From Fakes to Forced Labour

4. Testing the link: Econometric evidence on counterfeiting and labour abuse

Establishing a robust empirical link between illicit trade and labour abuse requires moving beyond descriptive statistics and anecdotal evidence. While previous chapters have shown clear correlations between poor labour standards and indicators of counterfeiting, the question remains whether these relationships persist once broader economic and governance variables are taken into account. The purpose of this chapter is therefore to test the statistical robustness of the observed associations using econometric techniques.

The underlying premise is that weak labour standard enforcement, the prevalence of forced or informal work, and low wage protection are directly connected to structural incentives that reduce production costs and facilitate illicit manufacturing and trade networks. To put it differently, in jurisdictions with poor law enforcement and a low regard for the law, labour abuses and IP infringement tend to coexist as criminals tend to disregard both. Conversely, countries with stronger labour rights, effective social dialogue, and higher minimum wages should—ceteris paribus—display lower exposure to counterfeit trade.

The econometric analysis presented here operationalises this logic by combining OECD‑EUIPO indicators of counterfeiting with a broad set of socio-economic and institutional variables, primarily sourced from the ILO and World Bank databases. The objective is to determine whether the presence of exploitative labour conditions is significantly associated with the intensity of counterfeit trade, after controlling for the level of economic development, trade openness, and institutional quality.

4.1. Modelling strategy and approach

The analytical framework relies on multiple linear regression models estimated through Ordinary Least Squares (OLS). For the purpose of this report, the analysis focuses on a single observation per country of origin for the year 2021. The dependent variable alternates between (i) the logarithm of the estimated value of counterfeit goods exported by a given economy and (ii) the logarithm of the number of customs seizures identifying that economy as a provenance of counterfeit products. These proxies collectively capture the magnitude and frequency of counterfeiting activities.

A systematic, iterative procedure was followed. The team first carried out exploratory data analysis to identify variable distributions, detect outliers, and assess potential multicollinearity. Subsequently, more than 50 model specifications were tested, combining different subsets of variables and transformations (logarithmic, quadratic) to ensure robustness. From this set, two models were ultimately selected for presentation in this report as they offered the best trade-off between explanatory power, statistical reliability, and interpretability. Other models generally confirmed the direction of the associations but added limited marginal insight; hence, for clarity and parsimony, only the two most representative are discussed below.

Diagnostic tests: Each model was subjected to tests for homoscedasticity, normal distribution of residuals, and absence of multicollinearity. Multicollinearity diagnostics (Variance Inflation Factor, VIF) guided the removal of redundant variables. Measures such as adjusted R² and F-statistics were also considered to provide an indication of overall fit and explanatory strength, alongside these other diagnostic checks.

Model selection: The two final models presented below were chosen for their interpretability, stability of coefficients, and statistical robustness. Although other specifications yielded similar trends, the two selected regressions most clearly illustrate the link between forced labour and counterfeit trade.

The model specification can be expressed generically as:

(Illicit Trade Indicator)_i = β0 + β1(Labour Variables)_i + β2(Control Variables)_i + Ꜫ_i

where i indexes the country. All continuous variables were transformed into logarithms where appropriate to normalise distributions and reduce heteroskedasticity.

4.1.1. Control variables

To isolate the effect of labour exploitation on counterfeiting, a set of control variables was included to account for differences in economic structure, institutional quality, and trade exposure.

• Exports value which represents the total monetary value of goods exported by a reporting economy (expressed in USD), as recorded in the United Nations Comtrade Database. It is based on official data reported by national statistical authorities and harmonised by the UN Statistics Division to ensure international comparability. This measures trade openness and global market integration. Economies with higher export volumes are more exposed to global supply chains and thus more likely to appear in customs seizure records—either as legitimate exporters or as transit hubs exploited by counterfeiters. For the purposes of the analysis, the logarithm of export values was used.

• GDP per capita expressed in constant 2015 USD, based on the World Bank’s World Development Indicators (WDI) database controls for the level of economic development. Wealthier economies typically display stronger regulatory frameworks and enforcement capacity, though they may still act as transit points for illicit goods.

• Population covered by at least one social benefit, produced by ILO measures the percentage of the total population that is legally covered by at least one social protection benefit. This indicator captures the breadth of social protection. Broader welfare coverage can reduce vulnerability to exploitation by offering safety nets that prevent coercive labour conditions.

• Poverty gap, produced by the World Bank, measures the intensity of poverty by calculating how far on average, the poor are from the poverty line. This reflects the depth of poverty below the national poverty line. Higher poverty gaps increase economic vulnerability and can push workers into informal or exploitative employment, indirectly supporting illicit production networks.

• Rule of Law Index from the World Bank’s Worldwide Governance Indicators (WGI) measures the extent to which agents have confidence in and abide by the rules of society, particularly the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence. Scores range from –2.5 (weak) to +2.5 (strong) governance performance. This indicator serves as a proxy for institutional quality and enforcement effectiveness. Strong rule of law is expected to constrain both forced labour and counterfeit production.

• Unemployment rate, which measures the share of the labour force that is unemployed, expressed as a percentage of the total labour force. It is sourced from the World Bank’s World Development Indicators (WDI) database and is based on national labour force surveys and official statistics: Reflects labour market pressure. High unemployment can weaken workers’ bargaining power and increase susceptibility to exploitation.

• The Statutory Nominal Gross Monthly Minimum Wage, published by the International Labour Organization (ILO), refers to the legally mandated minimum wage set by national legislation or regulations, expressed in nominal gross monthly terms. This indicator is used to capture the effect of wage protection. For the purposes of the analysis, the inverse of the minimum wage was used, which means that a lower minimum wage (higher inverse value) indicates cheaper labour and potentially stronger incentives for illicit production.

Collectively, these controls ensure that the estimated link between counterfeiting and labour exploitation is not simply an artefact of broader development differences or institutional weakness.

4.1.2. Regressions

This section presents the two most relevant regression models.

Model 1, presented in Table 4.1, examines how forced labour prevalence and wage conditions correlate with the scale of counterfeit trade, controlling for trade openness and income levels.

Results show that exports value (log) is strongly and positively associated with the value of counterfeit goods – confirming that economies integrated in global trade are also those where illicit goods can more easily circulate. GDP per capita and its squared term are statistically insignificant, although their signs suggest a mild inverted-U pattern – counterfeiting may rise with early stages of industrialisation but decline once governance capacity strengthens.

Table 4.1. Regression of the value of fakes (in log), Model 1

Dependent variable: Value of fakes (in log)
Exports value (log)	0.9260321***
	(0.146281)
GDP per capita	0.0000212
	(0.000034)
GDP per capita²	-0.0091348
	(0.010227)
Forced labour (%)	0.0000764***
	(0.0000249)
Minimum wages (inverse)	0.1192676***
	(0.012234)
Constant	-2.392026
	(3.200848)

Note: Data are shown on a logarithmic scale to better compare values that differ widely in magnitude. Number of observations: 66. Standard errors in parentheses; * p<0.1, ** p<0.05, *** p<0.01.

Crucially, the forced labour variable is positive and highly significant. This indicates that countries with higher prevalence of forced labour tend also to export higher values of counterfeit goods. The coefficient magnitude, though small in absolute terms, is economically meaningful, implying that even modest increases in forced labour prevalence are associated with measurable rises in counterfeit trade intensity. Indeed, the coefficient associated with forced labour indicates that, all else being equal, a one-percentage-point increase in the share of forced labour in an economy is linked to an estimated 0.0076% rise in the value of counterfeit trade. The minimum wage (inverse) variable also shows a strong positive and significant effect, demonstrating that lower wage protections correlate with greater counterfeit activity. This finding reinforces the notion that cost-cutting through labour exploitation is a key competitive strategy in illicit production.

In summary, Model 1 highlights two key drivers of counterfeit export intensity: participation in global trade networks and the prevalence of exploitative labour conditions.

Model 2, presented in Table 4.2, relies on the same dependent variable as in Model 1, but incorporating informal employment as an additional explanatory variable.

Table 4.2. Regression of the value of fakes (in log), Model 2

Dependent variable: Value of fakes (log)
Exports value (log)	0.7230599***
	-(0.160243)
GDP per capita	0.000016
	(0.0000525)
GDP per capita²	-0.0001556
	(0.020829)
Forced labour (%)	0.0000727***
	(0.0000247)
Informal employment (%)	0.0026392
	(0.010231)
Minimum wages (inverse)	0.1216203***
	(0.016562)
Constant	1.821787
	(2.92283)

Note: Data are shown on a logarithmic scale to better compare values that differ widely in magnitude. Number of observations: 56. Standard errors in parentheses; * p<0.1, ** p<0.05, *** p<0.01.

The introduction of informal employment allows to test whether the size of the unregulated labour market contributes to counterfeiting beyond the effects of forced labour. Although the coefficient for informal employment is positive (as theoretically expected), it is not statistically significant in this sample. Nonetheless, its inclusion leaves the other coefficients virtually unchanged – an indication of model robustness.

The persistence of the strong and significant coefficients for forced labour and low minimum wages confirms that the observed relationship is not merely a reflection of informality, but rather of structural coercion and the absence of labour protection. The stability of results across models suggests a consistent pattern: exploitative labour practices and weak wage enforcement remain key correlates of counterfeit trade, even when broader economic conditions are controlled for.

Across all specifications, forced labour emerges as the most consistent and statistically robust determinant of just how intense counterfeiting is. The positive relationship persists across different proxies, model forms, and control sets. The interpretation is straightforward: economies in which forced labour is more prevalent are systematically more likely to engage in – or be used as platforms for – the production or distribution of counterfeit goods.

Lower minimum wages and weak labour protections further amplify this association, as they lower production costs and create environments conducive to illicit manufacturing. The lack of significance for GDP per capita indicates that income level alone does not explain differences in counterfeiting exposure; institutional and labour governance factors matter more.

These findings align with broader qualitative evidence showing that illicit producers frequently rely on coerced or underpaid labour to maintain profitability in the absence of economies of scale or technological advantage.

The econometric analysis supports the hypothesis that labour exploitation is closely associated with illicit trade in counterfeit goods. Forced labour stands out as the strongest predictor, followed by wage repression and weak institutional enforcement. The data suggest a statistically consistent relationship, even after controlling for economic size, openness, and governance indicators.

Together, the results suggest that tackling forced labour and strengthening labour rights are not only human-rights imperatives, but essential components of trade-integrity and anti-counterfeiting strategies. Policies that improve wage protection, expand social coverage, and enhance enforcement capacity can simultaneously reduce both human exploitation and illicit trade.