Public policies promoting the informal economy: effects on incomes, employment and growth in Burkina Faso

Transcription

1 Publc polces promotng te nformal economy: effects on ncomes, employment and growt n Burkna Faso December 2015 Jean Abel Traoré * and Idrssa Moamed Ouedraogo Abstract: Ts paper ams to analyze te mpact on ncomes, employment and growt, of publc polces enancng te access to captal and lqudty for te nformal economy n Burkna Faso. Specfcally t assesses te drect effects of tese publc polces on te nformal economy outcomes, and te ndrect effects on te formal sector, te agrcultural sector as well as te economc growt. For tat purpose, polcy socks are smulated troug te PEP standard sngle-country and statc Computable General Equlbrum (CGE) model adapted to te structure of te Socal Accountng Matrx (SAM-2008) developed by te IFPRI for Burkna Faso. Our results sow tat a 10% ncrease n te stock of captal n te nformal sector results n a paradoxcal contracton of te nformal sector and te formal sector wle te agrcultural sector expands. Te employment rses n te nformal sector wereas t drops n te formal sector. Te ncomes mprove margnally for te own account workers and te farmers wle t decrease for te publc and prvate salared workers. Fnally te GDP at basc prces srnks weakly. Lkewse a 10% ncrease n publc transfers for te nformal ouseolds enances only te ncomes of te latter but causes an evcton effect on te publc nvestment leadng to negatve spllovers on all te sectors, te ncomes and te GDP. JEL Classfcaton: E26, E16, H81, O17 Keywords: Informal economy, Publc Polces, CGE model * Researc Fellow, Unted Natons Economc Commsson for Afrca (UNECA) and PD canddate, Unversty Ouaga II, Laboratory of Economc Polcy Analyss, Burkna Faso. abeltraore@uneca.org/ traoreeanabel@yaoo.fr Professor of Economcs, Unversty Ouaga II, Drector of te Laboratory of Economc Polcy Analyss, Burkna Faso. drss_mo@yaoo.fr. Te autors acknowledge te support from te Macroeconomc Polcy Dvson of te Economc Commsson for Afrca. Tey also tank Lacna Balma and Bourema Sawadogo for ter useful comments Ts model was bult by researcers from te Partnersp for Economc Polcy Network (PEP): Bernard Decaluwé, André Lemeln, Véronque Robcaud, and Hélène Masonnave Internatonal Food Polcy Researc Insttute

2 1. Introducton Ts paper ams to sed lgt on te effects on ncomes, employment and growt of publc polces overcomng captal and lqudty constrants faced by te nformal economy n Burkna Faso. Specfcally t evaluates te drect effects of tese publc polces on te nformal economy outcomes, and te ndrect effects on te formal sector, te agrcultural sector as well as te economc growt. Polcy nterventons are smulated usng te PEP standard sngle-country and statc Computable General Equlbrum (CGE) model adapted to te structure of te Socal Accountng Matrx (SAM-2008) developed by te IFPRI for Burkna Faso. Te results provde evdence tat enancng te nformal sector s access to captal and mtgatng lqudty constrants troug publc transfers for nformal ouseolds lead manly to negatve spllovers on te nformal and formal sectors, on te salared workers ncomes as well as te GDP. Te nformal economy 1 n Afrca as been consdered for a long tme as a range of margnal actvtes doomed to dsappear wt an mprovng economc development (Scneder and Enste, 2000). But durng ts last decade, te nformal economy placed tself as a potental source of ncomes by provdng after te agrcultural sector te most number of obs. In fact, accordng to te Internatonal Labour Organzaton (ILO) te nformal economy represents 42% of te GDP and occupes about 70% of te non-farmng workforce of Sub-Saaran countres 2. In Burkna Faso, te nformal economy employs 74% of te non-farmng workforce and contrbutes to about 25% of te GDP (Natonal Offce of Statstcs of Burkna). However, te man actors of te nformal economy are paradoxcally among te most precarous and vulnerable. In fact, most of te nformal actors are deprved from socal protecton, faced wt weak sklls and credt constrants wc undermne ter productvty (Traoré, 2013; Benamn and Mbaye, 2012). Furtermore, relatvely low levels of women s educaton and socal norms often lmt women to unpad domestc cleanng work and small ncome-generatng actvtes at ome or on streets from wc tey derve te most precarous ncomes (Traoré, 2013; Kuepe, Nordman and Roubaud, 2009; Cen, 2001). So te nformal economy s a concern for Governments of Afrca nterested n allevatng poverty and nequtes. Despte ts magntude te nformal economy contrbutes very weakly to te domestc drect tax revenues. Accordng to a recent study carred out n te Francopone West Afrcan countres, te nformal actvtes provde only 3% of te natonal drect tax effort (Benamn and Mbaye, 2012). Terefore, promotng nformal actvtes to rase ter productvty by 1 Te defnton of te nformal economy as been te subect of several debates (ILO, 1972; Hart, 1973; Weeks, 1975; Seturaman, 1981; Tokman, 1987; Lauter, 1994; Internatonal Labour Conference consensus, 2002). But nowadays tere s a knd of consensus on ts defnton. Te nformal economy encompasses two dmensons: te enterprse-based dmenson and te employment-based dmenson. Te enterprse-based dmenson refers to all te non-farmng frms wc don t comply wt te admnstratve and fscal rules requred by te legslaton (regstraton, wrtten accountng ). Ts dmenson s captured by te term nformal sector. Te employment-based dmenson refers to all (non-farmng) obs deprved from socal protecton, work contract and socal dalogue. Ts dmenson s captured by te term nformal employment. Troug te bot dmensons, tere s te noton of contnuum opposed to te dcotomy noton. In fact, some frms and obs comply wt some of te rules but escape to te oters. So tere are dfferent degrees of nformalty towards formalty wc requred te complance wt all te legslatve rules 2 Internatonal Labour Organzaton (ILO) : 12t Afrcan Regonal Meetng on Empowerng Afrca s peoples wt decent work ; Joannesburg, Sout Afrca, October

3 facltatng ter formalzaton represent an opportunty for Governments to broaden tax base necessary to moblze more domestc resources for development outcomes. Anoter relevant ssue related to te nformal economy s found n ts tendency to treaten te structural transformaton of Afrcan countres. Accordng to te ERA 3 (2014) : In Afrca, resources ave often moved from ger to lower productvty sectors, partcularly to te nformal sector, reversng structural transformaton and slowng productvty growt. Te ERA 4 (2015) glgted te large nvolment of te nformal sector n trade and servces - dentfed as engnes of structural transformaton- and stressed te need of approprate polces to arness te potental of te nformal sector n Afrca by expandng socal protecton systems, fscal and credt ncentves, skll development programmes, tecnology transfer and nfrastructure nvestment. In lne wt tese recommendatons, many Afrcan countres (Senegal, Gana, Ivory Coast, Burkna Faso, etc.) are mplementng a range of polces to enance te productvty of te nformal economy. In Burkna Faso, te Government establsed publc structures snce 1998 to support te nformal frms at tecncal and fnancal levels. Tese structures provde vocatonal tranngs to mprove manageral abltes and credt at low rates along wt cas transfers, n order to mtgate captal and lqudty constrants. However, some studes provded evdence tat subsdzed credt or vocatonal tranngs to promote te nformal sector productvty could be neffcent or armful for te nformal sector tself, for te formal sector, te agrcultural sector as well as te economc growt subect to some condtons (Arvn-Rad, Basu and Wllumsen, 2010; Ala et al, 2009; Baner and Sanay, 2007; Roy, 2006). Tese condtons pertan manly to te lnkages between te nformal and formal economes, suc as competton or cooperaton, mperfect substtuton between te nformal-based products and formal-based products and te saturaton of market. So t s mportant to know n te context of Burkna Faso: wat s te mpact of te polces enancng te access to captal and lqudty for te nformal economy on tself, on te formal sector and te agrcultural sector? Dd tese polces nduce postve spllovers on ncomes, obs and economc growt? In Burkna Faso, most of studes ave addressed te determnants and structure of te nformal economy, ts lnkages wt te formal economy and te barrers to formalzaton (Traoré, 2013; Grmm et al, 2012; Ouédraogo et al, 2011; Böme and Tele, 2011; Grmm and Günter, 2005; Zerbo, 2001; Ouédraogo, 1996). However researc to evaluate te effects of polces supportng te nformal economy s almost non-exstent. Tus te added value of ts researc s to fll ts gap n order to enlgten te polcy makers. Several econometrc analyses, for nstance, propensty-score matcng and randomzaton ave been used to evaluate te mpact of mcro-lendng and vocatonal tranng programs servng te nformal enterprses on earnngs, employment, and nequaltes n te context of 3 ECA (Unted Natons Economc Commsson for Afrca) and AUC (Afrcan Unon Commsson) Economc Report on Afrca 2014: Dynamc Industral Polcy n Afrca: Innovatve Insttutons, Effectve Processes and Flexble Mecansms. ECA, Adds Ababa. 4 ECA (Unted Natons Economc Commsson for Afrca) and AUC (Afrcan Unon Commsson) Economc Report on Afrca 2015: Industralzng troug Trade. ECA, Adds Ababa. 3

4 countres lke Madagascar and Cameroon (Ngumkeu, 2014; Mano et al. 2012; Berge et al. 2011; Glaub et al. 2011; Gubert and Roubaud, 2011). Tese studes n general found postve effects (often lmted) of suc programs on earnngs and employment wtn te nformal economy along wt a reducng nequalty. But te man lmt of te mcroeconomc analyses s found n ts falure to capture te ndrect effects on te formal and agrcultural sectors as well as on economc growt. Gven te strong nformal-formal lnkages wc could reverse te benefts of te scalng-up programs, a CGE model s well suted to glgt te transmsson cannels and capture te ndrect effects. In fact, ts approac as been used by recent researc addressng te nformal economy. For nstance, Erero et al (2014) and Erero et al. (2013) used a mult-sectoral computable general equlbrum model adapted to te economy of te Democratc Republc of Congo (DRC) to analyze te effects of reducng tarffs and wage subsdes for lower-sklled formal workers. Daves et al. (2011) also used a CGE model ncluded te economy wde lnkages between te formal and nformal economes, to address te effects of trade lberalzaton n Sout Afrca. However, recent studes usng a CGE model to analyze te effects of a polcy sock on te nformal economy are scarce, especally n te Afrcan context. Indeed, one of te most recent one s te study by Ala, Ndana, and Ngogue (2009) wc assessed te mpact of te ncrease n te nformal ouseolds nvestment under te program Integrated Support Program for Actors of te Informal Sector n Cameroon. However, n ter model te value added of te sectors (formal and nformal) are combnatons of captal and labor accordng to a Cobb Douglas functon wc constraned te value of te elastcty of substtuton between captal and labor to 1. Te captal s sectoral specfc and only one category of labor s consdered so gnorng te qualty aspect of labor. Our model s more flexble and realstc because all te sectors (nformal, formal and agrcultural) combne unsklled labor, sklled labor and captal troug a CES producton tecnology. Moreover, te ouseolds are assumed to ave a Stone-Geary utlty functon wc ncludes a mnmum level of consumpton for eac commodty and avod zero crossprce elastctes between all pars of goods, and a unt ncome-elastcty for all goods. Polcy experments deal wt a 10% n te demand for productve captal n te nformal sector and a 10% ncrease n te Government s transfers to te nformal ouseolds. Te remander of te paper s structured as follows. Secton 2 provdes a relevant lterature revew. Secton 3 addresses te man ponts of te CGE model wereas te followng secton presents te SAM-2008 used n our emprcal analyss. Secton 5 summarzes te motvatons of te polcy experments and Secton 6 presents te results along wt te related dscusson. Te last secton concludes wt polcy recommendatons. 2. Lterature Revew Inspred by te Harrs-Todaro model (1970), many autors ave ncorporated te dualty n te analyss of te urban labor market consderng te nformal sector as te lower urban segment able to absorb all or almost te surplus of labor at compettve low wage rates (Felds, 1975; Candra et Kan, 1993; Gupta, 1993; Mart, 2003; Mart et al, 2007; 4

5 Battacarya, 2011). Gven ts mportance n te economes of developng countres, te nformal sector occuped early a promnent place n development polces. Snce 1975, autors suc as Weeks (1975) advocated subsdy polces of captal and prces to boost ts sector. Follow-up tese recommendatons, oter researcers (Gos and Sarkar, 1989; Caudur, 1989; Gupta, 1993; Caudur, 2000) studed te effects on employment and ncomes of dfferent subsdy polces to te beneft of nformal sector. Tese studes can be classfed n two maor groups accordng to te assumed nformal-formal relatonsp. Te frst class of studes assumed tat te nformal sector produces only ntermedate goods for te formal sector. In oter words, te formal sector sub-contracts ntermedate stages of producton to te nformal sector. Te second class of papers allowed competton between formal and nformal sectors. For nstance, te nformal sector uses output of te formal sector as ntermedate nput to produce also fnsed goods consumed by te ouseolds. Informal sector as a producer of ntermedate goods for te formal sector Based on te sub-contractng nformal-formal lnkages, Sarkar and Gos (1989), Caudur (1989) provded evdence tat subsdzng te nterest rates for te nformal frms to easer access to credt tends to reduce producton and employment n te formal ndustral sector wle t expands te nformal sector. But n ter models te urban nformal sector s supposed to be able to absorb all te labor comng from te rural areas. Contrarwse, Gupta (1993) ncluded n ts analyss te exstence of an urban unemployment even n te presence of te nformal sector and found nuanced results. He concluded tat an ncrease of captal subsdy for nformal frms leads to a growt of te wage rate and a decrease n output prces wtn ts sector by reducng employment (so rsng unemployment) n te urban area consstng of formal and nformal sectors. Contrary, a prce subsdy for te nformal frms output tends to reduce te wage rate but to enance te output prce wtn ts sector by rsng employment n te urban sector. In order to compare tese two types of polcy Gupta (1993) used te defnton of socal welfare by Sen (1974), to sow tat an output prce subsdy granted to te nformal sector rses te socal welfare wle a captal subsdy polcy srnks t. Gupta (1993) went furter n s analyss by explorng te effects of subsdy polces n favor of te formal sector and te rural 5 sector. Ten e demonstrated tat subsdzng wages pad by te formal frms tend to lower te output prce and te wage rate n te nformal sector, but to ncrease employment n te urban area. As for te rural sector, prces or wages subsdes mples an ncrease n te prce of output and te wage rate n te nformal sector but a fall n employment n te urban area. Neverteless, Caudur (2000) glgted te man lmt of te model used by Gupta (1993). Indeed ts model gnored te role of te aggregate demand n determnng te level of producton and employment n te manufacturng and rural sectors. Tus, by ncludng te prncple of demand, Caudur (2000) confrmed te fndngs of Gupta (1993) relatve to te effects of captal subsdy granted to nformal frms on te nformal sector tself: ncrease n 5 Te rural sector represents agrcultural actvtes n te rural area 5

6 wage rate and decrease n output prce. But e found tat ts polcy also mpled a growt of producton and employment n te urban formal sector and te rural sector wc s contrary to Gupta (1993) conclusons. Furtermore, accordng to Caudur (2000), a prce or wage subsdy n favor of te rural sector rases te aggregate ncome of rural workers wc tends to stmulate demand for products n all sectors. Stmulated demand results n a growt of producton and employment trougout all te sectors (rural, nformal and formal). By te same aggregate demand mecansm, polcy promotng formal manufacturng sector nduces postve spllovers on all te economy. Rater tan analyzng te effects of polces tat cange te relatve costs of producton (captal, prce and wage subsdes) of nformal frms, oter autors ave studed te effects of polces tat affect ter productvty. For nstance, accordng to Arvn-Rad, Basu and Wllumsen (2010) a postve tecnologcal sock on te nformal sector represents any polcy suc as vocatonal and entrepreneursp tranngs, or access to captal or a new producton tecnology leadng to te mprovement of workers productvty n ts sector. In ter general equlbrum model based on four sectors - te rural, te urban nformal, te urban domestc formal and te foregn captal-owned sectors -, te ncrease n nformal frms productvty s captured by reducng te nput output coeffcents n te nformal sector. Tus tey demonstrated a postve tecnologcal sock on nformal sector tends to rase ncomes wtout alterng te output prce and te return to domestc captal wtn ts sector. Oter nterestng fndngs depend on te relatve factor ntenstes of te two urban formal sectors (domestc and captal-owned multnatonal) relatve to te rural sector. Assumng tat tose sectors are bot relatvely more captal ntensve n te nformal employment adusted gross sense tan te rural sector, ten te enancement of nformal workers productvty results n a contracton of te output of te nformal sector and te domestc formal sector but an ncrease n te rural sector output f and only f te elastcty of substtuton between labor and captal n te nformal sector s greater tan unty. Te output of te foregn captalowned sector remans constant. Subect to te same condtons, employment srnks n te nformal and te domestc formal sectors wle t rses n te rural sector. Employment remans uncanged n te foregn captal-owned sector. Besdes, te autors provded evdence tat a credt subsdy to te nformal sector results n te smlar effects on producton and employment. So tey ponted out te condtons under wc tecnologcal mprovement n te nformal sector can be damagng to te nformal sector tself and te oter sectors of te economy. However some surveys and stylzed facts confrmed tat n Afrca consumers faced wt fnsed goods produced by te formal and nformal sectors (1-2-3 Surveys, 2000; Benamn and Mbaye, 2012). Te nformal sector also uses part of te formal sector output as ntermedate nput and te two sectors are usually n competton. 6

7 Competton between te nformal sector and te formal sector Terefore, takng nto account te possblty of competton between formal and nformal sectors, Roy (2006) demonstrated tat subsdes granted to te nformal sector would be neffectve n growng te ncomes wtn tat sector subect to certan condtons suc as mperfect substtuton between nformal products and formal goods and te market saturaton. Indeed assumng mperfect substtuton between te nformal-based goods and te formalbased goods, Kelley (1994) provded evdence tat strateges desgned to promote te nformal sector lead to a reducton of te aggregate output and te ncomes of nformal producers n te context of Peru. Sanay and Baner (2007) use also an analytcal framework based on competton between formal and nformal but n wc te nformal sector s supposed to produce qualty goods less tan tose n te formal sector. Te formal sector as a low cost of captal but facng te g cost of labor due to te mnmum wage. In contrast, nformal frms face te g cost of captal due to credt ratonng, but as te low cost of labor due to wage flexblty. In oter words, te formal sector as a comparatve advantage n te producton of g qualty goods ntensve n captal wle te nformal sector as te advantage n te producton of low qualty goods and ntensve n labor. Wtn ts framework, Sanay and Baner (2007) proved tat fnancal support polces as mcrofnance to reduce te nterest rate for te nformal frms mprove te advantage of tose frms n te producton of relatvely low-qualty goods. Terefore, te producton and te sare of te nformal sector market expand at te expense of te formal sector wc srnks. Adoptng te same dea, Cogneau, Razndrakoto and Roubaud, (1996) ncluded qualty dualsm n a CGE model appled n te context of Cameroon. In ter model n case of a fallng ncome, te consumpton of formal goods decreases more tan tat of te nformal sector goods. So te ncome-elastcty of nformal products s lower tan tat of formal products. Te autors dd not study te effects of sock on nformal sector but tey found tat nformal sector and formal sector are n general counter-cyclcal n Cameroon. Contracton of te formal sector releases labor wc fnds refuge n te nformal sector and also favors te consumpton of less qualty nformal-based goods. Ts ncrease n te demand for nformal-based goods stmulates te producton and employment so tat te ncomes of nformal actors reman less senstve to te reducng pressure of te labor surplus released by te formal sector. Tey concluded tat a sustanable growt strategy must be based on polces tat could accelerate productvty n te nformal sector, and tus mprove te qualty of work and qualty of ts products. Ts polcy recommendaton found a strong support from te analyss of Montaud (2000) based on two approaces of te nformal sector n te context of Equator. Te frst approac treats te nformal sector as subsstence actvtes and restrcted operatng n a knd of margnalty, as falng to ft nto te rest of te economy. Te second approac consders a wde sector sze perfectly nserted nto te economy operatng n a compettve logc wt te 7

8 formal sector. Based on te frst approac, an enancement of nformal frms captal stock mproves ts output and te ncomes of ts workers. Te rse of ncomes mples an ncrease n te consumpton of formal and rural-based goods tus stmulates te producton and employment n tese sectors. Fnally subsstence actvtes declne but te ncomes and employment mprove n all te economy. Wtn te same framework an mprovement of labor productvty n te nformal sector generates smlar results wt a greater magntude. Follow-up ts study, Ala, Ndana, and Ngogue (2009) evaluate troug a CGE model te mpact of te rse of te nformal ouseolds nvestment troug te program Integrated Support Program for Actors of te Informal Sector n Cameroon and tey found nuanced results. Accordng to ter fndngs, 26% ncrease n nformal actors nvestment lead to a rse of te producton and te demand for labor by 0.43% wtn te nformal sector. Te ncomes also ncrease by 0.15% and tus stmulate te demand of nformal-based goods by 0.1%. But ts good performance n te nformal sector s followed by a weak contracton of te formal sector due to an evcton effect. Te reducton of output and ncomes n te formal sector generates unfortunately a decrease n te Government recepts by 0.31%. In fact publc recepts are manly based on taxes on te formal sector actvtes. All tese studes cted above weter teoretcal or emprcal sed lgt on te postve and paradoxcal effects of polces promotng te nformal economy. But ter man lmt s te lack of te labor qualty aspects n ter framework. However tere are strong proofs tat te sklls determne te sector n wc one can fnd ob. Sklled people n general prefer employment n te formal sector wc usually provdes te gest ncomes assocated wt socal protecton. But structural ssues n Afrca constrant te formal sector to provde obs for all te qualfed labor. So te remanng qualfed labor s absorbed by te nformal sector. However, te nformal sector values educaton because te sklled salared workers are relatvely better pad tan te unsklled one (Traoré, 2013). Consderng labor qualty ssues could lead to nuanced results wen mplementng polces to promote te nformal sector. For nstance, tere s strong evdence tat te elastcty of substtuton between unsklled labor and pyscal captal s ger tan pyscal captal and sklled labor wtn te nformal sector (Amaral and Quntn, 2006). In fact for nformal frms faced wt credt constrants but takng advantage on wages flexblty, unsklled labor may be a better substtute for pyscal captal tan sklled labor. And tanks to te prevous lterature revew, one knows ow mportant te effects of polces benefcal to te nformal economy are senstve to te degree of substtutablty between labor and captal. Zerbo (2006) ncluded labor qualty ssues combned wt socal protecton ssues n s analyss of urban labor market n Sub-Saaran Afrca, but e dd not focus on polces supportng te nformal sector. More recent studes also consdered labor qualty aspects n ter general equlbrum model analyss but tey focused on polcy experments related to mnmum wage, wage subsdy for lower-sklled formal workers, unemployment benefts, tax polces and trade lberalzaton (Carlot et al, 2015; Bosc and Esteban-Pretel, 2015; Erero et al, 2014; Erero et al, 2013; Megr et al, 2012; Daves et al, 2011). 8

9 So even f our study does not address te nformal workers productvty, t ncorporates a better realstc feature of te nformal sector and te nformal-formal lnkage n Burkna Faso s context. In fact our model consdered an nformal sector combnng unsklled labor, sklled labor and captal troug a two levels CES producton tecnology lke te oter sectors. Te elastcty of substtuton between captal and te composte labor s assumed to be ger tan 1 and te elastcty of substtuton between te dfferent types of labor less tan one. Furtermore, competton and cooperaton between te formal and nformal sectors are taken nto account dependng on te branc actvty. Fnally, te dden nformal employment (employment deprved from socal protecton) n te formal sector s addressed n our analyss. 3. Te Model In ts paper, te teoretcal framework s an adaptaton of te PEP standard sngle-country and statc CGE model: PEP-1-1, (Verson 2.1). Ts CGE model s based on te structure of te SAM-2008 developed by te IFPRI for Burkna Faso. Te model 6 can be summarzed nto four blocks of nterdependent equatons: ) Supply; ) Demand; ) Income dstrbuton; v) Equlbrum and closure 3.1. Supply Te local producton tecnology s represented by a nested functon wt several levels. At te top level, sectoral producton (nformal, formal and agrcultural sectors) combnes value added and total ntermedate consumpton n fxed sares followng a Leontef nput-output functon. At te second level, eac ndustry s value added conssts of composte labor and captal, followng a constant elastcty of substtuton (CES) specfcaton. Fnally, at te bottom level on te value added sde, te varous categores of labor are mperfect substtutes n aggregate labor accordng to a constant elastcty of substtuton (CES) tecnology. 3.2 Demand Based on te Armngton assumpton, composte products demanded on te domestc market are a combnaton of local and mported products accordng to a constant elastcty of substtuton (CES). Output demand for ntermedate consumpton for eac good s a fxed proporton of te aggregate nput consumpton by economc sectors. Demand of eac goods and servce for nvestment purposes s a fxed value sare of total nvestment. Inventory canges are assumed to be exogenous. 3.3 Income dstrbuton Houseold ncomes come from labor earnngs, captal earnngs and transfers receved from oter agents ncluded te Government. Eac ouseold category receves a fxed sare of te earnngs of eac type of labor. Lkewse, total captal ncome s dstrbuted between agents, ncludng ouseolds, n fxed sares. Government s ncomes nclude tax recepts on 6 Te detals of te model are provded n Appendx 9

10 ouseolds ncomes, tax revenues on economc actvtes, captal remuneraton, and offcal development assstance (blateral and multlateral ad) from abroad. 3.4 Equlbrum and closure Followng te neoclassc perfect compettve market, equlbrum between te supply and demand of eac commodty on te domestc market must be verfed. Also, te total demand for eac factor (labor and captal) must equal avalable supply (exogenous). Lkewse, total nvestment expendture must be equal to te sum of agents savngs. Te closure of te model s based on te neoclassc rule. Tat means total nvestment s endogenous and drven by total savngs wc s a functon of te full employment of resources ncome. Income dstrbuton s ten determned only from te supply sde. Once te full-employment ncome s determned, consumpton s derved ten savng and fnally nvestment. Te excange rate s defned as te numerare, wt government expendtures and te current account balance fxed. Captal s moble between ndustres. 4. Data and calbraton 4.1 Data Te SAM for Burkna Faso represents economc nteractons troug sx (06) groups of accounts: ) 28 accounts producton actvtes; ) 28 accounts for commodtes (eac actvty produces a sngle good or servce and eac commodty s produced by a sngle actvty); ) 7 factors of producton (6 types of labor and a sngle class of captal); v) 8 categores of ouseolds dependng on te actvty of te ouseold ead (3 salared, 3 agrcultural and 2 non-agrcultural); v) 2 oter nsttutonal agents: te Government and te Rest of te World; v) 1 savngs-nvestment account. Based on te obectves of our study, te 28 brances of producton were grouped nto fve brances: Agrculture; Formal ndustry; Informal ndustry; Formal servce; Informal servce. Te agrculture ndustry ncludes te agro-forestry-pastoral producton. Te dstncton between nformal and formal frms s based on te Natonal Offce of Statstcs (n Burkna Faso) operatonal defnton of te nformal sector 8. Based on te SAM-2008, we llustrate troug te followng graps, te sare of nformal actvtes n te total value added, n te total demand for captal, as well as te total employment and ter lnkage wt te formal sector. Accordng to Grap 1 below, n 2008, te economy of Burkna Faso was drven by gly productve formal servces (as modern trade; modern transport; post and telecommuncatons and fnancal servces) followed by te agrcultural sector. Te formal ndustres were embryonc wereas te contrbuton of te nformal sector (ndustres and servces) was about 19.05% of te value added wc s not neglgble. 7 A detaled descrpton of te SAM s provded n Appendx 8 Accordng to te Natonal Offce of Statstcs n Burkna Faso, te nformal sector refers to all te unts of producton wtout a fscal dentfcaton number and/or wc do not keep wrtten books 10

11 Grap 1: Repartton of te total value added per sector 50% 40% 30% 32.24% 39.50% 20% 10% 9.21% 11.49% 7.56% 0% Agrculture Formal ndustry Formal servce Informal ndustry Informal servce Sare n te total value added Source: SAM-2008 for Burkna Faso Concernng te demand for captal, te formal sector used about 64.43% of te total productve captal followed by te agrculture sector (24.08%) wereas te nformal sector ad access to only 11.49% (Grap 2). So te nformal sector seems to be constraned by te captal sde. Grap 2: Repartton of te aggregate demand for captal per sector 60% 50% 49.15% 40% 30% 20% 10% 0% 24.08% 15.28% 7.49% 4% Agrculture Formal ndustry Formal servce Informal ndustry Informal servce Sare n te total captal demand Source: SAM-2008 for Burkna Faso Accordng to Grap 3 (below), 37.48% of te total labor ncome was provded by te agrcultural sector wereas te formal servces account for 33.30%. Te nformal sector contrbuted substantvely to 23.91% of te total labor ncomes. 11

12 Grap 3: Repartton of te aggregate labor ncome per sector 40% 35% 30% 25% 20% 15% 10% 5% 0% 37.48% 33.30% 14.05% 9.86% 5.31% Agrculture Formal ndustry Formal servce Informal ndustry Informal servce Sare n te total labor ncomes Source: SAM-2008 for Burkna Faso Concernng te lnkage between te formal and nformal sector (Grap 4), about 49.06% of te total ntermedate consumpton of te nformal ndustres was provded by te formal sector (formal ndustry and servce). Lkewse, te nformal servces acqured about 70.86% of ts total ntermedate consumpton from te formal sector. So te producton n te nformal sector s strongly dependant on te formal sector output. Grap 4: Te sare of te formal sector n te total ntermedate consumpton of te nformal sector 40% 35% 30% 25% 20% 15% 10% 5% 0% 31.44% 17.62% Informal ndustry Formal ndustry 37.70% 33.16% Informal servce Formal servce Source: SAM-2008 for Burkna Faso Accordng to Grap 5 (below), te nformal ndustres represented only 2.44% of te total ntermedate consumpton of te formal ndustres were as nformal servces provded up to 16.53%. About 22.91% of te formal servces total ntermedate consumpton stemmed from te nformal sector. So, te producton n te formal sector s relatvely less relyng on te nformal sector output. 12

13 Grap 5: Te sare of te nformal sector n te aggregate ntermedate consumpton of te formal sector 20% 15% 10% 16.53% 11.91% 11% 5% 0% 2.44% Formal ndustry Informal ndustry Informal servce Formal servce Source: SAM-2008 for Burkna Faso 4.2 Te Calbraton Te calbraton s about te determnaton of numercal values of te parameters and coeffcents so as to reconsttute te equlbrum stuaton of te SAM. Some of te parameters as elastctes of substtuton are fxed accordng to te lterature and oters are computed based on te values provded by te SAM. Based on te lterature of CGE model appled for Burkna Faso economy (Sawadogo et al, 2015; Balma et al, 2010) we summarzed te values of some parameters n Table 1 below. Table 1: Intal values of some parameters Elastctes Elastcty of substtuton between captal and composte labor n sector Elastcty of substtuton between types of labor n sector Elastcty of transformaton between exports and local sales Elastcty of substtuton between local supply and mports Houseold Income elastcty Frsc parameter Prces Prce of te local commodty Prce of exported commodty World prce of mported product Wage rate of type l labor Rental rate of captal n sector Prce elastcty of ndexed ,: formal ndustry; formal servces; nformal ndustry; nformal servces; agrculture l:rural formal; formal unsklled; formal sklled; rural nformal; nformal unsklled; formal sklled 13

14 5. Polcy experments Snce te 1990s, Burkna Faso as been mplementng a number of specfc programs to promote employment creaton troug te fnancng and vocatonal tranngs for mcroenterprses. Among te maor employment programs are te Informal Sector Support Fund (FASI), te Employment Creaton Support Fund (FAPE), te Yout Intatves Support Fund (FAIJ) and te Vocatonal Tranng and Learnng Support Fund (FAFPA). Created n 1998, te prmarly obectve of FASI s to mprove access to credt for te nformal sector. Te Fund provdes credt (up to 1.5 mllon FCFA) wt nterest rate between 8% to 13% dependng on te proects sze and ts area of actvtes. Te nterest rates and guarantes requrement are lower tan te requrements of te mcrofnance nsttutons and te commercal banks. Practcally, te program s assgned to grant 700 mllon FCFA credt per year to fnance 1500 mcroenterprses. But n fact, ts target s not always reaced. For nstance, n 2010, te most recent data 9, te Fund provded loans n te amount of FCFA to 1244 enterprses wc generated 256 new obs and consoldated 2922 obs. Te FAIJ (created n 2007) obectve s to reduce poverty, unemployment and underemployment of yout n urban and rural areas. It lends loans up to 2 mllon FCFA wt nterest rate between 2% to 4% followed by entrepreneursp tranngs. Specfcally, te program s amed to reac 5000 Young per year. In 2010, te Fund fnanced 1297 proects for an amount of FCFA wc created about 5904 new obs for Young. So FASI and FAIJ granted FCFA to te nformal actors n However ts amount represents only 1% of te value of te demand for captal n te nformal sector accordng to te SAM Aware of ts nsgnfcance, te Government announced n 2014 te openng of a specal fund to grant up to 10 bllon FCFA credt for te nformal frms and women entrepreneursp. Based on te latter facts, we smulate a rse of 10% of te demand for captal n te nformal sector wc s equvalent to about 14.5 bllon FCFA accordng to te SAM Te process to get te fund or credt s not analyzed. Te nterest s focused on te fnal obectve of te mcro-lendng program wc s te ncrease n te demand for captal (nvestment) n te nformal sector. In te same lne, our second polcy experment s a 10% ncrease n te Government s transfers to te nformal ouseolds (te own account workers and te nformal sector salared workers). 6. Results 6.1. Smulaton 1: 10% ncrease n te demand for captal n te nformal sector Impact on employment, ncomes and producton wtn te nformal sector A 10% ncrease n te demand for captal by te nformal sector nduces an ncrease n te demand for all te varous types of labor. However, te ncrease n te formal and qualfed 9 Observatore Natonal de L Emplo et de la Formaton Professonnelle (2011). Annuare statstque du marcé de l emplo Document d analyse, Mnstère de la Jeunesse de la Formaton Professonnelle et de l Emplo, Burkna Faso 14

15 labor demand s ger tan tat of te nformal and unsklled labor (see Table 1 below). For nstance, te demand for rural formal labor, for urban formal sklled labor, and for te urban nformal sklled labor expand more tan 1% wereas te ones for rural nformal and urban nformal unsklled labor rase only by 0.65% and 0.82%. Furtermore, te overall enancement of te demand for composte labor tends to reduce te wages followng te flexblty of wages n te nformal sector (and n te model). All te wages of te dfferent category of labor decrease by at least 0.14% except te wage of te rural nformal labor and te urban nformal unsklled wc get te lowest ncreases n demand. Table 2: Impact on employment and wages n te nformal sector Demand for categores of labor Cange (%) Wages of categores of labor Cange (%) Rural formal labor 1.06 Rural formal labor Urban formal unsklled 1.11 Urban formal unsklled Urban formal sklled 1.11 Urban formal sklled Rural nformal 0.65 Rural nformal 0.12 Urban nformal unsklled 0.82 Urban nformal unsklled Urban nformal sklled 1.08 Urban nformal sklled It s wort to glgt tat te composte nformal labor (rural, unsklled and sklled) accounts for about 91% of te total labor used by te nformal ndustry and about 98% of tat of te nformal servces sector. So te nformal labor s most confned n te nformal sector. Moreover about 88% of te total labor ncomes of te farmers come from ter supply of nformal employment mostly from te rural nformal labor (72.8%). Lkewse about 58% of te total labor ncomes of te own account workers comes from ter supply of nformal employment mostly from te rural nformal labor (50.34%). So te combned mprovement of te demand for rural nformal labor and ts assocated wage lead to an upward pressure on te farmers labor ncomes wc ncrease margnally by 0.063%. But te own account workers total labor ncomes decrease by 0.068%. However, followng te new nflow of captal n te nformal sector, te ouseolds captal ncomes rse despte te lowerng of te rental rate and offset te fall of labor ncomes for te own account workers. Fnally te aggregate ncome enances for te farmers and te own account workers by 0.146% and 0.144%. Table 3: Impact on te ncomes n te nformal sector Houseolds Labor ncomes Cange (%) Captal ncomes Cange (%) Total Income Cange (%) Te Farmers Te own account workers Tese two categores of ouseolds are te largest consumers of te nformal sector output. Accordng to te SAM-2008, about 65% of te nformal sector output s consumed by te farmers wereas te own account workers consume about 12.37%. But wen ter ncomes mprove tese ouseolds reduce ter demand for te nformal goods and servces n favor of 15

16 te formal sector products. Indeed, te demand for te nformal sector output declnes by 0.91% for te farmers and by 0.97% for te own account workers. Despte te rse of ts prce, te demand for te formal goods and servces ncreases by 0.25% for te farmers and by 0.2% for te own account workers. Tese results are n lne wt te fndngs of Bome and Tele (2011) n te case of sx West Afrcan captals. Moreover, due to te ncrease n ts prce, te demand of nformal sector s output for ntermedate goods and for nvestment purposes drop by 0.41% and 1.82% respectvely. Tus te supply of te nformal goods and servces follow te declne of ter domestc demand accordng to te equlbrum constrant. Te aggregate productons of te nformal ndustry and te nformal servces fall down respectvely by 0.53% and 0.22% mplyng also a decrease n te total ntermedate consumpton and te value added of te nformal sector. In fact, te value added of te nformal sector declnes by 0.75%. So an ncrease n te nformal sector s stock of captal results n a paradoxcal contracton of ts total producton and ts value added. Impact on employment, ncomes and producton wtn te formal sector As t s llustrated n te Secton 4 te largest sare of te total ntermedate consumpton of te nformal sector s provded by te formal sector. So a reducton of te nformal sector output leads to a decrease n ts demand for te formal sector output as ntermedate goods by 1.5%. Even f te agrcultural sector s demand for te formal sector output as ntermedate goods rses by 0.3% te overall demand for te formal sector output as ntermedate goods falls down by 0.32%. Lkewse, te total demand of te formal sector s producton for nvestment purposes declnes by 0.078%. So te domestc demand for te formal sector output reduces by 0.16% even f te aggregate ouseold consumpton demand for formal goods and servces ncreases by 0.106%. Accordng to te equlbrum constrant te supply of te formal sector output also declnes by 0.16%. Obvously te reducton of te supply generates a downward pressure on te value added and te aggregate ntermedate consumpton wc declne proportonally by 0.16%. Followng te cut of te formal sector value added, te demand for all types of labor also decreases along wt te assocated wages wereas te demand for captal remaned uncanged. Te demands for rural nformal labor and for urban nformal unsklled fall down by 0.53% and 0.35% respectvely. Neverteless te lessenng s weaker for te demands for formal and qualfed labor (see Table 4 below). Table 4: Impact on employment n te formal sector Labor types Cange (%) Rural formal labor Urban formal unsklled Urban formal sklled Rural nformal Urban nformal unsklled Urban nformal sklled Te formal labor (rural, sklled and unsklled) represents 81.83% of te aggregate labor demand used by te formal ndustry and 66.08% of te one of te formal servces. So te 16

17 formal employment s most confned n te formal sector. Furtermore, about 75% of te publc salared ouseolds ncomes labor stem from te supply of te formal labor (rural, sklled and unsklled). Lkewse about 77% of te formal prvate salared ouseolds ncomes labor stem from te formal labor (rural, sklled and unsklled). About 50% of te nformal prvate salared ouseolds labor ncomes come from te formal labor (rural, sklled and unsklled). So te reducton of te labor demand and te wages n te formal sector affects more severely te former categores of ouseolds. Indeed, te publc salared ouseolds ncomes labor decreases by 0.15% and te formal prvate salared ouseolds ncomes labor falls down by 0.14% (see Table 5). But te necton of new captal troug te nformal sector tends to ncrease te captal ncomes of all te ouseolds. Ts rsng of captal ncomes offsets te declne of te labor ncomes but fnally te total ncome of te publc salared ouseolds, te formal prvate salared ouseolds and te nformal prvate salared ouseolds srnk weakly by 0.02%, 0.081% and 0.08%. Followng te decrease of ter ncomes, tese ouseolds reduce ter consumpton of goods and servces. Te formal-based goods and servces represent about 64% of te total consumpton of te publc salared ouseolds and 67% of te one of te formal prvate salared ouseolds. Ter consumpton of bot formal and nformal products decreases wc tend to renforce te reducton of te supples of te concerned sectors. Table 5: Impact on ncomes n te formal sector Houseolds Te publc salared ouseolds Te formal prvate salared ouseolds Te nformal prvate salared ouseolds Labor ncomes Cange (%) Captal ncomes Cange (%) Total Income Cange (%) Impact on employment, ncomes and producton wtn te agrcultural sector Te farmers are te largest consumers of agrcultural products. Tey consume about 83.65% of te total consumpton of agrcultural products. Te own account workers consume only 5.6% of te total consumpton of agrcultural products more tan te salared publc ouseolds wo consume 4.41%. Wt ter rsng ncomes combned wt te decrease n te agrcultural output prce, te farmers and te own account workers tend to ncrease ter demand for agrcultural products. So te total demand by ouseolds for agrcultural products enance by 0.21% even f some ouseolds as te formal prvate salared ouseolds reduce ter demand due to te decrease of ter ncomes. 17

18 Due to te contracton of te formal and nformal sectors te overall demand of agrcultural output as ntermedate goods falls down by 0.19% but te aggregate domestc demand of te agrcultural goods stll ncreases by 0.15%. Ts ncrease mples also a proportonal ncrease n te supply and te value added of te agrcultural sector by 0.15%. Wt te ncrease of te value added, te demand for captal remans constant wereas te demand for labor offered a mxed pcture (see Table 6). Indeed, te demand for rural formal labor falls down by 0.012% wereas te demands for urban formal unsklled labor and urban formal sklled rse by 0.012% and 0.013% respectvely. Te rural nformal labor wc represents 65.4% of te total labor demand n te agrcultural sector regsters te bggest decrease n ts demand around 0.21%. Te demand for urban nformal unsklled labor accountng for about 18.92% of te total labor demand n te agrcultural sector also declnes by 0.13%. In fact te rural nformal labor and te urban nformal unsklled labor are most confned n te agrcultural and nformal sectors. Wle te demand of tese two types of labor declne n te agrcultural sector, ter demands ncrease n te nformal sector. Knowng tat te demand for all types of labor srnks n te formal sector, one can assume tat te rural nformal labor and te urban nformal unsklled labor move from te agrcultural sector to te nformal sector. Table 6: Impact on employment n te agrcultural sector Labor types Cange (%) Rural formal labor Urban formal unsklled Urban formal sklled Rural nformal Urban nformal unsklled Urban nformal sklled Te farmers account for 99.69% of te total rural nformal labor and 50.03% of te total urban nformal unsklled labor. So te weak decrease n te demand of tese categores of labor n te agrcultural sector may mply a contracton of te labor ncomes of te farmers. But as t was sown above te farmers earnngs also come from te nformal sector were te demand for labor and te wages are mprovng. Fnally te total ncome of te farmers ncreases. As a summary at te sectoral level, te nformal and formal sectors contract wereas te agrcultural sector expands. Employment tends to move from te formal sector to te nformal sector manly. Tese fndngs are n lne wt te concluson of Arvn-Rad et al (2010) wc argues tat a credt subsdy to te nformal sector leads to a decrease n te output of te nformal sector and te domestc urban sector wle te rural sector expands f and only te elastcty of substtuton between te labor and captal s greater tan unty n te nformal sector. 18

19 Impact on tax revenues and te GDP Accordng to te SAM, te formal ndustry and te formal servces moblzed respectvely 83.3% and 7.35% of te total taxes on products. Te nformal ndustry and te nformal servces represent only 7.25% and 1.28% of te total taxes on products. Te contrbuton of te agrcultural sector s almost nsgnfcant, only 0.83%. Concernng te taxes and dutes on mports, te formal sector (ndustry and servces) accounts for about 94% wereas te agrcultural sector moblzed only 6%. In our model te nformal frms do not mport goods and servces due ter lack of complance wt te legal and admnstratve rules. Wt te contracton of te formal sector, te contrbutons of te formal ndustry and te formal servces to te total taxes on products decrease by 0.031% and 0.058% respectvely. Despte te reducton of te nformal sector output, te sares of te nformal ndustry and te nformal servces to te total taxes on products ncrease by 0.3% and 0.7%. Te agrcultural sector s sare declnes by 0.015%. Fnally, te total Government revenues from taxes on products remans almost uncanged (a rse by %). Te total Government s revenues from taxes and dutes on mports decrease margnally by %. Accordng to te varatons of ter mports, te sare of te total taxes and dutes on mports moblzed by te formal ndustry rses slgtly by % wereas te sares of te formal servces and te agrcultural sector declnes by 0.01% and 0.175% respectvely. Besdes, te salared ouseolds (publc, formal and nformal prvate sectors) provde about 83.7% of te total Government s revenue from ouseold ncome taxes. Te contrbuton of te farmers accounts for about 11.67% wereas te one of te own account workers s only around 1.1%. After te polcy sock on te nformal sector, te contrbutons of te salared ouseolds decrease accordng to te declne of ter ncomes. At te opposte, te contrbutons of te farmers and te own account workers ncrease by 0.146% and 0.144% respectvely. But as te former categores of ouseolds are not te largest provders, te total Government s revenue from ouseold ncome taxes srnk by 0.037%. Even f te overall Government s revenue from taxes decreases ts revenues from captal and external transfers tend to ncrease ts overall recepts and reduce ts defct (Table 7). Table 7: Impact on Government s ncomes Taxes revenues Cange (%) Government ncomes Cange (%) Total taxes on products Government captal ncome Total taxes and dutes on mports Government transfer ncome Total ncome taxes Total Government ncome Followng te contracton of te formal and nformal sectors, te real GDP at basc prces decreases by 0.048% but te real GDP at market prce rses weakly by 0.015% Smulaton 2: 10% ncrease n te publc transfers to te nformal ouseolds Te nformal ouseolds encompass te own account workers and te nformal prvate salared ouseolds. Te ncrease n te publc transfers tends to enance te total ncomes of te own account workers and te nformal prvate salared ouseolds by 0.15% and 0.12% 19

20 respectvely. Te ncrease n total ncomes mples te rse of ter total consumpton of goods and servces and ter savngs. But ts enancement s not suffcent enoug to pull up te aggregate demand for goods and servces especally as te consumpton of oter ouseolds s declnng. In fact te nformal ouseolds account for about 15.91% of te aggregate demand for goods and servces. Beyond te mprovement of ncomes for only te nformal ouseolds, te ncrease n te publc transfers causes an evcton effect on te publc nvestment spendng wle degradng te Government s budget defct. Indeed te fnal demand of formal goods, nformal goods and nformal servces for publc nvestment purposes decreases by 0.025%; 0.05% and 0.052% respectvely. Only te demand of formal servces for publc nvestment purposes enances by 0.053%. So te decrease of te publc nvestment spendng tends to pull down te domestc demand and te supply of te formal ndustry and te nformal sector (ndustry and servces). Ts results n te drop n ter ntermedate consumptons, ter fnal demand for prvate nvestment purposes wc affect negatvely on domestc demands, productons and valued added of te agrcultural sector and te formal servces sector (Table 8). Te decreasng value added of all te sectors remans te dstrbuton of te captal uncanged. But te demand of types of labor rses n all te sectors except n te formal servces sector. Tat results n te decrease n wages and labor ncomes for all categores of ouseolds (Table 9). Te publc transfers offsets te drop of labor ncomes for te nformal ouseolds only wereas te total ncomes of te formal ouseolds srnks ten pullng down te aggregate consumpton of goods and servces as well as te domestc demand (Table 10). Te ouseolds ncomes taxes follow te varaton of ter total ncomes but at te end te total Government s revenue from ouseold ncome taxes remans constant. Moreover wt te contracton of all te sectors te total government revenue from taxes on products and mports declnes by 0.023%. Fnally te real GDP at basc prces decreases by 0.09% along wt te real GDP at market prce by 0.088%. Table 8: Impact on value added per sector Value added Cange (%) Agrculture Formal Industry Informal Industry Formal Servces Informal Servces Table 9: Impact on employment and wages n all te sectors Labor demand Cange (%) Agrculture Formal Industres Informal Industres Formal Servces Informal Servces Wages Cange (%) Rural formal labor Urban formal