 Questions, comments, okay right back there and then yeah and then Imran and then up front and then Channing and then Yvonne, that's five people already. So please do say your name. Yeah my name is Christophe Muller, I'm from the University of Ex-Marseille in France. I found the presentation very very interesting with in-depth data analysis and obviously we can learn a lot with this type of exercise. I'm a little bit worried about one thing which is the unemployment. The fact that a lot of the effect of the economy on the poor is going to go through the fact that some people are just unemployed, they don't have so they don't have wage and you miss this part of the analysis with these techniques especially in South Africa where the level of unemployment is so high and in some communities it's still higher than the average that you have shown. So my first point is what's happening with unemployment and how does it interact with this analysis. Perhaps some of the effect that you see are through selection effect in and out an unemployment state. And the second remark is that would be really interesting to look at at the different committees, different ethnic group to see what happened especially you know that is South Africa so we are really interested in the difference between these people. Great, thanks for that Harun. Just one one thing that sort of doesn't end up for me is your story about about the automation in the trade union collective bargaining story so my sense of it would be it's precisely the it's precisely the group who do the automated routine jobs are precisely the group that are that are most sort of powerfully in the trade union movement and they likely to be the group that would most sort of managed to have secured some sort of premiere. Thank you very much Ivan Turok very interesting presentation just a quick question right towards the end when you talked about the outsourcing from United States and Europe South Africa could be a recipient of some of these jobs and I'm thinking particularly around the BPO call center type is that ICT category of tasks that you talked about the government claims this is one of the major beneficiaries of South Africa and there are advantages of same-time zone as Europe language issues for UK and also big government incentives I'm just wondering whether that could be a factor as well in the positive performance of that particular category of tasks. Thanks Arun I'm looking forward to reading the paper just when you were talking through the the first points with the share it was a lot of technology and then when you were talking with respect to the wage wage was a lot of trade and I if you just talk about that a little bit and then the other thing that if you happen to have thought about it you know we're observing you know tremendous penetration of South African enterprises into the rest of Africa and you really are I mean you're not highly skilled in doubt versus the United States but you certainly are skills in doubt versus versus Africa and I tend in the South African case I tend to think that trade would might be a big player bigger than what your first shares would seem to indicate and but I what your story was telling in the second part me is this FDI if you just talk a little bit about that I think my question is related to last questions but it's related but the basically story is a task wage premier is driven by trade in tasks and possibility of slicing up productions and is it really is it multi-nationals different multi-nationals American firms or European firms targeting South Africa and therefore South Africa wage premier in particular tasks is changing according to competition or its perception that that is fair it's really you know competition global competition and driving so without really moving around the production site actually task premium is reducing in particular segment of production process and the fed it is I mean related to chance question last question that is it to South African firms moving out or is it coming in to South Africa or what is the situation with respect to firms behavior in regions continent that's where the global says thank you sure thanks those are great questions there so the race the race effects are in the the quantiles but what I hear you saying which I think is correct is run them separately so see if there's anything different that's going on for Africans versus whites and see and I think that's a great suggestion actually the I think partly the bottom end is where some of the actions going on in terms of unemployment so I think the bottom end of these quantile distributions is where you see a competition for jobs that's actually not happening either because of trade union action or and then and then across the rest of the distribution you've got the structural mismatch and I think that's where it's not a case of necessarily institutional effects that prevent a reduction in the employment it's more structural barriers right so whether it's a skills mismatch or whatever the case may be that means almost that you've got a fairly heavily segmented labor market so you've got the excess supply of labor that that really doesn't allow much churning except at the bottom end between the employed and the unemployed across jobs Imran's point I think that's a really that's that's a clever point because in a sense are we saying that you've got the National Union of metal workers is is really should if they're doing their job correctly which they supposedly are because they're one of the strongest unions you should be seeing a wage distribution that looks the opposite of what we're showing there so one way to think about that is possibly that that we observing we observing wage bargaining as it happens at that point in time so the real question is over time have you seen a decline in the bargaining power of that trade union and I think that may be true I think you may be seeing a decline in the bargaining power of trade unions despite despite what the sort of strike strike data and annual bargaining around seem to suggest I think that's true for all private sector unions except the public sector so I think it's true for all unions except the public sector unions where I think they've actually grown in power and in strength over the period and I think there's data to show that that's true yes so teachers teachers are there so exactly teachers are in face-to-face thank you so if Ivan's points Channings and Machiko's if you don't mind I'll sort of combine them because they sort of similar right so there's a two-part answer one is the modeling approach is that there is something called the new approach called the recent influence function which I'm not clever enough to figure out yet but I know I'll eventually get there but that allows you to decompose across the quantiles the effects of those different forces so you can actually get at technology trade exogenous policy shocks and then all the other individual control so that's the if you like that's the technique answer so that's what we'll try and do I don't think it's as good as your stuff the input output analysis I think you really in order to separate out the technology trade story you need more than these these techniques that we have but but we'll give it a go all the same just in terms of a general reaction I think that's it's interesting yes South Africa is a big player but the in the rest of Africa the question really is what kind of trade and my sense is that it's primarily in the retail end so if you're looking at retail goods and final sort of produce goods that's where South African retailers have actually got a competitive advantage in the continent are they setting up large manufacturing like manufacturing firms I don't see that happening there's mining license action a little bit of primary sort of resource extraction going on and then the really big one is our comparative advantage which is financial intermediation and so if you like the probably the biggest story in South Africa is around China buying 20% of standard bank as a portal of entry into doing deals in Africa and I think that's that's particularly interesting but then that would reinforce these kind of trends it's time to zero minutes okay I'll stop thanks floor is open Carol over there even so I have three people so far okay okay and that's enough thank you for an interesting presentation a couple of small questions really I'm wondering in your data do you know how often people move and maybe it's the QLFS that you can use to detect this so is it the same people transitioning in and out and there's some people who are like in stable employment for the duration I think that might be interesting to look at another thing that was interesting and I don't know I don't think you commented on it was and the fact that for the female employees it seems that unions mattered and for the males the contracts mattered but not to the around so I was wondering if you had some comment on that thank you another another question in a country with huge spatial inequalities it's hard to measure lay spatial mobility but I wonder whether with the NIDS data you can do that and whether or not you factor that in at all do people is there also a corresponding degree of change in where people are working or where they live when they lose their jobs or when they gain a gain employment any views on that hi Carlos Radim from the University of Eagle I think the paper is very interesting because it deals with one very special issue in South Africa that is the lack of employment that is rising what is happening in the distribution one just one specific comment about NIDS data is that the special design of this data makes that people are not interviewing the same quarter in both ways so maybe you have you need to deal with possible seasonal effects of especially when looking at transitions and also that the time span between both interviews is different for for people so maybe also you are more likely to make a transition the longer the time between both interviews just a thought on the on the sort of time thing because I thought you said there was a there was a huge sort of drop in employment in that there's a change at the time of the crisis which then which then which then starts to starts to reduce with time is that right so the what might be going on is that it's quite tough in South Africa to lay to lay off workers but it's quite easy to do it if you can show that there's that there's some economic grounds for it so the kind of at the time of the crisis what you might be sort of picking up is some sort of pent up need to reduce the workforce and then the crisis hits and firms reduce a lot more than what they should ordinarily and a second thought just tying up your sort of paper with Haroon's I think it does show the important need for Haroon and them to to really think about the gender issues as well which which come out strongly in your paper just by way of a comment I mean one of the what are the related to him Ryan's thing point is that when the crisis struck there was a view particularly within the dispute resolution body that had to settle all these claims retrenchment packages that were offered that the majority of workers that were retrenched were those that had didn't have permanent contracts so I see there that you do have nature of contract but I was wondering if you could think of pushing that a little bit more maybe an interaction effect with sectors or something to see if there's something going on there because you may be underestimating that just because of the specifying at all and the specifying okay Dennis two minutes please yes oh well how often do people transition have no idea of that I think it would be possible using like duration models or something because in the QLFS you do have households well sorry you do have individuals that can be observed over a maximum of four quarters so it would be nice to see whether they switch from employment to non-employment or or it's always other people that do that so that's a very good suggestion about the unionization for females well actually I I I don't think well the the effect for males is also positive but it's it's not significant and in the QLFS I do find that the union membership which is only available in the two doubt from 2011 onwards that it's significant for both males and females so I'm not sure whether they are really any gender differences there about spatial mobility well I do include in all the regressions province effects but I haven't really studied in detail well some of them are significant but it's it's hard to find really a clear trend in that so so that's also something that I that I should look at and it needs you you you have data up to the district level so maybe I can do something with that dealing with interview data I think that's a very good suggestion because the interview data are available so so I can try and see if I reduce my sample or I adjust my sample to take that into account whether this makes a difference I think that's important makes it probably also better comparable to the QLFS then about the reducing buffering effects well I was thinking actually the same that that that as long as the crisis lasts it's it's easier to also fire those people that have higher skills or are better attached to the to the firm or have had more on the job training which are typically those with with higher education so that those with lower education that are more easily replaceable fly out first and then and then maybe then the others follow us as economic activity doesn't pick up so maybe that's but it's it's difficult to find really hard evidence for that and then about the interaction effects yes I think that's that's that's a very valid comment so I also some of the racial effects could perhaps be interacted with with industry effects with type of contracts actually in the QLFS if I control for for for the type of contract you see that the almost all sectors did did better than than agriculture because in agriculture so meaning that because in agriculture you you don't have you don't have those strong formal permanent contracts so I think the contract so part of at least part of the industry effects are due to the differences in contract I think contract effects are probably more important even then the specific sector so yeah thank you very much we have time for questions I'm Darrell sequera an independent environmental consultant I just have two questions one is you use the term Indian migrants where did they come from did they come from neighboring Kenya Tanzania or did they actually come from India or somewhere else that's one question the second one is you related the the higher profitability of Indian business enterprises to endowments and you mentioned what those endowments are but I was also but it seems to me perhaps I'm wrong that you did not include two factors one is the ability of the Indian businessman to speculate correctly how to promote their business in the future from from the present so correct speculation plays a role perhaps and secondly also the management of their ongoing businesses in terms of efficiency and profitability I just want to add on to that question for a minute were these Indian entrepreneurs ethnically Indian but possibly Ugandan citizens or were they actually newcomers to Kampala I always say recently if there is any xenophobic reaction in the country against this because in some countries these minorities that are doing well they had there is some xenophobic reaction against them and maybe this is the case this could explain also they are reluctant to give information that may be used against them so it's by ignorance and asking if there is the case in Uganda or not a question about age and whether family ownership I mean it looked like a very skewed sample from what you were saying you didn't get much response from many companies so was this with these sort of long established Indian companies that have been passed through from generation to generation sending their you know sibling their children to university this just looked like a very different kind of cohort so I can't remember all the slides but I had two questions one is if you had sex of dummies in there so to control for something else going on but related to that whether whether you you you looked at differences across the profit distribution so in a way you may be seeing something more interesting if you looked at low profit rather than high profit or low capital intensity and high capital stock firms and so maybe instead of an OLS you could go I don't know quantile or something just to see pick up something more interesting just one question and I've marked are these firms selling to the same type of markets like are some of these Indian firms exporting or do you know that are these all very small maybe that might make a difference okay so on the nature of the Indians and their business they're all real Indians coming from migrating from India and they're all first migrate first generation migrants so that during the Idi Amin area all Indians got expelled from Uganda actually that before there was an even larger Indian population so they all came back just in recent years and their businesses are still quite small so they do not export and yeah I think adding like management of the business or correct speculation I think yeah that would probably explain a lot the problem is I'm not sure how we don't have a good measurement of this I think but yeah maybe I need to further look into this the firm age is here so they are they they are not significantly older or well more well established than the than the Ugandan enterprise there's no significant difference in firm age and to your comment I think that's a very good idea I'm not sure how much I can do with a small sample size but I will definitely look into it I think that's a good suggestion all right thank you very much