Informality, Job Quality, and Welfare, in Sri Lanka Informality, Job Quality, and Welfare, in Sri Lanka Poverty and Equity Global Practice South Asia © 2020 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Contents Acknowledgments    7 Abbreviations    7 Executive Summary     8 1. Introduction      10 2. Informal employment: Measures and trends      13 Defining informal employment    14 Informal employment as a salient and persistent feature of the Sri Lanka labor market    16 Informal workers: Older, less educated, and primarily engaged in agriculture  and low-end services    18 3. What explains the high rate of informality in Sri Lanka?      21 Key labor regulations: Application in practice and implications for labor cost    22 High cost of compliance and complexity of labor regulations as motivating  factors of informality    26 Limited benefits of formalization in Sri Lanka    29 4. Informal employment, job quality, and poverty      31 Informality and job quality    32 Informal jobs and contracts and working conditions    32 Informal jobs and the risk of poverty and vulnerability    35 5. Conclusion and policy implications      42 Appendix A      47 Bibliography    55 Figures Figure 1 Trend in formal and informal employment, 2006–17: Number of people (thousands)  and share of total employment (%)    16 Figure 2 Workers with pension entitlements and contracts    17 Figure 3 Formal and informal workers by gender    18 Figure 4 Informal employment by age group    19 Figure 5 Educational attainment of formal and informal workers    19 Figure 6 Informal employment share by province (left) and geographic sector (right)    20 Figure 7 Cumulative compensation payable as severance and gratuity upon dismissal  (no. of weeks’ salary by length of tenure in years)    23 Figure 8 Comparison of redundancy costs across countries (weeks of salary)    24 Figure 9 Average weekly work hours and underemployment (%)    33 Figure 10 Distribution of weekly work hours    34 Figure 11 Monthly and hourly earnings of formal and informal workers    36 Figure 12 Distribution of log hourly earnings for formal and informal workers    36 Figure 13 Gini coefficient of monthly earnings    37 Figure 14 Monthly nominal earnings by province, formal and informal workers (SL Rs)    37 Figure 15 Formal-informal wage gap, 2006–17    38 Figure 16 Distribution of working-age population by employment status and income decile    41 Figure 17 Male and female labor force participation rate by age    45 Figure A.1 Male-female earnings gap (employees only)    54 Tables Table 1 ILO conceptual framework of formal and informal employment    15 Table 2 Severance pay by tenure    23 Table 3 Paid leave and holiday regulations    26 Table 4 Working hours, overtime, and payments    27 Table 5 Formal and informal workers by type of contract    33 Table 6 Entitlement to paid leave among paid employees (%)    34 Table 7 Formal-informal wage gap by industry (SL Rs)    37 Table 8 Oaxaca decomposition results on log of hourly earnings: Formal versus informal workers    39 Table A.1 Labor force participation rate, age 15 and over (%)    47 Table A.2 Summary of labor market indicators    47 Table A.3 Formal and informal employment as share of total employment (%)    48 Table A.4 Number of workers with secondary job (thousands) and share of total employed, 2006 – 17     48 Table A.5 Distribution of workers with formal/informal main and secondary jobs    48 Table A.6 Formal and informal workers by sector of activity (%)    49 Table A.7 Formal and informal workers by firm size (%)    49 Table A.8 Sectoral distribution of formal and informal workers (number and share)    50 Table A.9 Formal and informal workers by contract type and full-/part-time    50 Table A.10 Monthly earnings in real terms for formal and informal workers    51 Table A.11 Average male wages as % of female wages    51 Table A.12 Wage gap regression results    52 Table A.13 Quantile regression results: 95% confidence interval of coefficient on informal dummy    54 Informality, Job Quality, and Welfare in Sri Lanka 7 Acknowledgments This report was prepared by a team led by Yeon Soo Kim (Economist, World Bank) and comprising Cristina Chiarella (Consultant, World Bank) and Anna Luisa Paffhausen (Consultant, World Bank). A background paper on the scope and application of employment protection in Sri Lanka was prepared by Shyamali Ranaraja (Attorney-at-law of the Supreme Court of Sri Lanka and Practitioner in Employment and Labor Law) and served as a primary input for chapter 3. Idah Pswarayi-Riddihough (Country Director, Sri Lanka, Maldives and Nepal) and Benu Bidani (Practice Manager, Poverty and Equity Global Practice) provided overall guidance. The team is grateful to the peer reviewers, Samuel Freije-Rodriguez (Lead Economist, Poverty and Equity Global Practice) and Nisha Arunatilake (Director of Research, Institute of Policy Studies of Sri Lanka), for their comments. Anne Himmelfarb edited the report. Abbreviations DCS Department of Census and Statistics DoL Department of Labour EPF Employees’ Provident Fund ETF Employees’ Trust Fund ILO International Labour Organization LFS Labour Force Survey MSE micro and small enterprises OECD Organisation for Economic Co-operation and Development SOEA Shop and Office Employees Act TEWA Termination of Employment of Workmen Act WBO Wages Boards Ordinance Informality, Job Quality, and Welfare in Sri Lanka 8 Executive Summary Informal employment remains a salient and persistent feature of the Sri Lanka labor market, with around 70 per- cent of the work force informally employed. There are generally three reasons to be concerned about high informal- ity: poverty, productivity and public finance. This report focuses on the poverty and vulnerability aspect of infor- mal employment, by showing that informal jobs are more precarious in nature than formal jobs and are associated with inferior working conditions and lower earnings. The following three key messages emerge from the analysis. Message 1: The quality of informal jobs is much lower than that of formal jobs. Informal workers have more precarious employment arrangements and inferior working conditions. Their low earnings levels elevate the risk of poverty. The Sri Lanka labor force is segmented into three strands: (a) formal public sector workers who enjoy high remu- neration, shorter work hours, many holidays, paid leave as well as other benefits, and overall job security; (b) for- mal private sector workers who make up a middle group, with access to social security and some paid leave but longer work hours and an earnings distribution that is closer to that of informal workers; and (c) informal workers, a group that includes informal employees as well as informal self-employed workers, most of whom are in precar- ious employment arrangements. Few informal workers have a permanent contract or access to benefits, and they tend to work excessively long hours. There exists a persistent earnings gap between formal and informal workers, even when comparing workers of similar characteristics. Moreover, the risk of extremely low pay and thus pov- erty is significantly higher for informal workers. However, it should be noted that the overall earnings distribution of formal private sector workers is much more similar to that of informal workers than that of public sector workers. Message 2: Stringent labor laws, along with the high cost of compliance and complexity of labor regulations, have encouraged informality. But formalization does not necessarily ease other constraints such as access to credit, reducing the incentive to formalize. The Termination of Employment of Workmen Act (TEWA), which regulates dismissal conditions and compensation, has been long criticized for making it difficult and expensive to dismiss employees. In fact, Sri Lanka has the second-highest redundancy cost in the world. Gratuity payments, contributions to the Employees’ Provident Fund, Sri Lanka’s social security scheme, and the Employees’ Trust Fund, and generous paid leave and holidays add to the labor cost borne by the employer. Multiple and overlapping types of coverage for workers owing to different regulations create a com- plex operating environment for firms, making compliance costly. As a result, compliance tends to be incomplete, even for larger firms. Enforcement is generally weak due to manpower shortages in enforcing agencies, and the many lines of enforcement render the existing institutions inefficient and ineffective, while the labor dispute settlement system can lead to lengthy and costly processes. Moreover, there appear to be few benefits to formalization. The prevalence of the informal sector can be explained by a trade-off between firms’ costs for labor and capital: that is, informal firms have higher capital costs and lower labor productivity, but can avoid certain labor costs associated with mandated taxes and benefits. But evidence suggests that while being informal likely precludes firms’ access to formal credit institutions, for- malization itself does not guarantee easier access because of high collateral requirements and other bureaucratic pro- cedures. This means that streamlining the business registration process alone will not suffice to decrease informality. Informality, Job Quality, and Welfare in Sri Lanka Executive Summary 9 Message 3: Rather than reducing informality itself, the focus of reforms should be on increasing productivity and jobs growth. Regulatory reforms that aim to reduce the cost and increase the benefits of formality would further encourage the creation of formal jobs. The following reform components are suggested in line with recommendations from around the world: • Invest in human capital to promote productivity gains and reduce the wage gap. • Increase flexibility in labor markets and quality of jobs by reducing labor taxes and streamlining regula- tions, which would support job creation more broadly. • Build an adequate and effective social protection system. A well-conceived formalization strategy would aim at increasing the benefits and reducing the cost of formaliza- tion through a policy mix that tackles the interrelated causes and consequences of informality together, ultimately leading to the creation of more and better jobs. This is based on evidence from other countries that found that for- malization efforts focusing on easing entry regulations yielded limited success. The imperative for Sri Lanka to implement these reforms is driven, among others, by the rapid aging of the labor force. The employment protection legislation that imposes a prohibitively high cost on the employer for dismissals of employees with long tenure encourages high turnover and early exit from the labor market. The regulatory and incentive scheme needs to be restructured to encourage more working-age individuals to join and remain in the labor market productively. This would help meet the labor demands in a country that still needs to grow signifi- cantly but is undergoing rapid aging. 1. Introduction Informality, Job Quality, and Welfare in Sri Lanka 1. Introduction 11 The informal sector represents an important part of the economy in many countries and plays a major role in pro- duction, employment creation, and income generation. Informality can be described as “the collection of firms, workers, and activities that operates outside of the legal and regulatory frameworks or outside the modern econ- omy” (Loayza 2016). Informality in developing countries around the world accounts for 60 – 70 percent of employ- ment and about 30 percent of gross domestic product (GDP) (Gatti et al. 2014; Loayza 2016). There are broadly three reasons to be concerned about informality: poverty, productivity, and public finance. 1 The first, poverty, relates to the perception that informality is correlated with or a cause of poverty. A few studies also suggest that informal workers have lower subjective well-being than formal workers, as they tend to be less satisfied with their jobs (Kim, Matytsin, and Freije-Rodriguez 2018; Perry et al. 2007). In Sri Lanka, where pub- lic health care is universal and free, the main source of concern is the lack of protection and benefits, in particu- lar social security, afforded by the informal sector. As this report will show, the precarious nature, low quality, and low wages of many informal jobs lead to higher poverty and vulnerability. The second concern, productivity, relates to the resource misallocation that occurs as a result of a firm’s informal status: informal firms have limited access to credit and rely on informal networks for the sourcing and distribution of goods, leading to lower pro- ductivity and growth. Evidence has also shown that informal firms tend to remain suboptimally small as they try to avoid becoming subject to regulations; this could account for a significant share of the output per capita differ- ences between rich and poor countries (La Porta and Shleifer 2014; cited in Loayza 2016). Moreover, credit access by formally registered firms could be constrained by the presence of informal firms in states where the rule of law and control of corruption are weak (Distinguin, Rugemintwari, and Tacneng 2015). Finally, the public finance con- cern relates to the fact that informal firms operate largely outside of the state’s regulation, where tax evasion is the norm. Informality therefore undermines state revenue collection and fiscal sustainability. While informality raises a common set of concerns, it is a complex, multifaceted phenomenon. A fundamental characteristic of underdevelopment of the economy and its institutions, informality is an outcome determined by low levels of human and physical capital. These are inherent to less developed countries and influenced by the state-agent relationship — specifically, by the state’s regulation, monitoring, and provision of public services such as health care and social security (Loayza 2016). 2 In the public debate, informality is often regarded as something “bad” that countries need to reduce. However, informality also has positive aspects: for example, in the absence of better alternatives and of a formal unemployment insurance scheme supported by a well-functioning market for job matches, the informal labor market provides a viable source of livelihood for many poor and vulnerable groups and thus helps reduce unemployment and poverty. Informal jobs may offer a necessary survival strategy in coun- tries that lack social safety nets and have few formal jobs, especially when the job creation capacity of the econ- omy is constrained. From employers’ perspective, informality also has some advantages: though it limits employers’ access to police and judicial protection and to formal credit, it provides for flexible employment or remunera- tion arrangements that cushion adverse impacts on their operations in response to changes in market conditions. In countries with high rates of population growth or continued urbanization, the informal sector tends to absorb most of the expanding labor force in the urban areas. In such settings, indicators such as the unemployment rate and time-related underemployment are not sufficient to assess the quality of employment and to reflect the con- tribution of all workers to the economy; statistics on informal employment are also necessary. This is especially the case for women, many of whom work as unpaid family workers. 1. These are collectively referred to as the “three P’s” in Kanbur (2014). 2. Informal economic activities were first analyzed by an International Labour Organization employment mission in 1972. The term “informal sector” was coined by Keith Hart’s study of economic activities in Ghana (Hart 1973). Informality, Job Quality, and Welfare in Sri Lanka 1. Introduction 12 Informality is broadly understood to be either a result of “exclusion” or “exit.” The exclusion view maintains that workers and firms are excluded from the formal sector because labor markets are segmented and because burden- some tax and regulatory barriers prevent small firms from formalizing and encourage some large firms to remain partially informal, potentially at the expense of growth and efficiency gains. On the other hand, the exit view pre- sumes that informality is an outcome chosen by firms and workers as the optimal level of engagement with the state, especially when the cost of formalization outweighs the benefits. That is, informality is a large-scale opt- ing out of formal institutions by firms and individuals based on their valuation of the net benefits from formality and the state’s service provision and enforcement capability. These two viewpoints are generally best understood as complementary given that contexts differ across countries and that the informal sector itself can be tremen- dously heterogeneous, with some informal workers choosing to be informal and others being involuntarily excluded from the formal sector (Perry et al. 2007). Although informality in Sri Lanka has been a long-standing concern, few studies have sought to understand it. Information to measure and analyze informal employment in Sri Lanka became available with an expanded Labor Force Survey (LFS) in 2006. Gunatilaka (2008) is the first study (and one of very few to date) to analyze the extent and the determinants of informal employment and wages. ILO (2014), while published not long ago, uses only 2006 survey data (which is quite outdated) and focuses on the impact of labor laws on firms’ decision to for- malize. De Mel, McKenzie, and Woodruff (2013) uses a field experiment to investigate firms’ incentives to formal- ize. Overall, the literature on Sri Lanka is very scarce, particularly on informal employment (as distinct from infor- mal firms). Since informal workers represent a significant share of the labor force, the causes and consequences of informality need to be properly reflected in development policies. Currently, the evidence base for proper policy formulation is relatively small, and this report aims to fill some of this gap. This report shows that informal employment in Sri Lanka is persistently high and that the quality of jobs differs significantly for formal and informal workers. Informal jobs account for nearly 70 percent of jobs in Sri Lanka, and this figure has not changed much since 2006, the first year such estimates became possible. Workers who are informally employed can be classified into different groups, depending on factors such as access to social security, employment status, and the firm’s registration status. Once different types of formal and informal workers are defined, the quality of jobs is measured along two dimensions: (a) contracts and working conditions, which capture the nonpecuniary aspects of a job; and (b) the risk of poverty and vulnerability, which specifically examines the risk of extremely low pay and the incidence of working poverty. By these measures, the Sri Lanka labor market appears to be segmented into three strands: (a) formal public sector workers who enjoy high remuneration, shorter work hours, many holidays, paid leave as well as other benefits, and overall job security; (b) formal private sector workers who make up a middle group, with access to social security and some paid leave but longer work hours and an earn- ings distribution that is closer to that of informal workers; and (c) informal workers, a group that includes infor- mal employees as well as informal self-employed workers, most of whom are in precarious employment arrange- ments with little access to any benefits and low remuneration. A key message of this report is that reforms should not focus on formalization itself, but on regulatory reforms that aim to reduce the cost and increase the benefits of formality, as well as on measures that enhance productivity. The rest of this report is organized as follows. Chapter 2 describes measures of informal employment, recent trends, and the demographic and socioeconomic profiles of informal workers using comparable data over time. Chapter 3 analyzes the possible drivers of informality. Chapter 4 examines the relationship between informal employment and job quality. Chapter 5 summarizes the findings and proposes some key components for reform. Appendix A includes a set of detailed tables of labor market indicators for formal and informal workers. 2. Informal employment: Measures and trends Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 14 Defining informal employment Informal aspects of the economy are known to be difficult to measure, and specific definitions are determined by country context. There is no single correct definition of what is meant by “informal,” and different measures can be used depending on the context, purpose, and available data. There is broad agreement that several crite- ria must be considered, including registration, legal incorporation, size, coverage by statistical surveys, borderline of activity, illegal activity, location, and terms of employment. The word “sector” in this context refers to production activities and not to a grouping of institutional units in the System of National Accounts (SNA) sense. Importantly, both production and employment can be characterized as formal or informal. For example, it is possible for for- mal production units to have informal employees. There is also a distinction between a job and a worker, as one worker can hold several jobs. The focus in this report is on informal employment rather than the informal sector. If labor informality is meas- ured only with respect to those working in firms, many people who are self-employed or work in a household enter- prise would likely be missed. The primary reason for focusing on informal employment in this paper is that the information available from household surveys allows for identification of informal employment from the work- ers’ side. A comprehensive analysis of informality would also need to consider firm-side decisions and would thus require the use of firm-level data. However, such an analysis is outside the scope of this report. For the purposes of the analysis undertaken here, a legalistic definition of informal employment is adopted, which considers com- pliance with labor regulation and the right to social protection (e.g., pension, health care) in its definition of infor- mal workers. In comparison, a productivity-based measure categorizes workers by job type, such as unskilled self-employed, salaried in a small firm, or zero-income worker. Hence informal employment refers to “the total number of informal jobs, whether carried out in formal sector enterprises, informal sector enterprises, or house- holds, during a given reference period” (ILO 2016). 3 If the focus was on informal firms, a closer look at the trade-off between taxes and public service provision would be warranted. 4 In Sri Lanka, several informality-related definitions are currently in use. The Central Bank of Sri Lanka classi- fies as informal private sector organizations those organizations that do not contribute to the general Employees’ Provident Fund (EPF) or do not maintain their own provident funds with the approval of the Commissioner General of Labour. The Department of Census and Statistics (DCS) monitors informal sector employment (i.e., employment in informal firms), where the formal sector is defined as firms that are registered, keep formal accounts, or have 10 or more employees. Since the EPF Act requires only firms with more than 15 workers to pay into the EPF, the infor- mal sector definition adopted by DCS is somewhat less stringent than that of the Central Bank of Sri Lanka. DCS introduced its own definition of informal employment starting with the 2017 LFS report, which broadly follows the ILO guidelines that are shown in table 1. 5 The DCS definition includes (a) all unpaid family workers; (b) all employ- ers/own-account workers in the informal sector; (c) all paid employees who do not have a permanent employer; and (d) all paid employees whose employers do not contribute to pensions. The International Labour Organization (ILO) recently operationalized its own guidelines for measuring informal employment, which include suggestions for specific, step-wise procedures (ILO 2018). 3. For a full description of concepts and definitions related to informality, see ILO (2016), chapter 8 (“Employment in the Informal Economy”); see also UN DESA (2009), chapter 25 (“Informal Aspects of the Economy”). 4. See Loayza (2016) for a detailed review of the literature on informality. 5. Table 1 omits members of producers’ cooperatives that are difficult to identify in the Labour Force Survey. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 15 Table 1 ILO conceptual framework of formal and informal employment Jobs by status in employment Own-account workers Employers Contributing family workers Employees Production units by type Informal Formal Informal Formal Informal Informal Formal Formal sector enterprises Informal sector enterprises Households Source: Adapted from Hussmanns 2004. Note: The table provides a framework for categorizing different types of formal and informal workers, by type of production unit and status in employment. The cells shaded in orange do not exist by definition. The cells shaded in green refer to formal jobs. Unshaded cells represent various types of informal jobs. The main source of data for the analysis is the Labour Force Survey for the years 2006 to 2017. 6 Quarterly survey data are collected and combined into an annual survey that is representative at the district level. The surveys use detailed information — on labor force participation, individual and job-related characteristics, wages, and training received — to estimate key labor force statistics in Sri Lanka. The mode of data collection changed from paper-based to computer-assisted personal interviews in 2017. Due to the civil war, the Northern and Eastern Provinces were excluded from the survey in 2006 and 2007. The 2008, 2009, and 2010 surveys included the Eastern Province but not the Northern Province. Complete national coverage was achieved in the 2011 survey and has continued. Because the Northern and Eastern Provinces account for a small share of the population and thus the country’s labor force, the exclusion of these two provinces had a relatively small impact on national-level estimates. Questions related to informal employment were first included in 2006, and more detailed questions were added in 2013. In 2006, the survey asked whether the employer contributed to a pension scheme — a key piece of infor- mation in defining informal employment in Sri Lanka. 7 The more detailed questions added in 2013 asked (among others things) whether the worker had a written contract. In 2013, the survey also changed the response catego- ries for the firm size question, which allowed for a clearer identification of domestic workers. Specifically, the firm size question now clearly asks whether an individual is “working for household,” whereas previously the closest response for domestic workers was “no specific institution.” The revision in the responses likely provides a more accurate trend for domestic workers. Finally, an added question on the job status of a worker’s secondary job has allowed for a more granular measurement of informal employment and jobs. The definition of informal employment adopted in this report follows ILO guidelines and is closely aligned with the DCS definition. Consistent with DCS’s definition, a worker is employed in the formal sector if the firm is registered, maintains a formal account recording system, or has 10 or more employees. 8 All others are considered to be in the informal sector. Formality in job status for own-account workers and employers is determined by the formal or infor- mal nature of the economic unit they own. For employees, the distinction is based on whether the employer contrib- utes to a pension scheme or provident fund on their behalf. For the purpose of this report, workers are classified into 6. The LFS is conducted by the Department of Census and Statistics (DCS); annual surveys can be found on the DCS website at http://www.statistics.gov.lk/page.asp?page=Labour%20Force. 7. The word “pension” throughout this report refers to pension schemes (such as the EPF) or provident funds that disburse a one-time payment upon retirement. 8. In Sri Lanka, business registration at the DS (Divisional Secretariat) level does not in itself make the firm liable for taxes, as there is an annual exemption threshold of SL Rs 300,000 in annual net profits, and in theory taxes are payable regardless of registration status. A sur- vey conducted by ILO (2014) suggests that most micro and small enterprises had at least one registration to facilitate the operation of their busi- nesses, while only a very small percentage of surveyed firms was unregistered. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 16 the following groups: (a) formal public sector employees; (b) formal private sector employees; (c) formal self-employed; (d) informal self-employed; (e) informal employees working in formal firms; (f) informal employees working in infor- mal firms; (g) unpaid family workers; and (h) domestic workers. The self-employed include own-account workers and employers. The heterogeneity of informal workers also means that some discussions are outside the scope of this report. A large share of self-employed are characterized as informal (based on their firm’s status), but different labor regulations apply to them, and the survey does not record the same information for them as for employees. For this reason, the discussion on job quality in chapter 4 is mainly relevant for paid employees. Informal employment as a salient and persistent feature of the Sri Lanka labor market Labor force participation trended slightly downward in the last decade, despite a small improvement among women. In 2017, 54 percent of the adult population ages 15 years or above participated in the labor market — three percentage points less than in 2006. Having fallen continuously from 2006 to 2012, total labor force participation started to increase again after 2012, driven by improvements among females. Male labor force participation has been stagnant since 2009 after declining slightly (table A.1). Nevertheless, the labor market added more than 1.1 million jobs as the size of the working-age population con- tinued to grow. The decline in labor force participation was countered by the influx of labor market entrants. The working-age population ages 15 and above grew by almost 2.6 million over the last decade, which contributed to a net job creation of 1.1 million, an equivalent of 16 percent growth in the number of jobs. The share of public sector jobs saw a distinct increase between 2006 and 2008, after which it remained largely stable (table A.2). The majority of jobs created between 2006 and 2017 were informal. More specifically: of the roughly 1.1 mil- lion jobs created during this period, slightly more than half were informal. The largest contributor was informal wage employment, with about 21 percent of jobs com- ing from an increase in the number of informal employ- Figure 1 Trend in formal and informal employment, 2006–17: ees working in informal firms; another 17 percent came Number of people (thousands) and share of total employment (%) from an increase in the number of informal employ- 6,000 40 ees working in formal firms. Together, the total num- 35 ber of informal employees grew by more than 400,000. 5,000 The number of unpaid family workers, most of whom 30 Percentage of total employment Number of people (thousands) 4,000 work in agriculture, declined by about 10 percent over 25 the same period, in line with the structural transforma- 3,000 20 tion in the economy. 15 2,000 Informal employment has been persistent in the Sri 10 Lanka labor market, with nearly 70 percent of workers 1,000 5 employed in such arrangements. The share of informal jobs in total employment was 69.8 percent in 2006 and 0 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 remained fairly stagnant until it slightly decreased to 68 percent in 2017 (figure 1; see also table A.3). The public Informal employment (R) Unpaid family worker Formal private employees Informal self-employed + domestic sector is almost entirely formal, with only a few work- Formal public employee Informal employee formal firm ers reporting themselves as informally employed. These Formal self-employed Informal employee informal firm workers are captured as informal employees in formal Source: World Bank staff estimation using LFS data for 2006–17. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 17 firms. 9 When only private sector workers are considered, the share of informal employment amounts to nearly 80 percent. Of the 8.2 million Sri Lankans employed in 2017, almost 5.6 million were informal workers, including about 2.5 million self-employed workers (about 31 percent of total employment) and almost 2.4 million informal employees (about 29 percent of total employment). Unpaid family workers, most of them women, constituted al- most 8 percent of the workforce. Domestic workers, defined as those reported to be “working for a household,” ac- counted for a little more than 50,000 workers (or 0.7 percent of total employment). Detailed statistics with break- down by type of worker are presented in table A.2. Under a broader definition of informality that also considers the nature of the secondary job, the share of infor- mal workers increases by a small increment, from 68 percent to 69.8 percent. The number of workers who have a secondary job has steadily declined between 2006 and 2017, with the number in 2017 standing at around 692,000. This means that about 8.4 percent of employed have another job on the side (table A.4). Almost all of these second- ary jobs are informal. However, most workers who have an informal second job already have an informal main job, which explains the small increase in the share of overall informal employment even when secondary jobs are tak- en into account (table A.5). The majority are self-employed (60.4 percent in 2017) and about 18 percent work as un- paid family workers in their secondary job. While having a written contract and being entitled to formal pensions are typically closely corelated, an im- portant divergence is found among estate sector workers. Among paid employees in general, almost 90 per- cent have both pensions (including EPF) and a written contract, or have neither (figure 2). The anomaly comes Figure 2 Workers with pension entitlements and contracts from the roughly 282,000 workers who have pensions but no written contract. This group mainly consists of estate sector workers, who have historically had strong collective bargaining power and access to formal No pension, no contract Pension, no contract pensions. As generations of estate workers have spent No pension, but contract their lifetime on plantations, signing formal contracts Pension and contract may not have been common practice. 10 A small num- ber of employees have a contract but no pension, likely because they work in a small firm that is not obligated to enroll their employees in a formal pension scheme. Source: World Bank staff calculation using LFS data for 2017. The latest estimates show that domestic workers make up a very small part of the labor force. In this report, domes- tic workers are defined as those who work for a household; use of this definition became possible only in 2013, when the responses to the question on firm size (i.e., total number of employees at the respective place of work) changed. Prior to 2013, the options were “less than 5,” “5 to 9,” “10 to 15,” “16 to 49,” “50 to 99,” “100 or more,” “No specific insti- tution,” and “No paid employees/regular employees.” As suggested above, these responses made it difficult to distin- guish domestic workers from own-account workers. Replacing “no specific institution” with “working for household” allows for a clearer identification of domestic workers, who include paid employees and own-account workers. While it is difficult to assess a long-term trend in the number of domestic workers because of the change in the firm size question, the latest figure for 2017 indicates that there are slightly over 50,000 domestic workers in the labor market. 9. This explains the difference between this report and the official Labour Force Survey report in the number of public sector workers. The number of formal public sector workers in the latter report can be replicated with the assumption that all those working for the public sector are formal, regardless of whether they are entitled to pensions or not. 10. This distinction is similar to the categorization in Kanbur (2009), in which the introduction of regulation creates four categories of enterprises: (a) those that are covered by regulation and comply; (b) those that are covered by regulation but do not comply; (c) those that adjust out of the cover- age of the regulation (e.g., firms that are smaller than they otherwise would be); and (d) those that are outside of the coverage of the regulation. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 18 Despite their small number, domestic workers require attention because lack of protection and inferior working conditions render them especially vulnerable. Currently, domestic workers comprise only a small fraction of the la- bor force, but an overwhelming share of them are women, and they continue to be outside the scope of most labor law protection and the social security framework. Domestic workers are rarely viewed as employees with the same rights and privileges as other employees. The Wages Boards Ordinance (WBO) regulates fundamental working condi- tions for particular trades, yet no wages board has been set up for the regulation of domestic work in Sri Lanka. Given the ambiguity in the scope of the ordinance, it can be argued that the ordinance does not apply to work performed in private households, and thus that domestic workers are not entitled to any legal protection with regard to the pay- ment of wages. Domestic workers are excluded from social security based on both the Employees’ Provident Fund Act and Employees’ Trust Fund Act; the former applies to employers with 15 or more workers. The Employees’ Trust Fund was intended to serve as an unemployment insurance scheme but is currently used to provide life insurance, housing, and other welfare benefits to its members. Sri Lanka currently has no labor law that regulates working time in the do- mestic work sector. It also has no law that governs the living conditions of domestic workers, meaning that domestic workers are often forced to live in inadequate or unsafe environments. While the Factories Ordinance regulates the working conditions of factory workers, no equivalent protection framework has been established for domestic workers. They are also excluded from any compensation for injuries sustained during the course of work (Verité Research 2015). Informal workers: Older, less educated, and primarily engaged in agriculture and low-end services Informal workers comprise a very heterogeneous group, with stark differences by age cohort, gender, and other dimensions. The previous section categorized workers by various types of formal and informal jobs. This sec- tion explores in further detail the differences and similarities among the different groups of workers. In Sri Lanka, women are less likely than men to be working informally, which sets the country apart from many other devel- oping countries. Compared to formal workers, informal workers are on average older and less educated. Informal employment is common in industries with a high concentration of small firms and with low-value-added activi- ties that tend to be carried out by own-account workers. The types of informal jobs differ significantly by gender, and women are less likely than men to be in informal jobs. Informality rates for women and men are at 64 percent and 70 percent, respectively, as of 2017. This con- trasts with the trend globally, where women tend to be overrepresented in informal employment (Chen and Doane 2008; ILO 2013). The proportionately higher representa- tion of women in formal wage employment (34.2 percent Figure 3 Formal and informal workers by gender of women, compared to 25 percent of men) was helped 100 11.2 by a recent inflow of mostly highly educated women into 90 24.4 9.2 the public sector. In fact, female employment grew most 80 Domestic worker strongly in the formal public sector, with an addition- 70 24.7 9.1 Informal employee informal firm al 130,000 women working in this sector in 2017 com- 60 Informal employee formal firm Percent Informal self-employed pared to 2006. Overall, the distribution of women and 50 17.7 33.8 Unpaid family worker men differs among the different types of informal jobs: 40 Formal self-employed 30 2.6 Formal public about a third of men are in informal self-employment, 16.8 Formal private 20 11.5 and another quarter work as informal employees in in- 10 17.4 formal firms. On the other hand, women are more like- 13.5 0 ly than men to work as public sector employees or un- Female Male paid family workers (figure 3). Source: World Bank staff calculation using LFS data for 2017. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 19 Older workers are significantly more likely than Figure 4 Informal employment by age group younger workers to be informally employed, primarily 100 because they are disproportionately employed in agri- culture. Informal employment has a J-shaped relation- 90 ship with age: the informal employment rate starts out at around 70 percent for the 15-to-24-year-old cohort, 80 then initially falls, but increases again to over 80 percent for workers ages 55 and above. The general pattern has Percent 70 not changed much in the past decade, except that those in the age group 15 – 34 are less likely to be in informal 60 jobs now than a decade ago (figure 4). While one would need longitudinal data to precisely understand whether 50 this is a life-cycle pattern or a cohort-specific pattern, it appears that the significantly higher rates of infor- 40 mal employment among older workers could be due 15 – 24 25 – 34 35 – 44 45 – 54 55 – 64 65 plus Age to differences in the sector and type of informal activ- 2006 2007 2008 2009 2010 2011 ity where older workers are concentrated. Specifically, 2012 2013 2014 2015 2016 2017 older workers (ages 55 and above) are much more likely Source: World Bank staff calculation using LFS data for 2006–17. than younger workers to be employed in the agricul- Note: Year in legend indicates survey year from which age profile is drawn. ture sector and work as own-account workers. The like- lihood of being engaged in own-account work increases significantly with age: among workers ages 55 – 64, 80 per- cent are in informal employment; among those ages 65 and older, the share is 93 percent. On the other hand, almost half (49 percent) of younger workers (ages 15 – 24) work as informal employees in either formal or informal firms. There has also been an increase in formal public sector jobs among those in the 15 – 34 age group, likely as a result of higher educational attainment and extended queueing for those jobs. Informal workers are on average 45 years old, which is higher than the average age of formal workers (40 years). Low levels of education are highly correlated with working informally. Almost 75 percent of informal workers had primary education or less, compared to only 35 percent of formal workers. Less than 10 percent of informal workers had obtained A-level certification or above, compared to about 45 percent of formal workers (figure 5). Put differently, among workers with a bachelor’s degree or above, almost 90 percent held a formal job, whereas most workers with primary education or less were working informally. Public sector workers have the highest con- Figure 5 Educational attainment of formal and informal centration of highly educated workers, with 60 percent workers of them having obtained A-level certification or above. 100 0.9 14.5 8.9 90 Analysis by sector of activity shows that more than 70 80 15.7 percent of workers are informally employed in agricul- 70 29.9 ture, construction, commerce, and transport and com- 60 Bachelor's and above Primary completed munications. The most informal sector by far is the ag- 50 O-level passed Percent 20.8 ricultural sector, where 89 percent of workers are infor- 40 60.6 A-level passed Less than primary mal. The sector is characterized by a large proportion 30 of informal self-employed (48 percent) and unpaid fam- 20 31.1 ily workers (18 percent). Likewise, the commerce sector 10 13.9 and transport and communications sector have com- 3.6 0 paratively high proportions of informal self-employed Formal workers Informal workers workers (39 percent and 44 percent respectively), most Source: World Bank staff calculation using LFS data for 2017. Informality, Job Quality, and Welfare in Sri Lanka 2. Informal employment: Measures and trends 20 of whom are performing own-account work. Informal work is much less common in the public utilities sector and in the community and family-oriented services sector due to the predominance of public sector workers — 47 per- cent and 70 percent are civil servants, respectively. Informal employment is also much less prevalent in the finan- cial and business-oriented services sector (25.5 percent), which also has the highest share of formal private employ- ees (51 percent) (table A.6). Overall, sectors with higher productivity tend to have lower informality rates. This sectoral concentration of informal workers is not unusual, as activities in the relevant sectors are gener- ally low-value-added and are usually performed by low-productivity micro and small enterprises (MSEs). In fact, most informal employment (88 percent) is concentrated in smaller firms (with five or fewer workers). Whereas infor- mal wage workers are more likely to be found in smaller firms, formal firms of all sizes hire workers informally: around 16 percent of workers without pension are employed by firms with 100 or more employees (table A.7). The main sectoral drivers of growth in recent years — i.e., construction, transport, and manufacturing — have a high concentration of informality and a tendency to favor employment growth among men over women. Male employ- ment grew strongly in informal wage work during 2006 – 17, adding 370,000 jobs. This increase is in line with the growth over this period in the construction sector, which predominantly employs men, and on an informal basis. Manufacturing also expanded, contributing to growth and job creation, but many workers are informally employed without access to pensions, even among larger firms. Unlike the public sector, construction, transport, and man- ufacturing are also less conducive to the creation of jobs for women. Three sectors account for almost 60 percent of all informal workers: agriculture (26 percent), manufacturing (19 percent), and commerce (14 percent) (table A.8). Informal employment is high almost everywhere except the Western Province. The North-Central Province has the highest share of informal workers, at around 79 percent, followed by the Uva, Northern, North-Western, Southern, and Eastern Provinces, all of which have an informal employment rate of around 75 percent. The Western Province has the lowest share of informal workers at around 57 percent, which is still considerable but substan- tially lower than in most other provinces (figure 6, left). Analysis by geographic sector shows that the rural sector has the highest share of informal workers and the estate sector the lowest share — likely because most plantation workers are entitled to pension benefits (figure 6, right). Figure 6 Informal employment share by province (left) and geographic sector (right) 78.9% 71.7% 76.1% 76.0% 75.8% 74.3% 73.6% 68.1% 64.9% 57.0% 56.7% 46.3% Urban Rural Estate North-Central Uva Northern North-Western Southern Eastern Sabaragamuva Central Western Source: World Bank staff calculation using LFS data for 2017. 3. What explains the high rate of informality in Sri Lanka? Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 22 The high rate of informality has long been attributed to the stringency of labor laws in Sri Lanka.  1 1 The Termination of Employment of Workmen Act (TEWA) in particular, which makes dismissing employees difficult and expensive, has been at the center of the debate. It has been argued that the TEWA increases informality, as its provisions are difficult to enforce in the informal sector, and that it reduces job creation in firms with more than 15 workers, which may wish to avoid high visibility and monitoring by the authorities and unions (Abidoye, Orazem, and Vodopivec 2009; Heltberg and Vodopivec 2009; Rama 1999). Stringent labor regulations are associated with high rates of informality across countries (World Bank 2017). This chapter argues that a more nuanced understanding is needed regarding what motivates firms to operate and hire labor informally. The chapter begins with an overview of key labor laws and their application in practice to describe the extent of direct and indirect costs associated with overly stringent labor laws. Two key labor regu- lations that are particularly pertinent to female employment are also described, along with how they are applied in practice. Even though these regulations have become less stringent in practice over time, they present a chal- lenge to the hiring of women in general and create additional incentives to hire female workers informally. As the regulations themselves are described in length elsewhere, the focus is on understanding how the stringency of and high cost resulting from compliance with these regulations induce many firms to operate informally (or in some cases, operate formally but employ some workers informally). As described in more detail below, the high cost of compliance induces various evasions and avoidance behaviors. The impact of labor regulations depends not only on the rigidity of the regulations, but also on the level of mon- itoring and enforcement. While enforcement is overall low, litigation can be lengthy and very costly, providing incentives to firms to avoid it whenever possible. The prevalence of informality reflects a fundamental asymme- try between high labor costs (for formal firms) and high capital costs (for informal firms). In the Sri Lanka context, it appears that formal status itself would contribute little to easing access to credit — t hat is, the benefits to for- malization are too low for most smaller firms. From the workers’ perspective, an informal job becomes a viable option when most jobs offered are informal and when the alternative is unemployment without any income. Slow changes to a rigid institutional environment, one formed decades ago when circumstances were drastically differ- ent, have contributed to persistently high rates of informality in Sri Lanka. Key labor regulations: Application in practice and implications for labor cost The Termination of Employment of Workmen Act of 1971 regulates dismissal conditions, including compen- sation, and is considered highly restrictive.  1 2 TEWA applies to employers with 15 or more workers that seek to ter- minate the services of an employee for nondisciplinary reasons. It was enacted during a period of inward-looking economic policies in the 1970s, which was characterized by an import-substitution industrialization policy, strin- gent exchange controls, a wide range of price controls on many commodities, and the nationalization of firms. The statute aimed to mitigate the impact of these policies on employment, as many firms had been forced to reduce the number of workers. Even though the intention then was to prevent mass retrenchment, the statute itself does not call for such a limitation and also applies to individual terminations. This means in effect that the employer is required to seek either the written consent of the worker or the written authorization of the Commissioner 11. The negative association was first proposed by Lazear (1990). 12. This section draws heavily on Ranaraja (2019). Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 23 General of Labour for each individual layoff. “Prior consent” is not defined in the law, but decisions in previous court cases have excluded the following from the application of TEWA: voluntary resignation, written consent of the employee at the time of entering into contract for the termination upon a specific event (such as retirement upon reaching the age of 55 or the expiration of a fixed-term contract), and retirement under the terms of a collective agreement. The result is an arbitrary and lengthy separation process. While most employment contracts contain a clause stating that the employer may terminate the services of the worker for any reason with notice of a par- ticular period (generally stated as one month) or with payment in lieu of such notice, this clause is inoperative as it is not considered “prior written consent” within the meaning of TEWA. The amount of the TEWA-regulated compensation payable upon separation presents a heavy financial burden to em- ployers. The compensation is mainly a function of the worker’s tenure (table 2). The basis for the current formula, which dates from 2005 after several attempts at reform, reduces the arbitrariness of the compensation level. Under this sys- tem, a worker with two years of service would be entitled to a separation payment equivalent to five months of sal- Table 2 Severance pay by tenure ary, while a worker with 25 years of service would be enti- Maximum payment tled to 43.5 months of salary (figure 7). How does this com- No. of years of No. of months’ salary (cumulative) pare with other countries? The World Bank’s (2017) Doing service at date paid as compensation (no. of months’ compen- of termination for each year of service sation) Business report, which presents internationally compara- ble measures of redundancy costs, shows that Sri Lanka 1 – 5 2.5 12.5 ranks second from the top, after Sierra Leone, making 6 – 14 2.0 30.5 it one of the costliest and most restrictive countries in the 15 – 19 1.5 38.0 world to dismiss a worker (figure 8). The redundancy pay- 20 – 24 1.0 43.0 ments do not take into consideration mandatory contri- 25 – 35 0.5 48.0 butions made by employers as EPF and Employees’ Trust Fund (ETF) benefits. Source: Gazette Extraordinary no. 1384/07, March 15, 2005. Figure 7 Cumulative compensation payable as severance and gratuity upon dismissal (no. of months’ salary by length of tenure in years) 70 63.0 62.5 61.5 60.5 59.5 58.5 57.5 60 56.5 55.5 54.5 53.5 52.5 51.0 49.5 48.0 50 46.5 45.0 43.0 41.0 39.0 37.0 40 35.0 Percent 32.5 30.0 27.5 30 25.0 22.5 20.0 17.5 20 15.0 12.5 10.0 10 7.5 5.0 2.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Employment tenure (years) Severance pay Gratuity Source: Based on Ranaraja 2019. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 24 The difficulties and costs associated with dismissals un- Figure 8 Comparison of redundancy costs across countries der TEWA discourage formal employment. The biggest (weeks of salary) impediment imposed by TEWA is that an employer is una- 250 234 ble to dismiss an employee, except for serious disciplinary 200 infractions, without prior written consent by the employ- Weeks of salary 144 ee or prior written approval by the Labour Commissioner. 150 124 109 109 In practice, this provision makes it very difficult and very 98 96 87 100 expensive to restructure a firm’s workforce in response 63 to changes in market conditions. This provision has been 50 33 29 12 criticized for a long time for discouraging private sector 0 Sri Lanka Thailand Bangladesh Pakistan Nepal Vietnam Malaysia Afghanistan India Bhutan Maldives Singapore investments. The Investment Policy Review of Sri Lanka, prepared by the United Nations Conference on Trade and Development (the trade, development, and invest- ment arm of the United Nations Secretariat), called this Source: World Bank 2017. Note: Redundancy costs measure the cost of advance notice requirements and severance provision “an inappropriate imposition on private inves- payments due when terminating a redundant worker, in weeks of salary. tors to make their staff retrenchment decisions depend- ent on government approval” (UNCTAD 2004). Following this review, the TEWA was subsequently amended to remove the Labour Commissioner’s discretion to decide compensation on a case-by-case basis; but other provisions remain in place, including the requirement to seek approval for the nondisciplinary termination of even one employee. Gratuity payments were introduced as an interim form of social security but present yet another cost to employ- ers. The Payment of Gratuity Act of 1983 was enacted to provide a minimum level of income security to workers ter- minated from employment. The statute provides for half a month’s salary per year of service, subject to a minimum qualifying period of five years of continuous service. For example, an employee who ceases to be employed after 10 years would receive five months of salary as gratuity. It was originally designed decades ago to help reduce high turnover at enterprises in Export Processing Zones (EPZs), where very few employees completed a qualifying period of five years of service. However, the act is applicable to all private sector employers and was seen as an interim measure in the effort to create a social security system. Gratuity is payable even when employees are terminated due to misconduct, subject only to possible deductions in the event losses or damages led to the termination. Gratuity payments are compulsory for employers with 15 or more workers, but the threshold is not strictly binding. This is because any worker who does not get paid gratuity, including one from a smaller firm, can apply to the Labour Tribunal, which will rule on whether any gratuity is due and the amount of such gratuity. Therefore the 15-employee threshold is not strictly binding as it is for TEWA, and a worker may be deemed entitled to higher or lesser gratuity amounts as decided by the Labour Tribunal. Some companies opt to pay higher gratuity than required by law as a means of rewarding long-serving employees and reducing turnover. The cumulative compensation payable upon dismissal under TEWA and the Payment of Gratuity Act increases with tenure and presents a significant cost to the employer (figure 7). The Employees’ Provident Fund Act of 1958 established a contributory scheme for all private sector employees, but the law’s lack of clarity invites disputes on what constitutes earnings. The system is maintained through contribu- tions by employer and employee based on the employee’s monthly earnings. The minimum contribution is 12 percent of earnings by the employer and 8 percent by the employee. While the law prescribes what is included under earn- ings,  1 3 there is considerable uncertainty about which components of the emoluments should be included in the cal- 13. Earnings include (a) wages, salary, or fees; (b) cost of living allowance, special living allowance, and other similar allowances; (c) payment in respect of holidays; (d) the cash value of any cooked or uncooked food provided by the employer to employees in prescribed employments and any such commodity used in the preparation or composition of any food as is so provided; (e) meal allowance; and (f) such other forms of remuneration as may be prescribed. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 25 culation of the base wage for EPF contributions. For instance, it is not clear whether paid commissions or incentives should be added to the base wage for calculating EPF contributions. Other disputed areas include payments in respect of travel, insurance, and location as well as contingency payments; even when these are part of regular remuneration, some employers do not label them as such in an effort to evade or avoid higher EPF contributions. While all workers are required to be members of the EPF, own-account workers and entrepreneurs are not cov- ered, as they are not considered “workmen” under the act, leaving a large gap in coverage. The EPF scheme differs from the noncontributory pension scheme applicable to public servants, where payments are made upon retirement based on the last-drawn salary in employment. With the EPF, the entirety of accrued contributions can be with- drawn when workers cease to be employed — starting at age 55 for men and age 50 for women. The EPF Act allowed for privately managed pension funds provided that the contribution rates and benefits were not less beneficial than the EPF. 14 However, the Commissioner General of Labour has ceased to approve the establishment of such private provident funds, and all contributions are now mandated to be made to the EPF. The Employees’ Trust Fund was established to help provide unemployment benefits in the absence of social secu- rity and an unemployment insurance scheme, but it does not serve its intended purpose and presents an undue burden to employers. The ETF Act of 1980 was established to promote employee ownership and welfare through participation in financing and investment and to provide noncontributory benefits to employees on retirement. The act requires employers to remit 3 percent of the employee’s gross monthly earnings to the fund. The fund has grown rapidly in the past four decades and achieved a total size of about SL Rs 306 billion by end-2018. The ETF has more members than the EPF, as employers who operate a private provident fund are also required to contrib- ute to the ETF. The amount lying to the credit of an employee can be withdrawn once every five years if the person ceases to be employed (in order to help the person cope with a period of unemployment). However, the fund has never served this specific purpose and is currently used to provide members with a range of welfare benefits, such as life insurance, housing, and selected health benefits. Under the Employment of Women, Young Persons and Children Act of 1956, an employer seeking to hire a woman for work at night must seek formal permission from the Commissioner General of Labour. 15 Although the original law was established such that written consent and approval was required for every instance of night work, the Department of Labour (DoL) has over time reached a more practical arrangement, where requests for approval are submitted by the employer every six months. When the law was established, there may have been concerns about exploitation of women and children, but such protective measures are less relevant today and effectively discourage the employment of women. Female workers who are regulated by the Maternity Benefits Ordinance (MBO) and the Shop and Office Employees Act (SOEA) are entitled to maternity benefits, but the costs are entirely borne by the employer, discouraging the em- ployment of female workers, whether formal or informal. In Sri Lanka a female worker in the private sector is entitled to paid leave of 84 working days after childbirth, of which up to 14 days can be utilized before confinement. The ordi- nance covers all female workers who are employed in any trade 16 and receive wages, and the maternity provision in the SOEA applies to all female workers employed for a wage in a shop or an office. This excludes female domestic workers 14. The benefit of a private fund to the employee was that all accrued contributions could be withdrawn at the time the employee ceased to be employed, rather than at a specified age. For employers, the remission of contributions and the ability to manage the fund was perceived as an advantage. 15. Written consent is required in order to establish that the female worker is not employed on overtime against her will, as provided in section 2A of the act. 16. “Trade” refers to any industry, business, undertaking, occupation, profession, or calling other than an institution providing training to dis- advantaged persons. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 26 and self-employed or unpaid family workers from maternity benefits. Both statutes contain restrictions on working conditions during pregnancy and on termination of employment of a pregnant employee. Payment during maternity leave is to be made by the employer, with neither the worker nor the state making any contributions. This is in breach of the Maternity Protection Convention of the International Labour Organization (C183) ratified by Sri Lanka, which provides in Article 6(8) that “in order to protect the situation of women in the labour market, benefits in respect of the [maternity] leave . . . shall be provided through compulsory social insurance or public funds . . . An employer shall not be individually liable for the direct cost of any such monetary benefit to a woman employed by him or her.” 17 High cost of compliance and complexity of labor regulations as motivating factors of informality Most formal workers are entitled to generous paid leave and holidays, adding to the cost of labor borne by the employer (table 3). There are also requirements to provide an even higher rate of overtime pay for work on holidays, such as 1.5 times the hourly pay for the first eight hours of work on a public holiday and double the hourly pay for work thereafter, and an extra half-day’s pay for work performed on a Poya day 18 regardless of hours worked. Casual and sick leave tend to be taken as an entitlement and not on a needs basis. Notably, the multiple and overlapping types of coverage Table 3 Paid leave and holiday regulations for workers create a complex operating environment for Number of days Number of days Number of firms, making compliance difficult and even more cost- Types of holiday for employees for trades days for ly — that is, employers incur costs in addition to the di- & leave covered by SOEA covered by WBO public sector rect costs described in the previous section. In some cas- Holidays es, employers use overlapping coverage to deny benefits Weekly 78 a 52b 104c to workers when application and coverage are uncertain. For example, working hours are regulated under several Statutory 8 8 13 different laws, and there are different provisions applica- Poya 12 12 12 ble to payments for overtime and holidays. This creates dif- Total holidays 98 72 129 ficulties and confusion for employers if they have differ- ent categories of workers in one workplace: some workers Leave could be covered by the SOEA by virtue of working in the Annual 14 14 24 office of a factory, while the rest of the factory workers are Sick/casual 7 0 21 covered by the provisions of the WBO. Table 3 and Table 4 Total leave 21 14 45 show how employees governed by different regulations end up with different entitlements in terms of leave, holidays, Total nonworking 119 86 174 days per year and working hours. Employees covered by the SOEA have 246 working days a year, which is significantly lower than Total working 246 279 191 days per year the 279 days of those working in trades covered by a wag- es board. Public sector employees have the fewest working Source: Ranaraja 2019. days, at 191 days a year. The number of maximum working a. Under the SOEA , a worker who is employed for 28 hours a week is entitled to 1.5 days of paid weekly holidays, to be taken on any day of the week as required by the employer. hours, restrictions for overtime for males as well as females, b. The WBO provides for one unpaid weekly holiday, which is usually Sunday. and compensation for overtime also vary substantially. c. Public servants in Sri Lanka are entitled to two paid weekly holidays. 17. The full text is at International Labour Organization, “C183 — M aternity Protection Convention, 2000 (No. 183),” https://www.ilo.org/dyn/ normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:C183. 18. A Poya day is the monthly holiday observed to mark the full moon. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 27 Table 4 Working hours, overtime, and payments   SOEA WBOa Maximum 45 hours over five days and one short working day 48 hours over six days working hours Daily working hours Not more than eight hours, with a one-hour meal interval. Not more than eight hours, with a one-hour interval. Eight hours Eight hours can be spread over a period of 12 hours, can be spread over 12 hours if the interval is more than one hour. provided there is one break of not more than three Note that different wage boards may have different working hours (excluding the meal interval); this applies only to hours subject to minimums/maximums. employees in residential hotels, clubs, theatres, places Factories Ordinance: Not more than 4.5 hours without a break of of entertainment, or any shop or office of an airline in an at least 0.5 hours. airport. Overtime Maximum number of overtime hours per week is 12. No restriction. for males Overtime No overtime work is possible except on short working No restriction. for females days of up to 8 hours. But the Factories Ordinance provides that for females over 18, • Total hours worked (normal + overtime) cannot exceed 12 hours in a day and 60 hours in a week • Maximum overtime a month is 60 hours, except for pregnant women, who are exempt from working overtime for three months after childbirth, and nursing mothers, who are exempt for a year — unless they consent to waive the exemption Night work for Women cannot be employed before 6 a.m. or after 8 p.m., No restriction in WBO. females whether as normal hours or overtime, except as follows: But under the Factories Ordinance, (a) between 8 p.m. and 10 p.m. in hotels/restaurants; • Female workers cannot be required to work at night without or (b) before 6 a.m. or after 10 p.m. in airline offices at their consent an airport as a ground hostess, or in a residential hotel • Commissioner General of Labour must give permission in as a receptionist/ladies’ cloakroom attendant/ladies’ respect of every worker for a specific period lavatory/linen room attendant. • If female has worked at any time between 6 a.m. to 6 p.m., she   cannot be employed after 10 p.m. • Night work must be paid at 1.5 times the normal rate • No more than 10 days of night work are allowed each month • Female wardens, women’s restrooms, and refreshments must be provided Holiday pay and days in lieu of time worked Weekly holiday or 1.5 times the normal hourly wage for eight hours or Rate of overtime specified by each wages board, but if not speci- half-holiday normal working day fied, not less than 1.25 times normal pay with a holiday in lieu OR 2 times the hourly wage for any additional work AND 1.5 times with no holiday in lieu Paid holiday within next six days Statutory An additional day’s pay irrespective of time worked OR Each wages board differs, but usually eight public holidays are holiday Day off before the end of the year provided Full-moon Poya If monthly rated, additional half day’s pay irrespective of If monthly rated, additional half day’s pay irrespective of time time worked, but NO DAY OFF worked, but NO DAY OFF If paid daily, 1.5 times the daily rate If paid daily, 1.5 times the daily rate Source: Ranaraja 2019. a. Some key differences between the Wages Board Ordinance and the Factories Ordinance are also described in this column. The high cost of adhering to such labor laws incentivizes firms to avoid compliance with them. ILO (2014) char- acterizes the types of avoidance behavior as follows: (a) not maintaining a record of employment, such as let- ters of appointment; (b) concealing the employment relationship; (c) employing workers in nonformal, tempo- rary arrangements; (d) ensuring that the firm has less than 15 workers; and (e) maintaining a low basic wage rate. Some of the regulatory avoidance behavior is costly, as it can be a drag on productivity. For example, firms could be induced to remain arbitrarily under the 15-employee threshold to avoid a large hike in labor costs; or instead of continuing to grow, the firm might establish several smaller firms in order to remain below the TEWA threshold. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 28 The evidence on such behaviors for Sri Lanka is mixed (Abidoye, Orazem, and Vodopivec 2009; ILO 2014). In addi- tion to avoidance behavior, various forms of evasion are also practiced. ILO (2014) reports that of the MSEs sur- veyed that were above the threshold level, 84 percent reported that they had never made gratuity payment to their workers. The reasons for failing to do so included “Workers have not demanded it” (64 percent) and “It is legally required, but not enforced” (32 percent). Even when firms are large enough to be visible and operate with some degree of formality, compliance has been incomplete. Practices include registering only some employees to avoid coverage by labor laws, resorting to casual employment arrangements, periodically discontinuing employment relationships to avoid compliance, and find- ing ways of avoiding involvement with authorities, including through informal payments to officials (ILO 2014). Such practices are consistent with the finding that a notable number of informal workers are employed by firms that are formally registered. A lack of awareness also contributes to low compliance. For example, in a survey of MSEs, ILO (2014) found that a considerable proportion of 64 percent claimed that they were not aware of value added tax (VAT) regulations, with little difference across sectors. Only 18 percent of MSEs were aware that there were penalties for noncompliance with the retrenchment regulation, and a similar percentage knew what the penalties were. Moreover, 64 percent of all sole proprietorships and unregistered businesses reported that they were not aware of income tax regula- tions requiring them to pay income tax if their earnings exceeded SL Rs 500,000. About 36 percent of respondents in the same survey were not aware that they had to pay EPF/ETF. Enforcement is weak due to manpower shortages in enforcing agencies, and the many lines of enforcement render the existing institutions inefficient and ineffective. Stringent laws coupled with weak or discretionary enforcement lead to a gap between regulations and their practical applications, and in developing countries in particular very strict rules may have little effect if they are circumvented. The level of enforcement, which defines the gap between the rules and practice, is determined by several factors, including the institutional capacity of law enforcement agencies and the demand for enforcement, including litigation practices. In Sri Lanka, there are several agencies that enforce com- pliance with labor laws, but the key institution is the Department of Labour. Under the direction and guidance of the Commissioner General of Labour, the DoL has responsibility for administering most employment protection legisla- tion. It has only limited capacity for inspections, as there are only about 600 officials to carry out inspections and dis- pute settlement functions nationally. Moreover, enforcement is problematic, and inspections are not universal for sev- eral practical reasons, which include the lack of records, the transient nature of employment in many cases, evasion by the enterprise, or the location of the enterprise in a rural or remote area. Besides dispute settlements by the DoL, enforcement is also carried out through Labour Tribunal proceedings, where a worker can make a written application concerning the termination of services and any gratuity that is due; arbitration/industrial courts; appellate jurisdic- tion of the High Court, Court of Appeal, and Supreme Court (when a decision or order of the Commissioner General of Labour, Labor Tribunal, or Arbitrator/Industrial Court is challenged); and collective agreements (Ranaraja 2019). Any employment-related dispute can be brought before the DoL, and the labor dispute settlement system relies on litigation and adversarial resolution, leading to lengthy and costly processes. While the Industrial Relations Division under the Commissioner of Labour is mainly responsible for dispute resolution, the broad definition of an “industrial dispute” allows for any employment-related dispute to be brought before the Department of Labour. Those proceedings have no time limits or set procedures, so that their duration and cost are uncertain for both employers and employees. The labor dispute settlement system is focused on litigation rather than mediation or pre- vention of disputes. Litigation costs are high, as lengthy processes incur costs at several points. The cost of litiga- tion under TEWA includes the cost of retaining legal counsel on both sides; and when the application for termina- tion is made by the employer, the worker continues to receive wages while the inquiry is pending. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 29 Labor Tribunal proceedings can also be initiated with respect to dismissals and any benefits that are due, and these, too, can be costly — because of lengthy procedures as well as arbitrary rulings on benefits. An employee or a trade union can make a written application to the Labour Tribunal concerning the employee’s termination, specifically whether any gratuity (or other benefit) “as may appear just and equitable” is due. The “just and equi- table” provision is considered problematic because it leaves room for nontransparent and arbitrary decisions. The only guidelines on this provision are those established in decisions of higher courts when decisions of Labour Tribunals have been appealed. In effect, a contract of employment entered into between employer and employee, with the consent of both, can be ignored by the Labour Tribunal in deciding a complaint by the employee about his termination under that contract. At the time the Industrial Disputes Act was enacted in 1950, the goal was to ensure that the employer could not benefit from the disadvantaged position of an employee, whose bargaining power was very limited by the social and economic conditions that prevailed then. The circumstances are very dif- ferent today, when employees have a far more empowered bargaining position and can demand and obtain benefi- cial terms. An additional factor that increases the cost and uncertainty of Labour Tribunal proceedings is the lack of any procedure to agree on witnesses or evidence prior to a hearing or arbitration. While all labor regulations technically apply to all workers, in practice the protection generally does not extend into the informal sector, leaving an important gap in coverage. Regulations apply to all workers, employers, and workplaces as defined in each law, regardless of whether the economic activity takes places in the formal or infor- mal sector, and regardless of the enterprise’s size. The two exceptions are the TEWA and the gratuity, which are applicable only to enterprises that employ more than 15 people — though as described above, the threshold is not strictly binding for gratuity. In practice, the labor agencies are largely focused on workers and enterprises engaged in the formal economy, and many practical difficulties affect the extension of labor protection to workers engaged under nontraditional arrangements and in the informal economy. Limited benefits of formalization in Sri Lanka The benefits of being formal include easier credit access, participation in government contracts and programs, fewer fines, and less harassment by law enforcement officials, while the costs include the one-time registration cost required for formalization and the ongoing taxes, paperwork, and other costs associated with compliance (Bruhn and McKenzie 2014). The imposition of stringent labor regulations already poses a significant cost to formality, and there appear to be few additional benefits to formalization. The prevalence of the informal sector can be explained by a funda- mental trade-off between firms’ costs for labor and capital. 19 Informal firms have higher capital costs (as they are unable to enter contractual agreements due to their informal status) and lower labor productivity, but they can avoid certain labor costs associated with mandated taxes and benefits, such as labor taxes, minimum wages, firing con- straints, and social security. Formal firms face higher labor costs but enjoy lower capital costs (Loayza 2016). In other words, informality is a rational choice by firms that perceive the benefits of formalization to be low relative to the additional costs involved (Maloney 2004). This means that streamlining the business registration process alone 19. Loayza (2016) notes that the informality literature can be divided into two strands: one that takes a public finance perspective, with an emphasis on the trade-off between taxes and public services; and another that takes a labor perspective, focusing on the trade-off between labor and capital costs — t hat is, the perspective that informal firms avoid mandated labor costs (including minimum wages, labor taxes, fir- ing constraints) at the expense of higher capital costs (due to the firms’ inability to enter contractual agreements). For understanding informal labor, the second approach is more suitable. Informality, Job Quality, and Welfare in Sri Lanka 3. What explains the high rate of informality in Sri Lanka? 30 will not suffice to decrease informality. Reforms also need to focus on increasing the benefits of formality. In fact, there is some evidence that formalization alone will not substantially increase access to credit: for example, de Mel, McKenzie, and Woodruff (2011) find that many high-return firms are constrained by an inability to find a personal guarantor and by other bureaucratic procedures. Credit growth is constrained by the large needs of the public sec- tor, an inefficient banking sector, and an underdeveloped nonbanking financial sector. Accessing finance is particu- larly difficult for smaller firms that have to post collateral equivalent to 200 percent or more of the loan amount; this is an insurmountable obstacle for many firms (World Bank 2015). These findings imply that while being infor- mal likely precludes firms’ access to formal credit institutions, formalization itself does not guarantee easier access. Support for the conclusion that formalization alone may not lower the barriers to accessing formal credit comes from the 2011 Enterprise Survey on Sri Lanka (World Bank 2012). The survey ranks access to finance as one of the most important challenges to conducting business, second only after “practices of the informal sector.” Commercial banks still rely on traditional lending processes with different approval stages and requiring substantial collateral, rather than applying cash flow–based lending processes. The lengthy approval process and the absence of collat- eral and reliable book records of many SMEs make lending either unattractive or inaccessible to many creditwor- thy firms (Sirimanna 2011). According to de Mel, McKenzie, and Woodruff (2013), firms that did not respond with formalization even when offered increased financial benefits cited land tenancy as a major issue. Relatedly, most banks prefer land as collateral, but more than 80 percent of land belongs to the government, and the land owned by SMEs does not have the clear deeds needed to bank them as collateral (Dasanayaka 2011). Informal employment arrangements may also be preferable from the worker’s perspective. Social insurance benefits that typically come with a formal job include health insurance, pensions, and unemployment benefits. But in Sri Lanka, health insurance coverage is universal and does not depend on employment status, so this cov- erage is unlikely to be an incentive. There is also no formal unemployment insurance scheme in Sri Lanka, which makes unemployment untenable for many; the Workmen’s Compensation Ordinance applies only to public serv- ants (excluding armed forces and the police). 20 Instead, informal employment increases current net take-home pay, which is preferred by many workers over any benefits. At the same time, while many informal jobs can be consid- ered subsistence jobs, given the lack of job security and the low level of compensation, they still represent a viable alternative to unemployment. This is especially true during recessions or in contexts where there are few good jobs and otherwise no formal income support scheme, and where violations of minimum wage regulations are wide- spread. From the firm’s perspective, informal labor provides flexibility and reduced labor costs. In sum, stringent labor regulations make compliance difficult and costly, and neither firms nor workers per- ceive significant benefits to formalization. The many holidays and overlapping coverage by multiple regulations contribute to higher labor costs (including for compliance) in Sri Lanka. The procedures for dismissal are com- plex, and the associated costs are among the highest in the world. While enforcement of regulations is relatively limited, the potential cost of litigation is onerous and imposes a high risk to the firm. As a result, firms engage in a variety of avoidance behaviors, ranging from not maintaining employment records and concealing employ- ment relationships to maintaining a low basic wage to minimize social security payments. Moreover, the benefits of formalization to employers are likely not substantial, given that constraints to growth are weakly related to the informal status of the firm. Constraints to credit access, a major challenge to firm growth, are due less to infor- mality than to firms’ inability to put up sizable collateral or show proof of land tenancy. The role and size of the primary sector (mainly agriculture and fishery) is also a contributing factor, though an in-depth discussion of this topic is beyond the scope of this report. 20. It requires the payment of compensation in the event of an accident, injury, illness, or death suffered in the course of work. 4. Informal employment, job quality, and poverty Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 32 Informality and job quality As seen in chapter 2, the number of jobs has continued to increase, though the pace of job creation has not been fast enough to absorb proportionately more of the new labor market entrants. Having more people work in pro- ductive jobs is critical in a rapidly aging society — one whose demographic window is closing just when the country has transitioned to upper-middle-income status according to World Bank classifications. Considering that overall unemployment is low, it may seem that the urgency for job creation is low. But this is not true in a context where levels of informality are high and where most workers cannot easily afford open unemployment. Understanding the nature of informal employment is key to assessing the quality of employment and to understanding the full extent of the jobs challenge in the country. The key message in this chapter is that Sri Lanka’s economy needs to simultaneously create more jobs and improve the quality of available jobs. The public sector has played an important role in creating jobs over the past decade, but this trend is not sustainable. One problem is that better-educated youth are now queueing for pub- lic sector jobs because of better benefit packages (e.g., pensions), and as further shown in this chapter, public sec- tor workers are also better remunerated; all this has led to biased incentives toward the public sector. A second problem occurs when subsistence-level jobs are preferred over unemployment. In this situation, both the absolute shortage of jobs and a continuum of “insufficiencies” related to job quality need to be considered when it comes to assessing job creation. The remainder of this chapter provides an assessment of the quality of jobs, in particu- lar by comparing formal and informal jobs. The quality of jobs is assessed by looking at contracts and working conditions as well as the risk of poverty and vulnerability A two-pronged approach is adopted to analyze job quality. What constitutes a “good job” can vary, but this approach makes it possible to include common, fundamental indicators of job quality that are measurable in the Sri Lanka household survey data: • Contracts and working conditions. The metrics included here are indicators related to workers’ contrac- tual status, job-related benefits, and underemployment. Measures related to nonpecuniary aspects of job quality fall into this group. • Risk of poverty and vulnerability. This dimension of job quality considers the wage gap between formal and informal workers, the incidence of poverty among informal workers, and the risk of extremely low or inad- equate pay. These measures primarily capture the remunerative aspect of job quality. Informal jobs and contracts and working conditions Very few informal workers have permanent contracts, even when working full-time Job stability is a major attribute of a formal job, but nonstandard contracts prevail among informal workers. To cap- ture a worker’s contractual status in employment, a continuum of contract types is considered: (a) permanent, full- time; (b) permanent, part-time; (c) temporary, full-time; (d) temporary, part-time; (e) casual; and (f) no permanent Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 33 employer. 21 Table 5 shows the distribution of formal and informal workers among these different contract types for 2017. The contrast between formal and informal workers is stark: an overwhelming majority of formal workers hold full-time, permanent jobs (77 percent), with another 14 percent employed in part-time, permanent arrangements. Very few, in fact only 9 percent of formal workers, report having temporary or causal working arrangements. On the oth- er hand, 40 percent of informal workers are employed in full-time, temporary arrangements, with another 12 percent in part-time temporary arrangements. Only 4 percent of informal workers have permanent contracts; 16 percent re- port being in full-time casual employment (table 5). The trend over time shows that this segmentation has become stronger, as most new informal jobs have been accompanied by temporary contracts (table A.9). Table 5 Formal and informal workers by type of contract   PF PP TF TP Ca NPEF NPEP Formal workers Number 1,795,360 317,593 135,207 8,752 61,939     % formal employees 77% 14% 6% 0% 3%     % paid employees 38% 7% 3% 0% 1%     Informal workers Number 83,543 19,270 963,602 278,758 379,541 430,755 249878 % informal employees 3% 1% 40% 12% 16% 18% 10% % paid employees 2% 0% 20% 6% 8% 9% 5% Source: World Bank staff estimation using LFS data for 2017. Note: PF = permanent, full-time; PP = permanent, part-time; TF = temporary, full-time; TP = temporary, part-time; Ca = casual; NPEF = no permanent employer, full-time; NPEP = no permanent employer, part-time. Underemployment is an issue only for some informal workers, while excessively long hours pose a more significant problem The issue of time-related underemployment is relevant only for a small fraction of informal workers. Sri Lanka has rel- atively low unemployment rates, but underemployment has received little attention despite its relevance for productivi- ty and wages. Underemployment can take several forms, including underutilization of skills; but in this context time-re- lated underemployment is considered, as it is a visible measure of labor underutilization. Specifically, a person is consid- ered to be underemployed when she is (a) currently working less than 35 hours and (b) available for additional work. Figure 9 shows the distribution of work hours by type of worker. Across workers, the work hours are skewed toward the low- Figure 9 Average weekly work hours and underemployment (%) er end, with a good share of employed working less than 35 60 7 6 hours. However, once the willingness to take on additional 50 Hours per week 40 5 work is taken into account, the extent of underemployment Percent 4 30 3 is significantly lower. It tends to be higher among informal 20 2 workers (e.g., 5.9 percent for unpaid family workers and 4.3 10 1 0 0 percent for informal self-employed) and for those workers Unpaid family worker Informal, self-employed Informal employee, informal firm Domestic worker Informal employee, formal firm Formal, self-employed Formal, private Formal, public with shorter work hours (figure 9). With a steady increase in the number of working hours for informal workers, un- deremployment has been improving, declining from 7.7 per- Average weekly hours Underemployment (%) cent in 2006 to 4.1 percent among informal workers in 2017. Source: World Bank staff calculation using LFS data for 2017. 21. Part-time is defined as less than 35 hours of work per week. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 34 At the same time, many informal workers work excessively long hours. “Excessively long hours” is defined as 60 hours or more of work per week; according to the Wages Boards Ordinance, a normal working week cannot exceed 48 hours, and a maximum of 12 hours of overtime is permitted. Despite these regulations, the practice of work- ing very long hours is quite widespread, including among formal private sector workers. The share of such work- ers is similar among formal and informal employees, at 17 – 18 percent (figure 10). The highest incidence is found among self-employed (particularly employers). 22 Figure 10 Distribution of weekly work hours 70 60 50 40 Percent 30 20 10 0 < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours < 35 hours 35-50 hours 50-60 hours 60+ hours Formal, private Formal, public Formal, Unpaid family Informal Informal Informal Domestic self-employed employed self-employed employee, employee, formal firm formal firm informal firm Source: World Bank staff calculation using LFS data for 2017. Entitlement to benefits such as paid leave is much rarer among informal workers than among formal workers Very few informal wage workers receive paid leave, Table 6 Entitlement to paid leave among paid employees (%) in stark contrast to formal workers. Almost 97 percent Formal Formal Informal Informal of formal public sector workers and around 76 percent employee, employee, employee, employee, of formal private sector employees reported being en-   public private formal firm informal firm titled to paid leave as part of their job benefits. By con- 2006 86.9% 73.0% 8.0% 2.4% trast, only about 7 percent of informal employees work- 2007 89.9% 71.3% 5.4% 1.1% ing in formal firms and less than 1 percent of informal 2008 92.8% 70.9% 4.8% 1.3% employees working in informal firms reported the same 2009 92.9% 72.3% 5.7% 1.5% (table 6). This indicates not only a large gap in benefits 2010 96.6% 73.2% 5.9% 2.8% between informal and formal workers as groups, but 2011 96.2% 73.7% 7.2% 1.4% also differences among workers who may be working for 2012 96.4% 77.2% 6.0% 1.9% similar formal firms. 23 What is more, both types of for- 2013 86.8% 75.4% 6.7% 0.6% mal workers have seen their benefits improving over 2014 89.5% 75.6% 6.8% 0.8% time, as the share of formal workers with access to paid 2015 92.5% 73.6% 5.2% 0.5% leave has increased in the last decade; but the opposite 2016 94.7% 73.8% 7.1% 0.7% trend is observed among informal workers, where the 2017 96.7% 75.6% 6.5% 0.7% same share is declining. Source: World Bank staff calculation using LFS data for 2006 – 17. 22. International evidence shows that self-employed workers work much longer hours than employees; see for example OECD (2015). 23. Information on job-related benefits is collected only from paid employees. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 35 An important distinction is to be made between noncoverage and noncompliance. The relevant regulation (TEWA) pre- scribes that workers in firms with 15 or more workers are entitled to social security, regardless of the formal status of the firm. This means that smaller firms (fewer than 15 workers) that do not provide access to social security are not in vio- lation of regulations. Their actions would technically constitute noncoverage and not noncompliance. In 2017, about 40 percent of workers had no pension but were working in firms with less than 15 employees (“noncoverage”). A further 40.2 percent worked in larger firms and had access to pensions (“compliance”). On the other hand, some workers in larger firms do not have access to social security even though they are covered by the law — these firms are the biggest offenders (“noncompliance”). This group included 10.3 percent of workers. The remaining 9.5 percent of workers worked in small firms but had access to pensions. Over the last decade, the “noncoverage” group experienced the largest increase. This coverage gap in pension participation requires attention. A large number of these workers not covered by social securi- ty are employed in manufacturing and commerce. By number of workers, noncompliance with pension contributions is most common in the manufacturing sector. Such noncompliance could be reduced with closer and better monitoring. Informal jobs and the risk of poverty and vulnerability Public sector employees earn much more than formal private sector employees and nearly twice as much as informal employees Wages are the most important attribute of any job, whether formal or informal, and low wages significantly raise the likelihood of living in poverty. Previous analyses of the wages of formal and informal workers and the wage gap between them are very limited and outdated. The analysis in this section updates estimates by using the most recently available Labor Force Survey data and more complete information on earnings for different types of informal workers. Some descriptive analysis of wage differentials is first presented, followed by more in-depth analysis that controls for differences in firms’ and workers’ characteristics. In line with continued economic growth, real wages grew consistently between 2006 and 2017. Overall, real wages grew at an annualized rate of around 4.3 percent per year between 2006 and 2017 among informal employ- ees, and 3.3 percent among formal employees (table A.10). Formal and informal self-employed, for whom earnings data were not collected prior to 2013, experienced an annualized wage growth of around 10 percent and 9 percent, respectively, between 2013 and 2017. Both men and women benefited from broad-based wage growth during this period. Men’s wages (for employees only) increased by an annualized 3.7 percent between 2006 and 2017. The equivalent figure for women was 4.1 per- cent. Despite sustained wage growth for both men and women, there exists a distinct gender gap in wages (figure A.1). Wages are almost at parity in the public sector, with the difference largest among informal employees (table A.11). The recent widening in the gap is likely due to the difficulty of women finding good jobs in the private sec- tor, in part because of rigid labor market regulations that discourage the employment of women—a situation ex- acerbated by traditional norms that impose the burden of household chores and child/elderly care on women, and by the lack of alternative arrangements for the latter. A static wage profile shows that the monthly earnings of formal employees are significantly higher than those of informal employees — well over SL Rs 30,000 – 40,000 for the former, compared to slightly more than SL Rs 20,000 for the latter in 2017.  2 4 Formal public sector workers earn on average about SL Rs 40,000 every month. 24. The survey collects gross earnings data. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 36 These are followed by formal private sector employees, Figure 11 Monthly and hourly earnings of formal who receive about SL Rs 34,000 per month. Among all and informal workers employees, public sector employees earn about twice 120,000 600 as much as informal employees. Formal self-employed 100,000 500 have the highest average earnings, at over SL Rs 100,000 per month, though they represent only a very small frac- 80,000 400 Monthly earnings tion of the labor market (less than 4 percent in 2017). Hourly earnings Informal self-employed, who are primarily own-ac- 60,000 300 count workers, earn about SL Rs 26,000. Domestic 40,000 200 workers earn the least, around SL Rs 16,000 per month. Estimates of hourly earnings present a similar picture, 20,000 100 with the rankings among the workers unchanged (fig- ure 11). It should be noted, however, that while monthly 0 0 Domestic worker Informal employee informal firm Informal employee formal firm Informal self-employed Formal, private Formal, public Formal self-employed earnings of formal public sector and formal private sec- tor workers are not significantly different, hourly earn- ings are considerably higher among public sector work- ers. This is attributed to their shorter working hours; Monthly earnings (L) Hourly earnings (R) they work on average three hours less per week than pri- Source: World Bank staff calculation using LFS data for 2017. vate sector workers (43.4 versus 46.4 hours). Note: Monthly and hourly earnings are in 2017 nominal prices. But averages mask large variation among formal workers — the wage distribution of formal private sector employees is actually much closer to that of informal employees than that of public sector workers Mere averages mask the fact that the wage distribution Figure 12 Distribution of log hourly earnings for formal of formal private sector workers is actually much more and informal workers similar to that of informal workers than that of formal 1.5 public sector workers, which suggests that productivity levels remain broadly low. Figure 11 suggests that public 1.0 Density sector and formal private sector workers earn on average about SL Rs 42,000 and SL Rs 34,000 per month respec- 0.5 tively, whereas informal employees earn on average slight- ly more than SL Rs 20,000. However, the full distribution 0 2 3 4 5 6 7 of hourly earnings, shown in figure 12, reveals a different Log Hourly Earnings picture: the public sector wages are highly concentrated Formal, private Formal, public Informal employee around the mean with little dispersion, whereas the over- Source: World Bank staff estimation using LFS data for 2017. all earnings distribution of formal private sector and in- Note: Wages are spatially deflated, and 1 percent outliers are trimmed from both tails. formal workers is actually very similar, except for a few outliers that inflate the private sector earnings average. The earnings distribution of private sector workers is highly dispersed compared to that of public sector workers. The median earnings of formal private sector workers are much closer to those of informal workers. Despite some improvements, there is considerable variation in wages between workers, regions, and industries Earnings inequality among formal workers has overall decreased over time, while that among informal workers has remained largely unchanged. The Gini coefficient of monthly earnings among formal workers steadily decreased Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 37 until 2012, after which it fluctuated somewhat but re- Figure 13 Gini coefficient of monthly earnings mained around 0.33 (figure 13). The downward trend was 0.40 likely driven by an increase in public sector jobs in the 0.35 0.30 earlier years, among which the earnings distribution Gini coefficient 0.25 is more compressed. On the other hand, the Gini coef- 0.20 ficient of earnings among informal workers did not im- 0.15 prove from a decade ago. 0.10 0.05 0 There is considerable variation in wages across re- 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 gions, even among formal workers. Formal workers Formal Informal in the Western Province earn almost SL Rs 61,000 per Source: World Bank staff calculation using LFS data for 2006–17. month, which is significantly more than their counter- parts everywhere else. In fact, formal workers in the Uva, Figure 14 Monthly nominal earnings by province, Northern, Sabaragamuva, and Eastern Provinces earn formal and informal workers (SL Rs) only SL Rs 26,000 – 30,000. Similar differences are observed 70,000 among informal workers as a group, with those residing 60,000 Monthly earnings (SL Rs) in the Western Province earning almost SL Rs 33,000, com- 50,000 40,000 pared to less than SL Rs 20,000 for those in the Northern, 30,000 Central, and North-Central Provinces (figure 14). 20,000 10,000 Across industries, the formal-informal wage gap 0 Western North-Central Southern North-Western Central Eastern Sabaragamuva Northern Uva as a share of informal wages is the largest in construc- tion and commerce. In construction, formal workers make almost three times as much as informal workers on average, though a large part of this difference is like- All Formal Informal ly explained by differences in occupations. A large differ- Source: World Bank staff estimation using LFS data for 2017. ence is also present in commerce, where formal workers earn more than twice as much as informal workers. Significant gaps are observed in most other services sectors, in- cluding transport and community services. On the other hand, the gap is very small in agriculture (where most work- ers are informal) and financial services (where most workers are formal) (table 7). These are unconditional differences in wages and do not reflect differences in firm or worker attributes that could be driving some of the gaps. Table 7 Formal-informal wage gap by industry (SL Rs)   All Formal Informal Gap Gap, % a Agriculture 16,335 18,064 16,063 2,001 12% Mining 21,694 44,914 17,574 27,34 0 156% Manufacturing 28574 39,297 21,431 17,866 83% Public utilities 41,739 47,669 30,545 17,124 56% Construction 33,374 77,218 27,906 49,313 177% Commerce 4 5,190 76,586 33,141 43,445 131% Transport and communications 35,472 50,266 30,174 20,092 67% Financial and business-oriented service 45,479 47,209 40,287 6,922 17% Community and family-oriented services 40,451 43,346 26,013 17,333 67% Other services, unspecified 33,320 44,391 22,824 21,567 94% Source: World Bank staff calculation using LFS data for 2017. a. Indicates the formal-informal wage gap as percentage of average informal wages. All wages are in nominal values. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 38 The formal-informal wage gap persists even after adjusting for workers’ and firms’ characteristics The wage gap between formal and informal workers has been persistent, and it has continued to widen in recent years. Figure 15 shows the trend in real monthly wages of formal and informal employees for the period 2006 – 17. 25 It should be noted that information on self-employed earnings was not available until 2013. For this reason, figure 15 presents the longer-term trend for employees only. While wages have grown steadily since 2006, the gap between Figure 15 Formal-informal wage gap, 2006–17 formal and informal workers has widened in recent years: 40,000 the 2006 wage gap of around SL Rs 11,400 has grown over 35,000 time to reach SL Rs 14,300 in 2017. A similar trend con- 30,000 sisting of a distinct and widening wage gap is observed 25,000 when looking at hourly wages, or when considering all SL Rs 20,000 formal and informal workers (including self-employed 15,000 workers for whom there was no earnings information 10,000 prior to 2013) for the years 2013 – 17. 5,000 Formal regression analysis is employed to further in- 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 vestigate whether there exists a wage premium for for- Formal Informal mal workers after accounting for observable character- Source: World Bank staff calculation using LFS data for 2006–17. istics. This analysis would make it possible to account Note: Wages are in 2017 prices. for wage differentials that are due to observable individ- ual and firm characteristics. The latter includes worker’s gender, ethnicity, education, geographic sector and dis- trict of residence, experience, sector of activity, occupation, employment status, and firm size. Two sets of analy- ses are conducted, one that pools all data for 2006 – 17 for paid employees only (because information on earnings of self-employed workers was not collected until 2013); and another that pools earnings data from all workers for 2013 – 17. Within each set of regressions, either one indicator variable for all informal workers is added, or multiple variables are added to indicate the type of formal or informal employment of the individual. The adjusted wage gap remains at around 30 percent. Results suggest that there remains a wage premium associated with formal employment, even after controlling for worker and firm characteristics. Full regression results are shown in table A.12 in the appendix. The results in column (1) distinguish between different types of formal and informal workers, with formal private sector workers as the reference group. The earnings of the formal private sector work- ers and formal self-employed are not different in a statistically significant way. Formal public sector employees en- joy a wage premium even after controlling for worker and firm characteristics. The wage gap is the largest for domes- tic workers, followed by informal self-employed and informal employees in formal firms. The estimates in column (2) compare all formal workers against all informal workers. As a group, informal workers are subject to a wage penalty of about 30 percent after taking into account a range of observable factors. Being a woman, being less educated, and being less experienced lead to lower earnings. Across industries, earnings are lower in manufacturing and commerce, and higher in construction (relative to agriculture and after controlling for worker and firm characteristics). By eth- nicity, Sri Lanka Moors and other non-Tamil minorities earn on average more than Sinhalese workers. As the results by occupation suggest, earnings also fall with skill level, with the largest gap found among those unskilled workers in el- ementary occupations. Workers in larger firms tend to be better compensated; this empirical regularity, which is ob- served among otherwise observationally equivalent workers, is also referred to as the “firm-size wage gap” in the liter- ature. In fact, accounting for firm size absorbs a large portion of the wage gap between formal and informal workers. 26 25. All wages are spatially deflated and temporally adjusted using the monthly Colombo CPI (2017 = 100) unless specified otherwise. 26. For public sector workers the firm size is set to 100 or more. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 39 The wage gap appears to be larger at the bottom of the earnings distribution. Given the heterogeneity of infor- mal workers, it is plausible that the formal-informal wage gap varies across the earnings distribution. This possi- bility is investigated by means of quantile regressions (table A.13). The latter confirm that the adjusted gap is indeed wider at the top. Almost half of the wage gap is attributed to differences in group means, with the rest “unexplained.” The wage gap attributed to differences in group means is due to differences in characteristics such as education, experience, firm size, sector, occupation, and so on. The unexplained difference refers to the part of the wage gap that comes from differ- ences in the return to those observable characteristics. An Oaxaca decomposition is employed to estimate how much of the wage gap is owing to either differences in the group means or differences in returns to those characteristics (based on differences in estimated coefficients) that are otherwise unexplained. These results, which also take into ac- count selection of different types of individuals into formal and informal jobs, are presented in table 8. Of the total gap of 0.495 (differences in log of hourly earnings), 0.224 is explained by differences in characteristics, which is equivalent Table 8 Oaxaca decomposition results on log of hourly earnings: Formal versus informal workers Overall   Gap   100 Group 1: Formal workers 4.873 *** Explained 45.25 Group 2: Informal workers 4.378 *** Female -0.029 *** -5.86 Difference 0.495 *** Years of education 0.086 *** 17.37   Experience 0.008 *** 1.62 Threefold decomposition Firm size 0.079 *** 15.96 Total gap, % 11.31% Ethnicity -0.002 *** -0.40 Endowments 0.031 *** Sector -0.003 *** -0.61 Coefficients 0.0793 *** Industry -0.006 ** -1.21 Interaction 0.3843 *** Occupation 0.120 *** 24.24   Region -0.003 *** -0.61 Twofold decomposition Year -0.020 *** -4.04 Explained 0.224 *** Mills ratio -0.007 *** -1.41 Unexplained 0.271 *** Unexplained 54.75 Female 0.063 *** 12.73     Years of education 0.405 *** 81.82     Experience 0.131 *** 26.46     Firm size 0 0.00     Ethnicity 0.009 1.82     Sector -0.003 -0.61   Industry -0.084 *** -16.97     Occupation -0.012 -2.42     Region 0.025 *** 5.05     Year -0.005 *** -1.01     Mills ratio 0.130 *** 26.26 Observations 160,417 Constant -0.387 *** -78.18 Source: World Bank staff estimation using LFS data. Note: Decomposition with Heckman selection conducted on paid employees only. Significance level: * = 10 percent, ** = 5 percent, *** = 1 percent Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 40 to around 45 percent of the gap (in the case of a twofold decomposition). The rest of the wage gap is unexplained. 27 While these estimates are for employees only for the period 2006 – 17, qualitatively similar results are achieved when including all self-employed workers and conducting an estimation for the years 2013 – 17. Education accounts for an overwhelming portion of the unexplained wage premium for formal workers. In other words, the returns to obtaining more education are larger for employment in formal arrangements. Education is also the second most important contributor to the wage difference (after occupation) that is explained by differ- ences in group means. In fact, education accounts for most of the wage gap between formal and informal workers, whether this is owing to the difference in educational attainment or the returns to the same. Overall, closing the education gap could improve productivity and earnings and help close the earnings gap. The risk of extremely low pay is significantly higher for informal workers, who are also more likely to be experiencing in-work poverty The risk of extremely low pay and resulting vulnerability is measured using the country’s minimum wage as a benchmark and applied to employees only. The minimum wage was first defined by the National Minimum Wage of Workers Act of 2016, which set forth a national minimum wage at SL Rs 10,000 per month. In the same year, the Budgetary Relief Allowance of Workers Act mandated a SL Rs 2,500 allowance for all workers earning less than SL Rs 40,000 per month, setting the effective minimum wage threshold at SL Rs 12,500 per month (equiva- lent to US$79.28). While it is difficult to compare minimum wage levels across countries, the ratio of minimum wage to value added per worker is only 0.18 for Sri Lanka, which is much lower than that of comparator countries (World Bank 2016). The risk of extremely low or inadequate pay is significantly higher for informal workers. Open unemployment is not very common in Sri Lanka, and in the absence of a formal unemployment scheme, workers may be pushed into subsistence jobs with very low pay. As of 2017, 27.1 percent of informal workers earned less than the stipu- lated monthly minimum wage. Among formal workers, 5.8 percent fell under the same threshold. Since the regu- lation was enacted in 2016, the share has slightly decreased, from 31.6 percent to 27.1 percent for informal workers; this is not surprising given that the threshold is set at a nominal level and does not increase unless an amendment is passed in Parliament. 28 Ad hoc revisions can also create unforeseen anomalies on the part of the employer. This implies that the share of workers falling below this threshold is naturally set to fall over time as long as there is any positive wage growth in this segment. The legislation does not provide for any indexation of minimum wages, and benefits have already started to erode since 2016 (at the rate of general inflation). This decrease is also indicative of the extent of firms’ noncompliance with the regulation on minimum wages. Consistent with the above findings, the incidence of working poverty is higher among informal workers. “Working poverty” in this context is defined as being in the bottom 10 percent of the per capita household earnings distribu- tion. The latter is calculated using earnings from all household members and dividing them by the entire household size, including members of the household who are not working or are in unpaid jobs. Figure 16 shows the distribution of working-age individuals by per capita household income. The bottom decile consists largely of households with 27. A Nopo decomposition is used to check the robustness of the results. This nonparametric tool, devised by Nopo (2008), calculates match- ing comparisons to estimate wage differences by taking into account differences in the supports of the distributions of observable characteris- tics, thus providing further insights into the unexplained wage difference. According to the Nopo decomposition results, the size of the unex- plained gap shrinks somewhat but is still significant at 34 percent. For further details, see Nopo (2008). 28. An amendment to increase the minimum wage to SL Rs 16,000 is pending parliamentary approval at the time of writing. Informality, Job Quality, and Welfare in Sri Lanka 4. Informal employment, job quality, and poverty 41 few workers (mostly employed on own account as unpaid Figure 16 Distribution of working-age population family workers) and many nonworking individuals. Very by employment status and income decile few formal workers fall in the bottom decile of the earn- 100 ings distribution, making them the least likely to be poor. 90 In fact, the majority of public sector workers fall in higher 80 income deciles. Formal private sector employees are more 70 spread out across the distribution, though again very few 60 are likely to be in the bottom 10 percent of the distribu- Percent 50 tion (figure 16). 40 30 The analysis in this section suggests that the Sri Lanka 20 labor market is broadly stratified into three segments, 10 each with significantly different pay and job quali- 0 ty. Workers in the public sector enjoy high monthly pay, 1 2 3 4 5 6 7 8 9 10 Decile shorter working hours, a large number of holidays in ad- Not working Unpaid family worker dition to paid leave, formal social security, and a high lev- Domestic worker Formal self-employed el of job security. Formal, private sector employees enjoy Informal employee informal firm Formal public slightly lower monthly wages, but their earnings distribu- Informal employee formal firm Formal private Informal self-employed tion is in fact more similar to that of informal employees than to that of public sector workers. They are entitled Source: World Bank staff calculation using LFS data for 2017. Note: Working-age individuals are distributed according to their per capita household to social security but work longer hours, reducing their income. average hourly pay. Most but not all of them receive paid leave. The last group comprises informal workers, the most vulnerable group by any measure. Heterogenous as they may be, most of them have subsistence-level jobs (except for informal employers who report the highest earnings of all but make up only a very small share in the labor force) with low pay, few benefits, and little protection. 5. Conclusion and policy implications Informality, Job Quality, and Welfare in Sri Lanka 5. Conclusion and policy implications 43 Informal employment has remained persistently high in Sri Lanka. Between 2006 and 2017, the share of infor- mally employed workers has remained almost unchanged, at around 60 percent. Adopting a “poverty” perspective on informality, this report investigates the quality of formal and informal jobs, based on two dimensions: (a) con- tracts and working conditions, which capture the nonpecuniary aspects of a job; and (b) the risk of poverty and vulnerability, which examines the risk of extremely low pay and the incidence of working poverty. Overall, the Sri Lanka labor market is stratified into three segments according to the quality of the jobs: (a) formal, public sector workers who enjoy high remuneration, comparatively short work hours, long holidays, entitlement to paid leave as well as other benefits, and overall job security; (b) formal, private sector workers who make up a middle group, with access to social security and some other benefits but with an earning distribution closer to that of informal workers; and (c) informal workers, consisting mainly of informal employees as well as own-account workers, most of whom are in precarious employment arrangements with little access to any benefits and with low remuneration. Stringent labor regulations have long been criticized as encouraging informality, but there also appear to be few benefits to formalization. In Sri Lanka, dismissals are not only difficult but also expensive, resulting in redun- dancy costs that are the second highest in the world. These costs hamper firms’ adjustment in response to changes in market conditions. While enforcement of regulations is limited, nontransparent laws leave room for arbitrary decisions, making litigation costly and risky. Moreover, access to formal credit — in theory a major benefit of for- malization — is constrained by other structural factors, including stringent collateral requirements and land ten- ancy. These issues would not be resolved by formalization alone. Formalization efforts around the world have focused on easing entry regulations, but they have generally yielded limited success. Following a slew of research that showed that informality led to a large loss in potential tax rev- enues, resource misallocation, and unfair competition between the formal and informal sector, many countries implemented business entry regulation reforms to reduce informality. However, evidence to date suggests that easing entry regulations and providing information on the formalization process have only a very limited impact on the formalization of existing informal firms. While many regulatory reforms have taken place in the past dec- ades, the majority of businesses in developing countries remain informal (Bruhn and McKenzie 2014). A few stud- ies in Latin America have found that a large reduction in the cost and time needed for business registration has a modest impact on the number of formal firms (for example, Bruhn [2011] for Mexico). The effect of business entry reforms on the formalization of informal firms is also found to be rather small (Bruhn 2013). Interventions that combine information, waived costs, and enforcement also found little to no effect on tax registration (for example, De Giorgi and Rahman [2013] for Bangladesh). Larger impacts were found when business registration did not involve tax registration or when large changes to tax rates were implemented. Alcazar, Andrade, and Jaramillo (2015) found that offering a subsidy for the cost of a munic- ipal license in Peru led 10 – 12 percent of informal businesses to obtain one. On the other hand, reducing tax rates and lessening the complexity of taxes for small firms led to no significant overall effect in Brazil (Monteiro and Assuncao 2012). In Pakistan, moreover, a large increase in the tax burden led more than half of tax-paying firms to leave the formal system; those that remained had substantially reduced reported earnings (Waseem 2018). A field experiment in Sri Lanka by de Mel, McKenzie, and Woodruff (2013) found that providing information about the registration pro- cess and reimbursing up-front for registration costs had no effect on firm registration at the Divisional Secretary Division level, the relevant registration for tax purposes. These findings reinforce the argument that firms are rationally choosing to be informal because the costs outweigh the benefits of formalization (Benhassine et al. 2016). Given the limited success in getting firms to formalize, there have been few opportunities to directly meas- ure the benefits of doing so. De Mel, McKenzie, and Woodruff (2013) also did not find evidence of benefits accru- ing to firms — such as access to bank accounts or loans, participation in government training programs, increased Informality, Job Quality, and Welfare in Sri Lanka 5. Conclusion and policy implications 44 receipt of government contracts, or increased certainty over taxes — a s a result of formal status. Benhassine et al. (2016) offered a full package of supplementary benefits to firms, which boosted registration rates when formali- zation was linked with access to government training programs, support to open bank accounts, and tax media- tion services, in addition to information on registration. However, these services were perceived to be of limited benefit to the firms, indicating that even enhanced formalization efforts may not be sufficient to induce formali- zation at large scale. The available evidence highlights the fact that reducing informality should not be in itself a policy goal. Existing research finds that informality tends to decrease with development, but growth alone does not ensure a reduction in informality. In fact, deindustrialization and the expansion of the services sector has contributed to a rise in infor- mality in some countries. Sri Lanka has seen continued economic growth in the past 15 years, including a period of high growth, yet informality levels remain almost unchanged. On the one hand, the role and size of the highly informal primary sectors (mostly agriculture and fishery) also matter. On the other hand, even as the size of the primary sector shrunk, the additional jobs created in industry and services were predominantly informal, even when growth was high. To some extent, the persistence of informality also has to do with the sectors that gener- ate economic growth and jobs. Thus it is important to address the root causes and correlates of informality rather than strictly focusing on formalization alone. Strategies that promote aimless formalization can lead to higher poverty, unemployment, and crime levels. A well-conceived formalization strategy would aim at increasing the benefits and reducing the costs of for- malization through a policy mix that tackles the interrelated causes and consequences of informality together. Informality tends to be higher when the regulatory and tax burden is heavy, enforcement is weak, and the tax-benefit system does not favor formality (Loayza 2016). The government’s policy toolkit includes regulatory, credit, and labor market policies that all affect the decisions of workers and firms about whether to be formal or informal. Thus reform packages should be implemented in an integrated rather than piecemeal way, which would also make the reforms politically and socially more acceptable (Loayza 2018). The following reform components are suggested to make formalization more attractive in line with recom- mendations from the international literature. They respond to the diverse characteristics and causes of infor- mality, as well as to interrelated issues such as inferior job quality, lack of social protection, and low productivity (all of which require investment in human capital and access to credit and technology): • Increase flexibility in labor markets and quality of jobs by reducing labor taxes and simplifying compli- ance. Labor costs, especially those associated with dismissal, are high, effectively reducing the demand for formal labor. TEWA was introduced during a time when the economic and social environment was drasti- cally different from today, and it needs to be reconsidered so firms can adjust their workforce in response to worker productivity and changing market conditions. Maintaining both high severance pay and a for- mal social security (EPF) system can be prohibitively expensive for employers. Streamlining and simplifying labor regulations to lessen the law’s ambiguity, increase its transparency, and reduce overlapping coverage would make compliance easier and less costly. Enforcement efforts should focus more on mediation than on litigation, which can drastically increase uncertainty and costs for both workers and employers. There have been recent efforts to combine existing labor laws into a single labor law but the proposed legislation failed to obtain final approval. • Build an adequate and effective social protection system. Social protection reform can reduce the vulnera- bility to economy-wide and idiosyncratic shocks — yet if poorly designed it can work to reinforce high infor- mality. Thus it is critical to balance work incentives and workers’ protection. The ETF contributions could Informality, Job Quality, and Welfare in Sri Lanka 5. Conclusion and policy implications 45 be converted into an unemployment insurance fund to serve the ETF’s original intended purpose. Increased flexibility for workers, firms, and markets to adjust to changes in broader economic conditions, combined with an effective social protection system, will lead to greater economic growth. The increase in workers in smaller firms who are outside of lawful coverage is an issue that merits more attention. • Increase human capital to promote productivity gains and reduce the wage gap. Decomposition analy- sis showed that a large part of the wage gap is due to differences in educational attainment or the returns to it. Improving the skills of the labor force could help reduce the wage gap. To that end, education reforms are needed to align the curriculum to the demands of the labor market; the workforce also needs continu- ing training to upgrade skills, adapt to changing market conditions, and prepare for a longer working life than in the past. • Streamline regulations while strengthening monitoring. Strengthening monitoring alone could lead to more informality (as firms might more actively try to conceal unlawful employment); it could also drive higher unemployment and even smaller firms. At the same time, leveling the playing field is important — e.g., enforcement activities should extend to both formal and informal firms. (The latter are currently not cov- ered by such activities.) These reforms can be achieved in the short run. In response to long-standing challenges and complaints stemming from the multiplicity of laws regulating labor, a Unified Employment Law was approved in the cabinet in June 2018, but failed to obtain final parliamentary ap- proval. The law aimed to consolidate the SOEA, the WBO, the Maternity Benefits Ordinance, and the Employment of Women, Young Persons and Children Act into one labor law, without reducing any of the rights or benefits un- der the individual laws. Further, the draft law included definitions for statuses and concepts—such as casual labor, fixed-term contracts, a mandatory retirement age (which is a matter for contract at present), and probation in em- ployment—that were previously defined only in practice. It also extended coverage of employment protection leg- islation to all workers, including domestic workers and informal sector workers; prohibited avoidance strategies such as use of undocumented workers; and addressed inward migrant workers and employment placement agen- cies, which were hitherto not included. The law was also intended to encourage women’s labor market participa- tion, by providing a framework for flexible work arrangements, home work, part-time work, etc. Current provisions state that all workers must receive the same benefits regardless of whether they are employed full-time or not; this may be a disincentive to employ women, who may not be able to work full-time. Removal of restrictions on night work for women, overtime work in some sectors, and equal pay for equal work, as well as provisions to take action against sexual harassment, were all aimed at expand- ing employment opportunities for women and provid- Figure 17 Male and female labor force participation rate by age ing them with a safe and dignified working environment. 100 90 The imperative for Sri Lanka to implement these re- 80 forms is driven by its demographic context, specif- 70 Participation (percent) 60 ically the rapid aging — combined with early retire- 50 ment — of the labor force. The majority of men in the la- 40 bor force exit the labor market starting at around age 55; 30 the age is even lower for women (figure 17). Labor force 20 participation can be increased by increasing the number 10 of workers joining the labor force and facilitating longer 0 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 working lives. The employment protection legislation that Age imposes a prohibitively high cost on the employer for dis- Female Male missals, combined with a relatively low retirement age Source: World Bank staff calculation using LFS data for 2017. Informality, Job Quality, and Welfare in Sri Lanka 5. Conclusion and policy implications 46 by international standards (55 for men and 50 for women), introduces an incentive structure that leads most of the workforce to exit the labor market early. Most of those who remain part of the labor force beyond average retirement age work in agriculture. While there is no formal mandatory retirement age in the private sector, because of the “prior consent” clause in TEWA (see chapter 3) employment contracts typically specify 55 as retirement age for men (Ranaraja 2019). Despite the lack of income support, the elderly tend to exit the labor market rather prematurely. Pension is an en- titlement that is reserved for public sector workers and a few private sector employees. Reforms should attempt to make labor regulations compatible with the challenges posed by an aging labor force. The regulatory and incentive scheme needs to be restructured to encourage more working-age individuals to join the labor market and extend their working lives. This would help meet the labor demands in a country that still needs to grow significantly but is undergoing rapid aging. Needed reforms include raising the formal retire- ment age (including harmonizing the age for men and women), and structuring incentives to be consistent with the legislation. Firms need to have the flexibility to adjust their workforce according to skills and market needs, and to promote and pay workers based on merit. The latter is particularly a concern in the public sector. Reforms should also aim to bring down the cost of labor, which currently makes it prohibitively expensive to retain workers for an extended period of time. Adjustments are needed from both employers and workers: employers need to adapt to the aging of the workforce, and workers have to invest in expanding and upgrading their skills. Allowing elderly workers to remain longer in the labor market also allows them to maintain a decent living standard and prevents them from falling into poverty, especially when social security maintains limited coverage. Women are particularly vulnerable because many work as unpaid family workers and are much more likely than men to become depend- ent earlier in the life cycle. More research is needed to identify a set of concrete policies that strengthen the incen- tive to extend one’s working life while also allowing for flexible wages. It should be noted that this report addresses only the first of the “three P’s” mentioned in the introduction — pov- erty. The other two — public finance and productivity — a re interrelated with the first and are equally important aspects of informality. While they are beyond the scope of this report, they warrant thorough analysis to fully understand the role of informality in the economy. Informality, Job Quality, and Welfare in Sri Lanka 47 Appendix A Table A.1 Labor force participation rate, age 15 and over (%)   2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Total 57.1 55.5 55.1 54.1 53.4 52.9 52.5 53.7 53.2 53.8 53.8 54.1 Male 76.7 76.2 76.1 74.7 75.0 74.0 74.9 74.9 74.6 74.7 75.1 74.5 Female 39.5 37.0 36.6 36.2 34.4 34.3 32.9 35.4 34.6 35.9 35.9 36.6 Source: World Bank staff calculation using LFS data for 2006 – 17. Table A.2 Summary of labor market indicators   2006 2007 2008 2009 2010 2011 Population (age 15 and above) 13,265,380 13,456,542 14,625,844 14,876,264 15,166,284 14,976,135 Labor force participation rate (%) 57.1 55.5 55.1 54.1 53.4 52.9 Total employed 7,086,364 7,025,268 7,627,999 7,579,835 7,696,14 3 7,591,404 Formal (total) 2,142,041 2,088,519 2,351,606 2,236,973 2,239,280 2,216,138 Formal, private employees 1,000,177 950,362 1,003,009 950,450 974,563 976,076 Formal, public employees 885,579 899,657 1,083,163 1,085,553 1,022,744 1,019,483 Formal self-employed 256,285 238,501 265,434 200,970 241,973 220,579 Informal (total) 4,944,323 4,936,749 5,276,393 5,342,862 5,456,863 5,375,266 Unpaid family worker 727,718 712,117 784,226 789,705 796,221 804,880 Informal self-employed 2,259,143 2,238,959 2,429,209 2,396,692 2,566,314 2,564,051 Domestic worker a — — — — — — Informal employee formal firm 571,622 590,900 650,502 618,187 588,799 576,676 Informal employee informal firm 1,385,840 1,394,772 1,412,456 1,538,279 1,505,528 1,429,659   2012 2013 2014 2015 2016 2017 Population (age 15 and above) 14 ,857,153 14,958,006 15,134 ,4 83 15,281,945 15,448,679 15,843,735 Labor force participation rate (%) 52.5 53.7 53.2 53.8 53.8 54.1 Total employed 7,488,704 7,680,621 7,700,490 7,830,976 7,947,683 8,208,179 Formal (total) 2,250,968 2,261,302 2,328,580 2,393,995 2,4 08,174 2,623,417 Formal, private employees 970,958 940,489 994 ,177 1,052,431 1,090,722 1,220,832 Formal, public employees 1,036,553 1,064,399 1,090,222 1,100,388 1,088,116 1,101,729 Formal self-employed 243,457 256,414 24 4 ,180 241,176 229,336 300,855 Informal (total) 5,237,736 5,419,318 5,371,910 5,436,982 5,539,509 5,584,762 Unpaid family worker 654,249 695,962 683,360 660,064 623,141 655,655 Informal self-employed 2,541,954 2,449,283 2,427,588 2,532,900 2,499,630 2,506,270 Domestic worker — 59,392 59,150 47,843 57,505 53,572 Informal employee formal firm 641,324 684,693 721,176 696,092 695,74 0 752,212 Informal employee informal firm 1,400,208 1,529,988 1,480,637 1,500,083 1,663,494 1,617,053 Source: World Bank staff calculation using LFS data for 2006 – 17. Note: — = Not available. a. Due to changes in the questionnaire, domestic workers are measured only after 2013, and some are included as informal self-employed prior to 2013. Informality, Job Quality, and Welfare in Sri Lanka Appendix A 48 Table A.3 Formal and informal employment as share of total employment (%) Share of total employment (%) 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Formal Formal, private employees 14.1% 13.5% 13.1% 12.5% 12.7% 12.9% 13.0% 12.2% 12.9% 13.4% 13.7% 14.9% Formal, public employees 12.5% 12.8% 14.2% 14.3% 13.3% 13.4% 13.8% 13.9% 14.2% 14.1% 13.7% 13.4% Formal self-employed 3.6% 3.4% 3.5% 2.7% 3.1% 2.9% 3.3% 3.3% 3.2% 3.1% 2.9% 3.7% Informal Unpaid family worker 10.3% 10.1% 10.3% 10.4% 10.3% 10.6% 8.7% 9.1% 8.9% 8.4% 7.8% 8.0% Informal self-employed 31.9% 31.9% 31.8% 31.6% 33.3% 33.8% 33.9% 31.9% 31.5% 32.3% 31.5% 30.5% Domestic worker a — — — — — — — 0.8% 0.8% 0.6% 0.7% 0.7% Informal employee formal firm 8.1% 8.4% 8.5% 8.2% 7.7% 7.6% 8.6% 8.9% 9.4% 8.9% 8.8% 9.2% Informal employee informal firm 19.6% 19.9% 18.5% 20.3% 19.6% 18.8% 18.7% 19.9% 19.2% 19.2% 20.9% 19.7% Informal employment, total 69.8% 70.3% 69.2% 70.5% 70.9% 70.8% 69.9% 70.6% 69.8% 69.4% 69.7% 68.0% Source: World Bank staff calculation using LFS data for 2006 – 17. Note: — = Not available. a. Due to changes in the questionnaire, domestic workers are measured only after 2013, and some are included as informal self-employed prior to 2013. Table A.4 Number of workers with secondary job (thousands) and share of total employed, 2006 – 17 % share Secondary job: formal/informal 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 in 2017 Formal, private               5 1 1 1 1 0.2% Formal, public               1 1 0 1 0 0.1% Formal self-employed               20 20 17 25 22 3.1% Unpaid family worker               127 126 124 129 125 18.1% Informal self-employed               470 416 445 589 418 60.4% Informal employee formal firm               12 16 16 17 21 3.1% Informal employee informal firm               100 85 98 114 90 13.1% Domestic worker               1 1 0 2 13 1.9% Total with secondary job 759 683 799 766 750 713 516 736 667 700 878 692   Total employed 7,086 7,025 7,628 7,580 7,696 7,591 7,489 7,681 7,700 7,831 7,948 8,208   % of workers with second job 10.7% 9.7% 10.5% 10.1% 9.7% 9.4% 6.9% 9.6% 8.7% 8.9% 11.0% 8.4%   Source: World Bank staff calculation using LFS data for 2006 – 17. Note: Detailed data on secondary jobs are available only after 2013. Years with no estimate are in gray. Table A.5 Distribution of workers with formal/informal main and secondary jobs All employed Distribution in 2017 Formal main job, formal secondary job 0.2% Formal main job, informal secondary job 1.7% Informal main job, formal secondary job 0.1% Informal main job, informal secondary job 6.4% Only one formal job 30.0% Only one informal job 61.5% Informality when only main job is considered 68.0% Informality when main & secondary jobs are considered 69.8% Source: World Bank staff calculation using LFS data for 2006 – 17. Informality, Job Quality, and Welfare in Sri Lanka Appendix A 49 Table A.6 Formal and informal workers by sector of activity (%) self-employed self-employed Unpaid family informal firm formal firm % informal employee employee Domestic Informal Informal Informal Formal, Formal, private Formal worker worker public   Agriculture 8.1% 1.9% 1.0% 17.8% 48.4% 3.0% 19.4% 0.4% 89.0% Mining 4.4% 3.2% 6.9% 1.5% 15.9% 26.9% 40.8% 0.3% 85.5% Manufacturing 32.9% 1.2% 3.1% 6.0% 27.4% 12.8% 16.3% 0.2% 62.7% Public utilities 15.2% 46.9% 2.6% 0.6% 6.5% 19.9% 8.3% 0.0% 35.3% Construction 5.9% 2.8% 2.4% 0.5% 14.5% 7.1% 66.5% 0.3% 88.9% Commerce 12.2% 1.5% 10.8% 11.2% 38.8% 12.1% 13.3% 0.1% 75.5% Transport and communications 10.5% 10.7% 4.9% 1.0% 43.8% 10.4% 18.2% 0.5% 73.9% Financial and business-oriented services 50.5% 22.6% 1.4% 0.2% 6.3% 16.9% 2.0% 0.0% 25.5% Community and family-oriented services 10.4% 69.7% 2.5% 0.6% 3.5% 9.1% 4.0% 0.2% 17.4% Other services, unspecified 10.3% 33.4% 3.2% 3.3% 17.6% 10.8% 18.3% 3.1% 53.1% Source: World Bank staff calculation using LFS data for 2017. Table A.7 Formal and informal workers by firm size (%) Formal Unpaid Informal Informal Informal Formal, self-em- family self-em- employee employee Domestic   private ployed worker ployed formal firm informal firm worker Formal Informal Less than 5 4.5% 24.8% 6.9% 5.5% 21.2% 45.4% 0.0% 8.5% 19.3% 5 to 9 5.7% 8.1% 1.3% 0.5% 12.4% 7.8% 0.0% 6.1% 4.2% 10 to 15 8.9% 5.6% 0.9% 0.0% 21.6% 0.0% 0.0% 8.3% 2.8% 16 to 49 16.1% 1.7% 0.1% 0.0% 17.0% 0.0% 0.0% 13.3% 2.1% 50 to 99 11.7% 0.6% 0.0% 0.0% 7.8% 0.0% 0.0% 9.5% 1.0% 100 or more 52.6% 1.8% 0.1% 0.0% 16.2% 0.0% 0.0% 42.5% 2.0% No paid employee 0.6% 57.3% 87.2% 94.0% 3.7% 45.5% 0.0% 11.8% 67.0% Working for household 0.0% 0.1% 3.5% 0.0% 0.1% 1.3% 100.0% 0.0% 1.8% Source: World Bank staff calculation using LFS data for 2017. Informality, Job Quality, and Welfare in Sri Lanka Appendix A 50 Table A.8 Sectoral distribution of formal and informal workers (number and share) Informal employee Informal employee Domestic worker informal workers informality rate Formal, private self-employed self-employed Formal, public Unpaid family Total workers informal firm Share of all formal firm Informal Sectoral Formal worker   Agriculture 172,803 41,588 20,866 380,324 1,035,896 64,477 415,425 8,807 2,14 0,185 26% 89.0% Mining 2,767 2,029 4,343 973 10,043 16,948 25,697 214 63,014 1% 85.5% Manufacturing 520,322 19,445 49,754 95,419 432,589 201,847 258,068 3,492 1,580,936 19% 62.7% Public utilities 7,327 22,534 1,249 270 3,134 9,549 4,000 - 48,062 1% 35.3% Construction 37,719 17,867 15,239 3,057 92,522 45,588 425,446 2,045 639,482 8% 88.9% Commerce 141,832 17,376 125,369 130,023 450,433 139,905 154,068 1,438 1,160,4 4 4 14% 75.5% Transport 61,280 62,493 28,309 5,916 255,722 60,894 105,891 2,752 583,255 7% 73.9% and communications Financial/business-oriented 88,816 39,724 2,468 284 11,152 29,700 3,592 - 175,735 2% 25.5% services Community/family-oriented 77,910 521,821 18,756 4,120 26,459 68,411 29,868 1,319 748,663 9% 17.4% services Other services, unspecified 110,058 356,853 34,502 35,268 188,322 114,894 195,000 33,506 1,068,404 13% 53.1% Source: World Bank staff calculation using LFS data for 2017. Table A.9 Formal and informal workers by contract type and full-/part-time Formal Permanent Temporary Casual No permanent employer   Full-time Part-time Full-time Part-time Full-time Part-time Full-time Part-time 2006 1,393,708 294,490 102,280 10,133 58,334 12,004     2007 1,333,364 309,098 107,632 14,702 58,532 8,855     2008 1,417,847 419,055 133,280 15,220 59,384 7,298     2009 1,393,534 431,249 109,689 16,464 49,547 12,332     2010 1,403,292 356,799 123,774 20,857 52,188 8,965     2011 1,352,620 424,656 110,205 12,544 65,160 9,363     2012 1,494,476 315,901 114 ,181 9,166 53,060 8,436     2013 1,4 37,711 322,838 132,995 19,460 64,405 10,099     2014 1,484,800 400,851 113,870 13,627 49,286 8,635     2015 1,584,620 370,856 110,470 12,883 50,515 9,194     2016 1,690,611 295,953 113,906 9,496 49,020 6,362     2017 1,795,360 317,593 135,207 8,752 55,586 6,353     Informality, Job Quality, and Welfare in Sri Lanka Appendix A 51 Informal Permanent Temporary Casual No permanent employer   Full-time Part-time Full-time Part-time Full-time Part-time Full-time Part-time 2006 94,584 25,501 588,595 222,510 287,531 128,469 345,899 352,308 2007 84,292 17,981 692,785 264,399 261,751 104,921 367,390 309,934 2008 81,262 19,950 732,476 283,883 252,770 110,543 403,004 324,327 2009 67,4 49 20,624 772,459 301,536 286,100 115,754 412,490 341,820 2010 62,605 13,237 792,479 268,403 266,677 89,422 4 07,990 351,964 2011 82,272 12,251 747,319 284,085 235,064 89,733 439,442 286,146 2012 92,782 14,659 875,625 237,499 245,678 68,384 433,965 244,717 2013 67,563 20,439 813,744 262,364 314,253 86,968 428,331 269,075 2014 78,560 17,275 826,809 262,663 318,226 103,258 405,674 238,271 2015 69,441 17,018 861,410 281,375 273,032 100,529 384,529 239,567 2016 65,446 13,340 956,542 264,627 295,396 89,412 474,740 244,659 2017 83,543 19,270 963,602 278,758 282,767 96,774 430,755 249,878 Source: World Bank staff calculation using LFS data for 2006 – 17. Note: There are no formal workers who report having no permanent employer. Table A.10 Monthly earnings in real terms for formal and informal workers Annualized Monthly earnings (real, in 2017 prices) growth rate 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2006–12 2013–17 Formal, 20,355 20,531 19,159 20,230 21,055 22,048 22,340 25,589 25,780 28,999 31,715 31,740 2% 6% private Formal, 29,453 30,636 26,902 28,275 28,316 27,774 27,772 29,471 30,648 37,459 41,052 40,988 -1% 9% public Formal 46,048 48,941 56,275 81,955 95,133 20% self-employed Informal 16,686 18,200 19,939 24,739 24,888 11% self-employed Informal employee 14,085 14,228 13,950 14,350 14,528 15,112 15,987 16,572 17,469 18,539 20,435 21,761 2% 7% formal firm Informal employee 12,654 12,577 12,530 12,817 13,393 14,243 15,373 15,730 16,300 17,756 19,584 20,167 3% 6% informal firm Domestic worker 12,532 11,756 13,054 13,857 15,668 6% Source: World Bank staff calculation using LFS data for 2006–17. Note: Cells are empty when relevant information was not collected. Table A.11 Average male wages as % of female wages Formal, private 163% Formal, public 103% Informal employee formal firm 176% Informal employee informal firm 195% Domestic worker 140% All employees 122% Source: World Bank staff calculation using LFS data. Informality, Job Quality, and Welfare in Sri Lanka Appendix A 52 Table A.12 Wage gap regression results For 2013–17 For employees (1) (2) (3) (4) -0.296 *** -0.297 *** Informal (0.01) (0.01) 0.213 *** 0.209 *** Formal, public (0.01) (0.01) -0.125 Formal, self-employed (0.16) -0.349 ** Informal, self-employed (0.16) Informal employee, -0.303 *** -0.265 *** formal firm (0.01) (0.01) Informal employee, -0.151 *** -0.166 *** informal firm (0.01) (0.01) -0.423 *** -0.294 *** Domestic worker (0.03) (0.02) -0.367 *** -0.368 *** -0.273 *** -0.274 *** Female (0.01) (0.01) (0.00) (0.00) 0.0278 *** 0.0286 *** 0.0347 *** 0.0363 *** Years of education (0.00) (0.00) (0.00) (0.00) 0.0231 *** 0.0252 *** 0.0247 *** 0.0271 *** Experience (0.00) (0.00) (0.00) (0.00) -0.000316 *** -0.000339 *** -0.000320 *** -0.000347 *** Experience squared (0.00) (0.00) (0.00) (0.00) Firm size 0.0185 -0.0247 ** 0.00678 -0.0240 *** 5–9 employees (0.01) (0.01) (0.01) (0.01) 0.0606 *** -0.0517 *** 0.0257 ** -0.0598 *** 10–15 employees (0.02) (0.01) (0.01) (0.01) 0.0755 *** -0.0363 *** 0.0423 *** -0.0476 *** 16–49 employees (0.02) (0.01) (0.01) (0.01) 0.0872 *** -0.0186 0.0604 *** -0.0268 *** 50–99 employees (0.02) (0.01) (0.01) (0.01) 0.107 *** 0.0913 *** 0.0838 *** 0.0660 *** 100 or more employees (0.01) (0.01) (0.01) (0.01) 0.0703 *** 0.0871 *** 0.0669 *** 0.0765 *** 0 paid employee (0.01) (0.01) (0.01) (0.01) 0.101 *** -0.0536 *** 0.138 *** 0.0335 *** Working for household (0.02) (0.02) (0.02) (0.01) Industry -0.347 *** -0.342 *** -0.307 *** -0.295 *** Mining (0.03) (0.03) (0.03) (0.03) -0.0856 *** -0.116 *** -0.0748 *** -0.109 *** Manufacturing (0.01) (0.01) (0.01) (0.01) 0.104 *** 0.130 *** 0.111 *** 0.152 *** Public utilities (0.03) (0.03) (0.02) (0.02) 0.190 *** 0.197 *** 0.254 *** 0.261 *** Construction (0.01) (0.01) (0.01) (0.01) -0.0805 *** -0.0934 *** -0.0572 *** -0.0691 *** Commerce (0.01) (0.01) (0.01) (0.01) Transport and 0.0247 * 0.0337 ** 0.0445 *** 0.0589 *** communications (0.01) (0.01) (0.01) (0.01) Financial and 0.114 *** 0.111 *** 0.0535 *** 0.0502 *** business-oriented services (0.02) (0.02) (0.01) (0.01) Informality, Job Quality, and Welfare in Sri Lanka Appendix A 53 For 2013–17 For employees (1) (2) (3) (4) Community and -0.00925 0.0526 *** -0.00622 0.0742 *** family-oriented services (0.01) (0.01) (0.01) (0.01) Other services, -0.0222 * -0.00953 -0.0234 *** 0.00849 unspecified (0.01) (0.01) (0.01) (0.01) Employment status 0.802 *** 0.618 *** Employer (0.16) (0.02) 0.116 -0.0741 *** Own-account worker (0.16) (0.01) Ethnicity 0.0107 0.00126 -0.00285 -0.0127 Sri Lanka Tamil (0.01) (0.01) (0.01) (0.01) -0.0127 -0.00745 -0.0222 ** -0.0179 * Indian Tamil (0.02) (0.02) (0.01) (0.01) 0.0663 *** 0.0660 *** 0.0247 *** 0.0244 *** Sri Lanka Moor (0.01) (0.01) (0.01) (0.01) 0.141 *** 0.122 *** 0.125 *** 0.105 *** Other (0.03) (0.03) (0.02) (0.02) -0.0783 *** -0.0746 *** -0.0434 *** -0.0400 *** Sector: Rural (0.01) (0.01) (0.00) (0.00) -0.178 *** -0.185 *** -0.163 *** -0.172 *** Sector: Estate (0.02) (0.02) (0.01) (0.01) Occupation 0.226 *** 0.264 *** -0.0274 ** 0.0183 Professionals (0.02) (0.02) (0.01) (0.01) -0.0495 *** -0.0442 *** -0.312 *** -0.298 *** Technicians (0.01) (0.01) (0.01) (0.01) -0.143 *** -0.131 *** -0.398 *** -0.379 *** Clerks (0.02) (0.02) (0.01) (0.01) Service and market sales -0.298 *** -0.300 *** -0.629 *** -0.619 *** workers (0.01) (0.01) (0.01) (0.01) -0.422 *** -0.420 *** -0.551 *** -0.534 *** Skilled agricultural workers (0.02) (0.02) (0.02) (0.02) -0.229 *** -0.215 *** -0.519 *** -0.499 *** Craft workers (0.01) (0.01) (0.01) (0.01) -0.247 *** -0.251 *** -0.511 *** -0.503 *** Machine operators (0.01) (0.01) (0.01) (0.01) Elementary -0.324 *** -0.319 *** -0.638 *** -0.622 *** occupations (0.01) (0.01) (0.01) (0.01) -0.194 *** -0.158 *** -0.410 *** -0.371 *** Armed forces (0.03) (0.03) (0.02) (0.02) -0.488 *** -0.507 *** Others (0.07) (0.07) 4.452 *** 4.509 *** 4.454 *** 4.482 *** Constant (0.03) (0.03) (0.02) (0.02) Observations 136,104 136,104 162,370 162,370 R-squared 0.236 0.233 0.403 0.397 Source: World Bank staff estimation using LFS data. Note: Dependent variable is log of hourly earnings. All estimations include year dummies. Reference groups (omitted categories) are as follows: formal private sector workers in columns (1) and (3), and all formal workers in columns (2) and (4); agriculture (industry), employee (employment status), Sinhalese (ethnicity), urban (geographic sector), managers (occupation), year 2006 (year), and Colombo (district—admin level 2). Significance level: * = 10 percent, ** = 5 percent, *** = 1 percent Informality, Job Quality, and Welfare in Sri Lanka Appendix A 54 Table A.13 Quantile regression results: 95% confidence Figure A.1 Male-female earnings gap (employees only) interval of coefficient on informal dummy 35,000 Coefficient [95% Confidence Interval] 30,000 q25 -0.219 -0.227 -0.210 25,000 Earnings (SL Rs) q50 -0.199 -0.210 -0.189 20,000 q75 -0.187 -0.199 -0.176 15,000 10,000 Source: World Bank staff estimation using LFS data. Note: Specifications for quantile regressions are the same as in table A.12. 5,000 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 All female employees All male employees Source: World Bank staff calculation using LFS data for 2006 – 17. Informality, Job Quality, and Welfare in Sri Lanka 55 Bibliography Abidoye, Babatunde, Peter F. Orazem, and Milan Vodopivec. 2009. Gimpelson, Vladimir, Rotislav Kapelyushnikov, and Anna Lukiyanova. “Firing Cost and Firm Size: A Study of Sri Lanka’s Severance Pay 2009. “Employment Protection Legislation in Russia: Regional System.” Social Protection Discussion Paper 0916, World Bank, Enforcement and Labour Market Outcomes.” IZA Discussion Washington, DC. Paper 4484, Institute for the Study of Labor, Bonn. Alcazar, Loreana, Raul Andrade, and Miguel Jaramillo. 2015. “Panel/ Gunatilaka, Ramani. 2008. “Informal Employment in Sri Lanka: Tracer Study on the Impact of Business Facilitation Processes Nature, Probability of Employment, and Determinants of Wages.” on Enterprises and Identification of Priorities for Future ILO Asia-Pacific Working Paper Series, ILO Subregional Office for Business Enabling Environment Projects in Lima, Peru — Report South Asia, New Delhi. 5: Impact Evaluation after the Third Round.” Report to the Hart, Keith. 1973. “Informal Income Opportunities and Urban Employ- International Finance Corporation. Mimeo. ment in Ghana.” Journal of Modern African Studies 11, no. 1: 61 – 89. Arunatilake, Nisha, and Priyanka Jayawardena. 2010. “Why People Heltberg, Rasmus, and Milan Vodopivec. 2009. “Sri Lanka: Choose to Participate in the Informal Sector in Sri Lanka.” Indian Unemployment, Job Security and Labor Market Reform.” SSRN. Journal of Labour Economics 53, no. 2 (January): 225 – 49. https://ssrn.com/abstract=1208662. Benhassine, Najy, David McKenzie, Victor Pouliquen, and Hussmanns, Ralf. 2004. “Measuring the Informal Economy: From Massimiliano Santini. 2016. “Can Enhancing the Benefits Employment in the Informal Sector to Informal Employment.” of Formalization Induce Informal Firms to Become Formal? ILO Working Paper 53, Policy Integration Department, Bureau Experimental Evidence from Benin.” Policy Research Working of Statistics, International Labour Office, Geneva. Paper 7900, World Bank, Washington, DC. ILO (International Labour Organization). 2013. Women and Men Bruhn, Miriam. 2011. “License to Sell: The Effect of Business in the Informal Economy: A Statistical Picture. 2nd ed. Geneva: Registration Reform on Entrepreneurial Activity in Mexico.” International Labour Office. Review of Economics and Statistics 93, no. 1: 382 – 86. ———. 2014. Labour Law, the Business Environment and Growth of Micro ———. 2013. “A Tale of Two Species: Revisiting the Effect and Small Enterprises: Sri Lanka — Country Report. New Delhi: of Registration Reform on Informal Business Owners in Mexico.” International Labour Organization. Journal of Development Economics 103: 275 – 83. ———. 2016. Key Indicators of the Labour Market. 9th ed. Geneva: Bruhn, Miriam, and David McKenzie. 2014. “Entry Regulation and International Labour Office. the Formalization of Microenterprises in Developing Countries.” ———. 2018. Women and Men in the Informal Economy: A Statistical World Bank Research Observer 29, no. 2: 186 – 201. Picture. 3rd ed. Geneva: International Labour Office. Chen, Marty, and Donna Doane. 2008. “Informality in South Kanbur, Ravi. 2009. “Conceptualizing Informality: Regulation and Asia: A Review.” WIEGO Working Paper 4. Women in Informal Enforcement.” IZA Discussion Paper 4186, Institute for the Study Employment: Globalizing and Organizing. https://www.wiego. of Labor, Bonn. org/publications/informality-south-asia-review-0. ———. 2014. “Informality: Causes, Consequences and Policy Dasanayaka, S.W.S.B. 2011. “Global Challenges for SMEs in Sri Lanka Responses.” Cornell University Working Paper WP 2014 – 18, and Pakistan in Comparative Perspectives.” Business Review 6, Cornell University, Ithaca, New York. no. 1: 61 – 80. Kim, Yeon Soo, Michael Matytsin, and Samuel Freije-Rodriguez. 2018. DCS (Department of Census and Statistics). “Labour Force Survey “Informal Employment and Worker’s Well-Being in the Russian Annual Report.” Various years. http://www.statistics.gov.lk/page. Federation.” Policy Research Working Paper 8989, World Bank, asp?page=Labour%20Force. Washington, DC. De Giorgi, Giacomo, and Aminur Rahman. 2013. “SME’s Registration: La Porta, Rafael, and Andrei Shleifer. 2014. “Informality and Evidence from an RCT in Bangladesh.” Economics Letters 120, Development.” Journal of Economic Perspectives 28, no. 3. no. 3: 573 – 78. Lazear, Edward P. 1990. “Job Security Provisions and Employment.” De Mel, Suresh, David McKenzie, and Christopher Woodruff. 2011. Quarterly Journal of Economics 105, no. 3: 699 – 726. “Getting Credit to High Return Microentrepreneurs: The Results Lehmann, Hartmut, and Norberto Pignatti. 2007. “Informal of an Information Intervention.” World Bank Economic Review 25, Employment Relationships and Labor Market Segmentation no. 3: 456 – 85. in Transition Economies: Evidence from Ukraine.” IZA ———. 2013. “The Demand for, and Consequences of, Formalization Discussion Paper 3269, Institute for the Study of Labor, Bonn. among Informal Firms in Sri Lanka.” American Economic Journal: Loayza, Norman V. 2016. “Informality in the Process of Development Applied Economics 5, no. 2: 122 – 50. and Growth.” Policy Research Working Paper 7858, World Bank, Distinguin, Isabelle, Clovis Rugemintwari, and Ruth Tacneng. 2015. Washington, DC. “Can Informal Firms Hurt Registered SMEs’ Access to Credit?” ———. 2018. “Informality: Why Is It So Widespread and How Can Unpublished working paper. http://www.apdr.pt/pej2015/ It Be Reduced?” Research & Policy Brief no. 20, World Bank, papers/59.pdf. Washington, DC. Gatti, Roberta, Diego F. Angel-Urdinola, Joana Silva, and Andras Loayza, Norman V., Anna Maria Oviedo, and Luis Serven. 2005. “The Bodor. 2014. Striving for Better Jobs: The Challenge of Informality Impact of Regulation on Growth and Informality: Cross-Country in the Middle East and North Africa. Directions in Development. Evidence.” Policy Research Working Paper 3623, World Bank, Washington, DC: World Bank. Washington, DC. Informality, Job Quality, and Welfare in Sri Lanka Bibliography 56 Maloney, William. 2004. “Informality Revisited.” World Development UNCTAD (United Nations Conference on Trade and Development). 32, no. 7: 1159 – 78. 2004. Investment Policy Review of Sri Lanka. New York and Maurizio, Roxana. 2014. “Labour Formalization and Declining Inequal- Geneva: United Nations. ity in Argentina and Brazil in the 2000s: A Dynamic Approach.” ILO UN DESA (United Nations Department of Economic and Social Research Paper 9, International Labour Office, Geneva. Affairs). 2009. System of National Accounts 2008. https://unstats. Monteiro, Joana C. M., and Juliano J. Assuncao. 2012. “Coming Out un.org/unsd/nationalaccount/docs/SNA2008.pdf. of the Shadows? Estimating the Impact of Bureaucracy Simplifi- Verité Research. 2015. “Sri Lanka: Domestic Workers. An Analysis cation and Tax Cut on Formality in Brazilian Microenterprises.” of the Legal and Policy Framework.” https://www.veriteresearch. Journal of Development Economics 99, no. 1: 105 – 15. org/wp-content/uploads/2018/06/Sri-Lanka-Domestic-Worker Nopo, Hugo. 2008. “Matching as a Tool to Decompose Wage Gaps.” s-Legal-Policy-Framework-No.-1.pdf. Review of Economics and Statistics 90, no. 2: 290 – 99. Waseem, Mazhar. 2018. “Taxes, Informality and Income Shifting: OECD (Organisation for Economic Co-operation and Development). Evidence from a Recent Pakistani Tax Reform.” Journal of Public 2015. “Enhancing Job Quality in Emerging Economies.” Chapter 5 Economics 15: 41 – 7 7. in OECD Employment Outlook 2015. Paris: OECD Publishing. World Bank. 2012. “Sri Lanka — Enterprise Survey 2011.” https://data- Perry, Guillermo, William Maloney, Omar Arias, Pablo Fajnzylber, catalog.worldbank.org/dataset/sri-lanka-enterprise-survey-2011. Andrew Mason, and Jaime Saavedra-Chanduvi. 2007. Informality: ———. 2015. Sri Lanka — Ending Poverty and Promoting Shared Exit and Exclusion. Latin American and Caribbean Studies. Prosperity: A Systematic Country Diagnostic. Washington, DC: Washington, DC: World Bank. World Bank. Rama, Martin. 1999. “The Sri Lankan Unemployment Problem ———. 2016. Doing Business 2016: Measuring Regulatory Quality and Revisited.” Policy Research Working Paper 2227, World Bank, Efficiency. Washington, DC: World Bank. Washington, DC. ———. 2017. Doing Business 2017: Equal Opportunity for All. Ranaraja, Shyamali, 2019. “Scope and Application of Employment Washington, DC: World Bank. Protection Legislation in Sri Lanka.” Mimeo. Sirimanna, Bandula. 2011. “Sri Lankan Banks Ease SME Loan Terms, but Few Do for Collateral.” Business Times. June 19. http://www.sundaytimes.lk/110619/BusinessTimes/bt32.html.