Distribution Days and Fund Flows

Bottom Line: Over the past month, there has not been any “bad” days. Historically, this is better for the long-term than short-term.

I’ve been asked several times over the past week to look at how the market performed after seeing as many “distribution” days as we have seen lately. The concept of a distribution day is as old as technical analysis, and is popularly defined as a day that closed lower than the day before, and on higher volume. The most well-known proponent of distribution and accumulation days is the newspaper Investor’s Business Daily, which makes such days a cornerstone of their analysis.

I have never been comfortable with the assumption that volume just needs to be higher than the previous day in order to be considered a distribution day. That is too easy a hurdle. For our purposes, I prefer to use whether volume on the NYSE was more than 10% greater than its 50-day average. Over the past 21 days (the length of an average month), I cannot find even one distribution day using these criteria. You may be able to, depending on where you get your volume data, but for testing purposes it doesn’t much matter as long as you use a consistent source over time.

Not only do I not show any distribution days, but I also do not show any days where the market rallied on poor volume (“poor” being defined as volume less than 90% of the 50-day average). So, over the past month, the market theoretically has done nothing wrong – no dips on rising volume, and no rallies on declining volume (again, this is according to these criteria only).

The chart below shows a plot of the number of “bad” days seen over the previous 21 trading days. For these purposes, a “bad” day is considered any day where the S&P 500 closed lower than the previous day, and volume was at least 10% greater than average share volume over the past 50 days. Also, I have included in the definition of “bad” day any day where the S&P closed higher than the previous day, but on volume that was at least 10% lower than the 50-day average.

Not surprisingly, over the past few years it has been rare to see a 21-day stretch go by without any bad days. Even going back to 1997, there have been only 33 days which fit the criteria. On the other hand, there have been 92 instances of a large number of bad days, which I defined as having at least 10 bad days scattered over the past month.

So does all this mean anything? I think it does. Namely, after periods where there were no bad days over the past month, the S&P typically underperformed those times when there were many bad days, at least in the short-term. However, “good” periods outperformed “bad” periods by a large margin when looking out 30 to 60 days.

	5 Days Later	10 Days Later	30 Days Later	60 Days Later
Any time there were NO bad days over the past month…
Avg Ret	0.3%	1.6%	4.7%	8.2%
% Pos	50%	74%	88%	88%
Any time there were at least 10 bad days over the past month…
Avg Ret	1.3%	2.3%	2.7%	3.7%
% Pos	70%	83%	66%	63%

From the table above, we can see that after 5 days, periods of “bad” days outperformed periods of “good” days by a relatively large amount. The S&P was higher an additional 20% of the time, and with an average gain 1% higher. However, look at 60 days later – this is where good markets made up for themselves. After stretches of zero bad days, the S&P was higher 60 days later 88% of the time, and with an impressive average gain of 8.2%. On the other hand, when the market had suffered at least 10 bad days over the previous month, 60 days later the S&P was higher 63% of the time, and with an average gain of 3.7%. That’s certainly still respectable, but good markets far outshone the bad. To see if this relationship holds up over time, or whether it is just a recent phenomenon, I went back and checked these same criteria since 1964. After stretches of zero bad days, the S&P was still relatively weak up to 10 days later. However, after 30 days the positive bias returned, and far outweighed those periods where there were at least 10 bad days during a 21-day stretch.

The fact that the broader market has done nothing wrong (according to the criteria above) suggests that while gains over the next couple of weeks may be substandard, over the intermediate- to long-term, it is a positive indication of a healthy market. Since 1964, S&P 500 was higher 76% of the time after 60 days, and with an average gain of 4.5%.

Bottom Line: Investors poured near-record amounts into stock mutual funds in January, which has negative connotations. However, in this case it may not be that clear-cut.

On Friday, the Investment Company Institute came out with its mutual fund reporting data for the month of January. The amount of assets held in liquid assets, which I went over last week, rose from 4.3% to 4.4%. Adjusted for the prevailing level of short-term interest rates, we can conclude that mutual funds were in a “cash deficit” of about 0.5% at the end of the month - nothing extraordinary (see last week’s comment for background information on this). There was something interesting in the data, however, and it contains both good news and bad news. Let’s get the bad news out of the way first.

Stock mutual funds took in $43.8 billion in January, a tremendous surge that has been exceeded only twice in the past 20 years. Those two times were January and February 2000 – near the peak of the bubble. The influx is not terribly surprising given the blanket of advertising we’ve seen touting the great returns last year, as funds try to capitalize on the assumption that we’re back to the good ‘ol days. But we all know by now that when investors base their investment decisions on advertisements and brokerage statement inserts, it is usually very late in the trend. Let’s look at how the S&P 500 performed 1, 3, 6, 9, and 12 months after the 10 largest stock fund inflows and outflows.

10 Largest Stock Mutual Fund OUTFLOWS
MONTH	OUTFLOW ($ billion)	1	3	6	9	12
Jul 2002	-52,628	0.5%	-2.8%	-6.1%	0.6%	8.6%
Sep 2001	-29,509	1.8%	10.3%	10.2%	-4.9%	-21.7%
Mar 2001	-20,600	7.7%	5.5%	-10.3%	-1.1%	-1.1%
Jun 2002	-18,045	-7.9%	-17.6%	-11.1%	-14.3%	-1.5%
Sep 2002	-16,078	8.6%	7.9%	4.0%	19.5%	22.2%
Aug 1998	-11,202	6.2%	21.6%	29.4%	36.0%	37.9%
Feb 2003	-11,105	0.8%	14.6%	19.8%	25.8%
Dec 2002	-7,747	-2.7%	-3.6%	10.8%	13.2%	26.4%
Oct 2002	-7,709	5.7%	-3.4%	3.5%	11.8%	18.6%
Oct 1987	-7,476	-8.5%	2.1%	3.8%	8.0%	10.8%
AVERAGE		1.2%	3.4%	5.4%	9.5%	11.1%

10 Largest Stock Mutual Fund INFLOWS
MONTH	INFLOW ($ billion)	1	3	6	9	12
Feb 2000	53,555	9.7%	4.0%	11.1%	-3.8%	-9.3%
Jan 2004	43,758	N/A
Jan 2000	39,980	-2.0%	4.2%	2.6%	2.5%	-2.0%
Mar 2000	35,603	-3.1%	-2.9%	-4.1%	-11.9%	-22.6%
Apr 2000	33,769	-2.2%	-1.5%	-1.6%	-6.0%	-14.0%
Mar 2002	29,275	-6.1%	-13.7%	-28.9%	-23.3%	-26.1%
Jan 1996	28,665	0.7%	2.9%	0.6%	10.9%	23.6%
Jan 1997	28,516	0.6%	1.9%	21.4%	16.3%	24.7%
Apr 1998	26,555	-1.9%	0.8%	-1.2%	15.1%	20.1%
Jul 1997	26,505	-5.7%	-4.2%	2.7%	16.5%	17.4%
Apr 1996	25,543	2.3%	-2.2%	7.8%	20.2%	22.5%
AVERAGE		-0.8%	-1.1%	1.0%	3.7%	3.4%

It is clear from the chart and the tables that by the time investors (as a group) panic enough to actually pull money OUT of stock mutual funds, it has often been an excellent buy signal, or at least a heads-up that we may be seeing the latter portion of the downtrend that is usually in place. On the other hand, large inflows to stock funds have preceded some shorter-term difficulties for the market. With the current inflow the third-largest on record, it suggests that we have some work to do before assuming the broader market will take off from here.

That was the bad news - now let’s look at the good news. Mutual fund assets have grown steadily over the past 20 years. A combination of good returns, demographics, marketing, and a hundred other factors have made mutual funds a staple of American investing. Funds accounted for approximately 7% of total household assets in 1990, but make up closer to 18% today. That growth in fund assets makes looking at absolute inflows and outflows misleading. A monthly inflow of $30 billion is lot different when there is a base of $4 trillion (such as in the year 2000) than when there is a base of $200 billion (such as 1987).

So let’s look at the same monthly inflows and outflows as above, except this time expressed as a percentage of total stock fund assets:

Obviously, this changes the landscape. We can see what a huge impact the crash of 1987 had on investor psychology, as over 4% of total fund assets were pulled out in a single month. No other month has come close to even half that – July 2002 is the closest, and not even 2% of total assets were pulled out that month. January’s inflow of $43.8 billion, while huge on an absolute basis, on a relative basis is only 1.15% of total stock fund assets. This is still the largest relative inflow since February 2000, and July 1997 before that, but it is not nearly as large as other readings we saw regularly in the 1980’s and early 1990’s.

I want to show one other look at this data, and this is the difference between the year-over-year growth in total stock fund assets and the year-over-year growth in the S&P 500. Assets have been climbing steadily over the past 20 years, along with the market, so looking at the absolute growth in assets is kind of a moot point. By taking the difference between growth in fund assets and growth in a broad market benchmark, we can get an idea of whether investors are more or less enthusiastic than they “should” be given the performance of the market.

From January 2003 to January 2004, the S&P rose 32%, while total assets in stock mutual funds grew 46%. This difference of 14% is about average over the 20-year period, so this measure as well does not appear to support the idea that investors are more enthusiastic than they should be.

It is tempting to want to bet against this market. We have seen numerous negative breadth divergences on recent attempts towards the highs, and the broader market has been struggling to rally off the extreme oversold short-term sentiment conditions I mentioned earlier in the week. Both of those points are useful and consistent indications that the larger trend may be changing.

So why not just get massively short right now? Because we saw the same types of divergences many times in 1997, 1998 and 1999 and it would have been difficult (to say the least) to make money on the short side for anything other than a short-term trade. Divergences are just theory until there is some confirmation from price itself. If we break below last week’s low, that will give most traders confirmation enough that the divergences are valid at it should bring in more selling pressure. Like most everyone else, I am watching 1450 on the NDX and 1135 initially then 1125 most importantly on the S&P 500. If those indices slice through those levels, the price structure will have changed from what we have seen previously during the uptrend. Like I showed in April for the market changing to a larger uptrend, this would indicate to me that we are now in a downtrend where the short side is favored and bouts of enthusiasm should be faded. Unless and until that happens, I continue to slightly favor the long side for the short-term. Most of our measures are relatively neutral, but seasonality for the next few days is quite positive. I’ve said numerous times before that I consider seasonality nothing more than a gentle breeze for or against you, and it’s not enough of a reason on which to base trades (except perhaps for a few days each year), but for the next few days it will be pushing at the back of the long side.

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.