Managing Weather Risk
There are a few fundamental steps common to every weather risk management process for buyers with natural exposures to weather:
- Identify the critical weather variable or variables
- Identify the impact of the weather variables on revenues, margins, profits and/or costs.
- Identify a reliable, neutral source of historical data and current recordings of the weather variables (usually a government agency such as the National Weather Service in the U. S., MeteoFrance or the Japan Meteorological Agency)
- Identify the date period during which the weather variables’ influence is operative (e.g. hot weather influences air conditioning use primarily in the summer).
- Quantify the relationship between changes in the weather variables and changes in the financial parameter affected by weather.
- Establish sensitivity to the changes in the financial parameter and translate the sensitivity into terms of the weather variable.
This process is not always straightforward, and it normally is an iterative process. Data quality and availability often are major issues, and the weather exposures of each business segment – e.g. utilities, retail, transportation or municipalities – may have its own idiosyncrasies. Nonetheless, in virtually all cases these steps are fundamental to assessing the weather risk and structuring a weather risk management solution to manage a concern’s natural exposure to weather. They underlie the case studies below, which are offered generally to illustrate some of the structures under which weather risk is transferred.
Embedded in weather risk analysis is the recognition that in most cases weather risk is a volumetric risk. Variations in winter temperatures are the major determining factor in variations in the volumes of gas consumed for ambient heating. Amounts of rainfall are a significant determinant of crop yield. Snowfall reflects the masses of snow which have to be cleared from roads and runways. The economic limit is developed by applying costs – historical, projected or budgeted – onto the changes in volumes linked to the changes in weather.
Essentially there are three types of weather risk programs, based on:
- Aggregate measures of weather variables over a defined period, such as average temperatures, cumulative degree days total snowfall, total rainfall. Aggregate based programs respond to the total of the values of the weather variable – e.g. 40 inches of total snowfall in a season.
- Adverse Days, where an adverse day is defined in terms of a weather variable, such as days in which the average temperature is less than 0ºC, or rainfall totals exceed 1.5 cm. Adverse day programs respond to the number of days which are characterized by the adverse condition – e.g. 10 days in the winter in which the maximum temperature is less than 15ºF
- Adverse Events, in which the program responds to the occurrence of a weather condition – e.g. a day on which the maximum recorded windspeed exceeds 50 mph during the term of the agreement.
The analysis of weather variables and their relationship to financial parameters will identify the type of program which corresponds to the risk.
Aggregate Measures of Weather: Cool Summer and Ice Cream Sales
Aggregate measures of weather, precipitation and temperature preeminently, affect the demand for (and supply of) a variety of products. Much attention has been devoted to the role of temperature. Utilities, for instance, have been the main purchasers of aggregate temperature structures. For example, if winters are warm less natural gas is consumed for ambient heating. Cool summers similarly imply less electricity consumption for air conditioning. For utilities aggregate temperature usually is determined by the number of Degree Days: Heating or Cooling Degree Days being the difference, respectively, between 65°F and the day’s average temperature. The result of this daily computation can be summed – i.e. aggregated - over a defined period to establish whether a season is unusually warm or cool. Temperatures are measured at a pre-agreed weather station, in the U.S. usually one administered by the National Weather Service
Utilities are not the only sector for which demand is sensitive to weather. Food, beverages, apparel, chemicals, agriculture, leisure/sport are sensitive to seasonal aggregate measures of weather such as precipitation or temperature.
As an example let us consider the business of ice cream. Assume that a dairy in the Chicago area reviews its annual sales and sees that peak season sales in the summer are affected by weather. In cool summers sales drop off; in hot summers sales peak. Further analysis relates the consumption of ice cream to Cooling Degree Days (CDDs). The more CDDs the warmer the summer the greater the sales; conversely the fewer the CDDs the cooler the summer the less the sales.
If the dairy pegs its production to summer temperatures, it may find that its sales in cool summers become critically short if the summer is 10% cooler than average during the critical ice cream consumption period (taken to be June 1 – September 15). At this level of temperature revenues become insufficient to support the business and inventory costs become burdensome.
Further analysis determines that the dairy’s loss of revenues and increased costs are approximated by a lineal relationship between summer coolness (fewer CDDs), expressed in terms of $ddd/CDD. The dairy elects to take to market a structure, based on this analysis, which begins to pay the dairy when the seasonal CDD total is less than 680 CDDs (10% less than the 20 year average) and will pay at the rate of $ddd/CDD less than 680 CDDs until the seasonal CDD total amounts to 375 CDDs, beyond which oint it no longer will respond. The 375 CDDs represents the coolest recent summer (1992). Under this structure, for each seasonal aggregate CDD less than 680 CDDs the dairy would receive $ddd. The maximum payment would be (680 CDDs – 375 CDDs = 305 CDDs) x $ddd/CDD. In this way the dairy would receive a payment which offsets its loss of revenues and its increased costs due to adverse summer weather. If the aggregate total of CDDs during the summer exceeded 680 CDDs the dairy would receive no payment, reflecting the conclusion of its analysis that if the summer was warm enough to result in 680 CDDs during the critical ice cream consumption period the dairy would be able to meet its basic financial targets.
Aggregate Measures of Weather: Precipitation – Agriculture
Farmers in certain parts of the southern United States desire to plant crops as early as possible in the season. If they are unable to plant early in the season they have to switch their planting strategies and they lose the opportunity to plant two crops. The critical factor is rainfall. If there is excessive rainfall early in the season farmers cannot work the soil and may not be able to move equipment onto their fields because of the soft ground. Excessive rainfall creates additional expense and lost income.
A farm cooperative offered a program to its members backed by a weather risk structure. If rainfall exceeded 9.5 inches in total in the months of February and March the farmer would receive a single payment based on lost income and extra expense due to the inability to till and plant in the early spring.
Based on this structure the program would have paid to the farmers nine times in the last 50+ years.
From the two examples above it is clear that the logic of aggregate structures translates well to temperature and to precipitation exposures. Aggregate measures of weather risk apply to many other enterprises:
- Insufficient rainfall requires increased sprinkling of golf courses and other recreational facilities.
- Snowfall totals affect the cost of maintaining highways in the winter.
- Combinations of snowfall and temperature are critical to ski resort revenues and the sale of ski-related apparel and equipment.
- Average wind speeds relate to the general efficiency of wind turbine generated electrical power.
- The combination of rain and temperature play a meaningful role in determining the demand for lawn fertilizer and swimming pool treatment chemicals.
Adverse Day Count: Agriculture – Corn (Maize) Yields
Temperature is one of the weather variables to which corn (maize) yields are sensitive. Excessive cold at germination and excessive heat in the phases before harvest reduce yields. Any party dependent on the volumes of corn harvested – farmer, grain elevator operator, agricultural credit bank – may look to manage a significant portion of the yield risk by managing the temperature risk through a weather risk management structure.
Taking the various elements into account, such as local climate, farming practices and seed varieties the buyer brought together the following critical information:
- Season: April 15 – October 31, in which cold temperature risk prevailed up to the middle of June and hot temperature risk commenced towards the middle of June.
- Expected (Average) Cold Days: 5 days with average temperature < 32ºF
- Expected (Average) Hot Days: 3 days with average temperature > 95ºF
- Impact of early season cold: loss of x bushels of grain equivalent to $20,000 revenue iro harvested corn for each Cold Day.
- Impact of mid-late season heat: loss of y bushels of grain equivalent to $30,000 revenue iro harvested corn for each Hot Day
- Historical maximum number cold days: 15
- Historical maximum number hot days: 13
Type: Count of Adverse Days
Adverse Days: Section A: Cold Days: Days with average temperature < 32ºF
Section B: Hot Days: Days with average temperature > 95ºF
Risk Period: April 15 – October 31 Section A: April 15 – June 14
Section B: June 15 – October 31
Attachment (Deductible/Priority): Section A: 5 Cold Days
Section B: 3 Hot Days
Limit: Section A: 10 Cold Days
Section B: 10 Hot Days
Amounts Payable, in total up to a maxim of $500,000 subject to the limitations below: Section A: $20,000 per Cold Day after the fifth Cold Day in the Section A Risk Period, up to a maximum of $200,000 with respect to Section A.
Section B: $30,000 per Hot Day after the third Hot Day in the Section B Risk Period, up to a maximum of $300,000 with respect to Section B.
Measurement of Temperature As recorded by the NWS at the Weather Station ___ during the Risk Period and reported by the National Climatic Data Center.
Premium: t b d
The example structure would pay out as illustrated below.
In a hypothetical season the payments might take place as follows:
Section A: 8 Cold Days recorded
Section B: 4 Hot Days recorded
Section A: 8 Cold Days – 5 Cold Day Attachment = 3 Cold Days x $20,000 = $ 60,000
Section B: 9 Hot Days – 3 Hot Day Attachment = 6 Hot Days x $30,000 = $180,000 Total Recovery $240,000
This payment of $240,000 corresponds to the reduced volume of corn and related loss of expected revenue due to weather factors occurring during periods critical to the growth of the corn plant. This revenue could represent the reasonably expected loss of revenue to the farmer (reduced harvest), or to the elevator operator (reduced volume of grain shipped), or to the processor (less volume of grain available for milling, etc.) or to the bank (unpaid loan installments).
Adverse Day structures have a wide range of applications beyond agriculture, including:
Construction: increased costs and penalties incurred due to weather related construction delays: e.g. number of days in which the average temperature is less than 32°F, number of days in which precipitation exceeds ¼ inch.
Insurance Companies: weather conditions which result in losses but which are not picked up under traditional property catastrophe excess of loss covers (e.g. days in which the maximum temperature is less than 32°F and there is more than ½ inch of precipitation)
Transportation: inclement weather increases operating costs, maintenance costs and costs of delay, or increases revenues (e.g. if rainfall in the Upper Mississippi basin exceeds or is less than a given amount agricultural goods otherwise carried by barge will be carried by rail)
Utilities: heat waves or cold spells which require utility to purchase energy on the spot market at times of high demand (e.g. successive days in which the daily average temperature exceeds or falls below a given threshold).
Entertainment and Sports: cancellations and delays during a playing season or concert tour (e.g. more than y evenings on concert days in which there is significant rainfall).
Sales Promotions: weather risk programs can finance sales promotions for weather sensitive products, such as air conditioners, ski mobiles, or snow tires. The promotions can be constructed around a weather variable, such as
If there are fewer than z days with a maximum temperature greater than 85°F before July 4, customers who have purchased air conditioners before June 7 will receive a rebate of $mmm.
If snowfall totals less than 12 inches by December 31, snowmobile buyers - or snow tire buyers - who purchased their vehicles before October 30 will receive a rebate of $mmm.
The concepts of Adverse Events and Adverse Days are closely related.
Adverse Days operate on the assumption that a certain number of days with weather conditions defined as adverse will take place during a given period. The risk management practice is to make reasonable allowance for the occurrence of such Adverse Days and to secure protection from an unusual number of Adverse Days occurrences. The reasonable allowance usually is the deductible, and the purchase of protection against the excessive condition relieves the buyer from making physical or financial allowance for the contingency.
Adverse Events in the strictest sense imply that the buyer cannot tolerate the presence of the adverse condition at all. Examples:
Rainout of a day in a test cricket match, of a World Series game or of a semi-final Grand Slam tennis match means rescheduling with increased costs for the promoter and venue and loss of audience and advertising efficacy for advertisers.
High winds during a week in which a contractor is dependent on cranes to complete a work process.
Adverse temperatures and/or precipitation on the day when there is to be spraying of fertilizers, herbicides or insecticides.
Adverse Event programs are structured to respond by paying the reasonably established amount of the increased costs and/or revenue loss associated with the Adverse Event taking place.
The logic of Adverse Events brings some participants in the weather risk market to offer capacity for Extreme Weather Events, of which hurricanes and typhoons are examples of prime current interest. The thought process is similar to other adverse events, in that programs are structured with respect to storms of a given strength afflicting a pre-defined area. Payments may be in relation to the strength of the storm (e.g. wind speed or category) or the may be digital on the event itself. At present the capacity in the weather market for Extreme Weather Event risk is relatively limited.
Weather risk management structures based on Aggregate Measures, Adverse Days and Adverse Events encompass the great majority of program structures used for weather risk management purposes. Among the other structures are forecast risk protections, which provide compensation if the buyer organizes its activities according to weather forecasts which, in the event, turn out to be wrong. Newly emerging, particularly in Europe, are weather-indices. Parties with weather risk can buy or sell a position on the index, in a way analogous to the trading of positions on securities indices. Indices have been developed for specific European business segments, such as warm weather indices in the summer for brewers. Another important development is the appearance of bundled weather and commodity structures. Dual trigger programs with a temperature trigger and a commodity price trigger are a basic example of such structures. They respond when there is both an adverse weather condition and an adverse price environment and focus on combinations which present the gravest risk management challenges to buyers who are sensitive to both price and to weather. Dual trigger programs operate most effectively in environments in which the commodity is a liquidly traded commodity.
This discussion, and the examples supporting it, are not intended to be limiting. On the contrary, each segment of business or government, each exposure, each weather variable all have their individual characteristics around which weather risk management programs can be molded. As importantly, the weather risk management program can be tailored to meet the buyers pertinent risk management objectives which may different from one situation to another. For example: the sales of apparel retailers may be sensitive to seasonal weather, but weather conditions on critical sales days may be a significant exposure as well. Aggregate measures of temperature or precipitation over a period may respond most closely to the former concern, and an adverse day structure (heavy snow on weekends before Christmas) or an adverse event cover (weather on a critical sales day) may respond to the latter. Weather risk management programs also can serve to manage exposures that are less obvious. One of the largest weather risk transactions has been an adverse day structure based on cold days that provides support to a government mandated fund which pays the wages of construction workers who are unable to work because of inclement weather. These examples and this discussion are intended as a starting point for those interested in managing the weather risk affecting their business and civic responsibilities.
The market of risk takers offering to accept the transfer of native weather risk is made up primarily of insurance companies, banks, the desks of utilities active in the weather business and monoline weather risk trading operations. Buyers can also avail themselves of exchange contracts, including index based contracts. In some cases participants in the risk management segment of the market make their chief business out of assuming third party risk (e.g. insurers and banks). Others, such as utilities, may take on such risk as diversifiers to their own, native risk. Where capacity is syndicated hedge funds often contribute their support as well.
Transaction sizes vary widely in accordance with the risk parameters of the buyers. Small and medium sized commercial and industrial buyers may require limits in the tens of thousands of dollars or hundreds of thousands of dollars. Dealing effectively with such small limit transactions requires a high level of automation and systematic means of standardizing products which nonetheless are tailored to the buyer’s circumstances. For larger enterprises limits between $10 million and $30 million are common, and larger enterprises also are well positioned to purchase multi-year agreements. The largest transaction in the market is syndicated and has an annual limit well in excess of $100 million.
Risk transfer can take place under derivative agreements or under insurance policies. To be an insurance the risk transfer must meet basic criteria, such as demonstrable insurable interest and demonstrably close relationship between the structure of the risk transfer (limit and deductible/trigger) and the commercial consequence of the exposure to weather risk. The insurance form is (parametric) index based insurance with pre-valued settlement. Collection may require demonstration that the buyer suffered loss. An insurance policy, of course, can only be issued by a recognized insurance entity. Derivative transactions, although accounted differently, have less onerous qualifications but do require that the counterparties demonstrably be qualified financial institutions or individuals. Regulations vary from jurisdiction to jurisdiction so the parties to a weather risk transaction, as with all risk transfer transactions, should ensure that they are acting in conformance with applicable laws and regulations.