back start next


[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [ 96 ] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205]


96

In his book Event Tradmg, Warwidc shows the test results of some of the major maikets, allowing us to compare the effects of the sj-stematic reaction of U.S. bonds with that of the S&P 500 and other markets. Unexpected news that causes bond prices to rally sharply (causing rates to decline) would result in a rise in the stock market, unless the news was particularly bad for the long-term economy of the country. Therefore, a drop in the consumption of durable goods should be followed by lower interest rates and a slightly lagged upward response in the stock market This pattern is shown in Table I4-I where the bond and S&P results are side-by-side.

In general, we expect a price shock to impact interest rates first, then currency and equities markets. Tjpically, the interest rates react in a way to compensate for the economic effects and stabilize the other two markets; this preempts the action that would be expected from the Central Bank. The exception is when the news is much more extreme than usual, and it affects the stock market first. Investors, because of a lad; of confidence, will shift their ftmds from equities to interest rates in an effort to find a safe haven during uncertainty. This will cause a decline in the stock market and a corresponding rise in interest rate prices (lower yields), although normally the two will move in the same direction.

In Table I4-I we can see the way the markets move after both the interest rates and stock maiket initially react with an upward price breakout. When both markets continue in the same direction, such as with retail sales, GDP, and employment data, then the ftill impact of the data takes longer to be assessed by the market, or there is expectation of

6AnioldL.-ir., MeiTmmeni fa I. . 1 " in FumresPncine,Food I ese.-iTcli Institute £boieE 1 (no 3, November 1960, referenced in Warwick. Event Trading.

TABLE I4-I Results of Upward Breakout of U.S. Bonds (left) and S&P 500 (right) for the Period I989I994

CwilMence

Ptnod

Lent

hnod

Dunblc goods

Retail lals

Low confide

NAPM

some continued response by the govemment. The PPI and the National Association of Purchasing Managers (NAPM) do not show this consistency, indicating that the market takes this infonnation differently. When the pattern of confidence is low, the market may do a good job of reaching the best price level in immediate response to the news; therefore, there is no consistency in the price move that follows.

Table 14-2 shows the side-by-side test results of reactions that cause a drop in bond prices and a corresponding drop in the S&P 500. The frequency of low-confidence results may represent the conflict between the normal upward bias of the stock maiket or the unusually bullish trend during the test period. The GDP results, which show the most confldence, indicate that the stock market moves higher even when the interest rates rise in reaction to positive GDP data. Trade balance data, which is not alwajs the primarj focus of the market, has a very consistent pattem, but is not clearly related to the interest rate move as are most of the other statistics.

Although this work concenfrates on the short-term daily reaction to unexpected news, the long-term trend should not be overiooked. Economic data can eshibit consistency over long time periods, and the response by the Cenfral Bank is usually moderate but consistent. If the CPl shows early signs of inflation, then the monetarj authority will push rates up slitly; if this doesnt work, as seen by the next series of economic reports, they will move rates slightly higher again. This pattem allows for an underljing frend that can be used to filter the direction of trades and add confldence to the results.

Reaction to Unemployment Reports

Maikets treat unemployment data seriously; because it is a direct reflection of how the population is faring, the govemment is sensitive to unespected changes in the number of persons filing for unemployment benefits. Using LIM, it was found that when the maiket underestimated the unemployment figure by at least 0.2" o, there was an average return of 4.500 in the S&P 500 during the following 27 dajs from the close of the report day. Simiilarly, U.S. 30-year



Treasury bonds showed an average (from a long position) in nearly all cases of 3 2° during the period from 3 to 28 dajs following the report.

Biiiik.-, "Event-based Analysis," Futures (Afnl 1995

TABLE 14-2 Results of Downward Breakout of U.S. Bonds (left) and S&P 500 (right) for the Period 1989-1994

US. Bonds

S&P 500

Holding

Buyer

Con/idence

Holding

Buy or

Confidence

Period

Sell

level

Period

level

Durable goods

Low confidence

Retail sales

Low confidence

Retail sales

Low confidence

Low confidence

Low confidence

Low confidence

Low confidence

NAPM

Low confidence

Low confidents

Industrial production

Low confidence

>99

Industrial producuon

>99

>99

Low confidence

Trade balance

>99

Presidential Election Cjcle

of all the events that move the marfcet, the Presidential elections have been the most consistent. The pattems stem from the motivation of the incumbent partj to provide good economic news to the voters prior to the election year, and as far into the election year as possible. Stock marfcet action during the election year is ahvajs more erratic as parties battle over the value of each others actions.

Studies of the 4-year pattems have all yielded very similar results to those shown in Table 14-3. The year preceding the election (jear 3) posts the strongest gains for the maifcet, followed by a reasonably strong election year The 2 years after the election show retums below average as the reality of politics reasserts itself

There is the additional possibility that there is an 8-year cycle that should be watched; however, the 8-year period should be most informative if it represented only those years in which the same President was in office. Actions by a President who cannot be reelected are likely to be different from one who seeks another term; therefore, we should expect a different pattem. This can be made more infricate by studying the pattems preceding and following a change of parij, all of which have a ftindamental basis in the behavior of the politick parties and the voters.

More sophisticated computer software, such as provided by Logical Information Machines, a Chicago firm, can produce a very interesting, closer view of how voters respond to election politics. Table 14-4 breaks the election year into periods between the key events for those years in which the stock marfcet began the election year within 8°o of its 2-year high price (dajs refer to business days):

1. The retums of the year preceding the election year

2. The first 10 dajs of the new year, tjpically a strong period (dajs 0-10)

3. Through the State-of-the-Union address and the primaries (dajs 10-83)

4. Waiting for the conventions (dajs 83-161)

5. Preelection blahs-the actual campaign (dajs 161-195)

6. The election to year-end reaction (dajs 195-253)

7. Combined periods (2) + (4) + (6)



Combinmg the three periods (2), (4), and (6), whidi have strong upward biases, grves consistently positive results. Even if the newly elected partj fads to deliver on their campaign promises, traders could have alreadj converted those marketing gimmicks into stock market profits.

ArtclesbvA. 11 lite, "Tlie Eight-Year iTesi itiBlElectio iTesrleiitial Election ?7cle," Teciniical Analysis -i Lb-ck: I (Fettuarv 1991, all sh-wverv similar results for tlie 4ear ele

Pattern," Tecimical Analys Oiimodities fMircl,

auber 19 ), A IJemll, The

TABLE 14-3 The Presidential Election Cycle, 1912-1992, Based on the Percentage Retums of the Dowjones Industrial Averages

Preelection year

11.0%

Election year

Postelection year

Midterm year

Average year

6.3%

Sowre: Adam White,

ion Year Anal;

/sisforYe.-ir.-i

m Wluclitli.

e i,cklJ.irketE«% tlie Year wittui

ne>b"ftlielrevi

onsTw-YearHi

2 First

PrewDus

2 weeks

Primories

PreetecDon

to - }

Yeor

Year

(IlO)

f/0-83)

3-16 f)

f(6(-(95}

(f9S-2S3t

(h + l3i+l5i

1936

41.82

4.64

1191

-0 62

5.85

20.52

1944

19 45

t.63

-.84

9.27

3 14

14.04

1952

16 15

-2.82

8.02

-249

5.47

15 10

1956

2725

-1.78

5 42

3 16

-6 65

3 76

I960

-2.49

-5 75

-7 13

1964

18.89

3 03

0.06

1968

2003

-0 10

4 39

1972

10.82

1.41

3 39

5.30

-1 78

4.88

11.59

1980

1231

2.26

-5.42

18.57

-0 20

7.66

2849

1984

17 53

-501

0 63

6 17

1992

26 30

-2.72

-C23

5.87

Average

19-91

0.86

-044

6 37

-1 15

4.43

1166

commMENT of traders report

Drawing from more than 20 years of experience in analyzing the CFTCs Commitment of Traders Report, William L. jiler described a supplementarj approach to identifying a major trend in his 1985 CRB Commodity, Yearbook." With his usual thoroughness and clarity, jiler presents material that had previously been unsuccessfully interpreted.

Originally released the 1 Ith day of each month (biweekly since 1992), the Commitment of Traders Report summarizes the positions of reporting and nonreporting traders as of the last day of the previous month. Reported positions are those exceeding a minimum determined for each futures market (e.g., -50,000 bushels for com, 500.000 for soybeans. 500 conhacts for 30jear Treasurj bonds, and 300 contracts for the s&p 500). The report subdivides the open interest into positions held by hedgers (commercials) and speculators. By subtracting the reported positions from the total open interest, the total positions of small hedgers and speculators can be found. For grain reports (Figure 1-1-2), positions are further divided into old crop and other positions, which include new crop and intercrop spreading. To analyze the shifts in position, jiler has compiled tables of normal pattems. similar to a seasonal shidy (Figure 14-3), When the open interest of one group is significantly greater than their normal holdings, they express a definite opinion on the direction of the maiket- By tracking these changes for many years and observing the corresponding price



[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [ 96 ] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205]