SINGAPORE (Oct 29): As a financial adviser representative, a common complaint I hear from clients on regular investment plans is that dollar cost averaging is a little simplistic. Sure, investing the same dollar amount every period helps increase returns by buying more units of stock when prices are low. But surely one can do better? Shouldn’t there be a structured way to invest a higher dollar amount when prices are low, and fewer dollars when prices are high?
As it turns out, such an alternative does exist. Instead of fixing the dollar amount invested each period, value averaging works by fixing a target portfolio growth rate (called the “value path”). The dollar investment in each period is then the difference between the actual portfolio value and the value path.
This is best illustrated with a simple example. Suppose our friend Sean would like to invest $12,000 a year for 10 years in the MSCI World stock index with a target return of 7% a year, starting in January 2008. To calculate the value path for each year, we increase the previous year’s value by 7% and add another $12,000 (see the blue line in the chart). If the MSCI World indeed grew consistently by 7% a year, Sean would need to add $12,000 to his portfolio each year to stay aligned to his value path.
But the MSCI World does not grow consistently by 7% a year. For example, from January 2008 to January 2009, the index plunged 40% due to the global financial crisis. Value averaging then dictates that Sean not only invest his $12,000 for the year, but also top up the 40% loss and 7% unearned gain from the year before, to bring the investment back on the value path. And indeed, Sean would have been richly rewarded for overinvesting at the start of 2009, as the index returned a whopping 30% that year and 12% the year after.
From this example, we can easily see how the proponents of value averaging believe that it might do better than dollar cost averaging, since it is more responsive to market price changes. The idea has enough support that at least one popular investment platform in Singapore helps clients to implement value averaging automatically for regular unit trust investments, thus freeing investors from the need to track and calculate the value paths on their own.
Unfortunately, as with much else in investing, what seems intuitively correct is not necessarily so. To those who are considering value averaging, beware — it mostly fails to outperform dollar cost averaging in creating wealth for the investor.
Let us first acknowledge that value averaging does indeed generate a higher internal rate of return on invested cash than dollar cost averaging, most of the time. As an example, Table 1A outlines Sean’s investment cash flows and results for both methods. Value averaging generated an IRR of 10.9% on invested cash flows, against 10.4% for dollar cost averaging. But — and this is a big but — because value averaging invested less cash, the portfolio’s value is lower. Would you be happy about having a higher IRR but less money? I wouldn’t.
Also, Table 1A assumes that Sean has an infinite pool of interest-free cash ready to deploy. This is unrealistic. For a fairer comparison, we should see what happens if Sean starts with no spare cash but is able to utilise any uninvested cash from previous periods where necessary. Cash uninvested after 10 years is added to the final value. Once these adjustments are taken into account, IRR for value averaging drops to 10.1% and terminal wealth is $212,026 (see Table 1B), compared with 10.4% and $215,631 for dollar cost averaging. So, the investor does unambiguously worse under value averaging.
These findings are not a freak accident created by the investment time period chosen. What if Sean did not start investing in 2008, but in year 2000? 1990? Using actual price data from the MSCI World from 1970 to 2018, we have a total of 39 overlapping 10-year periods. Out of these 39 periods, the same pattern holds true: Value averaging ends up with a higher IRR 82% of the time if we consider only the invested cash flows, but generates a higher terminal wealth only 28% of the time. This conclusion is in line with academic studies pointing out how IRR can be misleading when used to measure the benefit of value averaging (see, for example, “Value Averaging and the Automated Bias of Performance Measures” by Simon Hayley in 2012).
In fact, by tweaking the assumptions in the above simulation, we find that the same conclusions hold true even if we allow Sean to an early cash fund (which is invested immediately for dollar cost averaging but held back for discretionary investments under value averaging). They are also true under different value path growth rate assumptions. Value averaging is only advantageous if we measure it by a very limited metric, which is the IRR on cash invested. Under more realistic conditions and performance measures, it is unlikely to create greater wealth than a simple dollar cost averaging plan.
The above discussion holds a few lessons for us as investors, even for those who were never considering value averaging in the first place. Some of these are obvious: ensure we have the right metrics; don’t always trust our intuition. The subtler but more important takeaway is that it rarely pays to delay putting investible cash into the market. The root of value averaging’s downfall lies in its tendency to invest cash later than dollar cost averaging. While it may be right to do so sometimes, it is not correct often enough to beat a simple strategy where we invest money as soon as it becomes available, because equities trend upwards on average. Market timers and buy-the-dippers would do well to take note.
Herbert Lian is a financial adviser representative at IPP Financial Advisers Pte Ltd. The views expressed here are solely those of the author in his private capacity. This article should not be regarded as professional investment advice or as a recommendation regarding any particular investment.
