 Okay thank you all for coming so I don't know what I'm supposed to say about the city but I joined wider three years ago and I was actually hired to work on the fiscal states project and also on the domestic savings project and three years ago I started writing the paper and you tell my colleague has been waiting for the paper for the last two years it is 85 percent complete I've just not been able to bring myself to finish a 15 percent hopefully with the comments today I should be able to do that by the end of the year and otherwise I might lose my job or something so thank you all for coming this is just to rose set the agenda of what we're trying to talk about here so I'm not going to delve too much into the statistics on domestic savings or lack thereof in Cameroon right I'll just show you in terms of the econometrics what I was able to find I added the role for institutions to the title I haven't incorporated any measures of institutions yet but it forms a 15% I just talked about earlier and hopefully I can get that over the line so in terms of the motivation we know of Kunal actually published a paper quite wild back on the importance of investment in determining economic growth and investment in itself is determined by savings one of the primary determinants of high investment rates in countries of those countries that you know have high savings and if savings are high then you are government can invest without having to borrow and if investment is high then growth is going to be higher so this is the paper with world development which was published in 2004 higher savings countries have faster or higher economic growth as I've alluded to earlier that's because those higher saving countries have more in terms of investment and they can invest and you know spur their growth but historically they've been low savings rate in savings rates in South Southern Africa which have been dropping considerably over the years Cameroon is one of those countries that has a high savings rate but a low growth in savings so in terms of the level of savings it's relatively high given the economic structure but the level has kind of stagnated over a while it's not been growing as such hence why we say there's in low growth in savings in Cameroon so the aim of the paper is to decide for the determinants of domestic savings that private savings and gross savings in Cameroon so a bit of an overview of the Cameroonian economy this isn't patented information by the way this can be found with the typical Wikipedia search so it's a low low and middle income country per capita GDP is at one thousand four hundred and ninety nine US dollars which was surprising to me when I saw it thought it was going to be much lower than that the population is approximately 27 million and the GDP growth rate for 2018 was at 4.1 percent which was quite impressive for you know the Central African countries as it's typical of developing countries especially in sub-Saharan Africa there's a huge informal sector which you know comprises about 90 percent of total employment if you can call it to the you know informal sector part of the employment sector and it's mostly people operating in the primary sector so agriculture and mining querying or just the very basic skills which are needed to produce in the primary sector the government is the largest employer in the country employing about 65 percent of the population the World Bank data shows that the growth of the population is greater than the poverty reduction rate and this is evident by the number of people who have been you know brought into poverty over the years exacerbated by the coronavirus pandemic and there is also some civil strife in Cameroon the marginalized anglophone regions where I manage from have been you know rowing with the government so it's also contributing to the general lack of opportunities and the poverty which has been increasing in the country so in terms of the fiscal policy framework as expected the government controls fiscal policy which has generally been expansionary for the period over on the review so Cameroon has quite significant government spending and to fund their spending they don't raise as much in taxes as is expected which is a you know key theme we have in this conference and they rely quite a lot on debt most of the debt is you know resource backed loans from China but they also borrow from you know bilateral and multilateral organizations as well as international capital markets being a former colony of francs they do get a lot of grants from francs as well and technical assistance but quite a significant amount of money in terms of loans the fiscal system is quite complicated so it has a multiplicity of taxes relatively high statutory rates and I think the statutory CIT rating Cameroon is 32% which is quite high for a country without as many companies but their effective tax rates which these companies pay sometimes goal goes lower than 15% because you know they benefit a lot from tax incentives and all kinds of tax exemptions the country also lags in terms of infrastructural development with a lot of uncompleted projects or projects which are carried out in a substandard manner which is you know the direct result of corruption in the country the monetary policy framework is controlled by what we call Bayek so in French is Bank des Etats de l'Afrique Central which is a bank for Central African states so it serves as the central bank for Cameroon as well as other countries in the Semarque region so that's Gabon which just had a coup by the way Equatorial Guinea Chad the Republic of Congo and Cameroon and Central African Republic as well so all these countries are being served by the Bayek which is headquartered in Youndé so the central bank conducts the monetary policy the manner that ensures there is internal and external value of the CFA which is the currency used by you know Cameroon and the other Semarque countries so it does that by adjusting money supplied through direct and indirect means and it has a very strong imposition on reserve requirements in terms of the financial sector policies there are high levels of liquidity in Cameroon and there's a very heavy concentration on deposit and loan activity and a lower level of financial innovation so the financial sector in Cameroon hasn't really tapped into the dividends or they're not ripping the dividends of you know the digitalized economy just yet there's so very much loans and deposit kind of schemes and a lot of what you have to do for example when civil servants have to get their salaries unlike at wider way you just get a bank notification from Nodia in Cameroon you have to queue up at the bank and then try to get your money in cash so usually the salaries are paid by the 20 feet of the month and you know civil servants get very popular at the end of the month but everybody has to queue up naturally and they've not really ripped the dividends of you know digitalization in the country the banking industry of course dominates the financial sector and there are 15 commercial bonds I think it's gone up to 18 now but there is a you know poor city of companies that are listed in the stock exchange which is the Duala stock exchange in Cameroon so the low rate of financial inclusion and increased informal financial services and credit unions are actually contributing to these low savings rates in the country so there are low interest rates small collateral requirements lower levels of bureaucracy but then the country is still not tapping into all the benefits of those low interest rates at the time and as I mentioned earlier there's low uptake in digitalization of financial services so there are ATMs and a lot of the activities are now going digital so you can actually do some transactions with your friends or with your collaborators without having to set food in the bank but it's still it's at the nascent stages you know and it's fraught with complications and there are very strong capital requirements in terms of what you can trade so for example I think the maximum of what you can trade in a day online goes up to about 300 euros and if you have to do anything above 300 euros then you have to go to the bank and take those long cues I was talking about so we try to do some time series analysis to look at the determinants of private savings and gross savings which we get from the World Development Indicators database the independent variables are following the literature so we have the public savings rate the interest rate broad money the log of GDP per capita GDP per capita growth inflation terms of trade population growth and domestic credit to the private sector what's important to point out is that each of these variables has been postulated to influence savings in the literature and there is really no consensus of the consensus view on the direction of effects per se for example terms of trade might have a positive impact on gross savings as well as might have a negative impact for things like the log of GDP per capita you would expect that you know as your country becomes more developed the financial sector expands and then you know there's an uptake of digitalization that's going to lead to an increase in all types of savings but for the other variables is not immediately clear if it's going to be a positive or a negative effect so the estimation strategy we do some econometrics here which I'm not going to delve into too much the equation is based on the life cycle hypothesis which again I'm not going to talk about quite much but it postulates that spending and saving habits of an economic agent that dependent on their future incomes over the course of their lifetimes and this life cycle hypothesis was actually postulated by Kunal and his co-author in the paper in 2004 which has now been used you know to talk about savings private savings and national savings so we employ what we call the autoregressive distributed lab technique I'm not sure how familiar the audience is with this technique but I've just given some three points on you know why I think it's a suitable technique for that but to avoid us having a very long and tangential discussion we can talk about the econometrics of a coffee but one of the attractive points is it's suitable for time series analysis covering you know a shorter sample our sample is from 1980 to 2018 so that's quite long but if it were shorter than that the ARDL will be suitable for that it's also very good because you have a mixture of variables what we refer to as I0 and I1 processes so for those who are familiar with the time series literature it means that it's a mixture of what we call stationary variables variables that revert to the mean after a few deviations and non-stationary variables which are typically trending over time so for example the tax to GDP ratio is typically a non-stationary variable because it's very sticky and it trends over time government spending as well as typically a non-stationary variable because it's always trending over time and it's very rare to find government spending reduced you know it typically increases over time and another attractive aspect of the method it permits distinction between the short short run effects and the long run effects of independent variables which you know is quite important so as a precursor to actually estimating the empirics we do some stationarity testing so we want to know if the variables are I0 or if they're I1 this is just so we know how to develop the long run model if for example all the variables are I0 which means that they are stationary then we don't have to worry about the long run model is everything reduces to a short run model but if we have a mixture of both then we need to have the optimal lag length which is chosen using information criteria and then the lag length is going to determine how much of a long run and short run relationship we're going to have so what we would find here where you see the stars it means that you know the variable is I0 without the stars and the variable is I1 so as you'd see there's a mixture of I0 variables and I1 variables which is perfectly fine if you use an alternative method like the vector auto regression you would naturally want all your variables to be I1 that means you'd want them to be trending over time such that you can reparameterize it into an error correction model but because we're using the ARDL we don't mind having I0 and I1 variables together in the equation. A very fundamental part of applying this ARDL model is testing for cointegration so you have the PESA run at R2001 bounce test for cointegration so the null hypothesis is that there is no long run relationship between variables and by that I mean there's no cointegration between variables the alternative or the alternative hypothesis is that there is a long run relationship between variables and what we want to be able to prove is that we can reject the null of no long run relationship between variables we want to be able to show that in the long run these variables move together or the variables are related so the interpretation of this bounce test you look at what we call the F statistic if the F statistic is greater than the critical values we reject the null hypothesis so if you look at this I hope you can see from the back the F statistic is consistently greater than the critical values at 1%, 5% and 10% for private savings which shows that we can reject the null of no long run relationship between variables which demonstrates that the variables are cointegrated. We'd see the same thing for the gross savings equation where the F statistic is consistently higher than the critical values at 1%, 5% and 10% again demonstrating that we can reject the null hypothesis of no long run relationship between variables and conclude that the variables indeed have a long run relationship hence they are cointegrated so this is the cointegration test for private savings the cointegration test for gross savings and then we look at the long run estimates so for each dependent variable there are four different specifications and various independent variables that are introduced progressively I don't know if you can see this from the back but the first independent variable is the adjustment term which is what we refer to as the error correction term because we are doing a long run and short run model that adjustment term is significant and negative which shows that if there is a shock to the system the country can always revert to its long run level. Public savings have a negative relationship on private savings so if your public savings increase it naturally leads to a reduction in private savings because you know people don't feel or there's no pressure exerted for them to save as much anymore. The interest rates are negative and significant at one point which of course is quite intuitive right if interest rates are quite high people don't get to save if the interest rates are lower than the people who get to save. GDP per capita is positive which when you include GDP per capita it's kind of a proxy of the level of development and you know improvement in financial services, improvement in digitalization and all that so it's kind of a catch all for everything which may not be quantifiable but which are characteristic of a country that's you know in development and it's expected to have a positive relationship. GDP per capita growth that's a bit different sometimes it's negative sometimes it's positive so if a country is growing again it might demonstrate that there are other aspects you know which are not quantifiable which are going to increase private savings maybe in terms of the number of banks which are now there in terms of the kinds of services which are offered but then it might also be a negative relationship such that the growth in the country is funded almost entirely by the government and then it leads to a reduction in private savings so that some kind of social fiscal contract where people don't get to save as much anymore because they can depend more on the government which is demonstrated by an increase in growth with the government you know as a proxy for the government improving as well. Domestic credit to the private sector is you know a proxy for financial sector development and as expected it has a positive relationship. So gross savings again you see that the adjustment term is negative and significant GDP per capita is also you know positive and significant. The idea about GDP per capita growth till applies here it can be positive which is to be expected which shows that the country has been improving and you know whatever that entails as well as it can be negative which shows a different kind of fiscal contract where people tend to depend a bit more on the government and then savings don't increase and the government doesn't save as much either because it has to spend a bit more. Broad money is also a proxy for you know money supply which of course increases savings. The interest rates again are negative so higher interest rates and you know lower savings and all. When you look at the things at the end so the diagnostic tests are used to us attain the you know stability of the model. So we have a dubbing Watson statistic and then we have the Bruce Godfrey test and both failed to reject the null hypothesis of no autocorrelation. So this is a time series model so autocorrelation is always going to be an issue and being able to reject the null hypothesis sorry being unable to reject the null hypothesis of no autocorrelation is quite important. We also have to test for homoscedasticity and for that we use the Bruce Pagan test and the test also fails to reject the null hypothesis of homoscedastic errors which is also quite important. In terms of structural stability tests you do the COSUM square test and the COSUM square falls within the 5% confidence bound hence the model estimates are stable. So these are the references from the African Development Bank and some of the papers published on Cameroon which might give you more of an idea of the issue and that's the end of my presentation. I've touched something crazy here.