 Thank you so much for joining us for this early morning session. I'm Nikita Sanghwan from Indian Statistical Institute. And today I will be talking about weather shocks, its impact on labor from a gender perspective. And the question is, we are talking about weather shocks in a climate session and that, and I'll soon come into why this is important. Is it fine if I stand here or should I say over here? Okay. Okay. So when we talk about climate, we generally talk about rising temperatures. But it's not just about temperature, because if we look at extreme weather events, the frequency of these events has also been going up over the year. Okay. Sir. Sorry for this. Okay. So we know from the IPCC report that not only are these extreme weather events going to become more frequent in the future because of climate change, but the intensity is going to go up. And here is a graphical depiction when I look at the Indian context. And what we have here is the frequency of the number of droughts that are being experienced across the Indian landscape from 1901 to 2017. And what we see here is that there is a clear upward trend when we look at the number of grids that are facing a drought. And drought here is defined as deficient rainfall. And I will soon come to the definition of drought that I'm using here. And what's important to keep in mind here is not only that the frequency of drought has been going up, but what is also happening is that the length of these droughts is going up. And the length of drought here is captured as the number of years in the last decade when a drought was experienced. Now that, see, the concern here is you experience a drought, you're experiencing more number of droughts. Additionally, the intensity of this drought is also going up. And the intensity here is being measured in terms of how much is the deviation from the long run average of rainfall that is experienced within a grid. And as we see that there is a clear upward trend in all these three parameters, this is a cause of great concern. And it's important to understand why should we be concerned about these three, you know, increasing trends in these drought measures. And the reason for that is that more than 75% of the world cropped area, even today, is dependent on rainfall. And since these areas that are dependent on rainfall are majorly coming from the low and middle income countries where the societies that are managing these agricultural lands are not equipped with the resources to cope with such shocks, they are experiencing a strong, you know, impact when it comes to the agricultural output that they're producing. The income of these agricultural households, also the food security comes under threat when they are faced with such extreme weather events. Also, I'm focusing in this paper on the labor aspect because a large proportion of the workforce continues to be dependent on agriculture. If you look at the Indian statistic, 40% of the population today continues to be supported by agriculture and agriculture dependent activities. Now, another aspect to consider when we talk about labor and its engagement in agriculture is the gender disparity that we are seeing in terms of engagement in these activities. Because to start with, we have very few women in the labor market. Only 35% women in rural areas are working in these activities while for men it's close to 90%. Now, given that there already exists a gender disparity, what happens when there are going to be such extreme weather events? And more of these events being experienced because of the unabated climate change that is happening. And that is going to be of interest to understand how the existing gender disparities are impacted when these drought events or flood events are affecting different regions. The reference for this work is coming from two work. One is looking at the gender effects of drought shocks, which looks at how production shocks which are negative productivity shocks to agriculture has an impact on the equilibrium labor responses of people who are in these rural areas, which is joint work with Farzana Fridi at Indian Statistical Institute and Kaneka Mahajan at Ashoka University. Another work from which I'm going to be drawing some results for today's presentation is coming from the impact of weather shocks on the usage of labor in agriculture. Okay, so what is it that we're trying to do in this study is to understand how individuals and households respond to these extreme weather events and the idea here is to make use of a high-frequency novel data set that is available to us and is collected by Ikrisat as part of the VDSA study. And this panel data allows us to address any unabsorbed heterogeneities at the individual level, which are very important when we are looking at the responses to such shocks. The idea also is to understand if the responses have a gender difference wherein men and women are responding differently to these shocks and if there are such differences by gender, what are the underlying mechanisms that can explain these gender responses. In terms of the existing literature, the focus has majorly been on the impact on agricultural output. But when we look from the labor perspective, we do know of studies that have documented that the earnings of households are impacted and if there is access to non-farm opportunities, then the loss of earnings is lower. But given that social insurance is almost absent in these societies and credit markets are incomplete, there is an impetus on diversifying the labor supply to non-farm sector from the farm sector. Also, there is documentation of permanent migration in response to the climatic shocks. But when we look at the short-term responses, we do not know much about these responses in the literature. Additionally, when we talk about the gender perspective, there is insufficient or very sparse literature that talks about these gendered responses. Also, the underlying mechanisms have not been explored to the best of our knowledge. Now coming to the novel data set that is really, one of the key features of this study is that this data set which is coming from the VDSA study allows us to look at a very high-frequency data which is collected at a monthly frequency. So it's not a retrospective data which has been used like in most of the studies, but every month this data is being collected from these households. And the sample is coming from 18 villages from the semi-arid tropics of India and 12 villages from the humid tropics of India. And if you look at the color codes here, they are representing the agro-climatic zones of India. And the sample is representative of the 11 of the 21 agro-climatic zones of India. So it's covering a large amount of variation and allows us to make inference for these two areas. Coming to the climate data, so for each of the 30 villages that we have, we have mapped it to the climate data that is available from IMD. And IMD gives us daily gridded data of rainfall at a high resolution. And we are using this data from 1971 all the way to 2015 to construct our measure of drought, which is taken as the monsoon rainfall in the months of June to September, falling in the bottom two deciles of the long-run average rainfall from 1971 to 2014. And the idea here is to look at the deviations from the long-run average rainfall for the same village because if a region has, over the period, experiencing very low levels of rainfall, then another year of low rainfall is not a shock to them. It's a shock to them only when they are seeing a deviation from what they are used to seeing. But the question is, is this a good measure? So this measure has been used in the study by Jayachandran and we also validate whether it's a good measure for our study. And we do this validation by looking at how the farm output in terms of rice, which is a water-intensive crop, is impacted by our drought measure. And we see that there is a significant reduction in the output of rice and the yield of rice, which assures us that a measure of drought is accurately capturing deficient rainfall. We additionally look at what is happening to the productivity on the farm by looking at the revenue of these farm households and the profit that these households are making by considering all the crops because it's possible that the households diversify in terms of the crops that they're cultivating. And here again, we see that there is a significant fall in these productivity measures in response to the drought shock. So clearly, our drought measure is valid and is accurately capturing scarcity of auto. And this brings me to the estimation strategy that we are using to estimate the impacts. And what we are doing here is to regress our variable of interest, which are going to be different labor market participation on our measure of drought. So the beta one here is our coefficient of interest. Additionally, since the drought measure here is about the bottom two deciles of the deviation from the long run average, I'm additionally controlling from for the top two deciles which are like flood like situations. But the idea here is to compare how deficient rainfall has an impact relative to years of normal rainfall. And in this specification aided by the highly detailed data is using the individual fixed effects, which allows us to address any unobserved heterogeneity at the individual level. We are able to control for season fixed effects and EO fixed effects to all any concerns of seasonality. And all our standard errors are clustered at the village season level. And as a robustness check, we cluster it at the village level also and the results are robust to that. Now finally coming to the results. So the idea first is to look at what is happening to overall labor market participation. So here the first column is going to be, the first row here is about the impact of drought. And the second one is capturing the gender differential between female and male responses. And if we look at the overall labor force participation, we do not see a significant impact of drought. But when we break down this overall labor force participation into employment and unemployment, we see that female are losing 15% of their employment days when they face a drought shock. There is no similar impact observed on men. And as a result, women are losing 19% more days of work when they are impacted by drought. When we look at unemployment, we see that for women, there is an increase of 14.4% of days when they are looking for work and are not able to find work. While for men, there is a negative impact of 15%. And as a result, we see that women are bearing a higher burden of the drought shock, wherein they are losing work and they are not able to find work unlike men. And it's because employment and unemployment were aggregated to form the labor force participation that explains the null effect that we are observing on the overall labor force participation. Now, since we are seeing that women are bearing a greater burden of the shock, the idea is to look at the impact by sector. How is the engagement in different sectors impacted? And what we find here is that for men, we see that there is a reduction in the participation in paid form work days by about 14% and this is taken up by increased participation in non-farm work. So men are diversifying from form to non-farm work, farm work, which is very vulnerable and unsecure because of drought shocks to the more secure non-farm works which are less sensitive to these drought shocks. We do not find similar diversification for women. Now, a natural question is if women are impacted more, they are losing more and they are not able to diversify like men, then what happens to the earnings that they're making in the market? And for today's presentation, I'm going to be focusing on the monthly earnings, conditional on being in the farm sector or the non-farm sector. And what I'm finding here is that when there is a drought shock, women are losing 38% of their monthly earnings if they continue in the farm sector. With no similar impact observed on men. While when we look at the non-farm sector, there is no significant impact on women but there is a reduction in the non-farm monthly earnings for men. Now, this goes back to our ECON 101 lecture, right? Because if there is a greater dependence on one sector while there is a negative productivity shock to that sector, it leads to a reduction in the earnings in that sector. And that's exactly what we are seeing here because women continue to be in a low productivity sector which was impacted by the drought shock. Their wage rates were falling and as a result, their earnings were lower by 38%. Men, since there were greater dependence on the non-farm sector of men, this led to a depression of the wages in the non-farm sector. But if we look at the overall monthly earnings that the men and the women are making, we see that men on the overall picture, they are gaining because of their increased participation in the non-farm sector and also the higher earnings in the non-farm sector. Now, the question really is to now understand what are the underlying mechanisms? Because to summarize, we find there are gender responses. As a result, women are losing in terms of their labor market participation as well as the earnings that they're making. So what can explain these gender responses? And the answer is to look at the mobility constraint because when you have to diversify from farm to the non-farm sector, it is possible that most of these non-farm work are available outside the village so you have to travel for work and that's why we look at what is happening to the location where these men and women are working. And what we find here is that in response to a drought shock, men are more likely to be working outside the village, they are more likely to migrate and they are traveling longer distances in search of work. With no impact observed on women, we age working outside the village, migration or distance to work. So clearly men are casting a wider network in search of work when they're impacted by the drought shock with no similar coping mechanism adopted by women. Now this also raises the question, is mobility really driving force for the observed differences? And to validate that the impact is really coming from this mobility restrictions, the obvious ways to check, if we provide work to these women close to where they're staying, do they take up this work and are able to cope with the shocks? And for that I look at what is happening when they are given access to social protection program through the employment guarantee scheme which is the National Rural Employment Guarantee Act in India that provides 100 days of guaranteed work close to where these people are staying. And what I'm finding here is that we are using the data set for which we have observed the employment outcomes. There was a positive coefficient for female as well as male with the coefficient being higher for female compared to male. But since this data on participation in this employment guarantee scheme was not available for all the villages but only for half of the villages, what I'm doing is supplementing it with the administrative data in terms of take up of employment using the employment guarantee scheme. And what I find is that in response to drought shock women increase their participation in the government scheme by 37% while men increase it by 33.5% with the impact being larger for women. So clearly when women do get access to job opportunities close in close proximity to where they are staying they are taking up these opportunities. And this substantiates the mobility constraint channel that I was talking about. Okay, in summary what we are really finding is that extreme weather events have gendered effects and this is coming from the restricted labor mobility that women in these areas are facing. While men were able to substitute and take up non-farm work in response to drought shocks women were not able to do so. And another potential mechanism for that could have been a skill deficit. It's possible that women have different kind of skills while men, the kind of jobs men are engaging in requires more skills compared to what women already possess. But we have ruled out that channel because most of the non-farm work that men were engaging in was the unskilled work and women were already being engaged in those kind of unskilled non-farm activities within the village. So clearly there is no stigma associated with these works and women could have taken these opportunities if they were available within the village. So, and that is why since these were not available within the village and required women to travel this led women to drop out of the labor force entirely. And those who continue to work in the labor market in the farm sector which is lower productivity sector because of the drought shock, they continued engaging in the riskier agricultural work and lost a large part of their earnings. And that is going to have significant implications when we talk about the intra household bargaining framework. So an inference from this study is that persistence of extreme weather events due to climate change can exacerbate the existing gender disparities because we are seeing not only previously very few women were in the labor market but whenever they are facing a drought shock they are forced to leave the labor market even more. So this, you know, this provides us an insight that because of climate change as these events are going to become all the more frequent the extent gender disparities are only going to widen. And this is where there is a role for policy to play because if policy makers are able to provide access to work opportunities to these women it can help women to cope from such shocks. And with this I conclude my presentation I will not go into the appendix. Thank you so much for your attention.