Cost-averaging has been one of the most popular approaches to invest in stocks or mutual funds worldwide and financial planners swear by it.
Pioneered in the 1940s in the U.S. (hence often referred to as dollar-cost averaging) and popularized by the 401(k) system, the approach--also commonly called the systematic investment plan (SIP) in India--requires an individual to invest set amounts into a portfolio of securities at regular periods.
In India, the SIP, found its way into the investing lexicon in the early 2000s but after the experience with the so-called “hot” retail money in the pre-2008 boom and the subsequent intemperate redemptions during the financial crash, mutual funds have started promoting SIPs in a big way in a bid to attract more long-term, sticky capital.
The idea of an SIP is simple and forceful: it, by default, not only paves the way for regular investments, but also, thanks to forcing a preset amount, eliminates the behavioural greed-and-fear cycle--which could otherwise make the investor over- or under-invest during up or down markets.
DCA underperforms LSI
In an eye-opening research that has taken the global financial-planning industry by a bit of a storm, a study by mutual fund giant Vanguard showed that lump-sum investing (LSI) outperformed dollar-cost averaging (DCA) two-thirds of the time. (Click here to read the complete study.)
In the simulation, researchers picked $1 million and invested it into a portfolio either via LSI or drip-fed the amount into the same portfolio through DCA over the course of 6, 12, 18, 24 and 30 months and both portfolios were held for a period of 10 years.
In order to create as many scenarios as possible, the researchers started with portfolios going back to 1926 in the U.S., 1976 in the U.K. and 1984 in Australia and tested a variety of allocations ranging from all-stock to 60:40-stock:bond to all-bond and looked up the result for rolling 10-year periods (that is, from January 1926 to December 1935 followed by January 1927 to December 1936 and so on).
The conclusion: LSI portfolios outperformed DCA portfolios 66% of the time for 100% equity, 67% for 60:40 stock:bond and 65% for 100% bond in the U.S. Similar results were seen across the U.K. and Australia data.
Takeaways
The results of the study weren’t surprising as markets have historically tended to move upwards over long durations of time, so the cash that waits to be invested in the DCA approach has a greater probability of losing out on gains rather than providing cushion in a downturn.
Not surprisingly, authors of the study concluded that DCA works best when the start of the investment horizon co-incides with market peaks as the investor is able to buy more investments cheaply as the market comes down compared to the lump-sum investment, which would face a bigger immediate loss from which it may take longer to recover.
Of the 1021 hypothetical portfolios they analysed in the U.S., 229 LSI portfolios faced the prospect of a loss at the end of 10 years, compared to only 180 for DCA, meaning the latter helps insulate better from downturns.
But they added that even as the short-term fluctuations were likely greater in LSI portfolios, investors must remember that the relative insulation from volatility could also result in sacrificing greater portfolio gains.
They finally stated that “if markets are trending upward, it’s logical to implement a strategic asset allocation as soon as possible because it should offer a higher long-run expected return than cash.”
Investor Return is the catch
While being an expansive study, the Vanguard research only does a historic analysis of various possible scenarios and should not become a reason to advocate LSI over DCA.
The simplest reason for this being: the biggest virtue of DCA is it eliminates the need on the part of the investor to try and guess where the markets are trending (to quote the authors) whereas LSI involves timing the market, something that most investors have been shown to historically do woefully. Thus, the angle which the Vanguard study does not take into account is capital-weighted return.
Let’s illustrate with an example. A fund with a NAV of Rs 100 rises 100% in one year and falls 50% the next. At the end of year two, the fund would still be at Rs 100 and would have made a 0% gain.
But if the fund’s asset size was a meager Rs 100 crore at the start of the period before it made a 100% gain but swelled to Rs 1,000 crore at end of year 1 (after the gain) and before it made a 50% loss, far more investor capital would have been lost in those two years, even as the statistical net return after two years was 0%. (At Morningstar, we call the concept Investor Return, which is calculated based on flows data--and in whichever countries Investor Return data is available, we have seen a considerable difference between funds’ stated returns over different time periods and actual capital-weighted Investor Returns).
So give an the average investor $1 million--or any amount--to invest via lump sum and he is more likely to put it to work at the wrong time, that is, around the peak of a euphoric market when valuations are rich and the investment is more likely to lose money rather than at the trough of a bear market when recent returns have been poor but compellingly-cheap valuations will likely result in greater future gains. (Inflows into stocks/stock mutual funds at the height of the 2007 bull run compared to outflows after the 2008 crash bear testimony to this point).
Had the Vanguard study taken into account capital-weighted return rather than stated returns, the picture would have likely been different and in favour of DCA and not LSI.
So even as the SIP may have held the investor from investing in one go at the height of the downturn in January 2009--with the market trading at a price-to-earnings ratio of only 10--and making a killing, it also stopped him from going all-in in January 2008 when stocks were above 25 times earnings. And, on average, investors are prone to taking the latter decision more often than the former.
