Impact on Households
Having described the perceptions of households about changing weather pattern sand their environment, we now turn to the question of whether households declare having been affected by specific extreme weather events, and in that case which events had the largest impact on them. The data can also be used to assess whether households suffered from specific losses due to such events. As shown in table 4.4, when asked if they have been affected by a weather-related disaster in the last five years, almost all households say that this is indeed the case, exception the case of Egypt where the proportion is smaller, but still high at 70.75 percent.When asked which adverse event had the largest negative consequences for them, households cited droughts first (30.9 percent of the overall sample), followed by excessive heat (which can be associated with droughts) and floods,both affecting about 8 percent of households. These two factors—droughts and excessive heat on the one hand and floods on the other hand—are closely related to the two factors that were obtained from the MCA in the previous section,although the MCA factors tend to capture a broader range of phenomena,including some of the consequences of changes in weather patterns, for example in terms of land fertility. Note that there are differences between countries in table 4.4. In Syria, all households declare having been affected by droughts,which are also considered as the most damaging adverse event. In Morocco by contrast, floods were the main adverse event affecting households. There are also some differences between households according to their quintile of wealth, but these are less pronounced. For example, the data suggest that households in the poorer quintiles are more likely to identify the adverse events that affected them the most, probably because they are more vulnerable to such events.The fact that the poor are more likely to suffer from changes in weather pattern sand the environment is confirmed by households’ responses to the other question about the effect of these changes on them. As mentioned in section two,households were asked whether in the last five years they suffered from lost income, lost crops, lost livestock or cattle, or less fish caught as a result of weather and environment patterns (the surveys do not provide data on the magnitude of the losses; they only inform us as to whether losses occurred). Table 4.5 summarise responses. More than half of all respondents said that changes in weather patterns and the environment led to a loss of crops, and more than a third reported a loss of income. About a fourth reported a loss of livestock or cattle,and 8.6 percent said that they caught less fish (this would be observed only for those households whose livelihood depend on fishing). The results differ again between the countries, with especially high frequencies of losses of crops in Syria(remember that virtually all households in the areas surveyed reported suffering from a drought), and lower frequencies in Egypt. Yet as expected, households belonging to lower quintiles of well-being were more likely than better-off
households
to declare having suffered from the various types of losses Do these results on the differentiated impact of adverse weather events on households depending on their welfare level still hold when conducting multiple regression analysis? The answer to this question is provided in table 4.6which displays the results of standard probity regressions on whether households declare having lost income, crops, livestock/cattle, or caught less fish. The marginal effects estimated at the mean of the sample are displayed, and the levels of statistical significance are based on robust standard errors. Many of the variables included in the regression have statistically significant impacts on the probability of losses. There are differences between countries in the likelihood of losses as well as the types of losses incurred, which is not surprising, given the differences in the local economies in the various areas. For example, losses in crops are most likely in the Republic of Yemen, which is also the country with the largest share of GDP accounted for by agriculture, while losses in income are most likely in Syria, which is the country in which more households reported adverse events.As expected, the two climatic conditions factors have statistically significant impacts on the likelihood of losses. The impacts are large. Recalling that the climatic factors are normalized to take a value between zero and one, going from the best conditions (value of zero) to the worst conditions (value of one) in the sample for the first factor related mostly to droughts as well as dryer and warmer weather increases the probability of losses by 42.4 percent for crops, 45.8 percent for income, 31.0 percent for livestock or cattle, and 10.8 percent for fishing.For the second factor, which is related mostly to floods and excess water, the impacts of going from best to worst conditions are of a similar order of magnitude,at 42.7 percent for crops (the same order of magnitude as that observed for the first factor), 27.6 percent for income, 34.3 percent for livestock or cattle, and15.1 percent for fishing. Thus, even if the occurrence of adverse events and environmental conditions related to the first factor are more frequent than those related to the second factor, once those conditions come into play, both types of changes in weather patterns and environmental conditions have large negative effects on the livelihoods of households.Also as expected, the probability of a loss is higher in many cases for poorer households. This is clear for crop and income losses, where in both cases household sin the bottom three quintiles of wealth tend to have an increase in the probability of a loss of about 10 percentage points as compared to households in the top quintile of wealth. On the other hand, losses in livestock and cattle as well as in fishing are highest in the fourth quintile of well-being, possibly because those households are more likely to be involved in these activities which tend to require more assets, while households in the top quintile tend not to be working much in agriculture. It could of course be that part of the relationship between welfare levels and losses associated with adverse weather events is due to heterogeneity issue, in that the lower level of wealth observed for the households who suffered from a loss may reflect the loss itself. Yet because of the way the questions are asked over a five-year period, and because welfare is measured through assets as opposed to income
No comments