With the Dow recently closing at a record high and exchange traded products (ETPs) kicking off 2013 with their strongest January-February inflows on record, there has been a lot of buzz about new cash coming into the market. If investors are finally deploying some of that cash that had been sitting on the sidelines, it raises an interesting question: What is the “right” amount of cash to hold in a portfolio? And is this a good time to take cash holdings and put them back into the market?
To answer these questions, I took a look at investor behavior around cash holdings during the last few economic cycles. In particular, I have taken data on individual investor cash holdings from the survey that the American Association of Individual Investors has been conducting since November 1987 (right after the stock market crash of 1987). Individual investors are asked to report on their allocation to equities, fixed income and cash during the month the survey is conducted. I’ve plotted that data on the chart below.
On the left axis, you’ll see respondents’ average allocation to cash as percent of the total individual investor portfolio. On the right axis is the cumulative return of the non-cash portion of the portfolio (shown as the value of a dollar invested at the beginning of the period).
It is interesting to see how the peak levels of cash holdings seem to correspond to the bottoms of market cycles, and the lowest levels of cash holdings seem to correspond with the top of market cycles. In other words, when stock prices were low, investors were holding on to a lot of cash. As stock prices rose, investors were putting their money back into the market. Indeed, cash holdings were high right after the 1987 crash and steadily declined as the market rallied in the 1990s. They bottomed out almost exactly when the tech bubble burst in early 2000.
We then saw investors pull back their investments and go back into cash as the market declined. This means investors ended up with peak cash holdings when the rally of the mid-2000s started around the middle of 2003. The cycle repeated itself into the peak of the market before the last crisis, with investors showing the highest level of cash holdings as the market bottomed in the middle of 2009.
This appears to be one more manifestation of trend-chasing often exhibited by investors. Indeed, the chart below shows that on months when the survey reports a reduction in cash holdings, the non-cash portion of the portfolio had, on average, much higher returns in the preceding period than in months when the survey reports increases in cash holdings1. In other words, when returns have been good investors reduce their cash holdings and invest more in risky assets, and vice-versa when returns have been poor.
How effective is this cash-timing behavior that investors seem to follow? Unfortunately, it has not shown to be very effective. The chart below shows that returns, on average, are about the same in the month after cash holdings are reduced compared with months when cash holdings are increased. Even more discouraging, average returns in the year after cash holdings are decreased are actually smaller than when cash holdings are increased.
So, what is the right level of cash holdings for an investor? That of course depends on each individual investor. But the charts above illustrate that many investors could benefit from setting cash holdings at levels to meet expected near-term expenditure needs, such as covering daily expenses or paying regular bills like mortgage or car payments. In other words, keep cash stable at levels that match expected near-term expenditures and then focus on managing the rest of the portfolio to fit longer-term investment objectives and risk levels rather than using cash as a market or risk timing tool. This particularly applies to investors with long-term investment horizons where the probability that a typical strategic portfolio (such as the traditional 60% stock/40% bond portfolio) will underperform cash over longer periods is small.
To illustrate this last point I have constructed a hypothetical 60/40 portfolio with data going back to 1925, and I’ve compared its returns over various horizons with the return of cash investments over time. The chart below shows the probability that over various horizons (from 3 months to 10 years) the return of cash will be higher than the 60/40 portfolio.
Across all the 10-year periods from 1925 to 2012, roughly only 1 in 10 showed cash with better returns than the 60/40 portfolio. Even for horizons as short as 1 year, only 1 in 4 of all 1-year periods since 1925 showed cash with better returns than the 60/40 portfolio. These statistics are, of course, no guarantee than the next year or next 10 years may not be one of those periods. But for most investors, time spent making sure their overall portfolio matches their risk tolerance and investment objectives will likely be more productive than time spent managing their cash holdings.
 The return of the non-cash portion of the portfolio is, for every month, the weighted average of the total return of the S&P 500 and the total return of 50/50 mix of intermediated Treasury bonds and long-term corporate bonds. The weights are given by the relative weights to equity and fixed income reported by the survey for that month. Data for the S&P 500 returns as well as government and corporate bond returns is from the Morningstar SBBI database.
 The 60/40 portfolio is constructed as the 60/40 weighted average of equity and fixed income monthly total returns. For equities I used equity returns as reported by the Morningstar SBBI database. For fixed income I use a 50/50 mix of the returns for intermediate government bonds and long-term corporate bonds, also from the Morningstar SBBI database. For cash returns I use 1-month Treasury bill returns as reported by the Morningstar SBBI database.