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]


7 Seasonality

Chapter 3 introduced prices as a time series and identified its four components as the trend, the seasonal pattem, the cycle, and the chance movement- it included -various ways of finding the trend using statistical analjsis and forecasting techniques. Chapter 4 then showed various wajs to calculate trends. Of all techniques, the trend is overwhelmingly the most popular foundation for trading sjstems. In this and Chapter 8 „We will turn our attention to two other principal components, the seasonal and Cjclic movements.

Seasonality is a cycle that occurs yearly, it is most often associated with the planting and harvesting of crops, which can directly affect the feeding and marketing of livestock. -Normally, prices are higher when a product is not as readily available, or when there is a greater demand relative to the simply as often occurs with food during the winter months. For grain, the season is dominated by planting, harvesting, and weather-related events that occur in between. Most crops have been produced in the Northem Hemisphere, but South -\me rican soybeans and orange juice have become significant factors since the early 1980s, as have Australian and New Zealand beef and lamb, resulting in a structural change in seasonal pattems. Globalization has not only affected financial mariets, but nearly everj-thing we purchase.

Consumer habits can cause a seasonal pattem in metals and stocks as Weather does for agricultural products. Passenger airiine traffic is much heavier in the summer than in the winter, and profits of that industrj, especially companies not diversified, reflect seasonality. Eastman Koolak once had a classic pattern caused by much more active picture taking during the summer months, when most workers take vacations. This seasonal activity also increased demand on silver, a major ingredient of film coating, which developed a codependent pattem before a substitute was found.


Even when the impact of seasonality on agricultural products is not clear from the price pattems, it is still there. Consider the following factors that do not change:

1. More com, wheat, and soybeans are sold during harvest than at any other time of year because there is not enough available storage to hold all of the new crop. Rental storage, when available, requires a minimum three-month chaise. Lad; of storage and the need for immediate income result in greater sales and cause prices to decline.

2.Because feedgrains are harvested only once each year, forward confracts include a storage cost as part of the total carrjing chaise. Therefore, each forward delivery price should be higher within the same crop year.

Sometimes the price pattem of forward months does not seem to reflect the added costs of carrj. Occasionally these markets will even invert and the nearest delivery will trade at a price higher than the deferred months, a situation familiar to crude oil and copper. The cost of carrj, however, still exists in an inverted or backwardation martlet. Rxtreme shorterm demand pushes the nearest delivery much higher, while the events causing price disruption are expected to be tenporarj The normal carrj is still there, it is just overw helmed by temporarj demand.

It is important to be able to identify seasonal pattems. The methods for finding ihem are simple, and made more so by the use of a spreaddieet and a computer. These will be discussed in this chapter along with some practical applications.


Seasonal patterns are easier to find than the longer-term cycles or economic frends, because they must repeat each calendar year. Although any 12-month period can he used to find the seasonal pattem, academic studies usually begin with the new crop year for grains, just following harvest, when prices tend to be lowest. This approach will make the carrjing chaises, which increase steadily throughout the new crop year, more apparent. For uniformity, the exanples in this chapter will alwajs begin with a calendar year, which assumes no knowlete of where the season starts, and can be equally applied to stocks, it will alwajs include carrjing chaises as an integral part of the markel price.

United States agricultural production is considered to be the standard for seasonal, even though a wheat crop is harvested continuously throughout the year in different parts of the world. Prices are expected to he lower during the

us. harvest and highest during the middle of the growing season. The influence of world stocks and anticipated harvest from other major producers in South America or Russia will cause an overall dampening or inflating of prices, rather than change the seasonal pattern. While the costs of transporting grain to the United States is not practical, except for special products, each purchaser in the world maitet will select a supplier at the best price. This fiingibility has a direct impact on the US. maitet, which must alter its local price based on export demand.

Industrial commodities have seasonal price variation based on demand. Silver, although increasingly used in electronics as a conductor, is also consumed for jewelry, and photography, and has served as a general hedge against inflation. It still shows a seasonal pattem of greater demand during the summer months. Almost half of all copper is used in electrical and heat conductivity, with much of it in the form of an alloy with nickel and silver, its seasonality is heavily related to the housing industrj, where it is required for both electrical and water sjstems. New sources of ore are introduced infrequently, and the possibility of discovery or expansion is rarely seen in price movement as short-term anticipation. The primarj supply problems in copper are related to labor as well as social and political ;s in producing counfries.

There are many businesses with finished products that have seasonal demand and their publicly traded stock prices will reflect that tendency- Because the shares in a company are far removed from bujing and selling the raw materials that they use, even major oil companies, such as Exxon, may not show a seasonal pattem similar to crude oil or its refined products. As with Eastman Kodak, these firms have thoroughly diversified, and the impact of a $ 1.00 increase in a barrel of oil or a $50 increase in an ounce of silver will have only a small effect on the profitability of the firm, and may be the cause of stock prices advancing or declining. Yet some industries, such as airiines, still show seasonal pattems, and the same procedures given here can be used to find them.


Seasonal pattems are most often calculated using monthly data, although some shidies have fried to pinpoint their periodic tums to specific dajs. As with most other analjsis, closer observation or shorter time periods also bring more noise and erratic results. With this in mind, the seasonal studies in this chafer will use monthly data and keep an eve toward the big picture.

There is one important caveat about the prices used in some of these examples: they were created using a computer program that averaged the prices of a TradeStation contin

uous contract. This contract is gap-adjusted to remove the price jumps in futures when delivery months change; therefore, in the earlier years prices may be higher or lower than they actually were at the time. In the case of soybeans you will see that the prices in the mid-1970s are about $2 per bushel higher. In some cases, such as interest rates, prices that are 25 years in the past can actually become negative because of gap adjusting. A precise seasonal analjsis will need to use continuous cadi prices, rather than a constructed series, to have valid percentages. Results of these exam les, however, show seasonality similar to p

shidies that use cadi data. The following terminologj and concepts will also be of help Basic Calculation Components Average Prices

Finding the seasonal pattem does not need to be complicated; however, some basic rules must be followed to get sound results. For most analysts, it is easiest to begin witti a spreaddieet, where the months are recorded in each column and the rows represent years (see Table "-1). The average monthly price, placed in each cell, can give a good indication of seasonal pattems by simply averaging each column and plotting the results as shown in the first of four summarj lines at the bottom. The major criticism of this technique is that it ignores the changing price levels over time. For example, a 25-year shidy of soybeans will use prices that varj from $6 to $15 per bushel; price changes at the $15 level could overwhelm other years.

Indexing the Data

A simple way to adjust for price differences over time is by indexing data, where each new entry is based on a percentage change from the previous value. This method win also work with seasonal shidies, but must use unadjusted cadi data. Because the monthly average prices for these examples were created from TradeStation continuous data

they cannot be used for this pu ose. Percentage Change

Over the many years needed for a seasonal study, prices may increase by a factor of two or three, and the magnitude of volatility at these different levels will varj significantly distorting the results. To correct for the problems associated with a long-term increase or decline in price, a monthly percentage change can be substituted for the average price. The use of percentages adjusts for most of the volatility dianges and, at the same time, removes any long-term trend bias. It is a simple and very credible method for finding seasonality, but its results can be more erratic than the sophisticated calculations found in the next section.


Removing the trend using a moving average, then comparing the percentage differences between the price and the corresponding trend value, is a classic approach to evaluating seasonality

Median Price

The median is the middle price of a series of values that have been sorted in numerical order. Instead of using an average monthly price, or the average of for example, all June average prices, using the median value should improve results. The median ignores the extreme prices that occur during some unusual years and gives a tjpical price. For example, if the average September price of heating oil was 50 cents per gallon for four years, but reached $1.00 per gallon during one year, then the average would be 60 cents and the median would be 50 cents. Later in this chapter you might note that a laie difference between the median and the average prices indicates that there were a few volatile years mixed with mostly normal pattems.
















im If it

[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]