The macroeconomic impact of informal employment
\nAs of 2017, latest year for which data is available, we estimate that 657,385 (15+ years old) Syrian
\nemployees did not have a work permit in Turkey. The consequences of informal work arrangements for
\nemployees are well known29 and involve denied labour rights and benefits as well as a higher degree of
\nuncertainty with respect to the future of their jobs.
\nThis research, even though concerned about the impact of informality at the individual level,
\nproposes to fill a gap in the literature on Syrian refugees by carrying out an assessment of the aggregate
\ncost of informal employment. Research so far has tended to focus on other macroeconomic indicators;
\nfor example, T\u00a8umen (2016), studied the effect of informal Syrian workers on the price level of goods.
\nHe finds that the goods whose production process intensely employs informal workers showcased a
\ndecline in their prices. He explains this phenomenon by arguing that Syrian workers are replacing
\nTurkish natives in informal jobs at a cheaper rate, passing some of the lower labour costs onto the
\ngoods\u2019 prices. In addition, the research carried out by Semih T\u00a8umen also finds that natives have lower
\nchances of finding an informal job, an extent that we confirm with the use of microdata in the textile
\nand garment industries, while Turkish natives\u2019 opportunities of finding a formal job increase due to
\nthe expansion in the provision of public services caused by the arrival of the refugees.
\nThe present section drifts away from the impact of Syrian refugees on the labour market to focus
\non the effect of informal employment on the income of the Social Security Institute, the Treasury,
\nemployers and Syrian refugees. The reason behind this approach is to stress the often overlooked
\ndamage which is brought by informal work to business competition and to the viability of public health
\nsystems. With respect to the former, informal labour constitutes an unfair advantage that diminishes
\n29See, as examples for the case of Syrian refugees in Turkey Eder & Ozkul (2016) or S\u00b8enses (2016). \u00a8
\n22
\nSyrian refugees in the Turkish Labour Market
\nthe competitiveness of compliant businesses while promoting and rewarding less efficient companies
\nand economic sectors. In a sense, hiring informal employees is nothing but a subsidy granted by the
\nwider society without their consent. Something similar happens with the Social Security Institute
\nof Turkey with regards Syrian refugees;30 informal work does not change their obligation to provide
\nhealth care but prevents them from the funds that would be needed to carry out such activities.
\nThe contribution of this study is twofold; first, we calculate the taxes and social security contributions that would have been due, were Syrians formally employed. We do this by taking advantage
\nof an identification strategy that separates Syrian refugees from other migrants in the HLFS, thus
\nobtaining access to the refugees\u2019 labour earnings. This, in turn, allows to quantify how much is saved
\nby Syrians\u2019 employers and how much income is foregone by the Social Security Institute. The second
\ncontribution assesses a number of future scenarios with regards to the prevalence of informality among
\nSyrian refugees in Turkey. In particular, we combine HLFS microdata with data from the DGMM to
\nestimate the percentage of Syrians under temporary protection that would not be eligible to apply for
\na work permit under the existing hiring constraints.
\nThe section is split into two parts; the first one explains the methodology utilized to calculate the
\nearnings of informal Syrian employees in a hypothetical setting that assumes they are formally hired.
\nThis section also provides some details about relevant pieces of the Turkish tax system as well as the
\nsocial security legislation. The second part calculates the macroeconomic impact of informality in a
\nnumber of plausible scenarios for the whole country as well as by subregion and by economic sector.
\nThe section concludes with an evaluation of the current constraints faced by those applying for a work
\npermit and with some policy recommendations.
\n3.1 Payroll accounting
\nThe formalization of a worker go hand in hand with changes in the amounts due and received by several
\neconomic actors, including employees, employers, the Treasury and the Social Security Institution of
\nTurkey. Two major triggers of these changes are taxes and social insurance contributions. Our aim in
\nthis subsection is to build a simplified version of the Turkish tax system of 2017, year the data31 on
\nearnings used for this analysis comes from, with the purpose of calculating the labour cost of Syrian
\nrefugees had they been hired formally.
\nSocial insurance contributions for private employees. Social security contributions are provided by the social insurance and universal health insurance Act 5510, which states the type of insurances and the extent of the contribution provided by each of the three economic agents involved,
\nemployers, employees and the treasury, see Table 1. Government contributions are calculated as one
\n30It should be noted that there exists no difference between an ordinary citizen\u2019s informality and a Syrian\u2019s in terms
\nof social security legislation. Informal employment is a problem that exists independently of whether the worker is a
\nTurkish citizen or a foreigner.
\n31Data from the 2017 Household Labour Force Survey. We gratefully acknowledge the provision of the HLFS data by
\nthe Turkish Statistical Institute under reference number 27964695-622.03-E.26626.
\n23
\nSyrian refugees in the Turkish Labour Market
\nquarter of the sum of employer and employee contributions towards the long-term, short-term and
\nhealth insurance (unemployment insurance is not included in this calculation, the treasury contributes
\n1 per cent of the gross income towards this item). In the past, short-term insurance primes were
\ndetermined locally, ranging from 1 per cent to 6.5 per cent but they are now fixed at 2 per cent
\nnation-wide.
\nThe term \u2018long-term insurance\u2019 in Table 1 refers to the invalidity, old-age and survivors insurance
\npremium. Crucial for the type of analysis being performed in this section is Act 5763, which amends
\nAct 5110 and states that 5 percentage points of the long-term insurance contribution originally covered
\nby the employer shall be now paid by the treasury as crisis relief. This discount is already reflected in
\nTable 6. Overall, and after applying the long-term insurance discount, Turkey divides the contributions
\nalmost equally among employees (15 per cent), employers (17.5 per cent) and the Treasury (14.125 per
\ncent), with employers still bearing an slightly higher burden than the other two agents separately.
\nTable 6: Social security contributions, 2017
\nPercentage (%)
\nType of risk Employer Employee State Total
\nShort-term insurance 2 0 0 2
\nLong-term insurance 6 9 10 25
\nHealth insurance 7.5 5 3.125 15.625
\nUnemployment insurance 2 1 1 4
\nTotal 17.5 15 14.125 46.625
\nIncome tax. In addition, employees are liable to pay income tax in accordance to appropriately
\nupdated thresholds provided in the article 103 of the Income Tax Act 19332. The income tax brackets
\nand the tax rates applicable in 2017 are displayed for the reader\u2019s convenience in Table 7. It should
\nbe noted that income taxes in Turkey are paid from the first Lira earned irrespective how low is the
\nincome earned.
\nEven though Turkey does not have a minimum amount of income exempt of taxes, it does have
\na so-called minimum living allowance (Asgari Ge\u00b8cim Indirimi, article 32, Income Tax Act 193), a
\nmonetary compensation that is added to the net income after paying social contributions and income
\ntax. This allowance depends on the personal circumstances of the employee; for a single employee,
\nthe allowance is 50% of the income tax paid by the minimum wage -which in practice means that
\nthe allowance is tied to the minimum wage level. The allowance increases by 10 percentage points
\nwhen the employee is married to a non-working partner and, in presence of dependent children, 7.5
\npercentage points for the first 2 kids, 10 percentage points for the third kid and 5 percentage points
\n32See Gelir Vergisi Kanunu, 193, 31\/12\/1960 and its amendments.
\n24
\nSyrian refugees in the Turkish Labour Market
\nfor the fourth and fifth kids can be added to the coefficient. The maximum allowance is set at 85%
\nirrespective of the family circumstances and if both spouses work, only one33 of them is entitled to
\nclaim an extra allowance because of the children. Last but not least, the Law understands by children
\nthose aged less than 18 under the taxpayer supervision or those aged less than 25 if they are still in
\neducation. Aside from the income tax, employees are also liable for the so-called stamp tax (Damga
\nVergisi) which in 2017 was set at 0.759% of the gross salary (not the tax base, which is the one used
\nfor the calculation of the income tax).
\nTable 7: Income tax, 2017
\nRange (TRY) Rate (%)
\n0-13,000 15
\n13,001-30,000 20
\n30,001-110,000 27
\n110,000+ 35
\nCalculating gross incomes. Assessing the impact of the formalization of Syrian refugees requires
\ninformation on the taxes and the social insurance contributions that Syrian refugees, their employers
\nand the Treasury would need to pay were the former to be registered in the SSI. In order to calculate
\nsaid contributions we need to reverse-engineer the payroll accounting so as to convert net incomes, N,
\nthe information that is available, into gross incomes. This is the opposite exercise most accountants
\ndo, though, since generally it is information on gross income, G, what is available. The equation
\nthat transforms gross to net income, f(G), as well as the equation showing the inverse function that
\ntransforms net to gross incomes, f
\n\u22121
\n(N), are presented in Appendix D, including a version that uses
\ndata from the 2017 fiscal year which de facto constitutes a simplified version of a payroll calculator
\nfor Turkey. These two equations are, then, used to calculate the different contributions and taxes paid
\nand received by the different economic agents under study.
\n3.2 The impact of formalization
\nThe idea of calculating the changes that public institutions as well as employers and employees would
\nface if Syrian refugees were to be hired formally is relatively simple and undoubtedly appealing on
\nthe grounds of labour rights compliance. However, in practice, such counter-factual analysis requires
\nassumptions on how \u2018formal\u2019 the employee is going to be, in other words, how many hours is the
\nemployer going to declare to the Social Security Institution. These assumptions implicitly refer to the
\nnegotiations between employers and employees and there are reasons to suspect some employers may
\n33In the HLFS microdata it is not possible to know which spouse claimed the dependency of the children and the
\nadditional points are allocated at random to one of them.
\n25
\nSyrian refugees in the Turkish Labour Market
\nbe buying their employees short since the phenomenon of misreporting hours worked is a relatively
\nwell-established in Turkey. According to the relevant labour legislation, workers cannot be employed
\nfor more than 2,600 hours34 a year, however, 25 per cent of the country\u2019s formal employees report in
\nthe 2017 HLFS having worked more than said limit.
\nIf a significant share of formal employees works more than it is legally allowed, it can be reasonable
\nto expect Syrian refugees would too. Several reasons support this line of thought; first, informal
\nworkers have less bargaining power and less support to make their rights prevail. Second, informal
\nSyrian refugees tend to be employed in similar jobs as the ones formal employees who work too much
\ntend to have. In sum, the structural reasons for which formal workers are asked to work overtime
\nare likely to make Syrian refugees work overtime too. This factor is augmented in the case of Syrian
\nemployees due to the fact that informal workers are less capable to claim their rights and avoid abusive
\nconditions. The information from the 2017 HLFS seems to support the previously-discussed ideas as
\n56.1 per cent of informally employed Syrians (more than double the share of Turkish formal employees
\n-24.0 per cent-) work more than 50 hours on an average week, while 35.3 per cent do so for 60 hours
\nor more.
\nScenarios. Based on the fact that some employers provide abusive working conditions even to formal
\nemployees, we design three scenarios under which the formalization of Syrian refugees could become a
\nreality. These scenarios vary on the capacity and the willingness of employers to bear the full cost of
\nformalizing Syrian employees with the assumptions behind each of the scenarios being summarized in
\nTable 8. The first scenario, labeled \u2018pessimistic\u2019, assumes that employers are able to almost fully evade
\nthe cost of formalizing an employee. In practice, this scenario is built with the assumption that the gross
\npay offered by the employers under a formal contract is made equal to the net salary currently earned
\nby informal Syrian employees. Such behavior entails two consequences; one is monetary, Syrians\u2019
\nlower disposable income would also lower their contributions to the Social Security Institute. The
\nother consequence has to do with the number of hours reported by the employer to the SSI; since most
\nSyrian employees are already (while being informally employed) earning less than the minimum wage,
\na reduction in their salaries would need to be accompanied by a proportional reduction in the hours
\nofficially worked.
\nThe second scenario is labeled \u2018conservative\u2019 and it is built on the assumption that employers would
\nformalize Syrian employees while keeping their net wages constant, i.e. they would absorb the due
\ntaxes and the social insurance insurance contributions. In this scenario, the declaration of working
\nhours by the employer to the SSI works in a similar fashion it does under the \u2018pessimistic\u2019 scenario;
\nas many as possible provided the hourly minimum wage is respected. The only difference comes from
\nthe fact that salaries are higher and, thus, the number of hours declared will increase proportionally.
\nThe last scenario proposes to make Syrian employes as formal as possible by making employers declare as many hours as their employees work to the SSI (within the legal limit of 50 hours a week). As a
\n34That is an average week of 45 hours + 5 hours of overtime.
\n26
\nSyrian refugees in the Turkish Labour Market
\nresult, the so-called \u2018optimistic\u2019 scenario, also implies higher salaries and social insurance contributions,
\nboth in line with the number of hours actually declared.
\nTable 8: Summary of assumptions in each scenario
\nScenario
\nVariable Pessimistic Conservative Optimistic Net income
\nDerived from gross income
\nusing Eq. D.7 formula.
\nSame as the net income
\nearned when informally
\nhired.
\nAs much as implied by the
\nnumber of hours actually
\nworked (cap at 50 hours
\na week) if hourly wage
\nis lower than minimum
\nwage. Does not affect
\nthose earning above the
\nhourly minimum wage.
\nGross inc.
\nSame as the net income
\nearned when informally
\nhired.
\nDerived from net income
\nusing Eq. D.9 formula.
\nDerived from net income
\nusing Eq. D.9 formula.
\nHours declared
\nAs many as the gross
\nincome allows to declare
\n(cap at 50 hours a week).
\nDoes not affect those earning above the hourly minimum wage.
\nAs many as the gross
\nincome allows to declare
\n(cap at 50 hours a week).
\nDoes not affect those earning above the hourly minimum wage.
\nEqual to the number of
\nhours actually worked
\n(cap 50 hours a week).
\nMeasuring the impact. Four monetary indicators, one per economic agent involved, are derived
\nfrom the calculation of gross incomes with the intention of assessing the effect35 of the hypothetical
\nformalization of -currently informal- Syrian employees\u2019 working arrangements. For the employees
\nthemselves, we focus on the take-home pay, also known as net salary. In the case of employers, we report
\nthe difference in labour cost (C) between what the employer pays to an informal employee, C = N,
\nand what the employer would pay once that employee has been made formal, C = (1 + sc)f
\n\u22121
\n(Nh
\n),
\nwhere sc is the share of the gross income the employer contributes to the SS on behalf of the employee
\n35It should be noted that the calculations performed in this section are an educated guess to the actual taxes and
\ncontributions that would had been paid by the different economic actors under several hypothetical scenarios using a
\nsimplified version of the 2017 Turkish tax system. There are several reasons why these figures are only an approximation,
\nfor instance, the lack of accuracy of the family links in the HLFS data (which affects the minimum living allowance),
\nthe lack of other income information (affecting the applicable marginal tax rate) and\/or not knowing the work history
\nof the person during the fiscal year (which affects the annual gross income).
\n27
\nSyrian refugees in the Turkish Labour Market
\nand f
\n\u22121
\n(Nh
\n) is a function that calculates the gross income of an employee from a hypothetical net
\nincome, see Equation D.3 for more details (Appendix D). As for the Social Security Institute, we sum
\nall the social insurance contributions collected, S = (se + sc + sg)f
\n\u22121
\n(Nh
\n), with se, sc and sg being
\nthe shares respectively applying to employees, companies and the government. Last but not least, we
\nfocus on the Treasury of Turkey which has both, a source of income (income tax, T) and an obligation
\nto pay (SS contributions on behalf of the employee). The statistic reported is the subtraction of the
\ntwo, D = T(Nh
\n) \u2212 sgf
\n\u22121
\n(Nh
\n), with the functional form of the taxes collected, T(Nh
\n), being available
\nin Equation D.8 (Appendix D).
\nTable 9: Changes induced by the formalization of Syrian employees (TRY per year)
\nFormality scenarios
\nStatistic Current Pessimistic Neutral Optismistic
\nEmployer cost -14,638 -17,200 -21,832 -27,067
\nHours declared to SS na 1,559 1,895 2,415
\nHours actually worked 2,729
\nEmployee gross income 14,638 14,638 18,580 23,036
\nSS contribution 0 -2,195 -2,787 -3,456
\nTaxes 0 -738 -1,155 -1,586
\nNet income 14,638 11,705 14,638 17,994
\nTreasury 0 -1,329 -1,469 -1,667
\nTaxes 0 738 1,155 1,586
\nSS contribution 0 -2,067 -2,624 -3,253
\nSocial security 0 6,824 8,663 10,740
\nEmployee 0 2,195 2,787 3,456
\nEmployer 0 2,562 3,252 4,031
\nTreasury 0 2,067 2,624 3,253
\nInformal employees (Syrian) 657,385
\nSource: HLFS 2017 and author\u2019s own calculations. Notes: The table shows the effect of formalizing the average Syrian
\ninformal employee on a number of statistics for four economic agents, employers, employees, the Treasury and Social
\nSecurity Institution. The changes are shown under three scenarios, a) a pessimistic one that makes the gross income
\nequal to the informal net, b) a neutral one that keeps the net income received by the employee constant and c) an
\noptimistic one that pays the minimum wage for as many hours as the average informal Syrian employee actually works
\n(with a cap at 50 hours).
\nAll four indicators are calculated (see Table 9) for the average informal Syrian employee in the
\ncurrent scenario (i.e. the one where Syrian refugees are employed without a work permit) and in
\nthe three hypothetical setups designed to build counterfactual scenarios where Syrian refugees are
\n28
\nSyrian refugees in the Turkish Labour Market
\nemployed with a work permit. For employers, the effect of formalizing Syrian refugees goes from
\nmild -an increase of 2,562 per employee (annual figures) in the pessimistic scenario-, to extreme –
\nan increase of 12,429 that almost doubles the original cost of the worker. The sizable differences in
\nlabour cost between the three scenarios are largely explained by the number of hours officially declared
\nby the employer to the SSI. The reason why employers cannot declare all the hours worked is given
\nby the fact that many Syrian refugees earn below the minimum wage while working more hours than
\nthe ones legally permitted. As a result, declaring a higher number of hours would be accompanied by
\na pay rise so as to prevent the hourly wage from falling below the minimum wage level. It should be
\nnoted, though, that some payroll fraud would exist irrespective of the scenario chosen for the analysis;
\nfrom 314 hours misreported per employee a year in the \u2018optimistic\u2019 scenario to an staggering 1,170
\nhours a year in the \u2018pessimistic\u2019 one, i.e. there would be work for the labour inspection even in the
\nbest36 scenario under consideration.
\nWhile it is clear that the formalization process would be costly for employers, it is also clear that
\nthe main beneficiary of such process is the Social Security Institute of Turkey. The income received
\nper worker is estimated in the range of 6,824 to 10,740 per worker, with the contributions coming
\nalmost in equal proportion from employers, employees and the Treasury of Turkey. The size of the
\nlatter contribution, sparked by the five percentage points relief the Treasury temporarily took care
\nof, yields a negative net income for the Turkish public administration. Indeed, the Treasury might
\nneed to \u2018pay\u2019 an average of 1,329- 1,667 for every worker formalized because the social contributions
\npaid by the government greatly exceed the taxes received from low income earners. It should also
\nbe noted that without the five percentage points (originally part of the employers\u2019 long-term social
\ninsurance contribution) that is being currently paid by the government, the Treasury would be close
\nto break-even -although it would still find itself in the \u2018red\u2019.
\nThe effect of the formalization process on the Syrian employees is, perhaps, the most unclear one;
\ndepending on the scenario, the change on Syrian employees\u2019 take-home pay goes from a loss, (-) 2,933
\nto a gain of 3,356 per year. The ambiguous effect of having a formal job on net income, which is
\narguably a major source of concern for this sub-population,37 may lower informal workers\u2019 incentives
\nfor demanding a work permit. Net income is definitely not the only benefit carried by the formalization
\nof work arrangements and there are other benefits like the right to be represented by a trade union, a
\nhigher level of certainty with respect to contractual conditions, access to a pension after retirement and
\nhealth insurance, just to mention the most prominent ones. However, the benefits that in a different
\nenvironment would constitute a powerful incentive for a regular worker to desire a formal job, turns
\nout being not-so advantageous for Syrian refugees who, first of all, already have access to health care
\ndue to the temporary protection regime and, second, might not want to retire in Turkey, thus, lowering
\nthe value attached to their long-term insurance contributions.
\n36By best it is meant the one with the higher number of hours declared.
\n37In many Syrian families there is a single income earner due to the low female labour force participation.
\n29
\nSyrian refugees in the Turkish Labour Market
\nMacroeconomic impact. The number of (+15) Syrian employees working without a valid work
\npermit is estimated at 657,385, that is 4.2 per cent of the total number of formal employees working in
\nTurkey in 2017. This percentage is large enough to expect a substantial impact in the budgets, profits
\nand disposable incomes of, respectively, the SSI, employers and employees. In order to improve the
\nvisualization of the macroeconomic impact, Figure 12 reports the changes that would have accrued
\nto each of the economic agents under all three scenarios for a varying share of Syrian employees
\nformalized. Some labels have also been added at the point where 80 per cent of currently informal
\nSyrian employees obtain a work permit; this point is chosen because it roughly corresponds to the
\ncurrent share of Turkish employees enrolled in the SSI. On top of that, some extra labels are used to
\nillustrate the full impact under the optimistic scenario.
\nFigure 12: Impact of formalizing Syrian refugees on four economic actors under three scenarios
\n(a) Social security, net income (b) Treasury, net income
\n(c) Employers, profit (d) Employees, take-home pay
\nSource: HLFS 2017 and author\u2019s own calculations. Notes: The figure shows the macroeconomic impact of formalizing
\ncurrently informal Syrian employees for four economic agents, employers, employees, the Treasury and the Social Security
\nInstitution. The impact is shown under three scenarios, a) a pessimistic one that makes the new gross income equal to
\nthe informal net income, b) a neutral one that keeps the net income constant and c) an optimistic scenario where Syrian
\nemployees are paid the minimum wage for as many hours as they actually work (with a cap at 50 hours).
\nThe biggest beneficiary of the registration of informally employed Syrian refugees is the SSI; the
\n30
\nSyrian refugees in the Turkish Labour Market
\ninstitution could expect 5,648 millions to be added to its budget if the share of formalized employees
\nmatched the current share of formal Turkish employees. SSI\u2019s foregone income is even higher; under
\nfull-compliance its revenue could go up to 7,060 millions or, put differently, the SSI budget could go
\nup by 2.2 per cent38 nation-wide. Moreover, it is well-known that Syrian refugees are not evenly spread
\nover the Turkish territory. In practice, they tend to populate neighboring regions with Syria (Adana,
\nGaziantep, Hatay, S\u00b8anl\u0131urfa) as well as some of the most-populated cities of the country (\u02d9Istanbul,
\nBursa, Konya and \u02d9Izmir). In some of these subregions the share of informal Syrian employees with the
\nrespect to the formal population of employees more than doubles the national average (see Table 10) as
\ndoes the income that the regional branches of the SSI could collect from the formalization of refugees.
\nFor example, the SSI could receive an extra income of up to 433 millions a year from Gaziantep, most
\nof which at the expense of employers. A slightly lower impact could be achieved in the subregion of
\nKonya (where informally hired Syrian made up to 8.1 per cent of the formal employment), Adana or
\nHatay but still much higher than in the subregions outside the top-8, where Syrian refugees only make
\nup for 1.2 per cent of all formal employees.
\nTable 10: Macroeconomic impact due to the formalization of all Syrian employees (optimistic scenario),
\nby subregion
\nInformal Syrian employees Impact (millions TRY 2017)
\nSubregion Total % of all employees Social sec. Treasury Employers
\n\u02d9Istanbul 288,219 7.2 3,161 -467 -3,288
\nBursa 59,754 6.0 645 -103 -786
\nAdana 58,536 8.6 605 -102 -854
\nGaziantep 41,470 10.8 433 -74 -576
\nHatay 39,245 8.7 406 -68 -524
\nKonya 33,672 8.8 373 -58 -454
\n\u02d9Izmir 21,875 2.1 234 -42 -256
\nS\u00b8anl\u0131urfa 21,330 6.8 242 -36 -312
\nRest of Turkey 93,284 1.2 962 -147 -1,121
\nSource: HLFS 2017 and author\u2019s own calculations. Notes: The table shows the change in income derived derived from
\nformalizing the work arrangements of all informal Syrian employees by subregion (NUTS-2) in the optimistic scenario.
\nIt also shows the number and the percentage of informal Syrian employees in each subregion.
\nOf all the economic activities benefiting from the presence of Syrian refugees, one stands out from
\nthe rest; the textile, clothing, leather and footwear industries39 (hereinafter referred to as TCLF),
\nwhich together employ almost a third of the informal Syrian employees. TCLF industries\u2019 attachment
\n38The budget of the Social Security Insitute stood at 312,734 millions in 2017, source: www.ceicadata.com.
\n39These industries are defined by the ISIC rev.4 codes 13, 14 and 15.
\n31
\nSyrian refugees in the Turkish Labour Market
\nto informal work is not new, though, and the reasons behind it need to be assessed from a national
\nand a global perspective. As it can be seen in Table 12, back in 2005 TCLF industries were already
\nemploying almost half million informal workers; by 2014 the informality phenomenon had decreased
\nin both, the total and the share, with only 379,132 informal workers adding up to 26.4 per cent of
\nall TCLF employees. The positive trend towards formality came to a halt with the arrival of Syrian
\nrefugees who not only crowded out nationals from informal jobs but also increased the number of
\ninformal employees, bringing it back to the 2005 level.
\nTable 11: Macroeconomic impact due to the formalization of all Syrian employees (optimistic scenario),
\nby economic sector
\nInformal employment
\nActivity Increase in payroll (mill.) Total Share
\nAgriculture 768 60,750 9.2
\nTCLF manufacturing 2,771 218,628 33.3
\nOther manufacturing 1,279 106,373 16.2
\nConstruction 1,170 99,505 15.1
\nTransport & comm. 61 6,099 0.9
\nTrade, hospitality 1,360 105,396 16.0
\nBusiness act. 81 6,277 1.0
\nEducation 105 10,401 1.6
\nHealth 58 5,567 0.8
\nOther services 517 38,389 5.8
\nTotal 8,171 657,385 100.0
\nSource: HLFS 2017 and author\u2019s own calculations. Notes: The table shows 1) the change in companies\u2019 payroll derived
\nfrom formalizing the work arrangements of all informal Syrian employees by economic activity (measured by the ISIC
\nrev.4) in the optimistic scenario and 2) the number and the percentage of informal Syrian employees working in each
\neconomic activity. TCLF refers to the ISIC rev.4 codes 13, 14 and 15.
\nHistorical reasons by themselves are not sufficient to explain the strong attachment of TCLF
\nindustries to informal labour. Another contributing factor has to do with the strong competition
\nTurkey-based garment factories face in the international market. Their products have to compete with
\nthose coming from Viet Nam, Bangladesh, Mexico or Thailand just to name a few markets where wages
\nare at the level or lower than the ones offered in Turkey. In fact, employment growth between 2005 and
\n2015 in the TCLF industries of 20 countries (including major producers) is highly correlated40 with
\nthe level of wages41 initially held in 2005, adding yet another argument to the hypothesis of highly
\n40Data from the ILO-Labour Force Micro-dataset v1.5.
\n41Wages are measured in 2011 constant USD to facilitate the comparison.
\n32
\nSyrian refugees in the Turkish Labour Market
\nmobile capital in response to international competition -also sparked by the expiration of the Multi
\nFibre Arrangement\u2019s successor in 2005.42 To sum up, some producers, especially those operating in
\nthe TCLF industries, may have strong incentives to keep an illegal workforce in operation.
\nTable 12: Type of employment in the TCLF industries, by year
\nYear
\nType 2005 2014 2017
\nInformal 494,230 379,132 491,278
\nof which Syrians 218,628
\nFormal 683,878 1,059,164 1,046,977
\nTotal 1,178,108 1,438,296 1,538,255
\nInformality rate 42.0 26.4 31.9
\nSource: HLFS 2005, 2014, 2017 and author\u2019s own calculations. Notes: The table shows the number of formal and
\ninformal employees in the textile, clothing, leather and footwear industries (ISIC rev.4 codes 13, 14 and 15) in Turkey
\nfor three years, 2005, 2014 and 2017. In addition, it shows the year and sector-specific rate of informal employment in
\nthe TCLF industries.
\nWhich scenario and why? The calculations use three intra-firm bargaining scenarios alongside a
\ncontinuum of formalized Syrian employees ranging from 0 to 100 per cent of the target population. The
\nmultiplicity of situations opens an obvious yet relevant question, where will Turkey be in the coming
\nfuture? The main difficulty when assessing the outcome of the negotiations between employers and
\nemployees with respect to the formalization of work arrangements lies on the fact that this negotiation
\nhas not happened yet, i.e. it is hypothetical. In spite of this difficulty, some light is thrown in by
\na similar negotiation that has already happened: the one between current formal Syrian employees
\nand their employers. The simplest analysis would compare the average net earnings of formal Syrian
\nemployees ( 1,937 a month) with that of those employed informally, 1,185, finding an increase of
\n752 after formalization. However, this analysis is not only simple, it is also simplistic insofar it does
\nnot take into account the likely differences in skill, educational background and sector of employment
\nbetween those employed formally and informally. If these factors were to be taken into account we
\nwould expect the earnings differential to be reduced due to formal employees being employed in sectors
\nwith higher labour productivity and\/or in highly skilled occupations.
\nThe results from a linear regression on monthly earnings43 after taking into account differences in
\nthe above-mentioned characteristics as well as others44 actually reduces the counterfactual differential
\n42In reference to the Agreement on Textiles and Clothing.
\n43R2 = 0.3515, survey weights used.
\n44Other characteristics include the year of arrival to Turkey, the number of hours worked, the occupation, the sector,
\nthe company size and the sex of the worker.
\n33
\nSyrian refugees in the Turkish Labour Market
\nto 341 a month. This estimate should be taken with caution due to the existence of other pieces
\nof information -most notably, the extent of their personal and professional networks- that could be
\ncorrelated with employees\u2019 earnings and not taken into account in this regression. For instance, in the
\nparticular case where information on networks is omitted we would expect the earnings differential to
\nbe brought further down. Still, and with this caveat in mind, the result from the regression points
\nout at the optimistic scenario (which assumes an increase of 275 a month) as being the most likely
\nscenario to happen of the three presented for the formalization of the employment arrangements.
\nAccording to the salary prime associated to the formality of the work arrangements it looks like
\nformal means fully formal, i.e. employers do not misreport a massive amount of hours so as to lower
\nthe cost of the employee. This brings the discussion to the second dimension under analysis, what
\npercentage of informal Syrian employees can we expect to receive a work permit in the coming years?
\nOn the one hand, it will depend on the individual incentives of employees and employers; unfortunately,
\nas things stand no increases in the effort of these agents can be expected. This is particularly so for
\nemployers, particularly those operating in historically informal sectors.
\nFigure 13: Work permit, non-eligible Syrian refugees under temporary protection
\nTotal non-eligible: 68.9%
\nAnother province
\n10% rule
\n19.8% 19.1% 30.0%
\n\u2205
\nSource: DGMM, HLFS 2017 and author\u2019s own calculations. Notes: The Venn diagram shows the percentage of Syrian
\nrefugees informally employed as salaried workers that are not eligible to receive a work permit because 1) they reside in
\na different province to the one they are registered in -blue, left-, 2) they work in a company with less than 10 employees,
\nthus violating the 10% rule -red, right- or 3) both reasons at the same time -combined colour, center-.
\nAn even more important dimension, at least in practical terms, lies on whether Syrian refugees
\nthemselves are eligible to apply for a work permit. According to an UNHCR (2018) information
\nleaflet, Syrians covered by the temporary protection regime (96.6 per cent of the Syrians that came to
\nTurkey since the war) can only work in the province where they were initially registered. Moreover,
\nan employment quota whereby those on temporary protection cannot exceed 10 per cent of the total
\nnumber of nationals in their company is imposed to all but agricultural business. In practice, 311,55345
\nSyrian refugees (49.1 per cent) work in non-agricultural companies of less than 10 employees which
\n45This figure is the result of assuming that informally employed Syrian refugees who are not under the temporary
\nprotection regime work in similar size companies as the rest.
\n34
\nSyrian refugees in the Turkish Labour Market
\nautomatically makes them ineligible for a work permit. To make matters more complicated, 1,196,894
\nof Syrian refugees under temporary protection (38.9 per cent) are estimated to be living in provinces
\nother than the ones where they were initially registered by the authorities.46 Obviously, some of
\nthe Syrians affected by the 10 per cent quota might also be living in a different province, making
\nthe sum of the populations affected an inaccurate predictor of the percentage of Syrian unable to
\nobtain a work permit. Unfortunately, information on the province where the person was registered is
\nnot available in the microdata (the number is a macroeconomic estimate) and a cross-tabulation is,
\nthus, not available. The solution we propose is to assume that working in a company with less than 10
\nemployees is independent of living in a different province than the one where the person was registered.
\nThe percentage of ineligible refugees based on the above-mentioned assumption is shown in Figure
\n13. According to these estimates, 68.9 per cent of all Syrian refugees would not even be eligible to
\nobtain a work permit; 19.8 per cent because they are working in a province other than the one they
\nregistered in, 30.0 per cent because they work in a company with less than 10 employees and 19.1 per
\ncent because of both reasons at the same time. As a result, and provided the current conditions do
\nnot change, Turkey should not expect more than 31.1 per cent of the informal Syrian employees to be
\nformalized in the short-run<\/p>\n","protected":false},"excerpt":{"rendered":"
The macroeconomic impact of informal employment As of 2017, latest year for which data is available, we estimate that 657,385 (15+ years old) Syrian employees did not have a work permit in Turkey. The consequences of informal work arrangements for employees are well known29 and involve denied labour rights and benefits as well as a […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[54,63],"tags":[],"class_list":["post-614","post","type-post","status-publish","format-standard","hentry","category-headline","category-arastirma-3"],"yoast_head":"\n