Trends in Crop Production Losses and Claims Costs: 1969–70 to 2007–08

48. 
    The 39-year analysis of crop losses in Bangladesh shows a high vulnerability of crop production in the country to natural catastrophes. Figure 3.6 shows the World Bank’s estimates of annual crop losses at the 100 percent insured yield coverage level by cause of major losses over the past 39 crop years (1969–70 to 2007–08).
    
49. 
    Several droughts, cyclone, and flood events with severe effects on crop production in the country occurred during the analyzed period, 1969–70 to 2007–08. The worst events in the analyzed period were the 1998–99 floods combined with a tropical cyclone and would have caused a loss on the eight crop portfolio at 100 percent insured yield coverage level of Tk 62,130 million (US$914 million or 15.52 percent of the total value of production (VAR). Other catastrophic years of crop production losses include the severe drought in 1979–80, which would have caused a loss of Tk 59,173 million (US$870 million); the drought in 1995–96, which would have caused a loss of Tk 56,195 million (US$826 million); and the floods of 1998–99, which would have caused a loss of Tk 54 billion (US$795 million).
    
50. 
    There is a slight decreasing trend in the frequency and severity of annual crop production losses over the past two decades in Bangladesh, but this trend is not statistically significant. During the period 1990–91 to 2007–08, the annual average losses were above 10 percent loss cost in three years only. Conversely, during the period 1969–70 to 1989–90, there were nine years with annual average losses above 10 percent loss cost. The main reasons for the reduction in yield variability over the past 20 years may include (i) the improvement in flood control infrastructure, (ii) the major expansion of the irrigation facilities, and which has been accompanied by (iii) the switch out of pre-Kharif or Aus paddy, which is very exposed to drought and cyclone, into the irrigated Rabi season Boro paddy, which exhibits much greater yield stability. Between 1969–70 and 2007–08 the area planted with irrigated Boro paddy increased from 19 percent of the gross cropped area to 37 percent of the gross cropped area. An extended network of river embankments and drainage channels was built in the country to mitigate the effect of floods. Nevertheless, in light of the number of the flood events which have occurred in recent years and their effects on crop production, some authors have argued that these kinds of flood mitigation measures have not reduced the level of crop losses due to floods.³²
    
51. 
    Flood events have been more frequent during the past two decades. The analysis of the annual value of crop losses shows that whereas drought events were more frequent during the seventies and eighties, flood events have been more frequent during the past two decades. The analysis of annual crop losses shows that, whereas seven drought events (with a return period of 1 in 3 years) occurred during the 21-year period from 1969–70 to 1989–90, only 2 drought events (with a return period of 1 in 9 years) occurred during the 18-year period from 1990–91 to 2007–2008. The same analysis shows that, whereas only 4 flood events (return period 1 in 5.2 years) occurred in Bangladesh during the period from 1969–70 to 1989–90, they occurred on 5 occasions (return period 1 in 3.6 years) between 1990–91 and 2007–2008.
    

Source: World Bank, from BBS Great District Yield Data and BBS Crop Damage Data.

1. 
   The analysis of 39-year regional yields for major crops—Aman HYV, Aman LBV, Aman LTV, Aus HYV, Aus LV, Boro HYV, Boro LV, and wheat—shows that 1998–99 was the worst loss year in this series, with total crop losses valued at Tk. 62,130,000,000 (US$924,000,000) or 15.52 percent of the total value of production. Although 1997–08 was a severe loss year in Bangladesh, even worse crop losses could occur in the future. From an insurance viewpoint, underwriters need to know with a high degree of confidence the maximum losses that they might incur (called the Probable Maximum Loss, PML³³) either 1 in 100 years, or if it is necessary to be even more conservative, 1 in 250 years. This information is critical to structure any insurance program and to determine how much capital is required to cover the PML loss year.
   
2. 
   Figure 3.7. and table 3.4. show the results of the World Bank’s PML loss cost analysis³⁴ for return periods of 1 in 2 years up to a maximum of 1 in 250 years for the Bangladesh eight-crop national portfolio 39-year yield data assuming a 100 percent insured yield coverage level. The analysis shows that:
   

• 
  The losses in 1998–99 with 15.52 percent loss cost at 100 percent coverage level, equate approximately to a 1 in 30-year return period;
  
• 
  The 1 in 100 year estimated PML loss cost is 23.63 percent at 100 percent coverage level, equivalent to a financial loss of Tk 94,000,000,000 (US$1.4 billion).
  

Source: World Bank from BBS.

54. 
    Under any future crop insurance program in Bangladesh, it will be very important to perform a refined PML analysis based on the portfolio of crop and locations to be insured in order to be used to structure the crop insurance company’s risk retention and reinsurance programs.
    

55. 
    Agricultural crop production in Bangladesh is exposed to weather risks, especially flood, tropical cyclones and droughts. This is evidenced by the great district average loss cost for a 39-years period, 1969–70 up to 2007–08, estimated at 6.37 percent of the total gross value of production and a calculated 1 in 100 year PML of 23.1 percent of the national crop gross value of production.
    
56. 
    Different crops grown in different seasons have different susceptibilities to natural calamities. In general, pre-Kharif crops, with a gross value of production weighted annual average loss cost of 7.4 percent, are more vulnerable to natural calamities than Kharif crops, which have a gross value of production weighted annual average loss cost of 6.9 percent. In turn, Rabi crops are less risky than Kharif and pre-Kharif crops, with a gross value of production weighted annual average loss cost of 6.0 percent. Boro HYV and Aman HYV, which are the main paddy crops sown in Bangladesh, show annual average loss cost as of 5.7 percent and 6.4 percent, respectively. Conversely, Aman LBV and Boro LB, which are opportunistic crops representing only 3 percent of the assessed portfolio, show high annual average loss cost as of 12.4 percent and 9.0 percent respectively.
    
57. 
    Risk exposure varies by region: Almost the whole of Bangladesh is prone to floods. The central regions of Bangladesh are particularly prone to floods due to drainage congestion in the Kharif season. There the annual average loss cost for Kharif crops, with average loss cost ranging from 8.8 percent to 10.0 percent, are higher than the national average of 6.9 percent for Kharif crops. Coastal areas of southern Bangladesh are prone to floods due to storm surges, which affect pre-Kharif and Kharif paddy crops; the annual average loss cost for Kharif crops in this region is 7.4 percent, which is slightly above the national average. Northern areas of the country, in particular Sylhet region, are prone to flash floods affecting Rabi crops; the annual average loss cost in the areas affected by flash floods is 8.40 percent, which is far above the national average loss cost of 5.7 percent. Northwestern regions are prone to drought in pre-Kharif and Kharif crops and for hailstorms and tornadoes in pre-Kharif and late Rabi crops.
    
58. 
    Flood is a regular, predictable, and foreseen event in certain areas of Bangladesh and in these areas cannot be considered as an insurable peril for crop insurance purposes. Particular care should be taken in the eventual design of an area-yield index product for flood-prone areas. In certain cases the risk of flood is a certainty. The provision of area-yield index insurance should be targeted to those areas where flood is an unforeseen peril.
    
59. 
    The analysis of crop losses shows a slight decreasing trend in the frequency and severity of annual crop losses in the past two decades. The main reasons that explain these phenomena are related to the improvement in risk mitigation infrastructure. The expansion of the irrigation facilities has changed the predominant cropping pattern to winter crops and aimed to mitigate the effect of droughts. An extended network of river embankments and drainage channels was also built in the country to mitigate the effect of floods. Nevertheless, in light of the number of the flood events that have occurred recently and their effect on crop production, it appears that these kind flood mitigation measures did not improve the reduction of crop losses due to floods.
    

60. 
    There are three main seasons recognized in Bangladesh: (i) the summer season (March to June), with high temperatures and erratic rainfall; (ii) the monsoon season (June to October), characterized by heavy rains (two-thirds of annual rainfall); and (iii) the winter season (November to March), with drier conditions, cooler temperatures, and low rainfall.
    
61. 
    The premonsoon or monsoon onset period, starting in March, is marked by northwestlerlies³⁵ and the occurrence of localized cells of high winds and heavy rainfall. Monsoonal rainfall starts in the middle of May and lasts until the end of October, reaching its peak intensity at the end of June, and accounts for 80 percent of the cumulative annual precipitation. The monsoons, or South-West Trades, form powerful moisture-loaded convection cells and lead to annual rainfall of between 1,200 mm in northwestern Rajshahi to over 5,000 mm in northeastern Sylhet, and 2,600 mm in southeastern Chittagong. The isohyets map of Bangladesh shows increasing cumulative rainfall from the northwest to the south and the northeast (map 3.7).
    

Source: Biodiversity Research Group of Bangladesh

62. 
    Monsoonal rains in most regions of the country peak a second time during September before rapidly dropping after the middle of October. Following this, the winter season, between October and March, has little rainfall, except some light winter rains, typically between 10 to 40 mm, between mid January and mid February. Annual rainfall patterns are shown in figure 3.8.