 Good day. Now, my paper is the titled double-edged sword heterogeneity within the informal sector. Now, despite a high unemployment rate and widespread poverty, South Africa's informal sector is relatively small compared to other African countries. And this is usually attributed to either high reservation wage or capital constraints providing barriers to entry into the informal sector. Now, this paper plans to explore the incentives that motivate the entry into or the exit from the informal sector. And okay, it's ongoing work. So so far what appears to be happening is that these incentives differ depending on whether the person was an active job seeker or an inactive job seeker. I used the labor force survey that which is a rotating panel data set collected by statistics South Africa and in the period September 2001 to March 2004, we were able to extract the panel component of the data set. Now, according to statistics South Africa, a person is defied as being in an informal sector if they are classified, if they are in an enterprise, which is not registered and also is has a size of less than five workers. Now, this is a rather, you know, broad definition. So I looked at the occupations that the people find themselves in. Note also that I restricted my sample to black females because I didn't want to get into problems of race or gender when interpreting the results. Okay, so most of the people about a little less than 92% are in elementary occupations. They are service workers and you know craft and related trade workers. Now, these are the small scale retail enterprises which don't earn a lot of money for the people. Now, to give an idea of the well-being of people who are in the informal sector versus people who are in broad unemployment, I constructed kernel densities of log of household expenditure and here you can see, okay, the dotted line is the people who are in the broad unemployment category and the solid line is people who are in the informal sector and just to look at this, we see that the people who are in the informal sector seem to have a lower household expenditure and if we look at how much they actually earn within the informal sector, now I have decomposed it into active and inactive work seekers. You see that they earn, you know, roughly the same amount of money. Now, in order to look at, you know, the rough transition matrices of how people move between sectors, well, between labour market outcomes, I see that active work seekers, okay, by the way, the people in the columns, the columns represent time t plus 1 and the rows represent time t, so you look at it this way. So the active work seekers are more likely to be formally employed than informally employed, whereas the inactive work seekers are more likely to go into the informal sector rather than the informal sector and if we look at the exit rates from the informal sector, these are quite high. I mean, the entry is only around 7% and here we have 16% to 15% who are exiting into the various forms of unemployment. Okay, so then I wanted to construct a theoretical model to see, you know, how this and everything is motivated and what I did was take the classical job search theory model and had an explicit informal sector there, so now what is usually had in the classical job search theory is just formal sector. Now I had an explicit informal sector and a formal sector and from that I could find a latent variable that was constructed by showing, you know, the difference between the income opportunities and the person's reservation wage, given their physical capital, their human capital and how much value they get from their home production. From, okay, this will help you. I think this will help you more for discussions to the end. Peters represents income opportunities and the fares represents the reservation wage effect. So now in this data set that I had, I didn't have a variable that explicitly said what physical capital was. So I used a multiple indicator instrumental variable approach to sort of solve this omitted variable bias. I used, as an indicator, I used the long households for capital expenditure and as an IV, I used the state given old age pension grant and to remove a further source of endogeneity, which would be from an observable effect, I just used fixed-effect estimation. Okay, now I want to, due to time constraints, I want to just concentrate on the two main columns here because this is where the story really lies for me. It seems as if the active seekers are pulled into the informal sector because here, if we look at the long household expenditure, and remember the long household expenditure, we're using to see what's happening with the physical capital. A person gains way more in terms of, you know, wages, which would be the income effect than, you know, they would raise their reservation wage if they were to get an increase in the household expenditure or the capital. Whereas the people who are inactive are actually pushed into the informal sector. This means that if you lower the household expenditure, the person, you know, has to, the reservation wage effect, if you heighten, sorry, if you increase the long household expenditure, the reservation wage increases way more than what they would have had an increase in terms of income opportunities they would have had with the same kind of capital. So you'd have to actually decrease the long household expenditure in order to push them into the informal sector. Now, okay, again, this is a bit of ongoing work, and I am a bit worried about what the long of household expenditure variable is saying because we know from literature, especially from Kingdom and Knight, that the people, the active work seekers and inactive work seekers are the same. In fact, Kingdom and Knight also say that the inactive work seekers are actually discouraged work seekers. So the results, having differing results is actually a bit, you know, controversial. So what I know the problem is in terms of the household expenditure variable in the instatistic South Africa is that it was underreported. And so, okay, to try to counter that, I tried to use some interval regression using the household income, extracting what the person's actual earnings were from that regression in order to see what, you know, a bit of a higher estimate. But it's still giving me very bad results. So I'd like to get some feedback into which is the most honest way to construct a household expenditure variable that would reflect what is happening in terms of capital for the individuals.