Cash Isn't King

Bottom Line: Mutual fund cash levels are historically low, but considering prevailing interest rates, they are about where they should be.

A topic I’ve seen discussed with increasing frequency lately is that of the amount of cash held by mutual funds. Taken as a percentage of total assets, as of December, cash (or equivalents) made up 4.3% of total assets controlled by mutual funds. This is of concern for a couple of reasons, with the most common argument being that if funds begin to get a large number of redemptions from investors, and they have a minimal amount of cash on hand, then they will need to sell stock in order to meet their cash needs, fueling the market decline that is most likely the reason for the redemptions in the first place. There are many reasons why a fund would hold a low level of cash:

It is on that last point that I want to focus. When short-term interest rates are high, mutual funds have something of an incentive to hold cash. The logic is sound - if you are a fund manager, and there is a risk-free investment that will pay you a 10% return, are you going to put your money there or are you going to risk it in the stock market where you may get an 8% return during a good year, but with a heckuva lot more heartburn? Most of us would surely switch to the risk-free opportunity. For our purposes, we will use the yield on 90-day Treasury Bills as the risk-free rate.

This logic is certainly supported by the numbers. The correlation between mutual fund cash levels and the 90-day T-Bill rate is 0.74, which means that the prevailing level of interest rates can theoretically explain 55% of why mutual fund cash levels are where they are. So with short-term T-Bills now yielding less than 1%, what incentive is there for mutual funds to hold a lot of cash (other than the bullet points mentioned above)?

Using statistical modeling, we can estimate how much cash mutual funds should have on hand given a certain level of interest rates. By getting such an estimate, we can then see whether the current cash percentage is too high or too low, without the distorting effects of interest rates. This should give us a better feel for the underlying sentiment of the fund managers and (theoretically) provide us with an improved market forecasting tool. In actuality, this turns out to be the case – doing this extra work improves the forecasting ability of mutual fund cash levels by a large degree.

Using this modeling technique, we can conclude that if interest rates were at 0%, then funds would normally hold 4.55% of their assets in cash. This is a normal level of cash that funds presumably need on hand in order to meet redemptions or what have you. When we add in the fact that T-Bill rates were around 0.91% at the end of December, the model tells us that funds should have had about 5.0% of their assets in cash. As stated above, they only had 4.3% of their assets in cash, so they were “deficient” by about 0.70%. This is about average since 1954, meaning that it appears as though funds are holding neither "too much" nor "not enough" cash at the moment.

This data has been quite effective in the past when cash levels became either too high or too low, after adjusting for interest rates. The table below shows the performance in the S&P 500 one year after the funds were in a cash surplus or a cash deficit. “Surplus” is defined as funds holding 2% or more cash than they should, given the level of short-term interest rates. “Deficit” is defined as funds holding -2% or less cash than they should.

S&P 500 One-Year Performance After Cash Extremes
	Surplus	Deficit
AVG RETURN	13%	-4%
% POSITIVE	85%	31%

We can see from this table that one year after a surplus, the S&P was higher 85% of the time, with a whopping average return of 13%. In stark contrast, one year after a cash deficit, the S&P was higher only 31% of the time, and suffered an average decline of 4%. Both of these results differ significantly from a random one-year period, suggesting that consulting this data can improve our chances of success in timing long-term investment decisions (or at least it would have in the past).

Below is a long-term chart of this formula, showing periods of cash surpluses and cash deficits.

Whatever the reason, the current mutual fund cash level of 4.3% is extremely low historically, as it one of the lowest 10 readings since 1954. Other than February 2003, the other recent times cash has been so low were March 2000 and April 1998, and we know what happened after each of those occurrences. As we now know, however, in 1998 and 2000 short-term rates were considerably higher than they are now, so funds were expected to carry more cash. Right now there is not a lot of impetus for funds to invest in risk-free assets (from a yield perspective, anyway), so we expect the percentage of assets in cash to be lower than they were in 1998 and 2000. According to our model, funds are holding about the right amount of cash given prevailing rates.

Since the market and the fund industry are both dynamic, there is no guarantee that historical relationships will continue. However, in the past, monitoring these cash levels would have improved our chances of successfully forecasting future market direction to a good degree, so it should continue to be watched. I agree with the logic that low absolute levels of mutual fund cash are troublesome - market history supports that argument. But I don't agree with the assumption that cash is low because fund managers are too optimistic on the market - I believe there are simply too few alternatives giving them the opportunity to earn acceptable yields for their investors.

On a side note, mutual fund cash levels (both the raw percentage, and the formula from above) are now available on the site in the Indicators section, along with margin data for the NYSE and Nasdaq.

Bottom Line: A look at how NYSE specialists are acting during down weeks in the market leads to bullish conclusions.

For the week ended February 6^th (the most recent data available), specialists on the NYSE accounted for 21% of total short sales. Looked at another way, the “public”, which includes everyone except NYSE members, shorted 2.5 times as many shares as did the specialists. Both of these figures are astounding – they are the most extreme readings seen in the 61-year history of the data, except for the week immediately following 9/11. I’ve discussed these numbers many times over the past year, touching on their bullish nature, but also noting the fact that there is no question that increased convertible debt issuance has affected the numbers to some degree. Even accounting for the possible affects of more convertible debt, I believe these numbers are still extreme.

Historically, specialists decrease their short activity when the market is falling and increase it when the market is rising. Since 1950, when the S&P declined during a given week, specialists decreased their shorting 67% of the time. Likewise, on any week when the S&P rose, specialists increased their shorting 65% of the time. Over the past year, however, when the S&P declined on the week, the specialists decreased their shorting an incredible 87% of the time; when the S&P rose, the specialists increased their shorting only 54% of the time.

Let’s focus on that last sentence about specialists decreasing their shorting activity 87% of the time when the market had a down week over the last year. There have only been two other times in history when specialists were so adamant about buying a down market – 1975 and 1991. After the instance in 1975, the S&P was 19% higher one year later. Coincidentally (or perhaps not), after the instance in 1991, the S&P was also 19% higher one year later. Looking at any year in which specialists increasing their buying during at least 80% of the down weeks, the S&P was higher one year later 85% of the time, with an average return of 14%.

The readings this data are giving are truly historic. In order to try to get a better handle on what it might be saying, I have studied the data from every angle I can imagine. I have looked at absolute levels of specialist shorting, I have looked at relative levels of public vs. specialist shorting, I have looked at the shorting as a percentage of volume, I have looked at the behavior of the specialists and the public during up and down weeks, and I could go on and on. I firmly believe there are some forces at work that are “artificially” toying with this data (such as hedge fund activity), but the fact remains that NYSE specialists – who have the best information available as to the supply and demand characteristics of their stocks – are not shorting stock heavily, and are behaving in a way that has consistently lead to higher markets in the past. It is impossible to ignore the bullish implications from this.

As I discussed last week, the day immediately after option expiration tends to have a definite negative bias, particularly when open interest in put contracts is so high relative to calls. That negative bias does not tend to extend much beyond the first day, however, so I don’t see much hindrance in the way of seasonality for the coming week. As I noted in the intraday comments, we have become extremely oversold in the short-term. Our STEM.MR model reached an extremely high 67% on Friday – the highest reading in over a year except for the 69% we saw at the low on January 29^th of this year. The intraday cumulative TICK indicator on the Nasdaq dropped below -4000 on Friday to become the most oversold it has been since near the absolute low on October 7^th, 2002. But extreme short-term oversold readings are one thing at the end of a prolonged decline – they are quite another barely a month after the market was making new highs.

My test for the market over the past eight months has been simple – if it cannot rally off short-term oversold conditions, then we know something is amiss. Up to this point, the market has done precisely what it has needed to do by rallying off these types of conditions. Now, we are dangerously close to breaking that pattern. Nasdaq-type issues should have rallied today, but I am giving it a little leeway considering that consistent post-expiration hangover. So the real test comes now, and the broader market should rally or we’ll have a good clue that there is a more substantial and longer-term decline in store. At this point, my preference is to look long, but we’re hanging by a thread here. If the market can’t get some “giddy up” going in the next day or two, I don’t want to have a whole lot of long-side exposure.

This disclosure is not intended as trading advice in any form. It is meant as a note to subscribers that the author may have a position directly affected by the market outlook reflected in the commentary. Although the author takes great pains to remain objective in any commentaries, it is only fair that readers should know that the author may have taken positions in accordance with his market outlook. Positions can and do change at any time, without notice to the reader.