The Impact of Payroll Taxes on Informality. The Case of the 2012 Colombian Tax Reform. Cristina Fernández Leonardo Villar

The Impact of Payroll Taxes on Informality. The Case of the 2012 Colombian Tax Reform. Cristina Fernández Leonardo Villar

Why this might be of interest to SSA and particularly to South Africa? Useful for South Africa in the discussion of increasing its payroll taxes. Useful for other countries in Africa that show high payroll taxes as the Cote d Ivoire, Sudan and Tanzania Methodology itself useful in analyzing the impact of labour policies and particularly (but not only) for those countries that do not household panel data.

Colombian Reform Characteristics. Contributions payable by the employer 3% 4% Honduras Chile 21% 20% 19% 19% 19% 19% 16% 16% 16% 17% 14% 14% 11% Peru Ecuador Guatemala Uruguay World LAC El Salvador Colombia Dominican R. Paraguay Bolivia Panama Nicaragua 25% Mexico* 32% 29% 29% Colombia 2013 Argentina Costa Rica 40% Brazil* Reduced payroll taxes from 29.5% to 16% of wages. It only affects contributions made by employers and not contributions made by workers. The fiscal source of the contributions was replaced by a profit tax under the assumption that it is better to tax the capital than work.

Reform characteristics. Despite this reduction, payroll taxes remain relatively high at international level. 45.0 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 Labor tax and contributions (% of profit) Brazil Costa Rica Argentina Colombia* Mexico Cote d'ivoire Panama Sudan Bolivia Paraguay Dominican R. Colombia Cameroon El Salvador Tanzania World Uruguay SSA Ghana Guatemala LAC Ecuador Congo D. R. Nigeria Uganda Peru Zambia Trinidad and T. Ethiopia South Africa Chile Kenya Source: World Bank * Before the Reform

3 HINTS ON THE IMPACT OF THE REFORM

Hint 1: Informality Declined. 65% 60% 55% 50% 45% 40% 2000 2002 Informality rates Different measurements. 2004 2006 2008 2010 2012 2014 Firm (new definition) Health contributions Pension contributions Millions of jobs 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 Health & pension contributions 4.0 Jan-09 Sep-09 Formal and informal workers. May-10 Jan-11 Sep-11 May-12 Jan-13 Sep-13 May-14 Jan-15 Sep-15 Formal Informal Source: Dane GEIH ECH 13 main metropolitan areas

Hint 2: Informality rates salaried workers and Salaried and employers 38% 37% 36% 35% 34% 33% 32% 31% 30% 29% 28% Mar-11 Informality rates: Self employment and other employment Jun-11 Sep-11 Dec-11 Mar-12 Jun-12 Sep-12 Dec-12 Mar-13 Jun-13 Sep-13 Dec-13 self-employers. Mar-14 Jun-14 Sep-14 Dec-14 85.0% 84.0% 83.0% 82.0% 81.0% 80.0% 79.0% 78.0% 77.0% 76.0% 75.0% Self - employment Salaried and employers Self employment Source: Dane GEIH ECH 13 main metropolitan areas

Source: GEIH y Fedesarrollo Hint 3: Relationship informality - growth Tasa de formalidad Formality Rate 54% 52% 50% 48% 46% 44% 42% 40% 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% -6.0% -8.0% Brecha Output del gap producto -10.0% Formality in Colombia is procyclical, except in the most recent period Correlations (2002-2013) (2002-2015) --- Formality Rate (OIT, 10). Informality Rate (OIT, 5). Output Gap 0.74** 0.46*

METHODOLOGY

Objetive: How much of the reduction in informality was due to the reform To isolate the impact of other variables affecting informality as: Growth A general tax increase, including the creation of the profit tax Increases in the state payroll (the share fell 3.9% to 3.7% ). The increase in the minimum wage (adjusted by productivity) which increased in 1.8 p.p. per year. (increases informality)

Differences in Differences. Before Treatment effect After It compares the change in the probability of being informally within the control group with the change in the probability of being informal in the treated group. Nets factors that simultaneously affects both groups such as growth.

Differences in Differences 30 Differences in Differences. Control cms Treatment cms Baseline 100 90 Follow up 120 140 Differences 20 50

In the case of the Colombian Reform. Treatment group (beneficiaries) Earn between 1 and 10 salaries. Do not work at universities / NGOs. More than two workers Control group (neutral) Earn less than the minimum wage or more than 10 minimum wages. University workers / NGOs. Self employment (alternatively excluded) Government and no reported income - excluded

Graphic Analysis. 28.0% Informality rates (13 metropolitan areas - 3 month MA) 80% 26.0% 78% Analysis period: Treated group 24.0% 22.0% 76% 74% Control group 2012 (before the reform) 2014 (after the reform). 20.0% 18.0% y = -0.000021x + 1.650880 y = -0.000017x + 0.944726 Approval!st waiver 2nd waiver 72% 70% Long moving averages (volatility). Mar-09 Aug-09 Jan-10 Jun-10 Nov-10 Apr-11 Sep-11 Feb-12 Treated informality rate Jul-12 Dec-12 May-13 Oct-13 Mar-14 Aug-14 Jan-15 Jun-15 Nov-15 No treated informality rate

Implementation of the econometric exercise. If you don t have a panel, create a panel, the model conducting this procedure is called differences in differences with matching developed by Heckman et al (1997).

Matching. Treatment 2012. Treatment 2014. P(t)=0.5 P(t)=0.4 P(t)=0.2 P(t)=0.51 Control 2012. Control 2014.

Matching. Treatment 2012. Treatment 2014. Control 2012. Control 2012.

RESULTS

Differences in Differences (13 areas). Including self-employment in the control group Control Informality Treatment Informality Baseline (2012) Follow Up (2014) 76% 28% 75% 23% Differences -1% -5% Differences in Differences -4.3% % treated 38% Impact of informality rate -1.6%

Differences in Differences (13 areas). Excluding self-employment in the control group Control Informality Treatment Informality Baseline (2012) Follow Up (2014) 61% 28% 62% 23% Differences 1% -5% Differences in Differences -6.8% % treated 38% Impact of informality rate -2.6%

Robustness Tests. Common Support (similar range of p-scores) Quality of Matching (Rubin s criteria hold) Placebo Test (2009-2012: no significant)

Results of past exercises. Anton (2014), the recent reform reduced the rate of informality between 2.9 p.p. and 3.4 p.p. Theoretical model. Kugler and Kugler (2009), an increase of 10 % in payroll taxes reduces formal employment between 4% and 5 %. Mondragón et al (2010), a 10% increase in payroll contributions increases the probability of being informal between 5 % and 8%. More recent work on the impact of recent reforms show that the number of formal workers increased between 3.1 and 3.4 p.p. (IDB) and 1 p.p 2014, 2 p.p. for 2016 and 5.6 p.p. 2018 (Central Bank-preliminar).

DISTRIBUTION EFFECT

Informality rate by income quintiles. They benefit more those who receive similar income to a minimum because a restriction carving more at this point where wages are not flexible downward shifted.

MDID by economic sector. Including self-employment Low Educated (Primary or less) Baseline (2012) Follow Up (2014) DID Control Treatment Control Treatment 92% 49% 92% 39% -10.4% High school 56% 25% 66% 22% -13% Tertiary education or higher Male 25-45 years 56% 14% 54% 11% -1.2% (n.s.) 75% 26% 75% 21% -5% Source: Own calculations, based on GEIH 2007-2015 *Male 25 45 years * Not significant. All the other results are significant 99%. Students with high-school education or less were most favored by the reform.

Conclusions. The 13.5 points reduction in payroll taxes decreased the informality rate in Colombia in between 4.3 and 6.8 p.p. that impacted the informality rate in between 1.6 and 2.6 points. Men in productive years with low levels of education (high school or less) were most favored by the reform.