Rogue Missives
Rogue Missives
| ROGUE ARCHIVES | |
Indexing Inefficiencies Investing in a mutual fund that simply tracks the Standard & Poor's 500 Index has become something of a no-brainer. With large cap issues in the U.S. doing well in 1996, the performance of the S&P 500 index outdistanced that of roughly 90% of all mutual funds by racking up a total return of 22.95%. Vanguard Group's Vanguard Index Trust 500 Portfolio, the index fund of choice, returned 22.86%, essentially mirroring the S&P. Why would anyone want to pay hefty management fees to underperform a market index that did so well and cost so little to own? This if-you-can't-beat-'em-join-'em attitude has begun to have a significant impact on professional money management in the U.S. On the one hand, investors, led by aging baby boomers, are now pouring more money into the market than ever before (in dollar terms), and an increasing percent of that money is flowing into index funds. Money under management at Vanguard soared last year, rising from $17.4 billion at the end of 1995 to $30.3 billion by this past December. Companies that once specialized in actively managed funds are now anxious to get a piece of this increasing action. For example, Fidelity Investments, whose Magellan fund was once home to premiere stock-picker Peter Lynch, has thrown in the towel and started offering new index funds. Also, active fund managers, who measure their success or failure against the yardstick of the S&P 500, have been pressured into mimicking the index to some degree. As Barron's columnist Andrew Barry suggested two weeks ago, the advent of a new "Nifty Fifty" of brand name stocks such as COCA-COLA <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: KO)") else Response.Write("(NYSE: KO)") end if %>, MERCK <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: MRK)") else Response.Write("(NYSE: MRK)") end if %>, and MICROSOFT <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: MSFT)") else Response.Write("(Nasdaq: MSFT)") end if %> may be the result of fund managers looking to pick the workhorses of the S&P. That's a strategy with some logic to it, since the S&P 500's strong performance owes much to just a few dozen stocks. About 15% of the index's surge this past January was due to a 23% increase in INTEL <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: INTC)") else Response.Write("(Nasdaq: INTC)") end if %> and Microsoft alone. A sizeable chunk of the S&P 500 is wildly underperforming these stand-out stocks.
On the other hand, at least some of these Nifty Fifty stocks have seemed so richly valued already that fund managers have kept their distance. Barry cited data showing that whereas 20% of the shares of your average S&P firm are owned by mutual funds, some consumer knockouts are seriously underrepresented in those funds. Mutual funds own just 4% of Coca-Cola, 5% of Procter & Gamble, 7% of General Electric, and 9% of Walt Disney. As funds flock to such behemoths to capture the performance they're missing elsewhere, the valuations are sure to get stretched further.
Index funds such as Vanguard's are premised on the efficient market theory, which in its semi-strong version suggests that stockpickers cannot outperform the market because all of the information that they could use to select one stock over another has already been figured into the stock prices. Vanguard's thrashing of active fund competitors appears to validate the efficient market theory.
The irony here is two-fold, however. First, both the rush to invest in index funds and, particularly, the rush to put money on the real workhorses of the index may be promoting new market inefficiencies. However that may be, such widespread tracking of an index that undergoes minor changes every year has definitely created measurable short-term market inefficiencies that can be predicted and traded. In this case, the individual investor still has an advantage over the active fund manager who remains constrained by the sheer size of his or her portfolio.
The chief point to realize is that stocks are added to and dropped from the S&P 500 every year. During 1995, for example, 33 stocks were tossed from the index with 33 new ones filling their spots. Last year, 24 stocks were replaced. Mergers or spin-offs of large subsidiaries contributed to many companies being dropped from the index, but many others were "clean" additions and subtractions.
Until October of 1989, Standard & Poor's Equity Services Group announced a change in the index after market close one day, and the substitution took effect the next day. As might be expected, stocks added to the index went up the next day (about 3%) and stocks dropped from the list fell, all on very heavy volume. Standard & Poor's decided that giving fund managers some advance notice might help even out trading in these stocks. Since 1989, S&P has typically announced a change in the index after the market has closed, and the change takes effect about a week later.
Still, imagine what happens when these changes occur. The $30 billion Vanguard fund, as well as all of the other new index funds, need to sell one issue and buy another, all within a matter of a few days. Some rather elementary rules of supply and demand come into play.
In a paper completed last summer, economists Anthony W. Lynch of New York University and Richard R. Mendenhall of the University of Notre Dame studied short-term pricing associated with such changes to the S&P 500 index. They found that from October of 1989 through mid-year 1995, the price movements have become even more pronounced that under the old announcement system.
Lynch and Mendenhall concentrated on "clean" data derived from just those 71 issues not associated with a merger or spin-off. They found that stocks being added to the index jumped about 3.2% on the day following the announcement and an additional 3.8% in the ensuing week before the change actually occurred. Changes take effect after market close on a given day, and trading on this change day typically leads to price and volume spikes, as the last fund managers rush to take their positions.
With this significant new demand satiated, the price of a stock added to the index typically drops about 0.75% the next day, with a further decline of about 1.25% over the next week. In effect, indexing itself promotes new and obvious inefficiencies as fund managers willingly pay a premium just to be able to add the stock to their portfolio prior to its addition to the index. Also, that influx of new long-term holders also effectively reduces the float of shares outstanding, which perhaps explains the fact that newly added shares end up trading 5% above those levels seen just prior to the S&P's announcement.
The effect on stocks dropped from the index are even more significant. Lynch and Mendenhall found on the first trading day after an announced deletion, stocks dropped about 6.26%, with a further decline of 12.69% in the following week before the change actually took place. With unusual selling pressure then out of the market, these stocks experienced a mild reversal of a few percentage points in the next few days. Nonetheless, "a significant permanent price effect is associated with deletion from the S&P 500." These economists pinned that permanent drop at 14.1% to 15.7%, depending on when they closed the study period.
Lynch and Mendenhall estimated that about 16% of the shares of the companies comprising the S&P 500 are held by index funds. An enormous influx of money into index funds since their study was completed, however, would suggest that these price effects should be even more pronounced.
A very unscientific look at recent additions and deletions suggest that might be the case. Rogue looked at recent changes to the S&P 500 index (8 stocks added and 8 stocks deleted) that met the "clean" criteria. A number of admittedly unrealistic assumptions were made regarding an individual's ability to trade these stocks during these readily defined periods of transition.
For example, we assumed that for a stock being added to the index, an investor could buy those shares at the low registered the day after the announcement was made and sell them at the high the day before the change officially took effect. Those are the parameters that ensure the maximum return. Similarly, we assumed that for a stock being dropped from the index, an investor could short the stock at the high registered the day after the announced change and cover at the low on the day after the change took place (though the change day low also looked attractive). We also assumed no commission charges had to be shelled out and that an investor could have reinvested the accumulating assets into the next S&P change stock. (This would have been impossible since some changes occurred simultaneously.)
Even with these significant caveats, the results are interesting. From August 1996 through January 1996, an investor could have traded the S&P additions as follows: HFS <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: HFS)") else Response.Write("(NYSE: HFS)") end if %> up 2.13%, SEAGATE <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: SEG)") else Response.Write("(NYSE: SEG)") end if %> up 4.45%, DELL <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: DELL)") else Response.Write("(Nasdaq: DELL)") end if %> up 4.87%, UNION PACIFIC RESOURCES <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: UPR)") else Response.Write("(NYSE: UPR)") end if %> flat, MBIA <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: MBI)") else Response.Write("(NYSE: MBI)") end if %> up 2.21%, AUTOZONE <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: AZO)") else Response.Write("(NYSE: AZO)") end if %> up 15.79%, FRONTIER CORPORATION <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: FRO)") else Response.Write("(NYSE: FRO)") end if %> up 8.93%, and THERMO ELECTRON CORPORATION <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: TMO)") else Response.Write("(NYSE: TMO)") end if %> up 25.47%.
The average long position would have offered a maximum return of nearly 8% in a matter of just 6 trading days. If an investor initially had put $10,000 into this trading strategy and sequentially rolled over the funds, this strategy would have resulted in trading profits of $8, 095, or an 81% cumulative return in about 6 months.
Shorting the deleted stocks in a similar fashion would have produced even better results, with trades in OGDEN <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: OG)") else Response.Write("(NYSE: OG)") end if %> gaining 2.63%, OUTBOARD MARINE CORPORATION <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: OM)") else Response.Write("(NYSE: OM)") end if %> gaining 8.46%, COMMUNITY PSYCHIATRIC CENTERS <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: CMY)") else Response.Write("(NYSE: CMY)") end if %> losing 1.45%, YELLOW CORPORATION <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: YELL)") else Response.Write("(Nasdaq: YELL)") end if %> gaining 11.1%, CONSOLIDATED FREIGHTWAYS <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: CNF)") else Response.Write("(NYSE: CNF)") end if %> gaining 16.66%, LUBY'S CAFETERIA <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: LUB)") else Response.Write("(NYSE: LUB)") end if %> gaining 10.67%, RYAN'S FAMILY STEAKHOUSE <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: RYAN)") else Response.Write("(Nasdaq: RYAN)") end if %> gaining 12.5%, and SHONEY'S <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: SHN)") else Response.Write("(NYSE: SHN)") end if %> gaining 11.29%.
The average short here would have led to a maximum average return of about 9% in 7.5 trading days. Sequentially investing $10,000 using this strategy would have led to $9,700 in profits, or a 97% cumulative return in just 6 months.
Though there were certainly exceptions, playing the post-change reversal would also have been profitable in many cases. For example, an investor could have bought AutoZone at $23 3/4 the day after the announcement that it was being added to the S&P. She then could have sold it at $27 1/2 the day before that the change took effect, shorting it at the same time. Five trading days later, the stock closed at $20 3/4. Other stocks added to or dropped from the index had less significant but still profitable reversals. Consolidated Freightways, for example, recovered from a deletion-day low of $20 5/8 to $22 3/4 four days later.
Whether an actual investor would prove so successful in trading such changes is, of course, unlikely. And whether a Foolish investor should even try should be obvious. Finding strong companies is simply easier than trading on temporary price shifts. The point, though, is that given the massive quantities of money flowing into index funds, trading these transition periods is not simply a matter of market timing. These are periods when particular stocks suffer from predictable price inefficiencies.
Moreover, the more recent additions and deletions show the largest percentage price shifts. Without comparing such changes to overall market returns during the transition period (as Lynch and Mendenhall did for their calculations), it's impossible to say that these dramatic shifts follow from the increased power of indexing. On the other hand, the economists relied on overall price changes, not tradeable changes as in the above examples. Such double-digit price changes are comparable to many fund managers' annual returns. In any case, it's difficult to argue with the evidence that the extraordinary popularity of indexing, which is itself premised on the efficient market theory, is actually introducing some curious new inefficiencies to the market. Whether those inefficiencies are any cause for concern is debatable. At the very least, they do present predictable blips that promise nimble traders some potentially huge returns. |
|
