d. Nonmembers' short sales (1) The round-lot customer As noted previously, round-lot short selling by nonmembers customarily accounts for less than half of all short sales, but this proportion tends to increase during a sharp decline. Considering the nature of short selling itself, it is probable that nonmembers' short selling consists mostly of speculation for the decline. As indicated in the NYSE's June 1947 study of short positions 531 technical short selling, including short positions "against the box" and hedging, was only about one-quarter of total nonmember short selling. Ordinarily, nonmembers' short selling in the aggregate is small compared with total sales by nonmembers, especially toward the end of bull markets when the ratio tends to fall below 1 percent. The ratio has risen to around 2 to 3 percent in market drops. Indeed, during the critical decline in May 1962, the ratio rose to more than 4 percent and almost to 7 percent during the further decline in June 1962. (2) The odd-lot customer Chart VI-m pictures the odd-lot short selling in total shares for each month since January 1958. The chart indicates that most of the short selling by odd-lot customers occurs during declines in the market and that peaks in such selling coincide rather closely with the lowest point of the decline. For example, the peaks in October 1959 and October 1960 both occurred at the low points of declines. Moreover, the chart throws into bold relief the relatively large amount of short selling that odd-lot customers effected during the sharp decline in May 1962. As the market dropped from 671 (DJI) on May 4 to 561 on June 29, the volume of short sales in odd lots progressed in weekly totals as follows: 40,000 shares; 46,000; 60,000; 92,000; 92,000; 112,000; 112,000; 153,000; 113,000; and 120,000. It is necessary to go back to 1938 to find comparable volume. The May-June 1962, volume figures were 4 to 8 times larger than comparable figures for the November-December 1961 period, when the market was near its peak. The odd-lot short seller has at least one advantage over the roundlot short seller, which takes on added significance in a declining market. While both must sell on up-ticks,532 the short seller in the roundlot market must take his turn in finding a buyer, whereas the odd-lot short sale must be taken by the dealer, no matter how many shares there may be in the aggregate at a particular price level, whenever the order(s) can be executed on the next appropriate round-lot sale. It is not surprising, therefore, that during drastic declines, as in May 1962, odd-lot short sales tend to increase in relation to short selling in general, as well as in relation to total odd-lot sales. A comparison between odd-lot short sales and total round-lot short sales may not be too meaningful, since, as has been seen, 40 to 70 percent of all short selling is effected by specialists. Chart VI-n, there 531 See sec. 1.c(1), above. 633 The restriction as to price on odd lots is by Exchange rule; the odd lot is exempt under SEC rules. Chart VI-n ODD-LOT CUSTOMERS' SHORT SALES ON N.Y.S.E. as Percent of Total Round-Lot Short Sales and as Percent of Non-Member Round-Lot Short Sales fore, compares odd-lot short sales to both total round-lot short sales and to short sales by nonmembers alone. Short sales in odd lots are not large on an aggregate basis in either comparison. However, oddlot short selling, which during periods of rising prices often amounts to no more than 1 percent of total round-lot short selling, rises to 4 percent during pronounced dips in the market and rose to more than 6 percent during the decline in May-June 1962. As compared with nonmembers' round-lot short selling, odd-lot short sales varied from less than 4 percent to more than 16 percent, the same proportionate range. The ratio of odd-lot short selling to total odd-lot sales increased (in the 412-year period shown) from a minimum level of less than 1 percent (during well established upward trends in stock prices) to above 2 percent during declines and, on the May 1962 precipitous fall, to over 6 percent (chart VI-o). 3. THE STOCKS IN WHICH SHORT SELLING OCCURS As mentioned, reports issued each month by the New York Stock Exchange provide aggregate figures in shares of the short interest in stocks, the number of issues in which any short interest was reported by member firms, and a list of the individual stocks in which the short interest amounted to 5,000 shares or more or changed from the previous month by 2,000 shares or more. This section presents an analysis of these reports in an effort to determine the extent to which short selling is concentrated in particular issues and the characteristics of such issues. a. Number of stocks in which a short interest is reported Generally about 60 to 65 percent of the stocks listed on the New York Stock Exchange 533 have a reported short interest. The ratio has varied in recent years from a low of 55 percent in October 1956 to a high of about 70 percent in April 1961. The number of stocks in which there was a short position, and the ratio of such issues to the total number listed, are shown in charts VI-p and VI-q, respectively. Chart VI-p indicates a rather wide variation in the number of issues with a short position. To adjust for the increase in the number of stocks listed during the period, the figures are expressed as a percent of the total listed issues for the corresponding month (chart VI-q). In contrast with some of the preceding charts in which short selling was shown to reach a peak at or near the lowest point of a decline, charts VI-p and VI-q indicate that short selling tends to spread to a greater number of issues after the lowest point and when the market has progressed well onto a new upward trend. In fact, during the three instances shown in the charts, the peaks occurred from 6 to 8 months after the new upward trend started. b. The concentration of stocks in which short sales occur Generally, the NYSE's monthly report of stocks in which there is a short position of 5,000 shares or more, or a change of 2,000 shares or more from the preceding month, consists of less than 200 issues, ranging from 126 in July 1956 to 188 in May 1962; in November 1962 The number of listed stocks has exceeded 1,500 since 1952, rising from 1,522 that year to 1,558 in November 1962. |