @Lost Cowboy: It's true that I didn't include math in my previous post, but that doesn't mean I didn't do any math. I pointed to Hayley's work because I read it and understood it. Your contention that "using simple logic, VA should beat DCA most times when you compare total out of pocket money to total profit" is based on a false comparison. People have money to place in the market when they have it, not when some strategy says they should have it. To have avoided leverage with VA over the last 5 years, an investor would have had to hold back a very large percentage of their portfolio in cash. Not knowing when the next big (or even bigger) crash is coming, should I hold back 50% of my portfolio in cash every time I start a VA cycle? The dismal return on cash has to be factored into VA returns. Either that or factor in borrowing costs and set a leveraging limit, and take into account the fact that just when VA would be most valuable, you can't follow it any more due to the leverage limit. When you do the math to account for these things properly, VA does not beat simpler strategies.

      Delete
  3. Another way to look at it is a constantly-shifting asset allocation that occasionally has a large weighting in cash. Like the people who stay out of the market waiting for the prices to be "just right", the costs can be very large because of the low returns on cash.

    I wonder what the results would be using something other than cash, for example bonds. Their correlation to stocks may be higher than cash but at least they have a higher return as well which might improve the results of the overall portfolio.

    Other assets (or a mix) could also work to provide the alternative such as real estate and commodities. You could even look at it as doing "value averaging" with all the asset classes, lowering the weighting of one when it starts to go above its target return. I suspect this would look a lot like regular asset allocation, that is if you got all the numbers right to end up with a sensible result (if you target a 1% return on stocks you probably won't get a good result).

    ReplyDelete
    Replies
    1. @Value Indexer: It would be interesting to try out these ideas and see how they compare to a portfolio that has the same average asset allocation. A tricky part is to choose a target return for VA with historical returns without taking into account future information. VA tends to look better when you guess the average return for the VA cycle in advance.

      Delete
  4. I see some merit in using VA as a way of maintaining a portfolio in retirement and managing risk.

    For example, given a starting allocation of 50% stocks, 50% bonds one could use VA with a growth rate of 2% to manage withdrawals from the 'risk' portion of a portfolio (stocks and other volatile assets) with the residual in bonds or cash. I think in this scenario, your concerns about leverage or needing to borrow to invest may not be as founded.

    ReplyDelete
    Replies
    1. @Andrew F: This scenario you described sounds quite dangerous to me. Suppose the retiree started this plan at the beginning of 2008. He bought stocks all the way down until March 2009 when he was at or close to 100% stocks. At this point he lost his nerve completely and sold everything into cash. He suffered a huge loss in only 15 months. Of course, if he had just stayed the course, he'd have been fine eventually, but he chose a 50/50 portfolio for a reason -- he can't handle huge volatility.

      Delete
  5. I have read your post that VA does not work, but you offer no proof that you have tested the concept yourself to show that it does not work. Why don't you provide us with the number crunching that you have done to disprove the theory. In addition you make the comment in the second paragraph above "you have to have a pile of cash on the sidelines available to pour into the market if needed. Either that or you have to borrow deeply if the cash isn’t available". This is false and it shows that you do not have a complete grasp of how you can apply VA.

    ReplyDelete
    Replies
    1. @Anonymous: That's a pile of nonsense. I was going to test VA claims myself until I read criticisms by others that explained the problems clearly. IRR calculations must be based on when people have money available, not when some investing strategy dictates that money should or should not be available. If you don't think the quote you pulled is true, then I don't think you understand VA.

      Delete
  6. How do you know that the "others" have tested VA properly. Do it yourself and then form your own judgements.

    ReplyDelete
    Replies
    1. @Anonymous: It's not the testing that convinced me -- it was the explanation of where the problem lies. In any case, the onus is on VA proponents to prove that VA works. All of the attempts at justification I've seen are based on bogus IRR calculations.

      Delete
  7. Hi Michael and all commentators

    I recently starting to research about VA and I'm planning to dig until get convinced how real and valuable is for small investors like us.

    Before finding your post, I found all these studies:
    http://valueaveraging.ca/va_research.html

    - They seem serious studies from different sources
    - They cover many different scenarios, asset classes, risk, etc.
    - One of them covers your sample of 2008-2009 (but 2008-2012). This method is not for short-term investments but in 4 years period still was able to beat the other methods (DCA,LumpSum,PropRebalance).

    I also added similar comment to your other post about this:
    http://www.michaeljamesonmoney.com/2013/03/value-averaging-experiments.html

    Michael, thanks for taking some time dedicated to this topic (VA). Your blog is great for all us.

    Please, if possible, keep covering this topic.

    Best regards
    Raul

    ReplyDelete