Day labouring in Southern Africa after the global economic and financial crisis: The case of Pretoria and Windhoek

Contribution/value-add: This study enhances the understanding of the structural vulnerability of the informally wage-employed in Southern Africa. Furthermore, it provides initial signs of the possible presence of long-term hysteresis unemployment within the informal labour markets as a result of a shock such as the GEFC


Introduction and rationale for the study
Since the turn of the century until the global economic and financial crisis (GEFC), a number of countries in the Southern African Development Community (SADC) region had made notable headway in stimulating economic growth and achieving economic stability (Chipeta 2011).The economic growth rate of South Africa averaged more than 6% over the 5 years before the advent of the crisis.Arguably the strongest economy in the region, South Africa, reached a peak in the business cycle in November 2007 after a 99-month upswing (Venter 2009).This upswing yielded an average growth rate in real gross domestic product (GDP) of 3% between 1995 and 2005 (Venter 2009).In a number of countries, inflation decreased to single-digit levels (before the food and fuel price increases of 2008) and reserves had been built up as well (Chipeta 2011).
The GEFC was an exogenous shock to the world economy that changed this picture drastically.The crash of the sub-prime mortgage market in the USA in 2007 resulted in a global recession during 2008 and 2009.It must also be noted that financial crises tend to disrupt economic activity to a greater degree in terms of its extent and depth than standard economic recessions (Colombo, Menna & Tirelli 2019).The crisis affected developed economies primarily through financial linkages.Emerging and low-income countries were impacted negatively in the areas of trade, investment and remittances channels (Chipeta 2011).Biyase, Fischer and Pretorius (2021) reported on the importance for remittances for material survival for recipient households as well as the positive impact on these households' subjective well-being.Richter and Patel (2022) furthermore concluded that in Brazil the self-employed (especially largely informal and often racial minority workers) suffered relative more from this worldwide health crisis.
The SADC countries did not escape the fallout.Different countries in the region were affected differently -depending on their economic, financial and trade structures (Chipeta 2011).Macroeconomic indicators reflected decreased export earnings, worsened balance of payments positions, lower tax collections and larger budget deficits (Chipeta 2011).
For South Africa, this shock was compounded by the negative economic results during the tenure of the then President Jacob Zuma, rising debt levels and junk status downgrades by all three leading rating agencies (Blaauw, Yu & Schenck 2021).For South Africa, as for the rest of the SADC, the eventual outcome of the crisis has been slower economic growth coupled with increasing unemployment and social pressure on limited resources (Chipeta 2011).
Increased unemployment figures viewed in isolation mask important underlying dynamics within labour markets in developed countries.Colombo et al. (2019) emphasise that financial crises are also followed by decreasing labour force participation rates as well as an increase in informal employment activities.Empirical evidence suggests that a financial crisis represents a large asymmetric shockresulting in a strong reallocation of economic activity between the formal and the informal sectors of the economy (Colombo, Onnis & Tirelli 2016).Furthermore, these reactions are both strong and persistent.
The notion of persistence of recessions and its impact on labour markets is not a new one in the literature (Cerra, Fatás & Saxena 2020).There is indeed still a suggestion that even many advanced economies have not properly recovered from the GEFC after more than a decade since.The GEFC left its mark on investors, firms, workers, and consumers who suffered its negative consequences (Cerra et al. 2020).This applies to the informally wage employed as well.This notion is supported by the statistics that show a recovery from the GEFC that was slow among all advanced economies (Cerra & Saxena 2017;Cerra et al. 2020).
Financial crises in less developed economies result in comparatively even stronger reallocation of inputs towards less efficient sectors.Such a crisis also renders a larger section of the population vulnerable to the adverse effects of informality (Colombo et al. 2019).
The SADC countries such as Namibia and South Africa are no exception in this regard.One of the chief manifestations of this move to informal economic activity is the activity of day labouring.Day labouring is becoming more frequent as a means to receive income in order to survive (Theodore et al. 2015(Theodore et al. , 2017;;Theodore 2020).It is a national and international occurrence and the number of people resorting to this form of informal wage employment for income is on the rise or re-emerging in the global north and south (Theodore et al. 2017).In South Africa, this informal wage employment strategy is referred to as day labouring or 'piece jobs' and in Namibia such workers are referred to as 'street unemployed' (Gonzo & Plattner 2003).Colombo et al. (2019) suggest that direct evidence of informal labour market activities is often scarce, although some studies suggest that in developing countries informal employment can be as high as 70% of total employment (Bosch & Esteban-Pretel 2012).Furthermore, it is important to focus not only on the macro-economic impact of exogenous shocks such as the GEFC, but also the impact on informal economic activities.External economic and health shocks often impact the informal sector disproportionately, especially in southern Africa (Bassier et al. 2020).This notion is confirmed in recent literature on the impact of the coronavirus disease 2019 (COVID-19) pandemic on South Africa's economy.Ranchhod and Daniels (2021) as well as Rakabe (2020) specifically pointed out that the informal sectors of the economy and labour market endured most of the immediate impact in the aftermath of the pandemicrelated lockdowns in South Africa.This strand of COVID-19 literature confirms the structural vulnerability (Richter & Patel 2022;Schenck, Blaauw & Matthee 2020;Theodore et al. 2015) of informal sector participants such as informally wage employed day labourers in the face of exogenous macro-economic shocks, for example, the GEFC -which is the topic of this study.Heeding the call of Colombo et al. (2019) for more evidence-based research on the labour market dynamics of informal sector activities therefore provides the rationale for the need for and importance of the ex-post analysis of the GEFC -presented in this study.
This study makes its contribution to the need expressed by Colombo et al. (2019) as contextualised in the preceding paragraph by providing evidence from two capitals in the neighbouring SADC countries of Namibia and South Africa in a pre-and post-crisis comparison of labour market outcomes of day labourers as a proxy for informal wage employment.
The choice of South Africa (Pretoria) and Namibia (Windhoek) as the study areas for this study is based, among other aspects, on the economic interdependence and interwoven economic histories (over a prolonged period of time) between the two countries.The Namibian economy has been historically integrated with South Africa through the Southern African Customs Union (SACU) -which Namibia formally became a member of immediately after its independence from South Africa (Mosikari & Eita 2020).Twenty-three years after gaining independence, Namibia's economy remains intricately linked to its southern neighbour (Hengari & Saunders 2014;Saunders 2016).Namibia's macroeconomic performance, fiscal and monetary policies as well as its trade relations with South Africa provide prima facie evidence of this (Hengari & Saunders 2014;Saunders 2016).Macro-economic events that influence the South African economy also filter through to the Namibian economy as well.This renders the two capitals as suitable case studies for the envisaged exploratory comparisonenvisaged for this study.Schenck and Blaauw (2008) conducted a comparative study into the differences and similarities of the day labour situation in Windhoek and Pretoria -the capital cities of Namibia and South Africa, respectively.Schenck and Blaauw's (2008) comparison laid the groundwork for this study to focus on the micro-economic situation as far as day labourers are concerned in the aftermath of the GEFC in the two cities.

The relevant theoretical framework and aim of the article
The aim of this article is to compare the dynamics of day labouring of two capital cities in Southern Africa, namely Pretoria and Windhoek, and to explore how the day labouring situation has changed before and after the GEFC in each of these capital cities.The overarching aim of the article is imbedded within the background and possible extension of hysteresis in unemployment to the informally wage employed day labourers.
According to O'Shaughnessy (2011), hysteresis as a concept in economics originated from its initial application to physical systems (particularly the study of electromagnetism).The conceptual foundation of hysteresis dates back to 1881 when an engineer described the behaviour of electromagnetic fields in ferric metals when exposed to magnetising cycles (Cross 1993;Schoeman & Blaauw 2009).The crux is that transitory causes may have permanent effects (O'Shaughnessy 2011).In terms of an unemployment time series, this would imply an auto-regressive structure in the data where dependence on an input or shock or memory of the shock will remain, even long after the initial shock or cause has been removed (Cross 1993;Schoeman & Blaauw 2009).The labour market may theoretically be influenced by non-economic considerations after a shock and prevent a return to a long-run Non-Accelerating Inflation Rate of Unemployment (NAIRU) (Long-run unemployment -natural rate of unemployment).This is referred to as the so-called remnant view on hysteresis (Schoeman & Blaauw 2009).
Traditional macroeconomic models often ignore hysteresis effects (O'Shaughnessy 2011).Blanchard and Summers (1987) offered the first definitive application of the concept of hysteresis to labour markets in the context of the contrasting behaviour of unemployment in Europe and in the USA following the two oil price shocks.Most of the earlier work on hysteresis in unemployment investigated whether the data are consistent with this model (O'Shaughnessy 2011).The results were often confirmatory and the work of Schoeman and Blaauw (2009) for South Africa also falls within this strand.More recent examples of its application in the literature (e.g.Gustavsson & Österholm 2006) are univariate models of unemployment hysteresis.
Practically, hysteresis unemployment can be regarded as a type of long-term unemployment, which occurs as a result of persistent high unemployment rates over a long time period.It can come to the fore when an economy experiences a period of prolonged weakness, such as during a recession after a shock.The GEFC fits this description.
During such a recession an increase in the actual rate of unemployment also leads to an increase of the underlying or equilibrium unemployment rate (O'Shaughnessy 2011).These long-term effects can cause a situation where the unemployment rate remains high even in instances where the economy recovers after such a shock.Many workers may not be able to find jobs again and if they do these may pay less than previous similar jobs.The adverse impact of unemployment shocks such as the GEFC may therefore have long-term consequences with serious economic and social consequences (O'Shaughnessy 2011).This may also be the case for the informally wage employed -which is the issue under investigation in this article.
The study provides key socio-economic and demographic differences between the situation faced by day labourers in Pretoria and Windhoek before and after the GEFC.Specific focus is placed on the employment history (e.g.previous formal sector employment and how long it was held before losing it), labour market outcomes, and the changes in income levels.
The next section describes the particulars of the research methodology that was used in conducting the fieldwork for the research.

Methodology
The basis for the formulation of the research methodology was laid by researchers such as Schenck andLouw (2005), Valenzuela Jr et al. (2006), Blaauw et al. (2006), Schenck and Blaauw (2008), Blaauw, Pretorius and Schenck (2016), Theodore et al. (2017).These researchers investigated various dimensions pertaining to the specific informal sub-sector used in this study, namely day labourers.
The research design employed for the purpose of this comparative study is essentially a mixed-method approach.A survey-based instrument was used with mostly quantitative items and some applicable open-ended qualitative elements.The surveys were administered through the use of structured interviews and the method used ensured that questions asked to the participants could not be misunderstood or misinterpreted (Bless & Higson-Smith 1995;Schenck & Blaauw 2008;Valenzuela Jr et al. 2006).
The questionnaire that was used consisted of various sections pertaining to demographic and key economic information of the day labourers, such as personal background, education, employment, employment-seeking history, language proficiency, income information, dependents' information, information on the sites they stand at when looking for jobs, and relationships and social network information.The questions asked in the two surveys in Pretoria and Windhoek in the post-GEFC era were exactly the same apart from the demographic questions where aspects such as predominant and/or official language in the two countries differ and hence the options pertaining to these also differ.The survey used in Pretoria in 2004 also contained essentially the same questions -with an adjustment here and there in 2015 as an improvement in response to the experiences of the fieldworkers regarding suitability and understanding of questions in the 2004 survey.
In line with previous studies (e.g.Blaauw et al. 2006;Theodore et al. 2017;Valenzuela et al. 2006), the research population was defined as day labourers who are waiting at informal hiring sites such as street corners, next to traffic lights, or in front of job-related businesses in Pretoria and Windhoek.Here they hope to obtain a temporary informal job opportunity for the day or a number of days.People who were merely standing around hiring sites and who were employed and only waiting for transport were not included in the defining criteria of the research population.
The next phase of the research focused on the sampling procedures which were guided by the research strategies used in a national survey of day labourers in the USA (Valenzuela et al. 2006) as well as the principles that guided the sampling technique used by Blaauw et al. (2006), Theodore et al. (2015) and Theodore et al. (2017).It was impossible to compile a complete list with names of all day labourers in Pretoria due to the constantly changing numbers and flexible nature of this informal activity.
In the light of the above, the aim was to cover all the known hiring sites in Pretoria and to interview a minimum sample of 10% of the total research population visible in the city.All the hiring sites known to the authors as well as the new sites encountered were visited.The envisaged sample of at least 10% of the day labourers present at each of these sites were selected for interviews on a voluntary basis following the established protocols of earlier studies in the same areas (Blaauw et al. 2006;Theodore et al. 2015Theodore et al. , 2017)).While we may not necessarily be able to claim absolute statistical representativeness of the sample, the sampling methods used were consistent and covered the research population in all geographical areas of Pretoria.
In the case of Windhoek, the number of day labourers in Windhoek, which is the capital of Namibia, is not documented or registered.This again made it difficult to establish the exact number of day labourers in the city.We were fortunate to be assisted by Dr Rachel Freeman from the University of Namibia to identify all the known hiring sites in the city.Windhoek is much smaller than Pretoria.As a result, the sampling strategy was adjusted to allow for all the willing day labourers to be interviewed.Each known hiring site was visited.All day labourers we encountered were given the opportunity to participate in the research.The sampling procedure we followed can be described as nonrandom in nature.We purposefully interviewed all day labourers who were present and willing to participate during the fieldwork.Very few day labourers were not willing to participate.We are therefore confident that the research population was covered as extensively as possible as each day labourer had the option to participate in the study should he or she wish to do so.
The differences in absolute sample sizes obtained in the rounds of surveys in Windhoek and Pretoria must be viewed within the context of the vastly varied sizes in the research population in the two countries and the capital cities.South Africa has, for example, a population well over 20 times that of Namibia (Hengari & Saunders 2014).Pretoria is much bigger geographically and in terms of population.The same protocols were followed in the two cities, that is, all previously known hiring sites in Pretoria we revisited and the same was done for all the known hiring sites in Windhoek.When new sites were encountered during the fieldwork, these were covered as well.All day labourers encountered in the two cities were given an opportunity to participate.Hence, the eventual sample sizes represent the outcome of the same protocols -but proportionally so in terms of the response rate.The absolute numbers between the different surveys in the different cities before and after the GEFC were also more or less the same -providing further credence to the above.
Fieldworkers were recruited from a database of qualified social work students who had not yet found permanent employment at the time in Pretoria, and in Windhoek, fieldworkers were recruited from the Department of Social Work from the University of Namibia (UNAM).In both cases, the principal researchers in the research team provided training before the fieldwork commenced.Training for the fieldwork was carried out according to the suggestions of Rubin and Babbie (1997:356-358).Demonstration and roleplay formed part of the training before the actual fieldwork began.The fieldwork to collect the data for Pretoria was carried out in 2015, and for Windhoek in 2017.All completed questionnaires were evaluated to ensure all the answers had been completed.Questionnaires that were deemed illegible or incomplete were discarded in order for the results to be as accurate as possible.The data were captured in Excel and the subsequent analysis was carried out in the statistical analysis program, EViews.In total, the sample used in this study for Pretoria consisted of 290 participants in 2015, and in Windhoek, 80 participants were interviewed in 2017.The corresponding samples for the pre-GEFC surveys were 242 in Pretoria ( 2004) and 160 in Windhoek in 2003 (see Gonzo & Plattner 2003).
The study adhered to strict ethical elements and principles.It received clearance from the ethics committees of the Northhttps://www.jefjournal.org.zaOpen Access West University and the University of the Western Cape.The respondents also had the right to choose not to be included in the interviews and the study, and if they took part, their identities were kept completely anonymous.This principle was already respected in the design of the questionnaire, where no information was asked that could identify the respondents.
The information obtained for Pretoria and Windhoek was analysed and compared with the results of the Schenck and Blaauw (2008) study in which similar survey instruments and research protocols were followed.The next section presents the empirical results of the analysed data.

A comparison of the changing demographic features and working conditions of day labourers in Pretoria and Windhoek after the GEFC: 2008-2015/2017
On the surface, the demographic features of the day labourers in the two capitals did not change much in the decade before and after the GEFC.It is still a male dominated activity in both cities.The ages of the day labourers were concentrated around 20-45 years for both Pretoria and Windhoek, which implies that they fell into the age group that is expected to be the most economically productive one in a nation's workforce.
The most notable change is the considerable proportion of foreign migrants that make up the day labour population in Pretoria in 2015 compared to 2004.The country of origin in terms of the day labourers in Windhoek did not show any significant changes pre-and post-GEFC.The situation in Pretoria was markedly different.
In 2015 there was a marked increase in the number of respondents from Zimbabwe whom were active in South Africa, compared to the 2004 findings.The reason for this situation could be that day labourers are migrating from Zimbabwe to South Africa in the hope of finding jobs due to the harsh economic circumstances in Zimbabwe.This can partly be ascribed to the advent of the GEFC in 2008.The most important factor in the Zimbabwean case was the economic meltdown in Zimbabwe as a result of the disastrous land reform policy adopted by the government of President Mugabe (Theodore et al. 2017).This led to a third wave of immigration to neighbouring countries with South Africa being the primary destination country (Crush & Tevera 2010).The impact of the GEFC could have caused day labourers from other countries such as Lesotho, Namibia and Angola to migrate in bigger numbers to South Africa in order to find jobs, as their countries were hit hard by the GEFC and economic circumstances after 2008.Wentzel et al. (2022) confirmed that macro-level factors played a role in migrants' decision to migrate and/or stay.
It can be incredibly stressful for day labourers to have to care for a larger number of dependants on the relatively small income they receive (Magidi 2022).Over the period of this study the number of dependants a day labourer has, did not really change.In Pretoria in 2015, a day labourer on average still supported about four people, with the number being closer to five dependents for day labourers in Windhoek.
Most of the day labourers in Pretoria as well as in Windhoek also had some secondary schooling with 46.82% for Pretoria and 56% for Windhoek.Very few of the day labourers in both cities had schooling past matric or some secondary level of education.In both South Africa and Namibia, at school level, the highest grade that can be obtained is Grade 12 (Schenck & Blaauw 2008).Very few of the respondents in Namibia had completed their secondary schooling with 32.58% of the respondents completing their secondary schooling in Pretoria in 2015 as opposed to 71.4% in 2004.However, it should be noted that the same day labourers were not necessarily interviewed in these aforementioned studies (Pretoria 2004 andWindhoek 2003).This constitutes a decrease of more than 35% from the level of schooling indicated among day labourers in the Schenck and Blaauw (2008) study.In Windhoek, only one day labourer indicated some form of formal qualification, having done an electrician certificate course.As reasons for not completing or furthering their education, the day labourers indicated having no money, leaving school to support their families, and not being interested in school.Not having any formal education or training makes it difficult for day labourers to cross the The next section analyses the dynamics of day labourers in Pretoria compared with their counterparts in Windhoek in terms of employment history, and a particularly important factor, which is income earned and future prospects.
The employment dynamics of day labourers in Pretoria and Windhoek Theodore et al. (2015) describe the possible relationship between macroeconomic conditions (especially unemployment) and the size of the day-labour workforce.Given the fallout experienced in terms of unemployment as a result of the GEFC, we followed the analysis of Theodore et al. (2015) and explored the possibility of hysteresis unemployment in the years following the GEFC in Pretoria and Windhoek.Figure 1 and Figure 2 illustrate the year in which workers began day labouring in the two cities and the unemployment rate in South Africa and Namibia, respectively.South Africa's official unemployment rate increased from just over 22% in 2009 to almost 25% in 2011 as the impact of the GEFC shock unfolded.Thereafter, the unemployment rate persisted around these levels.The number of day labourers joining this informal labour market grew exponentially since 2011 with the most significant numbers joining the market 3-4 years after the advent of the GEFC.It would seem that the shock still impacted unemployment a number of years after the broader macroeconomic impact the shock had on GDP started to fade -indicating the possibility of hysteresis in formal unemployment which reverberates through to the informal labour markets of the informally employed day labourers.This resonates with the findings of Theodore et al. (2015) which found a statistically significant correlation between the year the day labourer in South Africa joined the day labour market and the unemployment rate of 2 years earlier.This was a countrywide study based on date (2008) gathered before the GEFC and the authors found that the correlation coefficient between the year that a worker started as a day labourer and the South African unemployment rate at the time was 0.73 (Theodore et al. 2015).When the 1 9 8 0 1 9 9 0 1 9 9 3 1 9 9 4 1 9 9 7 1 9 9 9 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 2 0 1 4 2 0 1 5 2 0 1 6 2 0 1 7

Year started as day labourer
Unemployment rate Number of day labourers unemployment rate is lagged with 2 years earlier, the correlation coefficient is no less than 0.84.Following this protocol, we performed the same standard correlation analysis.In Pretoria and Windhoek, the increase in the number of job seekers becoming day labourers is positively correlated with the unemployment rate in the two countries.The correlation coefficient is however markedly lower for both cities, namely, 0.62 and 0.47 for Pretoria and Windhoek, respectively.Although no firm conclusions can be based on this, it does leave open the possibility of the presence of hysteresis filtering through to the market of the informally wage employed day labourers.The changes in equilibrium unemployment, as a result of the GEFC, seem to display signs of remnance, with no significant adjustment back to pre-crisis employment levels after the macro-economic recovery following the crisis.
Figure 2 illustrates a much more pronounced increase in unemployment in Namibia after the advent of the GEFCfrom 17% in 2012 to 22% in 2016.As a smaller economy, the macro-economic effect of the shock in Namibia seemed to be longer in duration with the new equilibrium unemployment rate only reached after a longer time period than in South Africa.The continuous increase in unemployment in the period under consideration also reverberated in Windhoek's day labour market with a significant portion of the day labour workforce joining the market after the economic fallout of the crisis as seen in the formal sector.The pattern again corresponds with the results of Theodore et al. (2015).Again, the possibility of an element of hysteresis unemployment is existing and requires further investigation.
Investigating whether the day labourer had a full-time job before being forced into day labouring can also be an indication of the negative impact of a shock such as the GEFC on the formal sector of the economy.Many day labourers were indeed in full-time employment before becoming part of the day labouring community.Of the Pretoria day labourers in 2015, 131 respondents (45.17%) indicated that they had had permanent employment before becoming a day labourer and in 2017, for Windhoek, 54 (67.5%) of the 80 respondents indicated that they had been employed full-time before engaging in day labouring.On average, the last time that one of the respondents in Windhoek had been employed on a full-time basis was in 2012, meaning that at the time of data collection in Windhoek most of them had been unemployed for 5 years since the advent of the GEFC.For Pretoria, the average year from which the labourers had been day labouring was 2011, which means that in 2015, most Pretoria labourers had been unemployed for 4 years.This means that the labourers in both cities were long-term unemployed individuals.These numbers correspond with the Schenck and Blaauw (2008) study where for Pretoria, 102 respondents had held full-time employment before becoming day labourers, whereas 61% of Windhoek labourers had previously held full-time employment (Gonzo & Plattner 2003;Schenck & Blaauw 2008).The sectors in which these labourers had been employed before becoming day labourers also had not really changed: they still worked mostly in the construction and services sectors for both Pretoria and Windhoek from 2003and 2004to 2015and 2017.With the above indicating the period of unemployment among the day labourers in Pretoria and in Windhoek, it is also important to know how these day labourers lost their previous full-time employment or why they left.
Figure 3 indicates the reasons for full-time employment job losses of the day labourers in Pretoria and Windhoek, respectively.
The overwhelming reasons for labourers losing their previous full-time employment are generally that their contracts were not renewed as well as layoffs.All of these are reasons linked intrinsically to the business cycle and economic performance of the respective macro-economies.The GEFC may therefore have contributed to the worsening position of day labourers' prospects of finding formal employment in both cities.

An income analysis and comparison of day labourers in Pretoria and Windhoek between 2003/2004 and 2015/2017
The day-to-day income that day labourers generate is very unpredictable and inconsistent.They never know exactly how much they are going to receive for the job done on a specific day (Pretorius & Blaauw 2015;Theodore et al. 2017).It is also very possible that they do not receive any income for the day or for a period of time due to the irregular hiring patterns of day labourers.
We investigated the hiring patterns by recording the number of days a day labourer was able to secure employment in the week before being interviewed.Comparing this to the number of days that each individual stood at the hiring site in the hope of securing employment, allows us to determine the success rate for each respondent.The average success rate in Pretoria in 2015 was 18.1% and in Windhoek (2017) it was 17.1%.In practical terms this equates to 1 day of finding a job for every 6 days a day labour incurred the travelling and other cost of engaging in the job search activity.The gravity of their situation is illustrated more vividly when the above analysis is reported by means of the median.The median success rate in both cities are 0%.In other words in both cities, the majority of the day labourers were not able to find any employment in the week before being interviewed.These statistics imply that within this informal labour market the rate of unemployment can be anything from 80% to a 100% -a state of affairs bordering on disastrous for lives and livelihoods as will be seen from the discussion on income earned that follows.
The irregular hiring patterns and low rates of success make it difficult to pinpoint the exact amount a day labourer receives on a monthly basis (Theodore et al. 2017).In an attempt to take this into consideration, respondents were asked to provide estimates of their income in what they perceived to be a good month, bad month and the previous month.
For a more accurate comparison (given the different time periods involved) the relevant values were expressed in real terms.The values for the amount of remuneration the Pretoria day labourers received were calculated in real terms with 2004 being the base year for Pretoria and 2003 for Windhoek.This means that the remuneration earned by the day labourers in 2015 in Pretoria and in 2017 in Windhoek were adjusted for inflation to be compared to remuneration earned in the Schenck and Blaauw (2008) study.These results are presented in Figure 4.
As mentioned earlier, the income that is earned by day labourers is a big concern and is very unpredictable.On average the day labourers in Windhoek received almost R400 more than the day labourers in Pretoria on a good month.The Namibian Dollar is pegged to the Rand in terms of purchasing value which means that ZAR1 = 1N$.It is cause for concern to see the amount of money that these day labourers receive in Pretoria, even in a good month.This statistic reflects worsening macro-economic conditions and possibly a resultant oversupply in the day labour marketdriving already low wages to new extremes of precarity (Theodore et al. 2017).This state of affairs continues to create major uncertainties for day labourers and adds to the stress not knowing how they are going to care for themselves and their dependants on a daily basis (Blaauw et al. 2022;Magidi 2022;Smith 2020).This data indicates that day labourers are still living in severe poverty and that the picture has not improved over the last decade.
In Pretoria, 25% of day labourers earned more than R1 900 in a good month, and in a bad month they earned between R0 and R699.These values were calculated in real terms with the remuneration being adjusted for inflation, with 2004 being the base year for Pretoria and 2003 the base year for Windhoek.In both cases that is very concerning because it is impossible to feed a family of four on so little money per month.This adds to the discussion around the volatility of wages due to the notable difference in income in a good and a bad month.
In 2004, the day labourers in Pretoria earned in the R0 to R699 range in a bad month just as in 2015, according to the data, and in a good month most earned in the R800-R1100 range (see Figure 5).One may be tempted to assume that the poverty situation of day labourers is still more or less the same as it was a decade ago, with an increase in the real earnings in a 3 0 0 -4 9 9 5 0 0 -6 9 9 7 0 0 -8 9 9 9 0 0 -1 0 9 9 1 1 0 0 -1 2 9 9 1 3 0 0 -1 4 9 9 1 5 0 0 -1 6 9 9 1 7 0 0 -1 8 9 9 good month in 2015, where in real terms most labourers earned more than R1 900 in a good month.This would be a positive finding to observe, but such earnings may not be enough to care for a large family or a sizeable number of dependents.This would, however, be an incorrect observation for another reason as well.Previous research in the day labour market across the world has confirmed that the good days and months are the exception and occur far less frequently than bad months.Moreover, in the past 2 years the COVID-19 pandemic has worsened the bad months that labourers experienced.Hence, a more detailed comparison of all available income measures is presented next (Table 2).
The average day labourer in Pretoria would struggle immensely to get by on a month-to-month basis when comparing their monthly income in a good and bad month to the minimum living level income in South Africa.In 2004, the average labourer earned between R800 and R1100 in a good month.In 2015 the average day labourer in Pretoria earned R1490.48 (in 2004 prices) in remuneration in what they considered a good month.Although this represents a substantial increase, it is important to remember that good months are the exception and that there are far fewer good months than average or bad ones.Day labourers now struggle even more to get by in the aftermath of the GEFC, as the bad months are for the most part their daily reality.
In 2015, the Food Poverty Line (FPL) for a person to afford the minimum required daily intake only for food was R441 per month per person, an increase from R199 in 2004; that excludes any other non-food and household items (Statistics South Africa 2019).To be able to afford food, non-food and household items a day labourer would need to get enough money per month, that is, R992 per person per month, which is an increase from R403 per person per month in 2004.This means that the average day labourer in Pretoria would only be able to afford food, non-food and household items for one person, and that would only be possible in a good month.It would not be possible in a bad month for one person.It is particularly worrisome because most day labourers have an average of four extra people they have to take care for.
Comparing the pre-and post-GEFC income for Windhoek proved slightly more complex.The benchmark study for Windhoek by Gonzo and Plattner (2003) in Windhoek reported that at the time day labourers there were paid on the basis of completion of a job or task.This was not based on their labour input or the amount of time it took to complete a task (Gonzo & Plattner 2003;Schenck & Blaauw 2008).Gonzo and Plattner (2003) described the situation in 2003 as follows: The financial situation of Namibian 'street unemployed' can best be described as desperate … Three quarters of the investigated 'street unemployed' earn less than N$302 per job, a shocking result, especially in light of the fact that the majority of the respondents either get only one job per week or nothing at all.In addition, 83% of the respondents declared that they have no other means of income.These income figures are better understood when one considers that Windhoek is one of Africa's most expensive cities and N$30 per week is by far not enough for food, let alone accommodation and other basic needs.(p.64) Although many jobs were for only a day, one cannot assume that to be the case throughout their sample -rendering a direct comparison with the 2017 data as one that must be treated with caution.However, the message of precarity and vulnerability after the GEFC is confirmed when looking at the data for 2017 in Windhoek.Figure 6 and Table 3 represent the results.
In 2003, 21% of the labourers in Windhoek earned more than N$1900 in a good month, and in a bad month earned between N$0 and N$899.In comparison with the Pretoria results, in a bad month the Windhoek labourers certainly earned very little money with 43% of labourers indicating that in a bad month they earned between N$0 and N$299.For both Pretoria and Windhoek, this reiterates the discussion around the volatility of wages due to the substantial difference in income in a good month as opposed to a bad month, because a family of four dependants cannot be fed with so little money.The results also indicate that when bad months are experienced, the situation is dire.
The day labourers in Windhoek would be able to afford slightly more, but not as much as those in Pretoria, making their outlook just as bleak, especially seeing that the earnings in a bad month were also less than N$500, with an average FPL of N$293.10 per person.This amount would not be enough to care for a family, and the upper bound poverty line (UBPL) of N$520.80 was just enough to buy the basic requirements for three people.These increases in the poverty lines for Pretoria and Windhoek as well as the increase of inflation, eat away at the purchasing power of the earnings of these individuals when they direly need the money to survive.As was the case in Pretoria, the advent of COVID-19 and the worsening economic outlook for Namibia render the informally employed in Windhoek also structurally vulnerable and in a precarious position in terms of their lives and livelihoods.
The real median income levels for labourers in Pretoria in 2015 were also calculated, with Pretoria labourers earning a real median income of R1336 in a good month, and R321 in a bad month, expressed in 2004 prices at the time of data collection.For the bad months, the earning levels were still less than R500 per month in real terms which indicates the same or worsening macroeconomic conditions since 2004 and the Schenck and Blaauw (2008) study.It was further important to indicate the real growth rate that these labourers experienced during the good months.As stated, in 2004 Pretoria labourers earned between R800 and R1100, and in 2015 they earned on average R1490 in 2004 prices.These calculations revealed that the real growth rate for what is perceived as a good month was between 35% (R1100 and R1490) and 85% (R800 and R1490) between 2004 and 2015.
As a last step in the analysis we ran a series of standard Mincerian-type wage equations using the available income variables for both cities as the dependent variable -linked with the typical independent variables forthcoming from the day labour and general labour market literature (Theodore et al. 2017).
The results of the best fitted models are presented in Table 4 and Table 5.
Both sets of regressions show that very few of the traditional human capital theory variables such as the numbers of years of schooling had any statistically significant relationship with income earned in a good month in the 12 months prior to be interviewed.The number of days that a day labour stood at the street corners also had no statistically significant link -which is no surprise when the low success rate described earlier is taken into account.Whether a day labourer had full time employment before becoming a day labourer was also statistically insignificant.
Age proved to be statistically significant in Windhoek but not in Pretoria.In Pretoria, the dummy variable distinguishing between being a local or foreign migrant was statistically significant -indicating the foreign day labourers were in fact able to earn more in a good month than their local counterparts.This advantage disappeared when income earned in a bad month was analysed -as times were then bad irrespective of origin.
The only statistically significant variable in both cities was the dummy variable representing whether or not the respondent received some specific training in one or more than one marketable artisan skills, for example, plumbing and tiling.These results confirm previous findings (Theodore et al. 2017) that indicated that day labourers who had these types of artisan skills can earn significantly higher wages and are hired far more often than their counterparts with less developed skill levels.
For the day labourers in both cities who do not have these skills the picture is bleak with unemployment rates of more  3 0 0 -4 9 9 5 0 0 -6 9 9 7 0 0 -8 9 9 9 0 0 -1 0 9 9 1 1 0 0 -1 2 9 9 1 3 0 0 -1 4 9 9 1 5 0 0 -1 6 9 9 1 7 0 0 -1 8 9 9  than 80% and a significant decrease in real income levels.For them, the slowdown in the economy was nothing short of disastrous, with bad months being their new normal at the time of the survey.The advent of the COVID-19 pandemic and the impact on economic activities have in all probability worsened their plight even more -something that needs urgent research attention.

Summary of findings, policy implications and limitations of the study
Shocks like the GEFC and COVID-19 pandemic have brought the plight of those engaged in the informal economy to the fore and highlighted the often precarious existence of many informal workers such as day labourers.Heeding the call of Colombo et al. (2019), this study endeavoured to provide more evidence-based research on the labour market dynamics of informal sector activities in the aftermath of economic crises such as the GEFC.The impact of a crisis such as the GEFC is often hits structural and vulnerable elements of the informal sector the hardest.This provides the rationale for the need for and importance of the ex-post analysis of the GEFC -presented in this comparative case study on day labouring in the capital cities of two integrated SADC economies.
The main demographic features of the day labour population in the two cities did not change much in the decade before and after the GEFC.The only notably exception is the changing composition of the day labourers in Pretoria in terms of the number of foreign migrants that increased significantly -especially from Zimbabwe as a result of questionable macro-economic policies pursued by the government in that country.
The cyclical impact of lower economic growth (partly as a result of the GEFC) is evident in both capital cities.A sizeable portion of the day labourers had been employed on a fulltime basis before becoming day labourers.The main reasons for their losing their full-time employment were that their contracts had ended and they had been laid off, which consequently forced them into day labouring.
Between 2004 and 2015, the actual earnings of the labourers increased in real terms in what they termed as a good month for Pretoria -but mostly for those with marketable artisan skills such as plumbing and tiling.The regression results presented give further credence to that fact that having these marketable skills is one of the few statistically significant contributors to their monthly income in a good month.In a bad month, the situation was still bleak.In fact, it was even worse with labourers still earning less than R500 in a bad month, indicating deteriorating macroeconomic conditions.As stated, the increase in poverty lines and inflation since 2004 for Pretoria does not help the case for the day labourers as those increases simply chip away at the purchasing power of the labourers' earnings.Even though the average day labourer experienced an increase in real income in the good months, bad months are indeed still the norm at present, and the reality for most day labourers.The average success rate of below 20% in finding temporary jobs and the fact that the majority of the respondents could not secure any employment in the week preceding the interviews bore testimony to a labour market that is still experiencing the remnant effects of the GEFC.The advent of long-term hysteresis unemployment in this informal labour market is a distinct possibility that warrants further investigation.
The informal economy is often seen as playing a countercyclical role as a shock absorber and a safe haven in times of recession or an economic shock such as the GEFC.These results therefore also question the theoretical shock absorber function often ascribed to the informal sector in time of financial strife.It furthermore places existing policy responses to exogenous shocks such as the GEFC or COVID-19 under scrutiny and requires governments to rethink the depth and level of support afforded to the informal sector in such times.The impact of the financial assistance from governments at the time of the COVID-19 pandemic is an especially crucial element of such a research agenda to further provide evidence-based research on this informal labour market.Other researchers such as Magidi (2022) have also recommended that labourers are in dire need of government assistance and organised networks to keep them safe and to make it easier for them to find private sector employment.Magidi (2022) also emphasises that civil society and community-based organisations should be encouraged to assist with issues such as skills development, training and retraining, as well as with removing potential barriers so that these labourers can re-enter the formal employment sector where possible.Revitalising organisations such as Men on the Side of the Road (MSR) in South Africa can be of key importance in this regard.
We duly acknowledge several limitations to our study and, this also explains our cautious approach when interpreting our results.We are focusing on the micro-labour market dynamics of the informally wage employed at the backdrop of macro-economic events.Although the evidence and literature point towards the eventual labour market effects of such macro-economic events we cannot and do not claim causality in terms of the direct effect at play.Our comparison revolves on two cross sectional surveys in each city.A panel study would have been ideal, but the nature and fluidity of the day labour market and protecting the anonymity of the respondents made a panel study, which interviews the exact same individual a number of years apart, not a realistic option for this particular study.Without such a panel analysis we cannot claim to be able to control for all other variables that may have influenced our findings and results.

Concluding remarks
The vulnerability of the informally employed to exogenous macro-economic shocks such as the GEFC is evident in the employment and income levels forthcoming from the research.The long-term remnant effects of this shock are still keenly felt by the informal sector -which often are disproportionally affected by such exogenous shocks.The aftermath of such exogenous shocks may indeed accentuate already existing inequalities and structural vulnerabilities.
The structurally disadvantaged and vulnerable sections of the labour force are those who have paid and are still paying the highest price for the crisis.At the same time these are also the workers who face the most difficulty in finding quality and decent formal employment again.
The long-term situation and prospects (or lack thereof) facing day labourers in Southern Africa are therefore worrisome.If the plight of day labourers and their vulnerability to exogenous shocks such as the GEFC are not placed higher on the research and policy agenda of academia and government, the impact of the next macro-economic shock may be even more calamitous for these people who could have turned to crime in order to survive.The advent of the COVID-19 pandemic has pressed home this message even more.

FIGURE 2 :
FIGURE 2:The year in which workers began day labouring and the unemployment rate in Windhoek (2017 data).

FIGURE 1 :
FIGURE 1:The year in which workers began day labouring and the unemployment rate, Pretoria (2015 data).

FIGURE 5 :
FIGURE 5: Earnings distribution of day labourers in 2015 for Pretoria in real terms.

FIGURE 6 :
FIGURE 6: Earnings distribution of day labourers in 2017 for Windhoek in real terms.

Table 1
provides a comparison of the demographic data for the research in Windhoek (2003) and Pretoria (2008) with the research in Windhoek (2017) and Pretoria (2015).

TABLE 1 :
Demographic data forWindhoek (2003)and Pretoria in 2004 compared to 2015 (Pretoria) and 2017 (Windhoek).formal sector employment, because they do have the necessary experience to be able to receive formal sector employment, but the employers require proof of training or education, which the labourers do not have.In both Windhoek and Pretoria, the vast majority of labourers indicated that they did jobs such as painting, carpenter work, construction and gardening in 2003/2004 as well as 2015 for Pretoria and 2017 for Windhoek.In Pretoria, another job that was mentioned frequently was removing rubble and rubbish.

TABLE 2 :
Comparison of the average monthly income of day labourers in Pretoria in real terms compared to the inflation-adjusted national poverty lines: 2004 and 2015.

Monthly income needed to afford the minimum required daily intake per person in South Africa (2004) Monthly income needed to afford the minimum required daily intake per person in South Africa (2015)
Statistics South Africa, 2008, Measuring poverty in South Africa: Methodological report on the development of the poverty lines for statistical reporting, pp.1-36, Government Printers, Pretoria; Statistics South Africa, 2019, National poverty lines, viewed 01 August 2021, from http://www.statssa.gov.za/publications/P03101/P031012019.pdf%20%20Accessed%201%20August%202021. Source:

TABLE 4 :
Regression analysis of a good month's income of day labourers in Windhoek (2017).

TABLE 3 :
Comparison of the average monthly income of day labourers in Windhoek (2017) in real terms compared to the inflation-adjusted national poverty lines: 2003 and 2017.