Structural unemployment, labour market dynamics and the transmission of monetary policy in South Africa

Abstract

English: This thesis begins by raising the following question: what is the relationship between the labour market structure in South Africa and the ability of the monetary authority to keep control of macroeconomic dynamics? The answer to this question is threefold. First of all, we investigate the empirical evidence of the existence of wage rigidities in the South African labour market. Second, we assess the impact such frictions may have on the conduct of monetary policy. Finally, we introduce public employment to analyze the e¤ects it may have on labour market dynamics in the peculiar case of South Africa. This thesis is therefore divided into three main chapters. Chapter 2 estimates a New Keynesian Wage Phillips Curve for South Africa to determine the respon- siveness of wages to employment conditions. First, we estimate the model using aggregate data. The aggregate estimation results show that private sector wages are not quite responsive to employment con- ditions, while they also reveal a certain sensitivity to in‡ation and quite a good correlation with in‡ation expectations. The relationship between private sector nominal wage in‡ation and employment is clearly weak for the whole sample, and it becomes insigni…cant at the end of the sample, indicating an increase of wage rigidities in the post-apartheid South Africa. On the other hand, the relationship between nominal wage in‡ation and price in‡ation is quite strong and robust for the whole sample but it becomes quan- titatively weak for the in‡ation targeting period. In this period, trade unions in‡ation expectations are instead strongly correlated to nominal wage in‡ation. In the second part of the chapter we look at the relationship between nominal wages, productivity and the reservation wage. The …ndings con…rm that nominal wage in‡ation has consistently outpaced the growth in productivity, even after correcting for price in‡ation. Furthermore, the results suggest that employment conditions had little e¤ect on wage dynamics. We also …nd that for the community, social and personal services sector which, mainly consist of public …rms, the response of wages to market conditions is remarkably strong. The overall picture that comes out from the analysis is that of a wage formation mechanism that is very insensitive to overall macroeconomic conditions. Chapter 3 therefore investigates the implications this picture may have on conducting monetary policy. Therefore, we use a DSGE model with unemployment and labour market frictions following the work of Blanchard and Gali (2010) to answer this crucial interrogation. In an economy with ‡exible wages, the divine coincidence holds since stabilizing in‡ation automatically leads to a stable unemployment (output gap). However when we account for wage rigidities and further labour market frictions, the results show that pursuing an objective of a stable in‡ation leads to undesirable and persistent ‡uctuations in unemployment; with a sacri…ce ratio that is typically higher the more ‡uid the labour market. The estimation of the model using South African data displays two main results. First, labour market parameter estimates present a picture with pervasive wage rigidities which thus con…rms the …nding in chapter 2. Second, the labour market appears to be characterized by large ‡ows in job creation and job destruction rates, with the latter dominating the dynamics. Finally, chapter 4 revisits the …nding in chapter 2 of a strong response of wages to market conditions in the public sector. Furthermore, the public sector in South Africa is for the most part skilled workers intensive, and has been expanding since the 2008 …nancial crisis while private …rms experienced massive turbulences. This chapter thus raises two questions: what could be the impact of public employment on overall labour dynamics? Moreover, could these features of public employment contribute in part in explaining the rigidities in wages observed in the private sector? To answer this question, we use a DSGE model with two sectors and two types of workers (depending on their level of skills), which is essentially a modi…ed version of the framework …rst introduced by Gomes (2013). In particular, we mainly investigate the e¤ects of a positive public sector wage shock on a calibrated South African labour market. For comparison purposes, we also assess the impact of a positive private sector productivity shock. We further take into account the di¤erent level of bargaining power and how the di¤erence in productivity between public and private sectors both a¤ect the responsiveness of variables. The …ndings show that an increase in private sector productivity produces more desirable results with an increase in employment for both skilled and unskilled workers which translates into a decrease in overall unemployment. Public sector wage shock on the other hand mainly crowds out private skilled labour which the …rms react to by substituting it with unskilled workers. Ultimately, the increase in public wages raises overall unemployment as supplementary skilled unemployed individuals queue for public jobs. Altogether, the e¤ects are more pronounced when the bargaining power of unskilled workers is raised.the more ‡uid the labour market. The estimation of the model using South African data displays two main results. First, labour market parameter estimates present a picture with pervasive wage rigidities which thus con…rms the …nding in chapter 2. Second, the labour market appears to be characterized by large ‡ows in job creation and job destruction rates, with the latter dominating the dynamics. Finally, chapter 4 revisits the …nding in chapter 2 of a strong response of wages to market conditions in the public sector. Furthermore, the public sector in South Africa is for the most part skilled workers intensive, and has been expanding since the 2008 …nancial crisis while private …rms experienced massive turbulences. This chapter thus raises two questions: what could be the impact of public employment on overall labour dynamics? Moreover, could these features of public employment contribute in part in explaining the rigidities in wages observed in the private sector? To answer this question, we use a DSGE model with two sectors and two types of workers (depending on their level of skills), which is essentially a modi…ed version of the framework …rst introduced by Gomes (2013). In particular, we mainly investigate the e¤ects of a positive public sector wage shock on a calibrated South African labour market. For comparison purposes, we also assess the impact of a positive private sector productivity shock. We further take into account the di¤erent level of bargaining power and how the di¤erence in productivity between public and private sectors both a¤ect the responsiveness of variables. The …ndings show that an increase in private sector productivity produces more desirable results with an increase in employment for both skilled and unskilled workers which translates into a decrease in overall unemployment. Public sector wage shock on the other hand mainly crowds out private skilled labour which the …rms react to by substituting it with unskilled workers. Ultimately, the increase in public wages raises overall unemployment as supplementary skilled unemployed individuals queue for public jobs. Altogether, the e¤ects are more pronounced when the bargaining power of unskilled workers is raised.

Sepedi: Thesese ye e thoma ka go tšweletša dipotšišo tše di latelago: naa ke kamano efe magareng ga sebopego sa mmaraka wa bašomi ka Afrika Borwa le bokgoni bja ba taolo ya ditšhelete go laola diphetogo tša ekonomi ye kgolo? Karabo ya potšišo ye e gararo. Sa mathomo, re nyakišiša bohlatse bjo bo hweditšwego bja go ba gona ga go se fetoge ga meputso ka mmarakeng wa bašomi wa Afrika Borwa. Sa bobedi, re sekaseka seabe seo mathata ao a ka bago le sona go maitshwaro a melawana ya ditšhelete. Sa mafelelo, re tsebagatša go thwalwa ga setšhaba mešomong go sekaseka diabe tšeo se ka bago le tšona go diphetogo tša mmaraka wa bašomi mo seemong seo se sego sa tlwaelega sa Afrika Borwa. Thesese ye ka fao e arotšwe ka dikgaolo tše tharo tše kgolo. Kgaolo ya 2 e akanya Mokgopamo wa New Keynesian Wage Phillips wa Afrika Borwa go tseba go arabela meputso mabapi le maemo a go thwala mešomong. Sa mathomo, re akanya mokgwa wa go šomiša tshedimošo ya bogareng. Dipoelo tša bogareng tša kakanyo di laetša gore meputso ya lekala la phraebete ga e tloga e arabela maemo a mešomo, mola ka go le lengwe di utolla gore go hlokomela ka tsela ye itšego go infoleišene le kamano ye kaone kudu le dikemelo tša infoleišene. Kamano magareng ga infoleišene ya moputso wa mathomong wa lekala la phraebete le go hwetša mešomo e tloga e laetša gore e a fokola go sampole ka moka, gomme e ba bohlokwa mafelelong a sampole, gomme se sa laetša gore koketšego ya go se fetoge ga meputso ka go lebakanako la ka morago ga kgethologanyo ka Afrika Borwa. Ka go le lengwe, kamano magareng ga infoleišene ya moputso wa mathomong le infoleišene ya ditheko e tloga e tiile kudu le go se tekateke go sampole ka moka eupša e fokola kudu ka bontši go lebakanako la nepišo ya infoleišene. Mo go lebakanako le, dikemelo tša infoleišene tša mekgatlo ya bašomi di tloga di amantšhwa kudu le infoleišene ya meputso ya mathomong. Ka go karolo ya bobedi re lebelela kamano magareng ga meputso ya mathomong, tšweletšo le moputso wa fasefase. Dikutollo di tiišetša gore infoleišene ya moputso wa mathomong e fetile kudu kgolo ya tšweletšo, le ka morago ga go phošolla infoleišene ya ditheko. Godimo ga fao, dipoelo di šišinya gore maemo a mošomong a na le seabe se sennyane go diphetogo tša meputso. Re hwetša gape gore go setšhaba, go leago le go lekala la ditirelo tša batho tšeo, di bopilwego kudu ke difeme tša setšhaba, phetolo ya meputso go maemo a mebaraka e tloga e tiilego ka fao go makatšago. Seswantšho sa kakaretšo seo se tšwelelago go tshekatsheko ye ke gore mokgwa wa tlhomo ya meputso wo o sego wa hlokomela kudu maemo a ekonomi ye kgolo ka kakaretšo. Kgaolo ya 3 ka fao e nyakišiša diabe tšeo seswantšho se se ka bago le tšona ge go dirwa melawana ya ditšhelete. Ka fao, re šomiša mokgwa wa DSGE fao go nago le dithulano magareng ga tlhokego ya mešomo le mmaraka wa bašomi ka morago ga mošomo wa Blanchard le Gali (2010) go araba potšišo ye bohlokwa ye. Ka ekonoming yeo e nago le meputso ye e fetogago, kamanommogo ye bohlokwa e ba gona ka ge infoleišene ye e dulago sekeng ka maitirišo e feletšago ka tlhokego ya mešomo ye e dutšego felogotee (sekgoba sa tšweletšo). Le ge go le bjale ge re lekola go se fetoge ga meputso le dithulano tše dingwe tšago se fetoge ka mebarakeng ya bašomi, dipoelo di laetša gore go šala morago maikemišetšo a infoleišene ya go se fetoge di feleletša ka go fetogafetoga fao go sa nyakegego le fao go ganelelago tlhokego ya mešomo; fao rešio ya tšhitelego ka tlwaelo e lego godingwana ka go fetoga ka mmarakeng wa bašomi. Kekišo ya mokgwa wo ka go šomiša tshedimošo ya Afrika Borwa e laetša dipoelo tše pedi tše kgolo. Sa mathomo, dikakanyo tša mollwane wa mmaraka wa bašomi di hlagiša seswantšho seo se nago le go se fetoge ga meputso fao go phatlaletšego e lego se se tiišetšago dikutollo tša kgaolo ya 2. Sa bobedi, mmaraka wa bašomi o tšwelela o tsebja ke dikelo tše kgolo ka dipalopalong tša tlhomo ya mešomo le tša tshenyo ya mešomo, fao tshenyo ya mešomo e tletšego ka diphetogong. Sa mafelelo, kgaolo ya 4 e etela leswa kutollo ya ka go kgaolo ya 2 ya phetolo ye e tiilego ya meputso mabapi le maemo a mmaraka ka lekaleng la mmušo. Godimo ga fao, lekala la mmušo ka Afrika Borwa ke leo le tletšego ka bašomi bao ba nago le bokgoni, ka ge le be le oketšega go tloga ka mathateng a ditšhelete a 2008 mola e le gore difeme tša phraebete di itemogetše tlhakatlhakano ye kgolo. Kgaolo ye ka go realo e tšweletša dipotšišo tše pedi: naa seabe sa go thwala setšhaba mešomong se ka ba sefe go diphetogo tša kakaretšo tša bašomi? Godimo ga fao, naa diponagalo tše tša go thwala setšhaba mešomong di ka tsenya letsogo ka seripa go hlaloša go se fetoge ga meputso fao go bonwego ka lekaleng la phraebete? Go araba potšišo ye, re šomiša mokgwa wa DSGE fao go dirišwago makala a mabedi le mehuta ye mebedi ya bašomi (go ya ka maemo a bona a bokgoni), e lego karolo ye bohlokwa ye e fetošitšwego ya motheo wo o tsebagaditšwego la mathomo ke Gomes (2013). Gabotse, re nyakišiša kudu diabe tša bothata bja moputso wa lekala la mmušo go mmaraka wa bašomi wo o kopantšwego. Mabakeng a papetšo, re sekaseka gape seabe sa bothata bja tšweletšokaone bja ka lekaleng la phraebete. Re lekodišiša gape maemo ao a fapanego a maatla a go rerišana le ka fao phapano ya tšweletšo magareng ga makala a mmušo le a phraebete bobedi e amago phetolo ya diphapano. Dikutollo di laetša gore koketšego ka tšweletšong ya lekala la phraebete e tšweletša dipoelo tše di nyakegago kudu fao go nago le koketšego ya mešomo go bobedi bašomi bao ba nago le bokgoni le bao ba se nago bokgoni gomme se se feletša ka phokotšego ya tlhokego ya mešomo ka kakaretšo. Bothata bja meputso ka lekaleng la mmušo ka go le lengwe bo dira gore go tlale bašomi ba bokgoni ka lekaleng la phraebete fao difeme di arabelago se ka go ba tšeela legato ka bašomi bao ba se nago bokgoni. Mafelelong, koketšego ya meputso ya setšhaba e hlatloša maemo a kakaretšo a tlhokego ya mešomo ka ge batho ba tlaleletšo bao ba nago le bokgoni bao ba sa šomego ba nyakago mešomo ya mmušo. Ka moka, diabe tše di bonagala kudu ge maatla a go rerišana a bašomi bao ba se nago le bokgoni a hlatloga gomme a fetoga kudu ka mmarakeng wa bašomi. Kakanyo ya mokgwa wo re šomiša tshedimošo ya Afrika Borwa e laetša dipoelo tše pedi tše kgolo. Ya mathomo, dikakanyo tša mollwane wa mmaraka wa bašomi di hlagiša seswantšho seo se nago le go se fetoge ga meputso gomme ka go realo se tiišetša kutollo ka go kgaolo ya 2. Sa bobedi, mmaraka wa bašomi o bonala o tsebja ka kelo ye kgolo ya mešomo ka go dikelo tša tlhomo ya mešomo le tshenyo ya mešomo, fao tshenyo ya mešomo e tletšego ka tshepedišong. Sa mafelelo, kgaolo ya 4 e etela leswa kutollo ya kgaolo ya 2 ya phetolo ye e tiilego ya meputso go maemo a mmaraka ka lekaleng la mmušo. Godimo ga fao, lekala la mmušo ka Afrika Borwa bontši ke la bašomi bao ba nago le bokgoni, gomme le gotše go tloga ka nakong ya mathata a ditšhelete ka 2008 mola difeme tša phraebete di itemogetše tlhakatlhakano ye kgolo. Kgaolo ye ka go realo e tšweletša dipotšišo tše pedi: na seabe sa go thwala setšhaba se ka ba sefe go diphetogo tša mmaraka wa bašomi? Godimo ga fao, naa dibopego tše tša go thwala setšhaba di tsenya letsogo ka seripa go hlaloša go se fetoge ga meputso fao go bonwego ka lekaleng la phraebete? Go araba potšišo ye, re šomiša mokgwa wa DSGE fao go nago le makala a mabedi le mehuta ye mebedi ya bašomi (go eya ka maemo a bona a bokgoni), e lego karolo ye e fetošitšwego ya motheo wo o tsebagaditšwego la mathomo ke Gomes (2013). Gabotse, re nyakišiša kudu diabe tša bothatakaone bja meputso ya lekala la mmušo go mmaraka wa bašomi wa Afrika Borwa wo o kopantšwego. Mabakeng a papetšo, re sekaseka gape seabe sa bothata bja tšweletšokaone bja ka lekaleng la phraebete. Re lekodišiša gape maemo ao a fapanego a maatla a go rerišana le ka fao phapano ya tšweletšo magareng ga makala a mmušo le a phraebete bobedi e amago phetolo ya diphapano. Dikutollo di laetša gore koketšego ka tšweletšong ya lekala la phraebete e tšweletša dipoelo tše di nyakegago kudu fao go nago le koketšego ya mešomo go bobedi bašomi bao ba nago le bokgoni le bao ba se nago le bokgoni gomme se se feletša ka phokotšego ya tlhokego ya mešomo ka kakaretšo. Bothata bja meputso ka lekaleng la mmušo ka go le lengwe bo dira gore go tlale bašomi ba bokgoni ka lekaleng la phraebete fao difeme di arabelago se ka go ba tšeela legato ka bašomi bao ba se nago bokgoni. Mafelelong, koketšego ya meputso ya setšhaba e hlatloša maemo a kakaretšo a tlhokego ya mešomo ka ge batho ba tlaleletšo bao ba nago le bokgoni bao ba sa šomego ba nyakago mešomo ya mmušo. Ka moka, diabe tše di bonagala kudu ge maatla a go rerišana a bašomi bao ba se nago le bokgoni a hlatloga.

Afrikaans: Hierdie proefskrif is dus in drie hoofstukke verdeel. Hoofstuk 2 maak ’n vooruitskatting van 'n nuwe-Keynesiaanse-loon- Phillips-kurwe vir Suid-Afrika om die responsiwiteit van lone op diensvoorwaardes te bepaal. Eerstens beraam ons die model met behulp van saamgestelde data. In 'n ekonomie met buigsame lone geld die goddelike toeval, aangesien stabiliserende inflasie outomaties tot 'n stabiele werkloosheid (uitsetgaping) lei. Uiteindelik verhoog die toename in openbare lone die algehele werkloosheid, aangesien aanvullend geskoolde werklose individue toustaan vir openbare poste. In die geheel beskou, is die uitwerking wat volg wanneer ongeskoolde werkers se bedingingsmag verhoog word, meer prominent namate die arbeidsmark meer vloeibaar is.