Shadow Economy Drivers in Bosnia and Herzegovina: A MIMIC and SEM Approach

Comments 1-3: First three comments referring to Abstract

1. The abstract does not match the scientific abstract requirements

2. Method, technique, important findings …ect, missing information into the abstract

3. Author/s need to review some empirical studies to follow the instructions to how to write an

abstract, and what must be included.

Response 1-3:

Agree with the comments and therefore we revised the abstract (11 to 26 line):

This study explores the drivers and evolution of the shadow economy in Bosnia and Herzegovina,  a transitional, post-conflict country facing persistent institutional fragility. Using the Multiple Indicators and Multiple Causes (MIMIC) model, an extension of Structural Equation Modeling, the paper estimates the size and dynamics of the shadow economy from 1996 to 2022. The model integrates macroeconomic indicators (employment rate, GDP per capita, tax revenues) and institutional variables (rule of law, control of corrup-tion), with data primarily sourced from the World Bank. The results show that institutional quality, tax burden, and labor market conditions are significant determinants of the informal sector. The model demonstrates strong statistical validity (CFI = 0.986, RMSEA = 0.05), supported by robustness checks including unit root tests, structural break analysis, and the exclusion of controversial benchmarking methods. The shadow economy responds markedly to major shocks such as the 2008 global financial crisis and the 2014 floods. Findings provide valuable policy insights: strengthening institutions, simplifying tax systems, and encouraging formal labor market participation can significantly reduce informality. The study supports evidence-based reforms to enhance transparency, resilience, and sustainable development in Bosnia and Herzegovina.

Comments 4-7:

·        The introduction stated some paragraphs need to be in the literature section

·        Author/s stated the methodology section here!

·        The authors need to declare the motivation for this study because no explanation is provided in this regard. The explanation has provided is not enough.

·        The contribution of this study is not clear!

Response 4-7:

Agree with the comments and therefore we revised introduction section:

(47 to 71 line)

The shadow economy (SE) - comprising legal but unregulated, untaxed, or unreported economic activities - presents a dual challenge for policymakers. While it can serve as a coping mechanism for individuals excluded from the formal labor market, it simultaneously undermines public finances, distorts competition, and erodes institutional trust. These tensions make the SE a particularly pressing issue in recent transitional economies, or one that have specific lag effect of transition processes, where institutional frameworks are often underdeveloped and enforcement is inconsistent.

Bosnia and Herzegovina (BiH) exemplifies this challenge, its one of the largest shadow economies in the Western Balkans. Although this is not unique in the region, BiH’s situation is exacerbated by its fragmented post-conflict governance structure, which has led to uneven regulatory enforcement and weak institutional oversight. While this context is important, the focus of this study is on how these governance weaknesses - alongside high tax burdens, labor market rigidities, and corruption - con-tribute to the persistence of informality. Sectors such as construction, retail, and agriculture are particularly prone to shadow economic activity in BiH. These sectors are also central to employment and GDP, making their vulnerability to informality especially problematic. The shadow economy in these areas not only reduces tax revenues but also weakens the formal economy’s structural integrity. Moreover, external shocks, including the 2008 global financial crisis, the 2014 floods, and the COVID-19 pandemic—have it reflection on informal activity, pushing more individuals and businesses outside the formal system.

Compared to neighboring countries like Serbia and Croatia, which have implemented targeted tax and regulatory reforms to reduce informality, BiH has lagged behind. This regional contrast underscores the need for updated country-specific empirical research to inform effective policy responses.

(118 to 142 line):

Despite the scale and implications of the SE in BiH, empirical studies remain limited—largely due to data constraints and the methodological challenges of measuring informal activity. This study addresses that gap by applying the Multiple Indicators and Multiple Causes (MIMIC) model, a well-established extension of Structural Equation Modeling (SEM), to estimate the size and evolution of the shadow economy in BiH from 1996 to 2022. The MIMIC model is particularly well-suited for this context because it treats the shadow economy as a latent variable—unobservable directly but inferable through observable causes (e.g., tax burden, unemployment) and indicators (e.g., institutional quality, corruption). This approach allows for a more nuanced and statistically robust analysis, especially when direct measurement is not feasible.

Furthermore, our approach tries to avoid common pitfalls such as arbitrary benchmarking and addresses concerns about variable stationarity and model specification. By doing so, it provides a more reliable foundation for understanding the drivers and dynamics of informality in BiH.

This paper contributes in three key ways:

·        It offers updated, methodologically rigorous estimates of the shadow economy in BiH.

·        It integrates both institutional and macroeconomic drivers using a latent variable framework.

·        It provides evidence-based insights to inform policy reforms aimed at reducing informality and strengthening governance.

This study is guided by the central research question: What are the main drivers and how have the dynamics of the shadow economy in Bosnia and Herzegovina evolved over time—and how can this understanding inform more effective, context-specific policy responses in transitional economies like BiH?

Comments 8-13:

1. Author/s must follow the literature of empirical study, it is unclear

2. Again, many statements explaining the methodology

3. Hypothesis of this study are not clear

4. Author needs to link this study with theories

5. The author (s) should recast the explanation of the research contributions to the body of

knowledge. Precise explanations need to be made on how this research advances the

literature.

Response 8-12 (171 to 228 line):

Agree with the comments and therefore we revised the literature review section:

Unformal sector comprising legal but unregulated and untaxed economic activities, remains a persistent challenge in economies that struggled and struggle with transition for decades. This struggle is immanent for Western Balkan (WB) countries and BiH is one of them.

In BiH, it is estimated to account for approximately one-third of GDP (Pasovic & Efendic, 2018), yet empirical research on its scope and drivers remains limited. This study aims to address that gap by applying MIMIC model to estimate the size and evolution of the SE in BiH from 1996 to 2022.

Several methods have been developed to estimate the SE, including direct approaches (e.g., surveys, audits) and indirect methods (e.g., currency demand, electricity consumption, labor market discrepancies). While these methods offer valuable insights, they often suffer from data limitations, respondent bias, or narrow applicability (Feige, 1979; Tanzi, 1980). In contrast, the MIMIC model—an extension of SEM—treats the SE as a latent variable inferred from observable causes (e.g., tax burden, unemployment) and indicators (e.g., institutional quality, corruption). This makes it particularly suitable for contexts like BiH, where informal activity is widespread and direct measurement is difficult (Schneider & Enste, 2002; Schneider, Buehn, & Montenegro, 2010).

The MIMIC model has been widely applied in developed and developing countires, including France (Buehn & Schneider, 2008), Jordan (Alfoul et al., 2022), Egypt (Mai & Schneider, 2016), and Serbia (Krstić et al., 2015). More recent studies have extended its use to transitional and developing economies, emphasizing its adaptability to complex institutional environments. For instance, Wang et al. (2024) examined the role of governance indicators in shaping the SE across developing countries, while Hassan (2024) explored how regulatory quality affects informal financial activity in Southern Africa. Similarly, Abou Ltaif et al. (2024) investigated the SE in Lebanon amid financial crisis, highlighting the interplay between global shocks and local institutional weaknesses.

Despite its strengths, the MIMIC model has faced methodological critiques. Scholars have questioned its reliance on arbitrary benchmarking to produce point estimates, its sensitivity to variable selection, and its vulnerability to non-stationary time series data (Breusch, 2016; Pissarides & Weber, 1989). Smrčková and Brůna (2024) further caution against implausible estimates when theoretical definitions and calibration benchmarks are misaligned. This study addresses these concerns by avoiding benchmarking, incorporating structural breaks, and rigorously testing for stationarity. These refinements enhance the model’s robustness and credibility, particularly in a transitional setting like BiH.

The theoretical foundation of this study draws on institutional theory and tax compliance theory. Institutional theory posits that weak governance—characterized by low rule of law and high corruption—creates incentives for informal activity (Johnson, Kaufmann, & Zoido-Lobatón, 1998; Schneider, 2010). Tax compliance theory suggests that high tax burdens, complex regulations, and inconsistent enforcement discourage formalization and push economic actors into the informal sector (Arsic et al., 2015; Bordignon & Zanardii, 1997). These frameworks are especially relevant in BiH, where fragmented governance and regulatory inefficiencies have historically contributed to informality.

Earlier studies on BiH’s shadow economy (e.g., Pasovic & Efendic, 2018) provided foundational insights but were limited by outdated data and did not account for recent structural shocks such as the 2008 global financial crisis, the 2014 floods, or the COVID-19 pandemic. This study fills that gap by using updated data from 1996 to 2022 and explicitly modelling structural breaks to capture the impact of these events.

Based on the reviewed literature and theoretical considerations, this study explores two central propositions. First, higher tax burdens are expected to correlate with increased informal economic activity, as individuals and firms seek to avoid excessive fiscal pressure. Second, the quality of institutional governance—particularly in areas such as rule of law and corruption control—is anticipated to significantly influence the extent of informality. These propositions guide the empirical investigation into how macro-institutional conditions shape the dynamics of the shadow economy in BiH.

This research contributes to the literature in several keyways. It provides updated empirical estimates of the SE in BiH over a long-time horizon, capturing the effects of major structural events. Furthe, it refines the MIMIC methodology by addressing known limitations, including variable stationarity and benchmarking. It also contributes to theory-building by linking empirical findings to institutional and tax compliance theories. Finally, it offers policy-relevant insights for BiH and other similar economies seeking to reduce informality and strengthen governance.

References used in literature review

References (red colour references are subsequently added)

1.      Abou Ltaif, S. F., Mihai-Yiannaki, S., & Thrassou, A. (2024). Lebanon’s economic development risk: Global factors and local realities of the shadow economy amid financial crisis. Risks, 12(8), 122. https://doi.org/10.3390/risks12080122

2.      Alfoul, M. A., Mishal, Z. A., Schneider, F., Magableh, K., & Alabdulraheem, A. R. (2022). The hidden economy in Jordan: A MIMIC approach. Cogent Economics & Finance, 10(1), 2031434. https://doi.org/10.1080/23322039.2022.2031434

3.      Arsic, M., Arandarenko, M., Radulov, B., & Jankovic, I. (2015). Causes of the shadow economy. In G. Krstic & F. Schneider (Eds.), Formalizing the shadow economy in Serbia: Policy measures and growth effects (pp. 21–46). Springer. https://doi.org/10.1007/978-3-319-13437-6_4

4.      Bordignon, M., & Zanardii, A. (1997). Tax evasion in Italy. Giornale degli Economisti e Annali di Economia, 56(2), 169–210. http://www.jstor.org/stable/23248315

5.      Breusch, T. (2016). Estimating the underground economy using MIMIC models. Canadian Tax Journal, 64(1), 41–73. https://journals.docuracy.co.uk/jota/article/view/137/161

6.      Buehn, A., & Schneider, F. (2008). MIMIC models, cointegration and error correction: An application to the French shadow economy. CESifo Working Paper No. 2200. https://www.econstor.eu/handle/10419/26245

7.      Feige, E. L. (1979). How big is the irregular economy? Challenge, 22(5), 5–13. https://doi.org/10.1080/05775132.1979.11470559

8.      Fahad, S., Almawishir, N., & Benlaria, H. (2023). Using the PLS-SEM model to measure the impact of the knowledge economy on sustainable development in the Al-Jouf region of Saudi Arabia. Sustainability, 15(8), 6446. https://doi.org/10.3390/su15086446

9.      Gërxhani, K., & Cichocki, M. (2023). Informal institutions and the shadow economy in Eastern Europe. Journal of Institutional Economics, 19, 1–18. https://doi.org/10.1017/S1744137423000032

10.   Hassan, A. (2024). Mobile financial services and the shadow economy in Southern African countries: Does regulatory quality matter? International Journal of Financial Studies, 12(4), 115. https://doi.org/10.3390/ijfs12040115

11.   Johnson, S., Kaufmann, D., & Zoido-Lobatón, P. (1998). Regulatory discretion and the unofficial economy. American Economic Review, 88(2), 387–392. https://www.jstor.org/stable/116953

12.   Krstić, G., Schneider, F., Arsić, M., & Ranđelović, S. (2015). What is the extent of the shadow economy in Serbia? In G. Krstic & F. Schneider (Eds.), Formalizing the shadow economy in Serbia (pp. 47–77). Springer. https://doi.org/10.1007/978-3-319-13437-6_5

13.   Mai, H., & Schneider, F. (2016). Modeling the Egyptian shadow economy: A currency demand and a MIMIC model approach. Journal of Economic and Political Economy, 3(2), 309–333. https://www.proquest.com/docview/1814929206

14.   Pasovic, E., & Efendic, A. (2018). Informal economy in Bosnia and Herzegovina. South East European Journal of Economics and Business, 13(2), 112–125. https://doi.org/10.2478/jeb-2018-0015

15.   Pissarides, C. A., & Weber, G. (1989). An expenditure-based estimate of Britain’s black economy. Journal of Public Economics, 39(1), 17–32. https://doi.org/10.1016/0047-2727(89)90052-2

16.   Schneider, F. (2010). The influence of public institutions on the shadow economy: An empirical investigation for OECD countries. European Journal of Law and Economics, 6(3), 441–468.

17.   Schneider, F., & Enste, D. H. (2002). The shadow economy: Theoretical approaches, empirical studies, and political implications. Cambridge University Press.

18.   Schneider, F., Buehn, A., & Montenegro, C. E. (2010). New estimates for the shadow economies all over the world. International Economic Journal, 24(4), 443–461. https://doi.org/10.1080/10168737.2010.525974

19.   Smrčková, A., & Brůna, K. (2024). Implausible results in MIMIC-based shadow economy estimates: A critical review and methodological recommendations. Economic Modelling, 131, 106–118. https://doi.org/10.1016/j.econmod.2024.106118

20.   Tanzi, V. (1980). The underground economy in the United States: Estimates and implications. PSL Quarterly Review, 33(135), 427–453. https://rosa.uniroma1.it/rosa04/psl_quarterly_review/article/view/12996

21.   Wang, Y., Antohi, V. M., Fortea, C., Zlati, M. L., Mohammad, R. A., Abdelkhair, F. Y. F., & Ahmad, W. (2024). Shadow economy and environmental sustainability in global developing countries: Do governance indicators play a role? Sustainability, 16(22), 9852. https://doi.org/10.3390/su16229852

Comments 13: Sample selection should be justified.

Response 13: [Type your response here and mark your revisions in red]

Thank you for pointing this out. We agree with this comment. The study utilizes a time series dataset from Bosnia and Herzegovina (BiH) covering the period 1996 to 2022, with data sourced from the World Bank.

Therefore, we have conceptually summarized the rationale for our sample selection in the data section as follows:

(72 to 76 line)
First, Bosnia and Herzegovina (BiH) represents a post-conflict, transitional economy characterized by a complex institutional framework and a persistently large shadow economy - estimated at 30–40% of GDP.

 (77 to 83 line)

Second, data prior to 1996 are unavailable due to the establishment of BiH as a sovereign entity following the Dayton Peace Agreement, which marked the beginning of its current institutional and economic configuration.

(84 to 89 line)
Third, in the post-COVID period, a notable number of individuals and entities began declaring previously unreported income or employment to access government subsidies and relief programs. This phenomenon introduced a temporary formalization effect that extended over more than five quarters, making it a critical period for inclusion.

We have added the following sentence (84 to 89 line):

“BiH, a post-conflict transitional economy with a complex institutional structure and a historically large shadow economy (30–40% of GDP), presents a unique case for analysis, particularly as the post-COVID period saw widespread declaration of previously unreported income due to subsidy eligibility—creating a temporary formalization effect—while data prior to 1996 are unavailable due to the country’s establishment following the Dayton Peace Agreement.”

Comments 14: Measurements of variables should be supported with prior studies or theoretical arguments.

Thank you for pointing this out. We agree with this comment. Therefore, we have used this approach in our model specification (462 to 481 line):

·        Employment rate (emplBH), GDP per capita (GDP_BH), and tax revenues (taxesBH) are widely recognized macroeconomic indicators in shadow economy research, theoretically linked to informality through labor market dynamics, income levels, and tax compliance behavior, as supported by studies such as Schneider & Enste (2002) and Krstić et al. (2015).

·        Rule of law (ruleBH) and control of corruption (CCB) serve as proxies for institutional quality and are widely recognized in the literature as key determinants of informality—since weak institutions often correlate with higher shadow economy activity—as emphasized by Williams & Schneider (2016) and Schneider (2010).

·        We refer in our paper to the work of Buehn & Schneider (2008), who in some model specifications use corruption and legal system quality as primary indicators.

To be specific we have added the following sentence into our Methodology section (478 to 481 line):

“Numerous studies justify the use of a limited number of indicators in MIMIC models, and some—such as Buehn and Schneider (2008)—demonstrate that even just two institutional indicators, like corruption and legal system quality, can effectively capture the dynamics of the shadow economy.”

Buehn, A., & Schneider, F. (2008). MIMIC Models, Cointegration and Error Correction: An Application to the French Shadow Economy. CESifo Working Paper No. 2200, Category 1: Public Finance. Munich: CESifo Group.

Comments 15: Author needs to link this study with theories (419 to 430 line)

Thank you for pointing this out. We agree with this comment. Therefore, we have employed the MIMIC model, a well-established extension of Structural Equation Modeling (SEM). The methodological choice is further validated by its extensive use in prior studies, including Buehn and Schneider (2008), who applied the MIMIC model with cointegration and error correction techniques to the French shadow economy, and demonstrated that even a limited set of institutional indicators—such as corruption and legal system quality—can yield robust results. Similarly, Breusch (2016) critically examined the assumptions and transformations in MIMIC applications, reinforcing the importance of methodological rigor. More recent applications, such as those by Alfoul et al. (2022) for Jordan and Mai and Schneider (2016) for Egypt, confirm the model’s adaptability across diverse economic contexts. These studies collectively support the continued relevance of the MIMIC framework in contemporary research on informal economies, particularly when data constraints or structural complexities necessitate parsimonious yet theoretically grounded modeling strategies.

Comments 16: “A latent variable is one that cannot be directly measured but can be indirectly identified through indicators”. Cited please? (382 to 383 line)

Thank you for pointing this out. We agree with this comment.

The statement:

“A latent variable is one that cannot be directly measured but can be indirectly identified through indicators”

This a foundational concept in SEM and latent variable theory. While the exact wording may vary, this idea is widely supported in the literature. One of many reliable citations for this concept could be:

Bollen, K. A. (1986). Structural Equations with Latent Variables. New York: Wiley.

Comments 17: Paragraph started from line 255 – 259, talking about theory! Which theory?

Thank you for pointing this out. We agree with this comment

The paragraph beginning at line 255 discusses the role of theory in the context of Structural Equation Modeling (SEM). Specifically, it states (400 to 410 line):

“Theory plays a crucial role in SEM by guiding model specification, hypothesis testing, and ensuring construct validity. It helps researchers understand the meaning of relationships between variables and assess model fit. Theory also supports the generalizability of findings and fosters the development of new models. Without a strong theoretical foundation, SEM analyses may lack coherence and relevance…….”

The theory referred to here is not a specific economic or sociological theory (e.g., institutional theory or labor market theory), but rather the theoretical framework underlying SEM itself. This includes:

• Latent variable theory: The idea that unobservable constructs (like the shadow economy) can be inferred from observable indicators.
• Measurement theory: Ensuring that indicators validly and reliably reflect the latent constructs.
• Causal modeling theory: The specification of directional relationships between variables based on theoretical assumptions.

4. Response to Comments on the Quality of English Language

Point 1:

Response 1:    Corrected.

5. Additional clarifications

Discussion (563 to 599 line)

This study contributes to the literature on the shadow economy by applying MIMIC model to BiH economical context. BiH is a country marked by a complex post-conflict institutional structure and widespread informality. The findings of this analysis are reflecting its aim to be consistent with institutional theory and tax compliance theory, both of which emphasize the influence of weak governance and high tax burdens on informal economic activity. The results reinforce these theoretical perspectives in certain extent, demonstrating that institutional quality and corruption control are significant determinants of the size and behavior of the shadow economy in BiH.

The model captures how the informal sector responds to both internal institutional weaknesses and external shocks. On the other hand, results do not offer conventional precision of models that have traditional regression framework, although in context of  BiH’s statistical such precision has huge potential of being significantly misleading.

Structural break analysis conducted using the Bai and Perron methodology (Bai, 1994; Bai & Perron, 1998), identifies key disruptions in BiH’s economic trajectory, including the post-war recovery period, the delayed impact of the global financial crisis, and the 2014 floods. These events are not some sort of technical statistical anomalies but represent real-world shocks that altered the incentives for informal economic activity. We should mention, that similar patterns have been observed in related studies (e.g., Pasovic & Efendic, 2018; Buehn & Schneider, 2008; Albala & Jose, 2015), underscoring the importance of institutional resilience in mitigating the expansion of informality during crises.

The informal sector of an economy, is best conceptualized as a responsive subsystem within the broader economic framework. It tends to expand when formal institutions are weak or disrupted and contracts when governance improves and economic stability is restored. This dynamic interplay highlights the need for integrated and adaptive policy responses.

In, this study we utilize time series that required transformation to achieve stationarity. The variables were found to be integrated of order two (I(2)), necessitating second differencing to ensure statistical validity. This step, emphasized by Breusch (2016), is critical for maintaining the integrity of MIMIC model estimates. Nevertheless, it is not unusual that some authors avoid this step.

Unlike many studies, this analysis deliberately avoids benchmarking the shadow economy as a percentage of GDP—a practice often criticized for introducing interpretive ambiguity and potential confusion (Giles & Tedds, 2002). Instead, the focus is placed on relative trends, which provide a clearer picture of the shadow economy’s evolution over time.

The latent variable derived from the model does not exhibit a deterministic trend, which aligns with the expectations of structural equation modeling and supports the internal coherence of the results. This further strengthens the robustness of the findings, particularly given the challenges associated with modeling informal economic activity.

Conclusion (643 to 671 line)

From a policy perspective, the implications are simple and straightforward, although the nature of SE constrains precisions that could be assumed in context of conventional macroeconometric aggregates. First, improving institutional quality—particularly in areas related to the rule of law and anti-corruption—can reduce the incentives for informality. Second, simplifying tax regulations and reducing the overall tax burden, especially for small and medium-sized enterprises, may encourage formalization. Third, labor market reforms that promote formal employment opportunities can help absorb workers currently operating outside the legal framework. Additionally, enhancing crisis preparedness and ensuring timely institutional responses during economic shocks are essential to prevent surges in informal activity.

Despite its strengths, the study has several limitations. It relies on secondary data produced in challenging statistical surrounding, which may be subject to measurement errors or inconsistencies, particularly in a fragmented administrative context such as BiH. The country’s statistical infrastructure is complex, with two statistical institutes operating at the entity level and one at the state level. Further, the use of latent variables, while methodologically sound, introduces a degree of abstraction that may obscure some nuances of informal behavior. Furthermore, the model does not account for all external influences, such as remittances, cross-border trade, or regional spillovers. Future research could expand the model to include these factors and explore cross-entity comparisons within BiH, although this could assume complex econometric toolset (e.g. Dynamic Stochastic Growth Equilibrium model) and fully developed national account statistical framework on state level. Fact that there are no input-output tables for BiH speaks for itself.

Nevertheless, this study aims to offer a relatively robust and accessible framework for understanding the drivers and dynamics of the shadow economy in BiH. Attempt to link empirical findings to theoretical insights and highlighting actionable policy recommendations, is challenging task, but in summary this analysis provides a solid foundation for informed decision-making aimed at managing informality and strengthening economic governance.

Comments 1-6: First six comments referring to Abstract

·        Explicitly state the primary question or objective of the study

·        Provide specific results or estimates.

·        Briefly mention how the MIMIC model was validated or how robustness was tested.

·        Mention how the findings can inform specific policies

·        Reinforce the practical implications and how the study fills gaps in the existing literature or provides insights for policymaking.

.

Response 1-6:

Agree with the comments and therefore we revised the abstract (11 to 26 line):

This study explores the drivers and evolution of the shadow economy in Bosnia and Herzegovina - a transitional, post-conflict country facing persistent institutional fragility. Using the Multiple Indicators and Multiple Causes (MIMIC) model, an extension of Structur-al Equation Modeling, the paper estimates the size and dynamics of the shadow economy from 1996 to 2022. The model integrates macroeconomic indicators (employment rate, GDP per capita, tax revenues) and institutional variables (rule of law, control of corruption), with data primarily sourced from the World Bank. The results show that institutional quality, tax burden, and labor market conditions are significant determinants of the informal sector. The model demonstrates strong statistical validity (CFI = 0.986, RMSEA = 0.05), supported by robustness checks including unit root tests, structural break analysis, and the exclusion of controversial benchmarking methods. The shadow economy responds markedly to major shocks such as the 2008 global financial crisis and the 2014 floods. Findings provide valuable policy insights: strengthening institutions, simplifying tax systems, and encouraging formal labor market participation can significantly reduce informality. The study supports evidence-based reforms to enhance transparency, resilience, and sustainable development in Bosnia and Herzegovina.

Comments 7-13:

·        Explicitly state the central research question at the end of the introduction.

·        Explain why this research is particularly important for BiH. Emphasize the lack of existing studies specifically focusing on BiH’s shadow economy and why the MIMIC approach is a suitable tool.

·        Strengthen the explanation of why the MIMIC model is appropriate for this study. Discuss how it helps address the gaps in understanding the drivers and dynamics of the shadow economy in BiH.

·        Condense the discussion of the political and governance context. While important, the level of detail on the post-conflict situation and governance structure could be more concise to maintain focus on the shadow economy itself.

·        More clearly link the mentioned drivers (e.g., tax burdens, institutional weaknesses, corruption) to the shadow economy. This will clarify how these factors contribute to the persistence of informality in BiH.

·        Expand on the regional comparison, particularly with neighboring countries, to show how BiH's shadow economy compares and what lessons can be drawn from other countries in similar situations.

Response 7-13:

Agree with the comments and therefore we revised the introduction:

    (47 to 71 line)

The shadow economy (SE)—comprising legal but unregulated, untaxed, or unreported economic activities—presents a dual challenge for policymakers. While it can serve as a coping mechanism for individuals excluded from the formal labor market, it simultaneously undermines public finances, distorts competition, and erodes institutional trust. These tensions make the SE a particularly pressing issue in recent transitional economies, or one that have specific lag effect of transition processes, where institutional frameworks are often underdeveloped and enforcement is inconsistent.

Bosnia and Herzegovina (BiH) exemplifies this challenge, its one of the largest shadow economies in the Western Balkans. Although this is not unique in the region, BiH’s situation is exacerbated by its fragmented post-conflict governance structure, which has led to uneven regulatory enforcement and weak institutional oversight. While this context is important, the focus of this study is on how these governance weaknesses - alongside high tax burdens, labor market rigidities, and corruption - con-tribute to the persistence of informality.

Sectors such as construction, retail, and agriculture are particularly prone to shadow economic activity in BiH. These sectors are also central to employment and GDP, making their vulnerability to informality especially problematic. The shadow economy in these areas not only reduces tax revenues but also weakens the formal economy’s structural integrity. Moreover, external shocks, including the 2008 global financial crisis, the 2014 floods, and the COVID-19 pandemic—have it reflection on informal activity, pushing more individuals and businesses outside the formal system.

Compared to neighboring countries like Serbia and Croatia, which have implemented targeted tax and regulatory reforms to reduce informality, BiH has lagged behind. This regional contrast underscores the need for updated country-specific empirical research to inform effective policy responses.

(118 to 142 line):

Despite the scale and implications of the SE in BiH, empirical studies remain limited—largely due to data constraints and the methodological challenges of measuring informal activity. This study addresses that gap by applying the Multiple Indicators and Multiple Causes (MIMIC) model, a well-established extension of Structural Equation Modeling (SEM), to estimate the size and evolution of the shadow economy in BiH from 1996 to 2022. The MIMIC model is particularly well-suited for this context because it treats the shadow economy as a latent variable—unobservable directly but inferable through observable causes (e.g., tax burden, unemployment) and indicators (e.g., institutional quality, corruption). This approach allows for a more nuanced and statistically robust analysis, especially when direct measurement is not feasible.

Furthermore, our approach tries to avoid common pitfalls such as arbitrary benchmarking and addresses concerns about variable stationarity and model specification. By doing so, it provides a more reliable foundation for understanding the drivers and dynamics of informality in BiH.

This paper contributes in three key ways:

• It offers updated, methodologically rigorous estimates of the shadow economy in BiH.
• It integrates both institutional and macroeconomic drivers using a latent variable framework.
• It provides evidence-based insights to inform policy reforms aimed at reducing informality and strengthening governance.

This study is guided by the central research question: What are the main drivers and how have the dynamics of the shadow economy in Bosnia and Herzegovina evolved over time—and how can this understanding inform more effective, context-specific policy responses in transitional economies like BiH?

Comments 8-19:

• Explicitly state how the study addresses gaps in the current literature.
• Focus more on the MIMIC model and summarize other methods briefly.
• Connect the reviewed studies to the current research’s goals.
• Discuss methodological critiques in more depth and explain how the study resolves them.
• Include more recent studies on the MIMIC model and SE measurement, particularly in transitional economies like BiH.

Response 14-19:

Agree with the comments and therefore we revised the literature review (171 to 228 line):

Unformal sector comprising legal but unregulated and untaxed economic activities, remains a persistent challenge in economies that struggled and struggle with transition for decades. This struggle is immanent for Western Balkan (WB) countries and BiH is one of them.

In BiH, it is estimated to account for approximately one-third of GDP (Pasovic & Efendic, 2018), yet empirical research on its scope and drivers remains limited. This study aims to address that gap by applying MIMIC model to estimate the size and evolution of the SE in BiH from 1996 to 2022.

Several methods have been developed to estimate the SE, including direct approaches (e.g., surveys, audits) and indirect methods (e.g., currency demand, electricity consumption, labor market discrepancies). While these methods offer valuable insights, they often suffer from data limitations, respondent bias, or narrow applicability (Feige, 1979; Tanzi, 1980). In contrast, the MIMIC model—an extension of SEM—treats the SE as a latent variable inferred from observable causes (e.g., tax burden, unemployment) and indicators (e.g., institutional quality, corruption). This makes it particularly suitable for contexts like BiH, where informal activity is widespread and direct measurement is difficult (Schneider & Enste, 2002; Schneider, Buehn, & Montenegro, 2010).

The MIMIC model has been widely applied in developed and developing countires, including France (Buehn & Schneider, 2008), Jordan (Alfoul et al., 2022), Egypt (Mai & Schneider, 2016), and Serbia (Krstić et al., 2015). More recent studies have extended its use to transitional and developing economies, emphasizing its adaptability to complex institutional environments. For instance, Wang et al. (2024) examined the role of governance indicators in shaping the SE across developing countries, while Hassan (2024) explored how regulatory quality affects informal financial activity in Southern Africa. Similarly, Abou Ltaif et al. (2024) investigated the SE in Lebanon amid financial crisis, highlighting the interplay between global shocks and local institutional weaknesses.

Despite its strengths, the MIMIC model has faced methodological critiques. Scholars have questioned its reliance on arbitrary benchmarking to produce point estimates, its sensitivity to variable selection, and its vulnerability to non-stationary time series data (Breusch, 2016; Pissarides & Weber, 1989). Smrčková and Brůna (2024) further caution against implausible estimates when theoretical definitions and calibration benchmarks are misaligned. This study addresses these concerns by avoiding benchmarking, incorporating structural breaks, and rigorously testing for stationarity. These refinements enhance the model’s robustness and credibility, particularly in a transitional setting like BiH.

The theoretical foundation of this study draws on institutional theory and tax compliance theory. Institutional theory posits that weak governance—characterized by low rule of law and high corruption—creates incentives for informal activity (Johnson, Kaufmann, & Zoido-Lobatón, 1998; Schneider, 2010). Tax compliance theory suggests that high tax burdens, complex regulations, and inconsistent enforcement discourage formalization and push economic actors into the informal sector (Arsic et al., 2015; Bordignon & Zanardii, 1997). These frameworks are especially relevant in BiH, where fragmented governance and regulatory inefficiencies have historically contributed to informality.

Earlier studies on BiH’s shadow economy (e.g., Pasovic & Efendic, 2018) provided foundational insights but were limited by outdated data and did not account for recent structural shocks such as the 2008 global financial crisis, the 2014 floods, or the COVID-19 pandemic. This study fills that gap by using updated data from 1996 to 2022 and explicitly modelling structural breaks to capture the impact of these events.

Based on the reviewed literature and theoretical considerations, this study explores two central propositions. First, higher tax burdens are expected to correlate with increased informal economic activity, as individuals and firms seek to avoid excessive fiscal pressure. Second, the quality of institutional governance—particularly in areas such as rule of law and corruption control—is anticipated to significantly influence the extent of informality. These propositions guide the empirical investigation into how macro-institutional conditions shape the dynamics of the shadow economy in BiH.

This research contributes to the literature in several keyways. It provides updated empirical estimates of the SE in BiH over a long-time horizon, capturing the effects of major structural events. Furthe, it refines the MIMIC methodology by addressing known limitations, including variable stationarity and benchmarking. It also contributes to theory-building by linking empirical findings to institutional and tax compliance theories. Finally, it offers policy-relevant insights for BiH and other similar economies seeking to reduce informality and strengthen governance.

References used in literature review

References (red colour references are subsequently added)

1.      Abou Ltaif, S. F., Mihai-Yiannaki, S., & Thrassou, A. (2024). Lebanon’s economic development risk: Global factors and local realities of the shadow economy amid financial crisis. Risks, 12(8), 122. https://doi.org/10.3390/risks12080122

2.      Alfoul, M. A., Mishal, Z. A., Schneider, F., Magableh, K., & Alabdulraheem, A. R. (2022). The hidden economy in Jordan: A MIMIC approach. Cogent Economics & Finance, 10(1), 2031434. https://doi.org/10.1080/23322039.2022.2031434

3.      Arsic, M., Arandarenko, M., Radulov, B., & Jankovic, I. (2015). Causes of the shadow economy. In G. Krstic & F. Schneider (Eds.), Formalizing the shadow economy in Serbia: Policy measures and growth effects (pp. 21–46). Springer. https://doi.org/10.1007/978-3-319-13437-6_4

4.      Bordignon, M., & Zanardii, A. (1997). Tax evasion in Italy. Giornale degli Economisti e Annali di Economia, 56(2), 169–210. http://www.jstor.org/stable/23248315

5.      Breusch, T. (2016). Estimating the underground economy using MIMIC models. Canadian Tax Journal, 64(1), 41–73. https://journals.docuracy.co.uk/jota/article/view/137/161

6.      Buehn, A., & Schneider, F. (2008). MIMIC models, cointegration and error correction: An application to the French shadow economy. CESifo Working Paper No. 2200. https://www.econstor.eu/handle/10419/26245

7.      Feige, E. L. (1979). How big is the irregular economy? Challenge, 22(5), 5–13. https://doi.org/10.1080/05775132.1979.11470559

8.      Fahad, S., Almawishir, N., & Benlaria, H. (2023). Using the PLS-SEM model to measure the impact of the knowledge economy on sustainable development in the Al-Jouf region of Saudi Arabia. Sustainability, 15(8), 6446. https://doi.org/10.3390/su15086446

9.      Gërxhani, K., & Cichocki, M. (2023). Informal institutions and the shadow economy in Eastern Europe. Journal of Institutional Economics, 19, 1–18. https://doi.org/10.1017/S1744137423000032

10.   Hassan, A. (2024). Mobile financial services and the shadow economy in Southern African countries: Does regulatory quality matter? International Journal of Financial Studies, 12(4), 115. https://doi.org/10.3390/ijfs12040115

11.   Johnson, S., Kaufmann, D., & Zoido-Lobatón, P. (1998). Regulatory discretion and the unofficial economy. American Economic Review, 88(2), 387–392. https://www.jstor.org/stable/116953

12.   Krstić, G., Schneider, F., Arsić, M., & Ranđelović, S. (2015). What is the extent of the shadow economy in Serbia? In G. Krstic & F. Schneider (Eds.), Formalizing the shadow economy in Serbia (pp. 47–77). Springer. https://doi.org/10.1007/978-3-319-13437-6_5

13.   Mai, H., & Schneider, F. (2016). Modeling the Egyptian shadow economy: A currency demand and a MIMIC model approach. Journal of Economic and Political Economy, 3(2), 309–333. https://www.proquest.com/docview/1814929206

14.   Pasovic, E., & Efendic, A. (2018). Informal economy in Bosnia and Herzegovina. South East European Journal of Economics and Business, 13(2), 112–125. https://doi.org/10.2478/jeb-2018-0015

15.   Pissarides, C. A., & Weber, G. (1989). An expenditure-based estimate of Britain’s black economy. Journal of Public Economics, 39(1), 17–32. https://doi.org/10.1016/0047-2727(89)90052-2

16.   Schneider, F. (2010). The influence of public institutions on the shadow economy: An empirical investigation for OECD countries. European Journal of Law and Economics, 6(3), 441–468.

17.   Schneider, F., & Enste, D. H. (2002). The shadow economy: Theoretical approaches, empirical studies, and political implications. Cambridge University Press.

18.   Schneider, F., Buehn, A., & Montenegro, C. E. (2010). New estimates for the shadow economies all over the world. International Economic Journal, 24(4), 443–461. https://doi.org/10.1080/10168737.2010.525974

19.   Smrčková, A., & Brůna, K. (2024). Implausible results in MIMIC-based shadow economy estimates: A critical review and methodological recommendations. Economic Modelling, 131, 106–118. https://doi.org/10.1016/j.econmod.2024.106118

20.   Tanzi, V. (1980). The underground economy in the United States: Estimates and implications. PSL Quarterly Review, 33(135), 427–453. https://rosa.uniroma1.it/rosa04/psl_quarterly_review/article/view/12996

21.   Wang, Y., Antohi, V. M., Fortea, C., Zlati, M. L., Mohammad, R. A., Abdelkhair, F. Y. F., & Ahmad, W. (2024). Shadow economy and environmental sustainability in global developing countries: Do governance indicators play a role? Sustainability, 16(22), 9852. https://doi.org/10.3390/su16229852

Comment 20:

Provide a brief justification for using the MIMIC model, explaining its relevance to the research question and BiH’s context.

Response 20

Agree with the comments and therefore we added following section but in the literature review part (171 to 186 line):

“Unformal sector comprising legal but unregulated and untaxed economic activities, remains a persistent challenge in economies that struggled and struggle with transition for decades. This struggle is immanent for Western Balkan (WB) countries and BiH is one of them.

In BiH, it is estimated to account for approximately one-third of GDP (Pasovic & Efendic, 2018), yet empirical research on its scope and drivers remains limited. This study aims to address that gap by applying MIMIC model to estimate the size and evolution of the SE in BiH from 1996 to 2022.

Several methods have been developed to estimate the SE, including direct approaches (e.g., surveys, audits) and indirect methods (e.g., currency demand, electricity consumption, labor market discrepancies). While these methods offer valuable insights, they often suffer from data limitations, respondent bias, or narrow applicability (Feige, 1979; Tanzi, 1980). In contrast, the MIMIC model—an extension of SEM—treats the SE as a latent variable inferred from observable causes (e.g., tax burden, unemployment) and indicators (e.g., institutional quality, corruption). This makes it particularly suitable for contexts like BiH, where informal activity is widespread and direct measurement is difficult (Schneider & Enste, 2002; Schneider, Buehn, & Montenegro, 2010).”

Also, we added (195 to 202 line):

“Despite its strengths, the MIMIC model has faced methodological critiques. Scholars have questioned its reliance on arbitrary benchmarking to produce point estimates, its sensitivity to variable selection, and its vulnerability to non-stationary time series data (Breusch, 2016; Pissarides & Weber, 1989). Smrčková and Brůna (2024) further caution against implausible estimates when theoretical definitions and calibration benchmarks are misaligned. This study addresses these concerns by avoiding benchmarking, incorporating structural breaks, and rigorously testing for stationarity. These refinements enhance the model’s robustness and credibility, particularly in a transitional setting like BiH.”

”

Comment 21:

Compare with other models (for example: Path Analysis, regression models) and explain why MIMIC is the preferred choice.

Response 21

We appreciate your comment and it si totaly reasonable considering that MIMIC incorporates path analysis as part of its structure. It includes both a causal path model (linking causes to the latent variable) and a measurement model (linking the latent variable to its indicators). This dual structure allows MIMIC to model both the underlying causes and the observable manifestations of the shadow economy, making it more comprehensive than standard path analysis.

Path analysis is a subset of SEM that models direct and indirect relationships among observed variables. It is useful for visualizing and quantifying complex causal chains, as we presented with our diagram, but, like regression, it does not handle latent variables unless extended into full SEM.

On the other hand, the causal part of the MIMIC model is essentially a regression model. It estimates the effect of observable causes (e.g., tax burden, unemployment) on a latent variable (e.g., the shadow economy). This allows researchers to retain the interpretability and statistical rigor of regression while extending its capabilities to unobservable phenomena.

Therefore we revised the Material and Methods section and insert as follows (405 to 410 line):

The MIMIC model was chosen over path analysis and traditional regression models because it allows for the simultaneous estimation of both the causes and indicators of a latent variable, such as the shadow economy. Unlike path analysis, which deals only with observed variables, and regression models, which cannot capture unobservable constructs, the MIMIC approach integrates causal and measurement components in a single framework, making it ideal for latent phenomena (Schneider & Enste, 2008)

Comment 22:

Provide detailed explanations of the equations (1) to (4) and the theoretical assumptions behind them.

Agree with the comment and therefore we inserted the following explanation after first relation (331 to 343 line):

The relation (1) stands for structural part of SEM representing the underlying relationship between the SE  and its observable causes. In this framework, the shadow economy is treated as a latent variable - something that cannot be directly measured but can be inferred through its associations with other measurable factors. These factors, known as exogenous causes, typically include variables such as tax burden, unemployment rates, and the quality of institutional governance. The model assumes that these causes influence the size and dynamics of the shadow economy in a linear fashion. The strength and direction of each cause’s influence are captured by a set of coefficients, which quantify how changes in the observable variables are expected to affect the latent variable. Additionally, the model includes an error term to account for other unobserved influences that might affect the shadow economy but are not explicitly included in the model. This error term is assumed to be normally distributed and uncorrelated with the causes, ensuring that the model remains statistically valid and interpretable.

Also we, have inserted this explanation after relation (4) (361 to 381 line):

Relation (2) stands for the measurement part of the proposed model establishes the connection between the latent variable, such as the shadow economy, and a set of observable indicators that serve as indirect measures of this hidden construct. These indicators might include variables like cash usage or discrepancies in labor market data, which are assumed to reflect the presence and intensity of informal economic activity. Each indicator is associated with a factor loading, which quantifies how strongly it is influenced by the latent variable. These loadings help determine the reliability and relevance of each indicator in capturing the underlying concept. Additionally, the model accounts for measurement errors, which represent random noise or inaccuracies in the observed data. The assumptions underlying this part of the model include a linear relationship between the latent variable and its indicators, the independence of measurement errors from both the latent variable and each other, and the idea that the latent variable is the sole common source of variation among the indicators. This structure ensures that the model can validly infer the unobservable phenomenon from the observed data. Further, relation (3) represents an algebraic rearrangement of the measurement model, aiming to isolate the latent variable from the observed indicators.  The key assumption here is that the matrix of factor loadings is invertible, which allows for this transformation to be mathematically valid. Also, the measurement errors are either small enough to be negligible or can be reasonably estimated. Finally, relation (4) combines the structural and measurement parts into a reduced-form equation that expresses the indicators (y) directly as a function of the causes (X)

Comment 22:

Discuss the latent variable structure, causal relationships, and the model’s ability to handle multivariate data.

Response 22

Agree with the comment and we address it in our reaction to comment number 21.  In addition, we made it clear through our whole paper reaffirmed ground of our theoretical foundation that include:

• Latent Variable Theory: The shadow economy is not directly observable, so it is modeled as a latent construct inferred from observable causes and indicators.
• Structural Equation Modeling (SEM): Provides a framework for modeling complex relationships between observed and unobserved variables.
• MIMIC Model: A specific form of SEM that includes multiple causes and multiple indicators for a single latent variable.

Comment 23:

Describe the data sources (for example: World Bank) and the selection criteria for the time period (1996-2022).

Response 23

Agree with the comment. The study utilizes a time series dataset from Bosnia and Herzegovina (BiH) covering the period 1996 to 2022, with data sourced from the World Bank.

Therefore, we have conceptually summarized the rationale for our sample selection in the data section as follows:

(72 to 76 line)
First, Bosnia and Herzegovina (BiH) represents a post-conflict, transitional economy characterized by a complex institutional framework and a persistently large shadow economy—estimated at 30–40% of GDP.

(77 to 83 line)
Second, data prior to 1996 are unavailable due to the establishment of BiH as a sovereign entity following the Dayton Peace Agreement, which marked the beginning of its current institutional and economic configuration.

 (84 to 89 line)

Third, in the post-COVID period, a notable number of individuals and entities began declaring previously unreported income or employment to access government subsidies and relief programs. This phenomenon introduced a temporary formalization effect that extended over more than five quarters, making it a critical period for inclusion.

We have added the following sentence (84 to 89 line):

“BiH, a post-conflict transitional economy with a complex institutional structure and a historically large shadow economy (30–40% of GDP), presents a unique case for analysis, particularly as the post-COVID period saw widespread declaration of previously unreported income due to subsidy eligibility—creating a temporary formalization effect—while data prior to 1996 are unavailable due to the country’s establishment following the Dayton Peace Agreement.”

Comment 24 and 25:

·        Explain how missing data were handled and provide justification for using imputation methods.

• Provide more detail on the imputation process, including the method used (e.g., multiple imputation), and explain how this affects the analysis.

Response 24 and 25

Agree with the comment, and as you know we used Multivariate Imputation by Chained Equations (MICE) method, implemented in the R package mice, to handle missing values in their dataset. This is a widely accepted and robust method for imputing missing data, especially in the context of multivariate analysis like SEM and MIMIC models.

Therefore in data section we state:

“All data are downloaded from the World Bank database, for the period from 1996 to 2022. According to the model specification data are organized in two parts. Also, imputation of missing data is done with the mice R package.”

MICE operates under the assumption that data are missing at random (MAR), meaning the probability of missingness can be explained by the observed data and full specification how it works is available in package documentation, On the other side brief elaboration would seem something like this:

MICE begins by making simple guesses to fill in missing values, providing a starting point for more refined estimates. It then treats each variable with missing data as a dependent variable in a regression model, using the other variables as predictors. This modeling and updating process is repeated for each variable in turn, forming a chain of equations. These cycles continue until the imputations converge to stable values. Rather than producing just one completed dataset, MICE generates multiple versions, each incorporating slightly different imputations. These are analyzed separately, and the results are pooled to account for the variability introduced by the missing data, leading to more robust and reliable statistical inferences.

Now, if would mean repeating something that is available in package description, and we decided not to include it, because it is usual way in R community.

Comment 26

Discuss any robustness tests conducted.

Response 26 (533 to 539 line)

We agree with the comment and therefore in the Data and Discussion sections, emphasize that:

• The time series variables used in the model were found to be integrated of order two (I(2)), meaning they required second differencing to achieve stationarity.
• They applied unit root tests such as the Augmented Dickey-Fuller, KPSS, and Box-Ljung tests, and also referenced the Pantula, Gonzales-Farias, and Fuller test to confirm stationarity after transformation.
• This transformation was done to ensure robustness of the model against critiques related to non-stationary data, which is a common issue in MIMIC model applications.

Comment 27

Explain how model fit was evaluated.

Response 27 (546-555 line)

We agree with the comment ad we used Model Fit Evaluation measures as follows:

1. RMSEA (Root Mean Square Error of Approximation)
• Reported value: 0.05
• Interpretation: This indicates a good model fit, as values below 0.08 are generally considered acceptable, and values below 0.05 are considered very good.
2. CFI (Comparative Fit Index)
• Reported value: 0.986
• Interpretation: This is a very high value, suggesting excellent fit. A CFI above 0.90 is typically considered acceptable, and above 0.95 is excellent.
3. TLI (Tucker-Lewis Index)
• Reported value: 0.972
• Interpretation: Also indicates strong model fit, with values above 0.90 considered good.

Comment 28

• Provide clear definitions and measurements for all key variables in the study, especially those related to the shadow economy (e.g., "rule of government," "control of corruption").

Response 28:

Agree with the comment and therefore we have inserted following in our data section (470 to 477 line):

“The "rule of government" corresponds to the WGI’s "Rule of Law" index, which captures perceptions of the extent to which agents have confidence in and abide by the rules of society, including the quality of contract enforcement, property rights, the police, and the courts. The "control of corruption" indicator reflects perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption. Both indices are measured on a scale ranging from approximately -2.5 (weak governance performance) to +2.5 (strong governance performance), and are reported annually.”

4. Response to Comments on the Quality of English Language

Point 1:

Response 1: Corrected.

5. Additional clarifications

Discussion (563 to 599 line)

This study contributes to the literature on the shadow economy by applying MIMIC model to BiH economical context. BiH is a country marked by a complex post-conflict institutional structure and widespread informality. The findings of this analysis are reflecting its aim to be consistent with institutional theory and tax compliance theory, both of which emphasize the influence of weak governance and high tax burdens on informal economic activity. The results reinforce these theoretical perspectives in certain extent, demonstrating that institutional quality and corruption control are significant determinants of the size and behavior of the shadow economy in BiH.

The model captures how the informal sector responds to both internal institutional weaknesses and external shocks. On the other hand, results do not offer conventional precision of models that have traditional regression framework, although in context of  BiH’s statistical such precision has huge potential of being significantly misleading.

Structural break analysis conducted using the Bai and Perron methodology (Bai, 1994; Bai & Perron, 1998), identifies key disruptions in BiH’s economic trajectory, including the post-war recovery period, the delayed impact of the global financial crisis, and the 2014 floods. These events are not some sort of technical statistical anomalies but represent real-world shocks that altered the incentives for informal economic activity. We should mention, that similar patterns have been observed in related studies (e.g., Pasovic & Efendic, 2018; Buehn & Schneider, 2008; Albala & Jose, 2015), underscoring the importance of institutional resilience in mitigating the expansion of informality during crises.

The informal sector of an economy, is best conceptualized as a responsive subsystem within the broader economic framework. It tends to expand when formal institutions are weak or disrupted and contracts when governance improves and economic stability is restored. This dynamic interplay highlights the need for integrated and adaptive policy responses.

In, this study we utilize time series that required transformation to achieve stationarity. The variables were found to be integrated of order two (I(2)), necessitating second differencing to ensure statistical validity. This step, emphasized by Breusch (2016), is critical for maintaining the integrity of MIMIC model estimates. Nevertheless, it is not unusual that some authors avoid this step.

Unlike many studies, this analysis deliberately avoids benchmarking the shadow economy as a percentage of GDP—a practice often criticized for introducing interpretive ambiguity and potential confusion (Giles & Tedds, 2002). Instead, the focus is placed on relative trends, which provide a clearer picture of the shadow economy’s evolution over time.

The latent variable derived from the model does not exhibit a deterministic trend, which aligns with the expectations of structural equation modeling and supports the internal coherence of the results. This further strengthens the robustness of the findings, particularly given the challenges associated with modeling informal economic activity.

Conclusion (643 to 671 line)

From a policy perspective, the implications are simple and straightforward, although the nature of SE constrains precisions that could be assumed in context of conventional macroeconometric aggregates. First, improving institutional quality—particularly in areas related to the rule of law and anti-corruption—can reduce the incentives for informality. Second, simplifying tax regulations and reducing the overall tax burden, especially for small and medium-sized enterprises, may encourage formalization. Third, labor market reforms that promote formal employment opportunities can help absorb workers currently operating outside the legal framework. Additionally, enhancing crisis preparedness and ensuring timely institutional responses during economic shocks are essential to prevent surges in informal activity.

Despite its strengths, the study has several limitations. It relies on secondary data produced in challenging statistical surrounding, which may be subject to measurement errors or inconsistencies, particularly in a fragmented administrative context such as BiH. The country’s statistical infrastructure is complex, with two statistical institutes operating at the entity level and one at the state level. Further, the use of latent variables, while methodologically sound, introduces a degree of abstraction that may obscure some nuances of informal behavior. Furthermore, the model does not account for all external influences, such as remittances, cross-border trade, or regional spillovers. Future research could expand the model to include these factors and explore cross-entity comparisons within BiH, although this could assume complex econometric toolset (e.g. Dynamic Stochastic Growth Equilibrium model) and fully developed national account statistical framework on state level. Fact that there are no input-output tables for BiH speaks for itself.

Nevertheless, this study aims to offer a relatively robust and accessible framework for understanding the drivers and dynamics of the shadow economy in BiH. Attempt to link empirical findings to theoretical insights and highlighting actionable policy recommendations, is challenging task, but in summary this analysis provides a solid foundation for informed decision-making aimed at managing informality and strengthening economic governance.

Comments 1: The paper assumes that "institutional weakness directly leads to the expansion of the shadow economy", but does not clearly distinguish the differentiated impacts of different dimensions of institutional factors on the shadow economy. The document mentions that the "complex governance structure" and "corruption" in Bosnia and Herzegovina work together, but the independent effects of the two are not separated through the model.

Response 1: Thank you for pointing this out. We agree with this comment. In the current MIMIC specification, institutional quality is proxied through composite indicators like the rule of law and control of corruption, which are treated as part of the measurement model. While these indicators are statistically significant and theoretically grounded, the model does not isolate their individual causal impacts on the shadow economy. Instead, they are used to reflect the latent variable rather than to explain it directly and that is a feature of MIMIC model as an extension of SEM. Measuring institutional quality is challenging task and current statistical landscape in BiH does not offer direct measurement of this dimension.

To address this limitation, future research could refine the model by incorporating separate latent constructs for different institutional dimensions. For example, one latent variable could capture administrative complexity (e.g., number of overlapping jurisdictions, regulatory burden), while another could represent corruption (e.g., bribery prevalence, enforcement gaps). These constructs could then be linked independently to the shadow economy, allowing for a more nuanced understanding of how different institutional weaknesses interact or diverge in their effects.

Comments 2: The paper takes "employment rate" as the driving factor of the shadow economy, but does not explain why "the increase in total employment" is positively correlated with "the expansion of the scale of the shadow economy" (the emplBH coefficient in Table 2 of the document is 1.62, p<0.01). This result may be related to the "tax evasion tendency of self-employed individuals", but the hypothesis derivation process lacks theoretical support

Response 2: Agree. The observation regarding the positive and statistically significant coefficient of the employment rate (emplBH) in Table 2 is valid and highlights a counterintuitive result that warrants further clarification. In theory, one might expect higher employment to correlate with a reduction in the shadow economy, as more individuals are absorbed into the formal labor market. However, in the context of BiH, this relationship is more complex.

One plausible explanation for this finding is the composition of employment in BiH, particularly the high share of self-employed individuals and informal workers. In many transitional economies, including BiH, self-employment often serves as a channel for tax evasion and informal activity. As such, an increase in total employment may reflect not only growth in formal jobs but also a rise in informal or underreported employment, especially among the self-employed. This could explain the observed positive association between employment and the shadow economy.

That said, we agree with the critique that this interpretation lacks direct empirical support within the current model. Specifically, the hypothesis that self-employment drives this relationship is not formally tested, and the model does not distinguish between types of employment. Moreover, there is currently no reliable or consistent measure of the "tax evasion tendency of self-employed individuals" in BiH, which limits our ability to empirically validate this mechanism.

Future research could address this gap by incorporating disaggregated labor market data—distinguishing between formal, informal, and self-employed workers—and by exploring micro-level survey data, if available. This would allow for a more nuanced understanding of how different employment categories contribute to the dynamics of the shadow economy.

Comments 3: The paper assumes that "tax burden directly boosts the shadow economy", but does not consider the moderating effect of tax collection and administration efficiency. The document mentions that there is "inconsistent tax enforcement" in Bosnia and Herzegovina, but the model does not include variables such as "frequency of tax audits"

Response 3: Agree. In our paper we impose that a higher tax burden contributes to the expansion of the shadow economy, a relationship supported by tax compliance theory and widely observed in empirical literature. However, we acknowledge that the model does not account for the moderating role of tax collection and administrative efficiency, which can significantly influence how tax burdens are perceived and responded to by economic agents. To be honest, we do mention the issue of "inconsistent tax enforcement" in BiH, but this aspect is not explicitly modeled due to data limitations. In particular, variables such as the "frequency of tax audits" or the effectiveness of tax inspections—which could serve as proxies for enforcement efficiency—are not available in a reliable or consistent format for BiH. This limits our ability to empirically test how enforcement moderates the relationship between tax burden and informality.

Comments 4: The causal part only includes three explanatory variables (employment, GDP, and tax revenue), citing the reason of "reducing the complexity of interaction among independent variables". However, this simplification may miss key drivers, such as the "cash demand" you mentioned, which was excluded due to the currency board system. Furthermore, the measurement part only uses two indicators (governance rules and corruption index), making it difficult to comprehensively reflect the multi-dimensional characteristics of the shadow economy.

Response 4:  Thank you for your comment. The decision to include only three explanatory variables—employment, GDP, and tax revenue—in the causal part of the MIMIC model was guided by the intention to maintain model parsimony and reduce the complexity of interactions among independent variables. This approach is consistent with a substantial body of literature that employs simplified MIMIC specifications, particularly in contexts where data limitations and multicollinearity risks are significant (e.g., Schneider & Enste, 2002; Buehn & Schneider, 2008). In fact, many empirical studies on the shadow economy adopt a minimal causal structure to ensure model stability and interpretability, especially when working with time series data in small or transitional economies.

Regarding the exclusion of cash demand as a causal variable, this was a deliberate methodological choice based on the specific monetary regime in Bosnia and Herzegovina. BiH operates under an orthodox currency board arrangement, which tightly constrains monetary policy and limits the central bank’s ability to influence money supply. As noted in the paper, this regime significantly reduces the relevance of cash-based indicators for capturing informal economic activity. Previous studies (e.g., Hanke & Schuler, 1994) have highlighted how currency boards, by design, suppress monetary volatility, making cash demand a less reliable proxy for shadow economy dynamics in such settings.

Comments 5: The result of the unit root test shows that the variable is a second-order simple integer (I (2)), but the document does not clearly explain the influence of data transformation methods (such as the number of differences, seasonal adjustment) on the model estimation. Furthermore, the missing values were filled using the mice package, but the parameter stability test after filling was not reported.

Response 5: Agree with the comment, and as you know we used Multivariate Imputation by Chained Equations (MICE) method, implemented in the R package mice, to handle missing values in their dataset. This is a widely accepted and robust method for imputing missing data, especially in the context of multivariate analysis like SEM and MIMIC models.

Therefore, in data section we state:

“All data are downloaded from the World Bank database, for the period from 1996 to 2022. According to the model specification data are organized in two parts. Also, imputation of missing data is done with the mice R package.”

Also, in the Data and Discussion sections, emphasize that:

• The time series variables used in the model were found to be integrated of order two (I(2)), meaning they required second differencing to achieve stationarity.
• That we applied unit root tests such as the Augmented Dickey-Fuller, KPSS, and Box-Ljung tests, and also referenced the Pantula, Gonzales-Farias, and Fuller test to confirm stationarity after transformation.
• This transformation was done to ensure robustness of the model against critiques related to non-stationary data, which is a common issue in MIMIC model applications.

To be honest, the critique regarding the handling of second-order integration (I(2)) variables and the imputation process is well taken. In this study, unit root tests—including the Augmented Dickey-Fuller, KPSS, and Pantula-Gonzales-Farias-Fuller tests—indicated that the variables used in the model were non-stationary and required second differencing to achieve stationarity. This transformation was necessary to ensure the validity of the model estimates, as emphasized by Breusch (2016), who highlights the importance of proper treatment of non-stationary time series in shadow economy modeling.

However, we acknowledge that the paper does not provide a detailed explanation of how the differencing process or potential seasonal adjustments may have influenced the model’s structure or parameter estimates. While seasonal adjustment was not explicitly applied—due to the annual frequency of the data—extension of the study could present and elaborate on benefit from a more transparent discussion of how data transformations affect the interpretation of the latent variable and its dynamics.

Comments 6: The paper avoids standardizing the shadow economy as a percentage of GDP on the grounds of "avoiding controversy", but this leads to a lack of policy operability in the results

Response 6 Agree. We have, accordingly, added a paragraph in introduction (128-131 line).

Furthermore, our approach tries to avoid common pitfalls such as arbitrary benchmarking and addresses concerns about variable stationarity and model specification. By doing so, it provides a more reliable foundation for understanding the drivers and dynamics of informality in BiH.

We appreciate the reviewer’s observation regarding the absence of standardized shadow economy estimates expressed as a percentage of GDP. This decision was made deliberately to avoid the methodological and interpretive challenges associated with arbitrary benchmarking, which has been widely debated in the literature (e.g., Giles & Tedds, 2002; Breusch, 2016). Benchmarking often relies on selecting a base year or external reference point that may not be empirically justified, especially in contexts like BiH, where data limitations and structural breaks complicate such calibration.

That said, we fully acknowledge that the lack of GDP-based estimates may limit the immediate policy operability of the results. Policymakers often rely on such standardized figures to assess the fiscal impact of informality and to compare across countries or time periods. To address this trade-off, our approach emphasizes the relative dynamics and structural drivers of the shadow economy, offering a more stable and internally consistent foundation for understanding informality in BiH. As noted in the revised manuscript, this method avoids common pitfalls such as arbitrary benchmarking and addresses concerns about variable stationarity and model specification.

Comments 7: The document lists the matrix expression of SEM (Formulas 1-3), but does not explain the parameter estimation method and convergence criteria

Response 7: Agree with the comment and therefore we inserted the following explanation after first relation (331 to 343 line):

The relation (1) stands for structural part of SEM representing the underlying relationship between the SE  and its observable causes. In this framework, the shadow economy is treated as a latent variable—something that cannot be directly measured but can be inferred through its associations with other measurable factors. These factors, known as exogenous causes, typically include variables such as tax burden, unemployment rates, and the quality of institutional governance. The model assumes that these causes influence the size and dynamics of the shadow economy in a linear fashion. The strength and direction of each cause’s influence are captured by a set of coefficients, which quantify how changes in the observable variables are expected to affect the latent variable. Additionally, the model includes an error term to account for other unobserved influences that might affect the shadow economy but are not explicitly included in the model. This error term is assumed to be normally distributed and uncorrelated with the causes, ensuring that the model remains statistically valid and interpretable.

Also, we have inserted this explanation after relation (4) (361 to 381 line):

Relation (2) stands for the measurement part of the proposed model establishes the connection between the latent variable, such as the shadow economy, and a set of observable indicators that serve as indirect measures of this hidden construct. Each indicator is associated with a factor loading, which quantifies how strongly it is influenced by the latent variable. These loadings help determine the reliability and relevance of each indicator in capturing the underlying concept. Additionally, the model accounts for measurement errors, which represent random noise or inaccuracies in the observed data. The assumptions underlying this part of the model include a linear relationship between the latent variable and its indicators, the independence of measurement errors from both the latent variable and each other, and the idea that the latent variable is the sole common source of variation among the indicators. This structure ensures that the model can validly infer the unobservable phenomenon from the observed data. Further, relation (3) represents an algebraic rearrangement of the measurement model, aiming to isolate the latent variable from the observed indicators.  The key assumption here is that the matrix of factor loadings is invertible, which allows for this transformation to be mathematically valid. Also, the measurement errors are either small enough to be negligible or can be reasonably estimated. Finally, relation (4) combines the structural and measurement parts into a reduced-form equation that expresses the indicators (y) directly as a function of the causes (X)

4. Response to Comments on the Quality of English Language

Point 1:

Response 1:    Corrected

5. Additional clarifications

Discussion (563 to 599 line)

This study contributes to the literature on the shadow economy by applying MIMIC model to BiH economical context. BiH is a country marked by a complex post-conflict institutional structure and widespread informality. The findings of this analysis are reflecting its aim to be consistent with institutional theory and tax compliance theory, both of which emphasize the influence of weak governance and high tax burdens on informal economic activity. The results reinforce these theoretical perspectives in certain extent, demonstrating that institutional quality and corruption control are significant determinants of the size and behavior of the shadow economy in BiH.

The model captures how the informal sector responds to both internal institutional weaknesses and external shocks. On the other hand, results do not offer conventional precision of models that have traditional regression framework, although in context of  BiH’s statistical such precision has huge potential of being significantly misleading.

Structural break analysis conducted using the Bai and Perron methodology (Bai, 1994; Bai & Perron, 1998), identifies key disruptions in BiH’s economic trajectory, including the post-war recovery period, the delayed impact of the global financial crisis, and the 2014 floods. These events are not some sort of technical statistical anomalies but represent real-world shocks that altered the incentives for informal economic activity. We should mention, that similar patterns have been observed in related studies (e.g., Pasovic & Efendic, 2018; Buehn & Schneider, 2008; Albala & Jose, 2015), underscoring the importance of institutional resilience in mitigating the expansion of informality during crises.

The informal sector of an economy, is best conceptualized as a responsive subsystem within the broader economic framework. It tends to expand when formal institutions are weak or disrupted and contracts when governance improves and economic stability is restored. This dynamic interplay highlights the need for integrated and adaptive policy responses.

In, this study we utilize time series that required transformation to achieve stationarity. The variables were found to be integrated of order two (I(2)), necessitating second differencing to ensure statistical validity. This step, emphasized by Breusch (2016), is critical for maintaining the integrity of MIMIC model estimates. Nevertheless, it is not unusual that some authors avoid this step.

Unlike many studies, this analysis deliberately avoids benchmarking the shadow economy as a percentage of GDP—a practice often criticized for introducing interpretive ambiguity and potential confusion (Giles & Tedds, 2002). Instead, the focus is placed on relative trends, which provide a clearer picture of the shadow economy’s evolution over time.

The latent variable derived from the model does not exhibit a deterministic trend, which aligns with the expectations of structural equation modeling and supports the internal coherence of the results. This further strengthens the robustness of the findings, particularly given the challenges associated with modeling informal economic activity.

Conclusion (643 to 671 line)

From a policy perspective, the implications are simple and straightforward, although the nature of SE constrains precisions that could be assumed in context of conventional macroeconometric aggregates. First, improving institutional quality—particularly in areas related to the rule of law and anti-corruption—can reduce the incentives for informality. Second, simplifying tax regulations and reducing the overall tax burden, especially for small and medium-sized enterprises, may encourage formalization. Third, labor market reforms that promote formal employment opportunities can help absorb workers currently operating outside the legal framework. Additionally, enhancing crisis preparedness and ensuring timely institutional responses during economic shocks are essential to prevent surges in informal activity.

Despite its strengths, the study has several limitations. It relies on secondary data produced in challenging statistical surrounding, which may be subject to measurement errors or inconsistencies, particularly in a fragmented administrative context such as BiH. The country’s statistical infrastructure is complex, with two statistical institutes operating at the entity level and one at the state level. Further, the use of latent variables, while methodologically sound, introduces a degree of abstraction that may obscure some nuances of informal behavior. Furthermore, the model does not account for all external influences, such as remittances, cross-border trade, or regional spillovers. Future research could expand the model to include these factors and explore cross-entity comparisons within BiH, although this could assume complex econometric toolset (e.g. Dynamic Stochastic Growth Equilibrium model) and fully developed national account statistical framework on state level. Fact that there are no input-output tables for BiH speaks for itself.

Nevertheless, this study aims to offer a relatively robust and accessible framework for understanding the drivers and dynamics of the shadow economy in BiH. Attempt to link empirical findings to theoretical insights and highlighting actionable policy recommendations, is challenging task, but in summary this analysis provides a solid foundation for informed decision-making aimed at managing informality and strengthening economic governance.

Comments 1: Global Relevance and Replicability

Response 1: Thank you for pointing this out. We revised the abstract and introduction to emphasize the broader applicability of our findings and the potential for replication in other national contexts, as you rightly pointed out. Also we make some adjustment in discussion section:

Discussion (563 to 599 line)

This study contributes to the literature on the shadow economy by applying MIMIC model to BiH economical context. BiH is a country marked by a complex post-conflict institutional structure and widespread informality. The findings of this analysis are reflecting its aim to be consistent with institutional theory and tax compliance theory, both of which emphasize the influence of weak governance and high tax burdens on informal economic activity. The results reinforce these theoretical perspectives in certain extent, demonstrating that institutional quality and corruption control are significant determinants of the size and behavior of the shadow economy in BiH.

The model captures how the informal sector responds to both internal institutional weaknesses and external shocks. On the other hand, results do not offer conventional precision of models that have traditional regression framework, although in context of  BiH’s statistical such precision has huge potential of being significantly misleading.

Structural break analysis conducted using the Bai and Perron methodology (Bai, 1994; Bai & Perron, 1998), identifies key disruptions in BiH’s economic trajectory, including the post-war recovery period, the delayed impact of the global financial crisis, and the 2014 floods. These events are not some sort of technical statistical anomalies but represent real-world shocks that altered the incentives for informal economic activity. We should mention, that similar patterns have been observed in related studies (e.g., Pasovic & Efendic, 2018; Buehn & Schneider, 2008; Albala & Jose, 2015), underscoring the importance of institutional resilience in mitigating the expansion of informality during crises.

The informal sector of an economy, is best conceptualized as a responsive subsystem within the broader economic framework. It tends to expand when formal institutions are weak or disrupted and contracts when governance improves and economic stability is restored. This dynamic interplay highlights the need for integrated and adaptive policy responses.

In, this study we utilize time series that required transformation to achieve stationarity. The variables were found to be integrated of order two (I(2)), necessitating second differencing to ensure statistical validity. This step, emphasized by Breusch (2016), is critical for maintaining the integrity of MIMIC model estimates. Nevertheless, it is not unusual that some authors avoid this step.

Unlike many studies, this analysis deliberately avoids benchmarking the shadow economy as a percentage of GDP—a practice often criticized for introducing interpretive ambiguity and potential confusion (Giles & Tedds, 2002). Instead, the focus is placed on relative trends, which provide a clearer picture of the shadow economy’s evolution over time.

The latent variable derived from the model does not exhibit a deterministic trend, which aligns with the expectations of structural equation modeling and supports the internal coherence of the results. This further strengthens the robustness of the findings, particularly given the challenges associated with modeling informal economic activity.

Comments 2: Scientific Contribution

Response 2: Agree. We added paragraphs in introduction and literature review section to clearly articulate our original scientific contributions and how our work advances the existing literature on the shadow economy, particularly in the context of transition economies.

Comments 3: Citation Format

Response 3: Agree. We carefully reviewed and revised all references to ensure full compliance with MDPI citation and formatting guidelines.

Comments 4: Study Limitations and Future Research

Response 4: Agree. We revised our conclusion section:

Conclusion (643 to 671 line)

From a policy perspective, the implications are simple and straightforward, although the nature of SE constrains precisions that could be assumed in context of conventional macroeconometric aggregates. First, improving institutional quality—particularly in areas related to the rule of law and anti-corruption—can reduce the incentives for informality. Second, simplifying tax regulations and reducing the overall tax burden, especially for small and medium-sized enterprises, may encourage formalization. Third, labor market reforms that promote formal employment opportunities can help absorb workers currently operating outside the legal framework. Additionally, enhancing crisis preparedness and ensuring timely institutional responses during economic shocks are essential to prevent surges in informal activity.

Despite its strengths, the study has several limitations. It relies on secondary data produced in challenging statistical surrounding, which may be subject to measurement errors or inconsistencies, particularly in a fragmented administrative context such as BiH. The country’s statistical infrastructure is complex, with two statistical institutes operating at the entity level and one at the state level. Further, the use of latent variables, while methodologically sound, introduces a degree of abstraction that may obscure some nuances of informal behavior. Furthermore, the model does not account for all external influences, such as remittances, cross-border trade, or regional spillovers. Future research could expand the model to include these factors and explore cross-entity comparisons within BiH, although this could assume complex econometric toolset (e.g. Dynamic Stochastic Growth Equilibrium model) and fully developed national account statistical framework on state level. Fact that there are no input-output tables for BiH speaks for itself.

Nevertheless, this study aims to offer a relatively robust and accessible framework for understanding the drivers and dynamics of the shadow economy in BiH. Attempt to link empirical findings to theoretical insights and highlighting actionable policy recommendations, is challenging task, but in summary this analysis provides a solid foundation for informed decision-making aimed at managing informality and strengthening economic governance.

4. Response to Comments on the Quality of English Language

Point 1:

Response 1:    Editing done.