Growth, Distribution, and Poverty in Cameroon: A Poverty Analysis Macroeconomic Simulator s Approach

Transcription

1 Poverty and Economic Policy Research Network Research Proosal Growth, istribution, and Poverty in Cameroon: A Poverty Analysis Macroeconomic Simulator s Aroach By Arsene Honore Gideon NKAMA University of aounde II aounde Cameroon With the Collaboration of The irection of Forecast Ministry of Finance and Budget Personal Address S/O Lazare BELA, irector of Forecast Ministry of Finance and Budget P.O. Box 312 aounde Cameroon Telehone: (237) ahgnkama@yahoo.com

2 1. Research Question and Core Research Objective After enjoying an extended eriod of roserity from the indeendence to the mid 80s, Cameroon faced a severe crisis that lasted nearly a decade. The crisis had serious consequences for the country s socio-economic trends. It led notably to a drastic decline in ublic and rivate revenues; a decline in agricultural roduction for exort revenues; the closing of many enterrises leading to high unemloyment; the reduction of rivate and ublic salaries The combined effect of all these events led to the deterioration of living conditions. Poverty became widesread and deeer. The ercentage of eole living under overty line increased from 40 er cent in 1984 to 50.5 er cent in Even if the triennial rogram suorted by the World Bank and International Monetary Fund IMF ( ) and the IMF Poverty Reduction and Growth Facility (PRGF) have had slightly ositive effects on overty reduction between 1996 and 2001, more needs to be done in Cameroon in the context of overty reduction. In fact, looking at overty long trend evolution, average er caita consumtion at constant rices reached the eak in 1983 and comared to its 1964 level, one realises that the average Cameroonian was in 2001 oorer than he was in The imortance of Cameroon overty varies from rural (very high) to urban (relatively high) areas. Since the end of the 80s, suorted by international institutions (notably the World Bank and IMF), Cameroon began to imlement adjustment olicies that aim at reducing overty among its oulations. With a deeening overty and unsustainable debt service burden, Cameroon satisfied a three-year track record of imlementations under the suervision of the World Bank and the IMF from 1997 to Success in this three-year rogram took Cameroon to the ecision Point meaning qualification as a Heavily Indebted Poor Country (HIPC country). The Government of the Reublic of Cameroon reared an Interim Reduction Strategy Paer (I-PRSP) during the first half of The I-PRSP was reared in accordance with a articiatory rocess that gave the oorest inhabitants an oortunity to freely exlain their ercetion of overty, to list the major determinants of overty and to recommend measures to alleviate it. The full PRSP is awaited in January

3 The central question of the study is to know what has been the imact of such shocks and olicies on Cameroonian living standard and which messages can we foresee in the future. The obvious objective of the study is to evaluate the imact of shocks and olicies on the living standard of Cameroonians and roject over a chosen time eriod the overty and distributional effects of macroeconomic and structural olicies on Cameroon oulations. This central objective can be divided into three secific objectives: - udate overty rofile in Cameroon - roject Cameroon s overty indicators over a time eriod ( ) - comare rojected indicators with the millennium develoment goals 2. Scientific Contribution of the Research Poverty is a situation in which eole lack food, shelter and clothing, healthcare, education Those falling in such a difficult situation always fear for their future, thinking and worrying about how will be the next day or even the next hour of their life. As ointed out by the World Bank (2001), overty is even more than that because oor eole are often treated badly by the State s institutions and excluded from ower by these institutions. Poor ersons are characterised by low (and even no) income, exlaining their low level of consumtion. At this level, a oor erson is a erson whose income or consumtion level is below a minimum necessary to meet basic needs. The concet of Poverty Line is used to indicate this minimum level. Living conditions and habits vary over time and from one lace to another. For international comarisons, the World Bank uses reference lines set at one US dollar and two US dollars er day in 1993 urchasing ower arity. Studies that had been done in Cameroon in the context of evaluating the country s level of overty are based on the following national surveys: EBC83/84, ECM93, ECAM I (1996) and ECAM II (2001). These surveys emhasised the imortance of income 2

4 overty. Income overty is measured by a consumtion-based overty line that can be defined as the amount of exenditure required for the acquisition of minimum food for an adult-equivalent (2400 calories and 2900 calories er caita er day were considered resectively in 1996 and 2001 estimations) and the basic necessities. Peole living under this estimated line are considered to be oor. In 1996, according to ECAM I, Cameroon overty line was estimated at CFA Francs er caita and er year. Peole living with less than this amount were estimated at 50.5 er cent of total oulation. In 2001, according to ECAM II, eole living with less than CFA Francs are considered to be oor and reresent 40.2 er cent of total oulation, slightly the same ercentage than in 1984 (40 er cent). esite the recent ositive evolution in the fight against overty, one can observe that in a long run, Cameroon imoverished. This imoverishment can be observed by the longterm trend of er caita consumtion in aounde, which shows that in 2001, aounde inhabitants became oorer than in Table 1: aounde, er caita consumtion in 1983 CFA Francs ears Per caita consumtion Sources: comuted from EBC 83/84, ECM 93, ECAM I (1996) and ECAM II (2001) The above-mentioned national surveys led to a multitude of overty rofile and overty assessment studies in Cameroon. In fact these studies are based above-mentioned national surveys. Among these studies are the World Bank (1995), UNP (1998) and several other reorts on the toic. The last overty rofile and overty assessment were done using ECAM I data. Since then, ECAM II was roduced. The final PRSP will be officially ready in January The two documents rovide additional imortant information about Cameroon overty level as well as its strategies of alleviating it. It is now worth to roduce a new synthesis of existing information concerning the oor, overty and inequality to udate the former overty rofiles. A second and more 3

5 imortant oint is that, resently, no study considers all these surveys together with government macroeconomic olicies to roject overty evolution in Cameroon over time. Such activity would lead to more efficiency in overty reduction olicies and olicies recommendations for tackling overty in Cameroon. It is therefore very worth roducing information set on this area. This is an advantage for the resent study to consider these two asects of overty in Cameroon. 3. Policy Relevance Cameroon is actually struggling against widesread overty among its oulations. It will be with great interest to inform government and international donors about the outcomes of this battle and question about the ossibility of wining this battle that can be materialised by reaching the millennium develoment goals. This activity would hel adjust olicy instruments in the case those goals are not likely to be achieved. 4. The Methodology Aart from udating Cameroon overty rofile, the main oint of the methodology will focus on the rojection of the country s overty indicators. To roject overty over time, I will consider the Poverty Analysis Macroeconomic Simulator (PAMS) develoed by the World Bank in 2002 and alied to Burkina Faso. For any change in outut growth desegregated by sector, PAMS simulates changes in income of various Reresentative Households (RHs). The technique is based on the idea that changes in overty can be decomosed into changes due to the uniform growth of income and changes due to changes in relative incomes. PAMS redicts the consequences of any olicy affecting aggregate outut growth on overty under the assumtion that the examined olicy will be distribution neutral or assuming that the distributional change can be secified using a quantifiable form. 4

6 The technique consists in extending the relationshi between macroeconomic outcomes such as GP growth, consumer rice, inflation and emloyment and the income of various grous in the economy. This income is broken down into various socio-economic grous and economic sectors. The solution that is a distributional dynamic rocess between several tyical socio-economic grous used the RH hyothesis. Each RH being emloyed in a different economic sector, it is ossible to desegregate the roduction side of the economy. The labour market reflecting the skill comosition of the labour force, the dichotomy between the rural and the urban sectors, the effect of sectoral outut growth and real wages on the demand for labour is another imortant oint of PAMS. General features of PAMS are: 1) a macro model; 2) a base year household survey; 3) a labour market model and 4) a household survey simulator. The macro model rovides PAMS macro consistency. This can be a general equilibrium model or sohisticated macro-econometric model. The direction of Forecast of the Ministry of Economy and Finance, with the hel of the World Bank and UNP has develoed such model for Cameroon. The coefficients and relationshis of the model are estimated using the country s time series data. The macro model gives national account consistency in both real and nominal terms. It ensures that at an aggregate level, the budget constraints of economic agents are resected. The base year household survey (HHS) rovides information about initial levels of income and exenditure by economic sector of emloyment, by skill levels, by rural or urban location, and by formal and informal activities. The average wage and non-wage income of workers in each RH comes from the latest available household survey. In Cameroon, ECAM II (2001) will be the base year HHS. The labour market model simulates the labour market linked to the consistency of the macro-model. In fact, one can model labour demand and suly and estimate elasticities with country time series. PAMS breaks down the economy into two basic comonents: the rural and the urban areas. Within each comonent, the formal and informal sectors are distinguished. Each sector is broken down into two sub-sectors, the roduction of 5

7 tradable and the roduction of non-tradable, allowing linking each sub-sector of the roduction side of PAMS to each comonent of the labour market. emograhic elements as well as labour and skill categories and exogenous migrations affect Labour suly. The demand of labour is broken down by economic sector, skill level and urban/rural location and is deendent uon relevant sectoral demand and real wages. Hence the modules determines wage income broken by socio-economic categories, skill levels and location. A feature on taxes, transfers and social exenditures hel making average transfers or average taxation of the secific reresentative household. It also rovides simulation about the cost of attaining certain socio-economic goals. 4.1 PAMS secifications Production In PAMS, roduction is broken down as following. Gross omestic Product (GP) hereby refereed by is taken from the macro model. It is therefore exogenous. is broken down between rural ( ) and urban ( ). Rural GP is made u of two sectors: the roduction of cash cros ( ) - tradable goods for exorts - and subsistence agriculture ( ). Urban GP is divided between a formal sector and an informal sector ( ). Private sector and ublic sector (,PUB ) are the main comonents of the formal sector. The urban formal rivate sector is divided into a tradable sector (,PRIV ) and non-tradable sector (,PRIV ). The informal urban sector is viewed as a rivate-non-tradable goods sector. So in total, there are 6 sectors from which one gets 6 reresentative households. These reresentative households are: rural workers of tradable goods sector; rural workers in the non-tradable goods sector; urban workers in the non-tradable formal rivate sector; urban workers in the nontradable formal rivate sector; urban workers in the tradable sector; urban civil servants. Assuming all rices are normalised to one, one can write: = + = ( + ) + [( PRIV +, PRIV ) +, PUB + 6, ]

8 The exort sector is broken down into agricultural exorts and non-agricultural (urban) exorts. At this level, roduction is exogenous, deendent uon foreign demand ( * ) and the resective real exchange rates for each sector (domestic and foreign rices ratio). = ( +,PRIV ) = *, E *, = *, PRIV, E *, There are many otions to determine outut of the rural economy. The first one is to take agricultural roduction as given by the macro model, what will be done in the resent study. Another consists in modelling rural roduction searately. At this level, the simlest way is to calculate using a constant elasticity of outut to the rural labour (δ ). Some more comlicated secifications can also be adoted. For the two simlest secifications, one can write: δ = and = κ. L When cash cros for exorts and rural outut are known, non-tradable rural outut can be calculated as a residual. Urban outut is obtained with the roduction of tradable exorts goods (,PRIV ) and the ublic sector roduct (,PUB ) that is exogenous and fixed., PRIV, PUB = (, PRIV + ) In the urban area, for simlification needs, a fixed coefficient relation to the ratio of investment (I = I,PRIV ) to outut can give the outut s growth rate. 7

9 , PRIV, PRIV, 1 = σ I., PRIV, 1, PRIV, 1 Once determined the five sectors, its is ossible to determine the outut of the informal non-tradable goods as a residual Labour Market, Emloyment and Migration With the additional dimension of the two tyes of labour (skilled and unskilled), emloyment determination in the labour market model follows the same breakdown of the economy. In the model, the total labour force reresents a fraction of total oulation. Each sector of roduction is assumed to hire only one tye of labour (skilled labour or unskilled labour). Unskilled labour is emloyed in the rural sector in both subsistence and exorts sectors and in the non-tradable sector of the urban economy. Unskilled labour emloyed in the informal sector and ublic sector are assumed to be skilled. Skilled labour is emloyed in the urban formal exorts sector. Labour demand is decomosed as follows:,,, SK, SK, L = L + L = ( L, + L, ) + [( L, + L, + L, ) + L, G ] where subscrits and stand for rural and urban resectively, and SK for unskilled and skilled labour resectively, for tradable, for domestic informal nontradable, for domestic formal non-tradable, and G for ublic sector. i) Emloyment, Migration in the Rural Economy Rural economy emloys only unskilled workers. Emloyment is assumed to follow the Lewis tradition of unlimited suly of unskilled workers that grows with the oulation growth rate η(pop). One should also consider the rural-to-urban migration. 8

10 L S, S, = L.[1 + η ( POP) MIGR], Labour demand in the rural sector is a ositive function of both comonents of the rural outut with an elasticity ω. It is a negative function of real wage with an elasticity α. Real wages in cash cro sector ( w, ) are higher than real wages in subsistence agricultural sector ( w, ) so that workers will ot to join first the sector offering a higher real wage. A fixed minimum sectoral subsistence wage w, constitutes the basis for real wage setting. It can adjust if wages are assumed flexible. The nominal wage is the roduct of the real wage by the sectoral roducer rice. L, w,, = κ,. w W, = w, w, = ( 1+ λ1). w, 1 α,, Unskilled labour suly in the informal subsistence is the residual of labour suly minus labour emloyed in the cash cro sector. The same wage determination rincile is adoted in the subsistence agricultural sector. L S, S,, = L, L,, L w,, = κ,., w U, α, w, = ( 1+ λ 2). w, 1 MIGR = MIGR UNSk W, = w, Rural-to-urban migration at the rate MIGR is assumed without any cost. It is a function of the wage differential between rural sector and urban sector, which deends greatly on the robability of finding an unskilled job in the urban economy. 9

11 ii) Emloyment, Ugrading Skills in the Urban Economy In the ublic sector, labour demand is exogenous. Workers are assumed oting first to be hired in the ublic sector because of some benefits associated with ublic sector emloyment., SK L, G = L, G In the rivate sector (the rest of the urban sector), unskilled labour are assigned to formal non-tradable goods sub-sector and skilled labour to the tradable goods sector. Labour demand for unskilled workers deends ositively on urban rivate outut while labour demand for skilled labour deends ositively on urban tradable outut with their resective elesticities ( ϖ and, SK ϖ ). The former deends negatively on the real wage rate with an elasticity α while the latter deends negatively on the real wage SK with an elasticity α. L,, α ϖ, = κ, PRIV, PRIV w, L SK SK α, SK SK ϖ SK, = κ, PRIV. w, The к are the shares of each sector to total GP. The suly of skilled and unskilled labour in the urban economy grows with the rate oulation growth. The nominal wage rate in the ublic sector is exogenous W = W G G iii) Wage determination for unskilled workers 10

12 There is a minimum subsistence wage level set by the institutions. The informal sector is a residual in both roduction and emloyment. The wage rate of unskilled workers in the urban economy can be secified using information of the HHS Prices The GP rice level of the macro model alies to the aggregate roduction. Exorts rices are exogenous (small country hyothesis). If E stands for the nominal exchange rate and * the foreign currency rice of exorts, x = E.*. In rural areas, the weighted average (contribution to agricultural GP) of cash cro rices and subsistence agriculture rices hel calculate the rice index. An increase over the cost elements which include a weighted average of the minimum subsistence wage for the informal rural sector and the formal rural wage cost recas the change in the rice index in the subsistence agriculture. * = E. +, ( 1+ ( θ w + θ w )) θ + 1, = (, 1) 1, 2, 1 θ2 = The same rocedure is adoted for the urban areas where = E * ( + ), ( ( ) ) [ + ( w + µ w )] µ + 1, = (, 1) 1 µ 1, 2, 1 µ 2 = In nominal terms, GP ( y y ) can be decomosed into nominal agricultural roduction ( ), nominal exorts exressed in local currency ( x ) and the nominal value of non-tradable goods ( ). =

13 =, +, =, +, +, Rural and urban rice levels will hel rojecting the overty lines in rural and urban areas resectively Income and exenditures of reresentative households i) Wages and income of reresentative households Wages and income of each reresentative household can be calculated as follows:, Rural unskilled workers of tradable goods sector: W. L, i 1, =,, Rural unskilled workers in the non-tradable goods sector: W. L, i 2,, = Urban unskilled workers in the non-tradable formal rivate sector: W L, i 3,., = Urban skilled workers in the non-tradable formal rivate sector: W L, i 4 SK SK Urban skilled workers in the tradable sector: W L, i 5 Urban civil servants (skilled): W. L,, i = 6 G G,., =,., = Finally, non-wage income of caitalists and renters (rofits) is aroached by the difference between total income and wage income. PROF = i W L In such economy where the distribution of income by grou is known, it is ossible to outline indicators of inter-grou income inequality. i i ii) isosable income, taxes and transfers 12

14 An income tax alies on each RH at a category-secific average τ while each RH receives lum sum budgetary transfers T (on er caita basis) from Government s budget. INC INC = i T i ( 1 τ )( + ) + i Wi Li PROF Li 7 i= 1 INC i iii) Public and rivate exenditures For each secific item, the sums of both rivate and ublic exenditures should resect rivate and ublic budget constraints. j j INC i κ i INC and G κ 1 j = 1 j i= j G G where the к stand for the shares of each secific item in rivate and ublic budgets. 4.2 The household survey simulator of the PAMS The basic rincile of the simulator is an extension of Bourguignon 2002 decomosition rule to several RHs. (confer World Bank 2002). INC k, RH1 t = (1 + g RH1 t ). INC k, RH1 t 1 The rogramme rojects the INC of each household according to its growth rate given by the labour market simulation for the grou to which belongs the individual household. In fact, the result is the shift in the distribution of INCs. This shift may be to the right or to the left, deending on the result of the olicy or shock. The result can also lead to the change in the distribution of income (more flattened or more stretched curve). Gini coefficient measuring disosable income inequality is an imortant tool. The Gini measures the distance between the erfect equality curve (Gini=0) where each grou 13

15 receives a share of income exactly roortional to its oulation size and the Lorenz curve obtained by the actual cumulative incomes of each grou. Figure 1: Changing in the INC of each RH After difining overty lines for the rural and urban sectors, the labour market generates growth rates of disosal income for each grou of 6 reresentative households and the caitalist renters grou. For more details in the techniques and hyothesis, see the World Bank (2002). 5. ata Requirements and Sources It is clear that households articiation in the economic activity determines the imact of shocks and olicies on their living standard. So the status of emloyment, the level of earnings and the level of exenditure stand to be the key determinants. emograhical information is also an imortant element on which the analysis should be based. This information can be found in the household survey. ECAM I (1996) and ECAM II (2001) rovide such information in the context of Cameroon. The forthcoming Cameroon s Final Poverty Reduction Strategy Paer will be another source of information. The macromodel is available at the irection of Forecast of the Ministry of Economy and Finance. 14

16 6. issemination Strategy I will work with the irection of Forecast of the Ministry of Finance and Budget where I use to work as consultant. It is therefore clear that not only the irection will rovide me information (macro-model for examle) and additional cometency, but will also use the results of the study for its forecasting activities. This can also be seen as the roof of the likelihood that government will use the results of the study. A ublication in an international review is another objective. 7. Key References Luiz A. Pereira da Silva, B. Essama-Nssah and Issouf amake (2002): A Poverty Analysis Macroeconomic Simulator (PAMS); Linking household surveys with macromodels; raft. The World Bank. United Nations eveloment Programme UNP (1998): La Pauvrete au Cameroun, Raort sur le eveloement Humain, aounde. MINPAT/SNC (1984): Enquete Budget-Consommation 1983/1984 (ECB83/84), aounde. MINEFI/SCN (1994): Enquete sur la Consomation des Menages 1993 (ECM93), aounde. MINEFI/SCN (1996): Enquete Camerounaise Aures des Menages 1996 (ECAM I), aounde. MINEFI/SCN (2002): Enquete Camerounaise Aures des Menages 1996 (ECAM II) Resultats Preliminaires, aounde. MINEFI (2000): Reublic of Cameroon: Interim Poverty Reduction Strategy Paer, August 2000, aounde. MINEFI (2002): Reublic of Cameroon: Final Poverty Reduction Strategy Paer, forthcoming. The World Bank (2001): World eveloment Reort 2000/2001 The World Bank (1995): Cameroon: iversity, Growth, and Poverty Reduction Reort CM, Washington C. 15

17 8. Prior Training and Exerience of Researcher in the Issues and Techniques Involved - June-August 2002: Consultant, the World Bank, Cameroon Country Office. I worked with a senior economist who is a overty analyst exert. It is from him that I got the methodology. I also heled him suervising two students from Ivory Cost High Scholl of Statistics and Alied Economics. Based on PAMS, the two students were writing their final year dissertation on overty. - March-June 2002: I worked as consultant in a articular asect of Cameroon s Poverty Reduction Strategy Paer. My contribution was about industrial cometitiveness and sustainable economic growth March 2002: Conference on Understanding Poverty and Growth in Sub-Saharan Africa ; University of Oxford, - Saint Catherine s College. uring the conference, I got accustomed to the concet and methodological aroaches of overty. 9. Exected Caacity Building for Researchers and their Institutions As the leading researcher, I will familiarise with overty concet as well as its methodological aroaches. In this way, I will gain exerience that would hel me train and teach students in my university. A ublication in an international review will be an imortant asset for my career, as it will hel me being romoted in the future. Two research assistants, one from the university and one from the Ministry of Finance and Budget (a male and a female) will be involved in the study so that both the university and the ministry will befit in terms of caacity building. 10. Any Ethical, Social and Gender or Environmental Issues or Risks Which Should Be Noted By the moment, I do not foresee any social, gender or environmental issue that we can mention. 16