 So Finn sent me an email and said, come and do a presentation on South African labor markets, which for something like me is quite daunting because it means condensing lots of academic work into a storyline that hangs together in 20 minutes and then also perhaps to pick out some common themes. So I've sailed fairly close to the wind in terms of Finn's request, but I'm hoping that we can see some common stories at least for those of you from developing countries and in particular middle income countries. So what I'll try and do is give you firstly just to position South Africa for those of you not aware of anything about the country, then using the entry point of growth path dynamics because I think if you look at the dynamics of South Africa's growth path at least since the end of apartheid we get a strong sense of the nature of the labor market challenges which then is the next section and then I'll go on to just concentrating on three areas of what I think explains the poor labor market performance in the South African economy. So we have a population size I think approximately the same as South Korea, we're an upper middle income country, you've got a current account deficit which in fact explains much of our short term cyclical movements and in fact our growth path dependence and something I'll come back to quite a common theme in South Africa is a very low share as a percentage of GDP in terms of manufacturing and then this outlier statistic is the Gini coefficient. So essentially a middle income country with high levels of income inequality and a fairly undynamic manufacturing sector. I find these pictures incredibly useful to explain in sort of one visual the nature of an economy's growth path. So this is a growth incidence curve for those of you and I think most of you in the audience know about these but they plot essentially the lowest to the highest percentile of households over the period I've done the first 15 years of democracies and 95 more or less to 2010 and it asks the question what has been the growth rate across these percentiles. It's a very strong visual because it suggests very clearly an unequal growth path. So what you've seen is growth gains for those at the top end of the distribution and the state has then used the growth gains in the form of a revenue pool to finance social security at the bottom end. And essentially that's the growth story for South Africa and anti pro and anti poor or lack of pro pro growth gains to the top end in terms of organic growth filtering to the bottom through social assistance but very little going on in the middle of the distribution. So I'm missing middle story in South Africa is really about about this story. I want to concentrate on this and ask the question where is the labor market in the story and that's really the theme or the sub theme in the talk is what's happening to the labor market. If you look at the bottom end and I remove from the graphic which is our income and expenditure server I may have switched the variables. If I remove social assistance from the graphic that's what happens. So this essentially is about regular income right from economic activity. And essentially you've got a labor market that doesn't operate and doesn't function at the bottom end of the distribution. And that's a very very unusual story for developing country and for middle income country such as South Africa because in most other economies if you remove social assistance you don't see this type of collapse and something I'll argue later about in terms of why that happens in South Africa. This is just another pretty picture the Gini coefficient right. So the story you don't really see from the growth incidence curve for the previous two slides is what happens to the Gini. This is literally the Gini coefficient across the two years. So that's 1995 and 2005 but as soon as you remove social assistance the Gini coefficient increases sharply. So essentially what the social assistance which is old age pension system child support grants something France who's in the audience is going to talk about tomorrow. What it's done is to suppress income inequality. So essentially you've got a growth dynamic that is no evidence of proper growth rising income inequality something I didn't show but it's risen in the 15 years since apartheid has ended. And effectively you've got a growth policy debate which says that you've got this implicit contract that's broken down because the informal sector is nowhere to be seen and I'll talk a little bit more about that. Neither the non unionized and neither essentially a minimum wage is excluded from a growth process that is essentially skills biased in nature. Why are we such an outlier? Well here's the first statistic I just picked a few from the World Bank data. Essentially upper middle income countries I unfortunately don't have Vietnam here but I do have it in the remaining slides so apologies for that. But essentially they're South Africa. So we've got an ILO defined unemployment rate in excess depending on the quarter of 25% but a complete outlier in the upper middle income country set and even if you widen the set to include lower middle income countries. I want to argue that in most economies upper middle income countries the following two slides the share of self-employed and the share of agriculture in total employment closes the labor market that in most upper middle income countries is unable to create a sufficient number of formal sector wage jobs. So in most developing countries if you like unemployment rates are suppressed arguably alright through large numbers of people in self-employment or various forms of segmented informal sector employment and a fairly dynamic high employment agricultural sector. In South Africa you don't see that. So this is literally the share of self-employed in total employment in country estimates and there's Vietnam you can see there's a nice comparison over 60% of workers in the Vietnamese labor market are in self-employment in South Africa it hovers between 12 and 15% and again an outlier there I'd welcome some views on Russia but that is an odd result in the middle income country set. If you look at agricultural employment say if there's no urban informal sector employment because of low formal sector jobs a low number of formal sector jobs often the agricultural sector provides yes low quality jobs with jobs are being provided it's not the case in South Africa and if you look at Vietnam that's an again another interesting comparison. We do though have in terms of employment some industry employment that is comparable to other middle income countries. This is a quick overview of all the numbers that you've seen before but just a few additional ones is our very high unemployment rates unsurprisingly and I can explain that later we don't have time a very low labor force participation rate very low employment in agriculture some industry employment but a fairly large public sector labor market. So not only are we unable to create a sufficient number of formal sector jobs but it seems that most of these formal sector jobs are actually being created in the in the public sector and I'll come back to that. So I want to talk about why we have such a poor labor market performance in South Africa and this is a completely truncated discussion focusing on my pet subjects namely the labor reform issues but I'll talk a little bit about what I think is driving this poor labor market performance namely a growth path dependency and I should stress and emphasize that the whole series of other factors that are important. So here's a window into this growth path dependency this is that towards as just on the cusp of apartheid ending which is 1993 the share of GDP by sector and then I go to 2012 and look at the numbers. What happens in most developing countries that are creating jobs and growing which South Africa did over this period is that the share of GDP accounted for by manufacturing in fact grows. In South Africa it stays constant and depending on the period actually declines. In this case it's declined from 19 to 17 percent. In contrast the share of GDP accounted for by financial and business services actually increases dramatically and that's another way of saying that you've got a growth trajectory based essentially on creating jobs in financial and business services and an undynamic fairly capital intensive large-scale manufacturing industry that's not creating a sufficient number of jobs. In addition and I'll show that later the mining industry has contracted and agriculture has done nothing. And so beyond this growth trajectory that some would argue is capital intensive is based on uncompetitive and undynamic manufacturing sector that's never going to reduce sufficiently reduce the double digit unemployment rates that we see. Another way to think about it is that if you do have a sectoral strategy or growth path that is labor intensive if you imagine a 45 degree line here and we've got gross value added and employment here a labor intensive growth strategy would be one that's somewhere located over there above the 45 degree line where the proportion of GDP or value added growth is exceeded by the growth in employment. You don't see that in any sector in South Africa and that's the growth trajectory that's the growth trajectory we've been on since 1993 and certainly even for this period 2001 to 2012. I'll come back to what's happened in agriculture under the labor reform issue. So the employment story of that GDP version is that you we've destroyed jobs in agriculture and mining mining's been because of the nature of our mining capital intensive mining but arguably also because of a strong trade unions and most of the jobs have been created in the public sector accounting for 42% right. These are sort of relative employment shifts relative to the aggregate shift and in financial and business services. It turns out that and it's not shown here that within financial and business services it's the work we did for Martin for the WDR a lot of those jobs are temporary employment service provider jobs. So in fact a lot of those jobs are employers trying to get around and circumvent the labor regulatory regime and I'll be interested in the audience from the different countries whether this sector is actually growing in the developing country context I think it is which is this notion of trying to avoid the regulatory net. But what we've also done within that structure of growth these are simple as I warned you these are all sort of I think five or six different published papers drawing in different tables. These are based on these cats and Murphy decompositions which look at between and within sector shifts and ask the question to what extent the technology influence changes in labor demand patterns. We essentially find that technology has changed our labor demand preferences and our relative labor demand shifts and essentially what this points to though is an increased preference over the time period here for highly skilled workers. So amidst this lack of a lack of dynamism in the manufacturing sector the growth that's happened is not only capital intensive in nature but also skills intensive and that for me is probably the growth trajectory that we conceivably going to be on for the next 10 years and it's the growth trajectory that's not going to create the kind of jobs we want. So because of time I'll skip just saying what I've just told you again that's in the slide. Let me turn quickly to the labor reform story and take pick on two big big issues around labor reform. First is to look at you know we use again the doing the doing business of the World Bank and you find that if you take your EPLs your employment protection legislation and you and you you find the measure for your particular country and you look at a global distribution and I've got the upper middle income country distribution. South Africa's labor regulatory regime for an unemployment rate of 25% is not particularly high. Gary and I had this discussion earlier so at the 62nd percentile or they're about 65th I think it is for 25% unemployment rate I'd expect us to be much higher. So there's either something wrong in our interpretation of the ranking or labor legislation doesn't matter. I think it's the limitations of the doing business survey rankings but I'm happy to talk about that further. So legislation and labor reform and the labor regulatory net does have bites. This is based on work that Ravi and myself and Ben Stanix have done and I've just literally taken out the difference and difference estimator which asks the question this is the really important guy here right. Ask the question what has been the impact of the minimum wage right relative to some control group in agriculture right. So there we go. It's in agriculture and it's negative and significant and our results based on even the descriptive numbers show that the minimum wage in agriculture had a negative and significant impact on employment in that sector and that's the little bubble you saw right at the bottom in despite a growing value added in GDP or in growth in agriculture you saw contraction in employment and so we've thrown everything at this model whether it's exports where they exchange rates GDP all sorts of things and this result is robust. So there you have a very particular labor reform intervention that has a negative impact on jobs right. So the massive job losses we saw right at the beginning in that slide in agriculture was as a consequence of labor reform. It's not that simple though because when we do the same thing difference and difference same technique same estimation same data as it turns out for all the other minimum wages that we've had South Africa as a sectoral minimum wage system it turns out that the employment effects are fairly benign in fact they're insignificant right. So at the extensive margin there's some action on the intensive margin where hours of work in one or two sectors like taxi industry and so on are actually adjusted but in most cases there isn't a significant disemployment effect from minimum wage laws and so that then begs the question also about how one thinks about minimum wages in what is a contested terrain often between interest groups unions on the one hand and employers on the other hand right. So just incidentally whenever I speak to government luckily I think my national treasury friend is not here. They hate this result and they love this result and it's interesting always to see the but they always say no this is great and then when you show them the agriculture one they say you but what about this and didn't you think about that and so it's very interesting to see people's results people's reactions. One of the things we often talk about is trade unions and institutions and how they matter so I'm moving on to the second labor reform story here and it turns out that again in the South African context we've written a separate paper for Justin's handbook. These are estimates of wage premium using earnings functions right and I've taken Freeman's estimates from his study of different developing countries and I've inserted ours and South Africa's that wage premium coefficient is not that high so the pure sort of narrow wage gains offered by the trade union movement isn't particularly high. Neither are wage union density levels right. Neither as it turns out are sort of strike intensity levels but what does matter is our inefficient and ineffective dispute resolution system so I'm going through this fairly quickly what I'm getting at is the notion that trade unions don't seem to matter at least in a narrow earnings function estimate. They may matter in a political economy sense but that's a different debate but what does seem to matter in the labor regulatory regime is what happens in terms of dispute resolution and this is an estimate based on what happens right in terms of the dependent variable being the number of employed by province or district over time and what happens if you try and resolve some of you in our crowd have seen this already what happens if you try and resolve these disputes right through various means whether it's arbitration conciliation and and and and and other means right. What we essentially find in the specification three is the important one right is that institutional inefficiency pertains so resolution of disputes is negatively associated with employment shifts and what it what it points to and alludes to is the fact that it's not so much trade union power or labor legislation but rather inefficiency in the courts of law whether it's the labor courts the dispute resolution systems of a country that seems to be at least in the South African context the binding constraint in terms of employment generation both in time t but also t plus one as employers look to want to hire workers in future periods seem to be constrained primarily by labor market institutions so what where do we come up with in terms of our labor reform story for me the political power of trade unions in the South African context at context met matters more than the pure wage premium effects dispute resolution bodies seem to be the institutional binding constraint rather than anything else I'll get about the strike impact I think it's overstated based on the proportion of workers on strike based on strike intensity length of strike and so on if we look at international numbers but certainly mining I think is a special category and we can we can debate that I'm happy to do that so in conclusion it is about reform this is a smorgasbord of policy reform packages that you can think about in a high unemployment middle income country setting that that if the like doesn't have that the perhaps is highly concentrated there's stuff I think which is really really important about growing the informal sector I think we don't emphasize enough the role played by public sector procurement in in micro enterprise growth also think there's there's a lot to be done about risk mitigation for the informal sector they're also I'm almost on last slide they're also issues around reducing the cost of search for the unemployed and I think public and private employment services in developing countries don't work efficiently enough and effectively together and then a whole bunch of things about transport vouchers and so on I'll end off by these three labor reform packages which I think are worth looking at in in at least in the developing country sense and with the notion that at least in my conversations with Gary realizing that the search for common themes is a starting point in terms of how labor reform can assist in creating employment or preventing employment losses but certainly I think country specific factors are really crucial in trying to understand the dynamics of employment generation and preventing job losses thanks