 My name is Apia Kubi Johnson and I'm doing the study with my other senior colleagues, Robert Isaac and then Charles and also with Juka from University of Tampa. At this point it's basically using the gap tool to estimate poverty rates for Ghana using the various datasets available. By role of history, Ghana has about 25 million population, the first country in South Africa to gain independence in 1957 and also currently a new middle-income country with per capita GMP above 1500 US dollars, which is quite controversial because a lot of people say they don't fit in their pocket. So in about two decades GDP per capita has doubled which is quite interesting from just about 1200 to 2100 US dollars in 2005. In terms of human development in Ghana it's about 135 with life expectancy about 65 years and also 7 years of schooling so that's quite interesting. By role of growth rates you can see that from this graph Ghana is quite improving in terms of growth rate. Even though initially there were decreasing growth rates but now it's improving and with the discovery and new production of oil also. The new growth rates are very interesting with 14% the current 2011 figures, 40% very high and very encouraging. And most of these nice growth rates we see may have been accounted by the group purposes and also anti poverty policies that have been adopted. So by way of review there have been some policies adopted and implemented and that may have contributed to the improvement. One programme was called PAMPSCAT, it was a programme to mitigate the adjustment in the economy since 1989. Another programme was a vision 2020, it's a very ambitious strategy but failed to improve higher growth rates in the past. Two other interesting ones are the GDP rates 1 and 2, targeting poverty actually. The first one was 2002 was initially done to help Ghana benefit from the HIPID initiative and that was interesting because a lot of programmes benefited from it which we also achieved macro stability through that and that's mostly the time that we started seeing faster growth rates. Also the GPRS S2006 2009 also targeted poverty so it was quite interesting. Other programmes included in the GPRS 2 was the LEAP which was the first cash transfer programme actually targeting the poor since 2008 with inclusion categories being very vulnerable, single parents, orphan children, extremely poor people, people who are incapable of working for themselves and currently the inclusion category of households in the scheme is increasing which is quite interesting in Ghana. Now by way of this study we use the data sets collected by the National Task Force Service from 1991 to 2006 and there are three data sets collected by the National Task Force Service from 1991 to 2006 and there are three data sets collected by the National Task Force Service Service from 2005 to 2006. They have similar methodology so the data sets are quite comparable for this kind of analysis and we also augmented the analysis by one recent data set called the ISAEL panel. That's the baseline survey that in 2009-2010 and even though the methodology is quite a little different they are quite similar in terms of expenditure figures and one advantage of the Yale baseline is that it contains quantities whilst the Yale analysis data did not contain quantities, they did not capture quantities but they captured a expenditure and also prizes so we could deduce the quantities from the expenditures and the prizes. Now by way of empirical approach with this some things a little different from the original gap pool, one thing is we used because we didn't have actual quantities in the Yale analysis data sets we used the prizes as the cluster level prizes to divide the quantity, the values, the expenditures to get the quantities and also unlike the standard tool kits whereby the prizes divide the values to get the quantities here we used and the prizes are then used to deflate the actual values by your prize index. We used directly the original food prize index in the data sets which was collected so we relied on it and initially we experimented with ten regions because there are ten regions in Ghana but the results were not very encouraging so we narrowed down to three regions that is across the capital being one region and then other ever centers being one region and also all rural communities being one region so we experimented with three regions and I think the results were quite encouraging and we also experimented with just one region for the whole country to see how the results would be. Now by way of results if you compare the official rates with the gap code rates you are quite interesting. For the first data sets the gap computed rate is about 20% even though the official rate is about 52% which is quite a big gap and for the second data sets the updated rate is 29% whereas the official rate is about 40% it is the last data set which is quite interesting the official rate is 29% and the gap code is 34% but if you compute the correlation between for the ten regions and as you can see on the map the correlation between the official rates and the gap code rate I think is quite high for the first data sets the gap code rate is about 0.7% and for the two other data sets the gap code rate is about 0.9% so even though the levels are different in terms of the ranking per region you can see that it is quite reliable if the rates are being used as a tool for distribution of wealth and also targeting for regions I think the two data sets are comparable and the two methodologies can be relied upon even though the levels of the property predicted are different so that's one thing that we were happy about the results now we... the last data set that we used that is the yield baseline I think the property predicted rates were just about 20% which is the lowest in the computations and that is not so surprising because as the economy is growing and there are so many properties and property measures we'll be going down but the extent and the change in the property is supposed to be confirmed because the methodology for the yield baseline is a little different from the latest data set so we can't say that it's a change of 10% no, we can't say that but we are still experimenting with the results and see what why the gap code property lines are so lower than the national property lines