Intensity of Regulations as a Cause of the Informal Sector

Abstract

It is argued in the literature that the intensity of regulations and control in an economy is a determinant of the informal sector which however is ignored in most of its estimates. This article uses a new variant of the currency demand approach where ‘unemployment’ and ‘intensity of government control’ are used to estimate a shadow economy, alongside a the traditional tax variable. We choose Pakistan since it has a significant share of its activities in the informal sector along with the history of various political and dictatorial regimes. Further, there are examples of bureaucratic control leading to corruption in the economy. It provides an opportunity to study the nexus between regulation intensity and informal economy and present a case study for other developing countries exercising control over the economy through the large size of its public sector. The results show that the intensity of the control variable has statistically and economically significant role in increasing the shadow economy, almost equivalent to the tax coefficient. Once the yearly variation in our estimates is mapped with various political regimes, it seems that the validity of estimates is reinforced considering policy inconsistencies and prominent events of each regime.

Keywords

Informal sector shadow economy currency demand approach

Introduction

The average size of an informal sector in developing countries is higher than that in the developed economies (Schneider, 2005), as most of the developing countries have complex business processes and bureaucratic formalities. Bureaucratic hunger for power leads to an increased size of the public sector to exercise more control over the economy, which may increase corruption (Buehn & Schneider, 2009; Dreher et al., 2009; Tanzi, 1998). This may place restraints on regular economic activities through formal procedures, license fees and permits, and eventually may increase the shadow economy. These bureaucratic procedures represent the increased intensity of government control over the economy.

Most of the developing countries rank much lower on ease of doing business, with Pakistan being one of them. Regulations, their intensity and overlapping procedural formalities make Pakistan a difficult country for doing businesses. In 2017, Pakistan slipped down three places on the world’s ‘Ease of Doing Business’ index to 147th rank among 190 countries.¹ The institutionalization of red-tapism and the extent of corruption are evident from an experimental study of Khan et al. (2015) where the authors showed that even though performance-based incentives to tax officials raised the revenue, it simultaneously increased bribe requests by 30 per cent. Apart from this, there is a history of tax amnesty schemes and land grabbing entities legally paying off to gain ownership.² These examples of institutionalizing the intensity of regulations and its link with corruption make Pakistan a compelling and interesting case for empirical investigation.

The estimates of the informal sector in Pakistan have been computed largely through the most commonly used indirect methodology–the currency demand (CD) approach of Tanzi (1980, 1983). Past estimates assume tax as the only determinant, except the study by Arby et al. (2010) which included the unemployment rate in its CD equation. Further, most of the estimates for Pakistan, for example, Ahmed and Ahmed (1995), Shabsigh (1995), Aslam (1998), Iqbal et al. (1998), Kemal (2003), Yasmin and Rauf (2003), and Ahmed and Haider (2008) did not account for cointegration. OLS estimates without accounting for stationarity and cointegration are not consistent and are mostly from the time when the use of error correction models was not a common practice. Although Kemal (2007) also used the vector autoregression model, and Arby et al. (2010) and Kiani et al. (2010) used the autoregressive distributed lag (ARDL) model. Further, despite the fact that there has been discussion in theoretical and empirical literature about the intensity of government regulations and control in the economy as one of the causes of increased informality (Dell’Anno et al., 2007; Johnson, Kaufmann & Shleifer, 1997; Loayza, 1996; Schneider & Dreher, 2006;), it is mostly ignored due to difficulty in quantification, and none of the aforementioned studies on Pakistan accounted for the effect of intensity of regulations on the growth of the informal sector.

This article thus attempts to estimate the shadow economy of Pakistan with the novelty of quantifying the effect of intensity of regulations and control in the economy. Just like all preceding authors, we use a tax variable, and following Arby et al. (2010), we also include the unemployment rate. But most importantly, we include a proxy for ‘intensity of regulation and control’ by the government in the economy. In a nutshell, we provide the latest estimates from 1975 to 2015 with these variables reflecting improved specification. Also, this is for the first time in the case of Pakistan that we link our estimates with the political regime at the time to validate if our results are in sync with the events which have been the legacy of that specific regime. This study holds implications for developing economies exercising similar controls through the extended size of their public sector.

The remainder of this article is structured as follows. In the ‘Theoretical Considerations’ section, we discuss the definition of a shadow economy and theoretical considerations. The ‘Methodology and Data’ section describes the methodology and explains the variables. The ‘Empirical Results’ section presents the econometric results, and the ‘Size of the Shadow Economy of Pakistan’ section provides the size and development of the shadow economy of Pakistan for the period of 1973–2015. Finally, the last section draws some conclusions and policy implications.

Theoretical Considerations

A broader definition (OECD (2002)) is: ‘The shadow economy is defined as those activities that are productive and legal but are deliberately concealed from the public authorities to avoid payment of taxes and social security contributions or complying with regulations’. It needs to be acknowledged that incomes from illegal activities like smuggling and prostitution also go unreported due to the fear of prosecution; hence, they are the part of the shadow economy. Additionally, indirect the methodologies depend on the indicator and/or cause variables which cannot distinguish between legal and illegal activities. For the purpose of this research, it is plausible to assume that estimates of shadow economy through CD approach could include all unregistered economic activities relying on cash-based transactions—legal or illegal—that can contribute to the officially published GDP.

The published empirical literature shows that main causes for the existence of shadow economy³ are taxes, unemployment and excessive government control over the economy which leads to increased bureaucracy and regulations, although most of the studies, especially in Pakistan, have relied only on tax rates for their estimates. Higher tax rates result in lower disposable incomes and reduce the incentives for workers to work in the official sector (Drummond et al., 1994; Schneider, 2006; Schneider & Enste, 1998, 2000; Spiro, 1993; Tanzi, 1980, 1983). Further, when a country has no or minimal social welfare system, the need to meet basic needs may drive officially unemployed to informal means for livelihood (Arby et al., 2010; Dell’Anno et al., 2007). Despite comprehensive discussion in literature (Johnson et al., 1997; Schneider & Dreher, 2006), an appropriate measure for the intensity of regulations and bureaucratic control is debatable in general and missing altogether in the case of Pakistan and other developing countries. Since the most important contribution of this article is to us the intensity of regulations, it is necessary that the economic link with the informal sector is elaborated.

Intensity of Regulations

Developing countries often have bureaucratic government structures where formalities cause operational constraints reducing the freedom of choice to work in the formal sector. Additionally, they often lead to bribery and kickbacks which are directly a part of the informal income, although illegal. Restrictions by the government such as permits and licenses increase prices of goods and services due to additional costs which present opportunities to informal sector firms and workers. For example, unregistered construction workers and contractors who might not be paying taxes have a lower cost of service delivery. The same applies to a large number of service providers and small-scale household manufacturing establishments. There is also evidence in literature that the intensity of regulations increases the shadow economy (Dutta et al., 2013; Schneider & Dreher, 2006). Some authors argue that with decentralization, which reduces the intensity of control and increases the efficiency, the shadow economy is reduced (Buehn et al., 2013). Others highlight that increased enforcement of laws instead of new regulations is more favourable in reducing activities in this sector (Johnson et al., 1997; Loayza, 1996; Nagac, 2015; Schneider & Enste, 1998, 2000). Governments, however, prefer to increase regulations and laws when trying to reduce the shadow economy, mostly because this leads to the increased power of bureaucrats and a higher rate of employment in the public sector.

We summarize the following hypothesis: The greater the intensity of regulation, higher the size of informal sector, ceteris paribus. In addition to this, we also hypothesize that increased taxes and unemployment lead to an increase in the informal sector.

Methodology and Data

Measuring Shadow Economy

There are multiple approaches to estimate a shadow economy. Effectiveness of an approach mostly depends on two aspects: first is the part of hidden economy being measured and second is data availability. Generally, estimation approaches are classified into direct and indirect approaches. Both have their own suites of pros and cons. Direct approaches include survey questionnaires and tax auditing; however, the reliability of response in both is questionable. There are multiple indirect approaches as well like the approach of estimating discrepancy in national accounts where the difference in the income and expenditure sides of national accounts can be treated as shadow economy. However, as argued by Schneider and Enste (2000), the national accounts statisticians will be anxious to minimize this difference, and hence the published national accounts would not be a consistent measure. The electricity demand method is another indirect approach which is based on the assumption of unit elasticity between GDP and consumption of electricity. The growth in total electricity consumption is an indicator of growth of total GDP, including the official and unofficial sectors. Therefore, the difference between GDP based on simulations from electricity consumption and official statistics is the shadow economy. However, this approach will at best measure small-scale household manufacturing because many informal activities do not require use of electricity. Furthermore, due to electricity load shedding in Pakistan, many small-scale businesses are dependent on diesel or gas generators which complicate the modelling and inference from this approach. Another indirect approach is the multiple indicators multiple causes (MIMIC) model which is a latent or an unobserved variable approach. Despite its advanced ability to include more than one cause and indicator variables, it has some shortcomings. Schneider (2006) highlights the volatility in its estimates with changes in sample size and specification as one possible caveat of this approach.

The indirect approach used in this article is the CD approach. This approach follows the theoretical construct of literature on the informal sector where it is assumed that taxes are the main cause of increased informality (used by Tanzi, 1980, 1983). Hence, most of the transactions in the shadow economy take place with hard currency to avoid detection. Based on these assumptions, a CD equation is estimated using theoretically accepted official sector variables, which influence demand for currency, while also including a tax variable as one of the determinants/indicator variables. Then assuming the period with the minimum tax rate as a base year which has no shadow economy, the equation is simulated to estimate the total shadow economy as a percentage of GDP. These simulations are further explained in Section 5. This approach has been most commonly used an indirect approach for estimations in various countries and is known to produce acceptable results.⁴ Schneider and Hametner (2007) selected a variant of the CD approach in which two tax variables (direct and indirect taxes) were included. Ardizzi et al. (2012) also modified the traditional CD approach by using the ratio of the value of cash withdrawn from bank accounts to the value of total payments settled by instruments other than banks, as a dependent variable.

Data and Variables

In this article, we estimate the CD equation on yearly data for the period 1973–2015 having a currency deposit ratio (C2DD) as an endogenous variable. The data has been collected from the publications of State Bank of Pakistan. GDP per capita (GDPPC), household consumption expenditure (HHCONPC) and the rate of inflation (INFL) are used as the determinants of official demand for currency based on monetary theory considerations. Determinants for unofficial sector demand for currency include the unemployment rate (UNEMP) and the intensity of regulations and control over the economy (INTENSITY), in addition to the traditional tax variable (TAXGDP). The intensity variable is represented by the government’s public administration and defence expenditure. Greater the public sector employment and operational expense, greater the influence on and monitoring of the household and firm decision-making, limiting their freedom of choice.

Modelling and Methodology

We establish a cointegrating relationship between the currency deposit ratio and explanatory variables. The estimated equation is used to simulate the size of the informal sector in Pakistan’s economy. Even though the main purpose of this study is to present estimates of a shadow economy, we do not ignore econometric modelling concerns. Since the variables are of time-series properties, first we address stationarity and cointegration. Variables are tested for a unit root as suggested by Dickey and Fuller (1979), Philips and Perron (1988) and Kwiatkowski et al. (1992). The number of lags in our models has been included using the Akaike information criterion (AIC). Further, two different error-correction methodologies are used to estimate the model which accounts for possible cointegration among variables. The reason for using two different econometric approaches rests in comparing the estimates of the shadow economy for ensuring robustness of estimates. First, we use the Engel–Granger (EG) two-step approach suggested by Engle and Granger (1987). In case variables are non-stationary, the EG model is first estimated at level and residual is tested for the unit root. If residuals are stationary, the existence of cointegration is established. The second model is estimated at differences where the lag of residual from the first model appears as an error-correction term. Therefore, in this approach, the coefficients with differenced variables express the short-term relationship, while the lagged residual from the level model establishes long-term cointegration and shows error correction. We estimate the following two models using this approach:

\begin{matrix} Δ L n C 2 D D_{t} = γ_{0} + {\sum^{​}}_{n}^{i = 1} γ_{1 i} Δ L n C 2 D D_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{2 i} Δ L n T A X G D P_{t - i} \\ \begin{array}{l} {+ \sum^{​}}_{n}^{i = 0} γ_{3 i} Δ L n U N E M P_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{4 i} Δ L n I N T E N S I T Y_{t - i} \\ \begin{matrix} {+ \sum^{​}}_{n}^{i = 0} γ_{5 i} Δ L n G D P P C_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{6 i} Δ L n H H C O N P C_{t - i} \\ + {\sum^{​}}_{n}^{i = 0} γ_{7 i} Δ L n I N F L_{t - i} + γ_{8} S B_{t} + γ_{9} μ_{t - 1} + ε_{t} \end{matrix} \end{array} \end{matrix}

(1)

\begin{matrix} Δ L n C 2 D D_{t} = γ_{0} + {\sum^{​}}_{n}^{i = 1} γ_{1 i} Δ L n C 2 D D_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{2 i} Δ L n T A X G D P_{t - i} \\ \begin{array}{l} {+ \sum^{​}}_{n}^{i = 0} γ_{3 i} Δ L n U N E M P_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{4 i} Δ L n I N T E N S I T Y_2_{t - i} \\ \begin{matrix} {+ \sum^{​}}_{n}^{i = 0} γ_{5 i} Δ L n G D P P C_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{6 i} Δ L n H H C O N P C_{t - i} \\ {+ \sum^{​}}_{n}^{i = 0} γ_{7 i} Δ L n I N F L_{t - i} + γ_{8} S B_{t} + γ_{9} μ_{t - 1} + ε_{t} ​ ​ ​, \end{matrix} \end{array} \end{matrix}

(2)

where C2DD is an endogenous variable. GDPPC is GDP per capita, HHCONPC is household consumption expenditure and INFL is the rate of inflation. All these variables are used as the determinants of official demand for currency based on monetary theory considerations. Further determinants for unofficial sector demand for currency UNEMP is the unemployment rate, INTENSITY is the intensity of regulations and control over the economy and TAXGDP is the tax-to-GDP ratio. The models also contain a dummy variable for structural break due to change in definition of demand deposits (SB⁵) taking a value 1 from 2006 to 2015 and 0 otherwise. The variable of interest in this study is INTENSITY, which proxies the government’s control over the economy. As already highlighted, it is represented by the government’s public administration and defence expenditure. Larger public sector employment and operational expenditure lead to greater influence on and monitoring of household and firm decision-making, limiting their freedom of choice. However, the difference between the two models is INTENSITY and INTENSITY_2 variables, where the first is log of the government’s public admin and defence expenditure, while later is the same in the per capita form. If the two measures are a true proxy of the same variable (excessive control), the size of coefficient should not change drastically with change in their calculation.

The second error-correction technique used is the autoregressive distributed lag (ARDL) bounds testing approach suggested by Pesaran and Shin (1996) and Pesaran et al. (2001). ARDL allows the use of both stationary and non-stationary variables and presents long-term and short-term relationships of each variable. Pesaran and Shin (1996) showed that ARDL-based estimators are super-consistent, and valid inferences can be drawn on the long-run parameters using standard normal asymptotic theory. Although the ARDL bounds testing approach does not require all the variables to be integrated of the same order, I(2) variables cannot be included, since computed F-statistic the under bounds testing approach are based on the assumption that variables are either integrated of order 0 or 1, that is, I(0) or I(1). While using ARDL for estimation, in addition to our three previously used shadow economy inducing variables, we also include an interaction term between the dummy variable dictator and CD (CD_DIC) as another proxy for the intensity of control and regulation in the economy. It is generally agreed that dictatorship regimes have more stringent control over the economy as against elected democracies and lead to decreased economic growth and higher inflation in the long run along with the weakening of institutions (Overland et al., 2005; Papaioannou & Zanden, 2015). The dictator variable has a value of ‘1’ during the period of dictatorships, while it is equal to ‘0’ during democracies. We estimated the following model with ARDL:

\begin{matrix} Δ C 2 D D_{t} = β_{0} + β_{1} C 2 D D_{t - 1} + β_{2} T A X G D P_{t - 1} + β_{3} U N E M P_{t - 1} \\ + β_{4} L n I N E T N S I T Y_{t - 1} + β_{5} L n G D P P C_{t - 1} + β_{6} I N F L_{t - 1} \\ \begin{array}{l} + β_{7} L n H H C O N P C_{t - 1} + β_{8} C D_D I C_{t - 1} + β_{9} S B_{t - 1} \\ \begin{matrix} {+ \sum^{​}}_{n}^{i = 0} γ_{1 i} Δ T A X G D P_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{2 i} Δ U N E M P_{t - i} \\ {+ \sum^{​}}_{n}^{i = 0} γ_{3 i} Δ L n I N T E N S I T Y_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{4 i} Δ L n G D P P C_{t - i} \\ \begin{array}{l} {+ \sum^{​}}_{n}^{i = 0} γ_{5 i} Δ I N F L_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{6 i} Δ L n H H C O N P C_{t - i} \\ \begin{matrix} {+ \sum^{​}}_{n}^{i = 0} γ_{7 i} Δ C D_D I C_{t - i} + {\sum^{​}}_{n}^{i = 0} γ_{8 i} Δ S B_{t - i} + μ_{t} . \end{matrix} \end{array} \end{matrix} \end{array} \end{matrix}

(3)

In all the three models, our cause variables represent additional demand for cash owing to informal activities; therefore, it is expected that they would appear with positive signs. Additionally, we expect positive signs with all the independent variables to explain the official demand for cash, except the structural break and error-correction terms. The variable SB is included in the long-run and short-run equations since the currency deposit ratio would decrease after the inclusion of time deposits in the value of demand deposits; hence, it should come up with a negative sign. We also expect negative signs with the lagged residual in the EG models (Models 1 and 2) and with the lagged dependent variable in ARDL (Model 3), since they represent the error-correction term and must have negative significant signs for the models to be meaningful.

Empirical Results

Engle–Granger Two-Step Approach

All the variables were I(1) as tested by Dicky Fuller, Philips Perron and Kwiatkowski–Phillips–Schmidt–Shin tests (Table A.1). The Akaike information criterion (AIC) has been used to include the optimal number of lags. The results of Models 1 and 2 from the second step after including the lagged residual from the first step of the EG approach are placed in Table A.2 which also contains the unit root test for the residual from the first step of the EG approach and the Error-correction term. The most important results to be discussed are placed in Table 1. The unit root test for the residual from the first step of the EG approach shows that it is stationary. Hence, the long-run cointegration among the variables is established. The ECT shows that disequilibrium is corrected by 42.8 per cent and 43.6 per cent annually in the case of Models 1 and 2, respectively. The diagnostics tests are presented in Table A.3. The null of ‘no serial correlation’ from the Breusch–Godfrey LM test and the ‘constant variance’ of the Breusch–Pagan/Cook–Weisberg heteroskedasticity test cannot be rejected. The difference between two models is basically the form in which ‘INTENSITY’ appears in each equation. The small difference in estimated coefficients despite difference in their composition shows that the variable indeed measures the extent of government control and regulations in the economy. All the variables in Models 1 and 2 have expected signs except GDP per capita, which is statistically insignificant.

Table 1.

Currency Demand Equation Using the Engle–Granger Two-Step Procedure (Models 1 and 2)

Cause Variable	Model 1	Model 2
Taxgdp	0.465*	0.439*
Unemp	0.227**	0.221**
Intensity	0.465**	–
Intensity_2	–	0.449**
R-square	0.5870	0.5876

Note: * p < .1; ** p < .05; *** p < .01.

Table 2.

Currency Demand Equation Using ARDL (Model 3)

Cause Variable	Model 3
Long run
Taxgdp	0.108***
Unemp	0.061**
Intensity	0.601**
Short run
Taxgdp	0.077***
Unemp	0.0438**
Intensity	0.428**
R-square	0.668

Note: * p < .1; ** p < .05; *** p < .01.

Most important for this study are the variables for informal activities, namely the tax-to-GDP ratio, unemployment rate and the intensity of regulations. All these variables appear with the expected sign and confirm our hypotheses. Hence, increased taxation, unemployment and intensity of control in the economy result in increased shadow economic activities, which is in line with the literature (Arby et al., 2010; Johnson et al., 1997; Loayza, 1996). A 1 per cent increase in the tax-to-GDP ratio increases the currency deposit ratio by 0.46 per cent in Model 1 and 0.44 per cent in Model 2. Similarly, an increase in the unemployment rate by 1 per cent causes the currency deposit ratio to increase by 0.22 per cent. Interestingly, the economic significance of the intensity of government regulation is the same as taxes. It significantly increases the shadow economy (Johnson et al., 1997; Loayza, 1996; Schneider & Dreher, 2006). This shows that a combination of dictatorships and democracies with varying level of intensity of control has had its effect in increasing the shadow activities. This result has serious implications for past estimates since an important variable with economic significance has not been considered by those estimates.

Autoregressive Distributed Lag Model

Model 3 was estimated using the ARDL bounds testing approach. Estimation results are presented in Table A.4. None of the variables is I(2). One of the advantages of this technique is that we get long-as well as short-run relationships among the variables. The statistics for tests of the serial correlation by the Breusch–Godfrey LM test and heteroscedasticity by the Breusch–Pagan/Cook–Weisberg test are presented in Table A.6 which show that there is no problem of serial correlation or heteroscedasticity in the model.

Results show that all the variables appear with the expected signs except CD_DIC in the short run, which is insignificant. The variables to measure the informal sector, namely TAX GDP, UNEMP and INTENSITY, also appear with the expected signs showing that rising taxes, unemployment and increased intensity of government regulations and control in an economy lead to a greater demand for cash and hence increased participation in the informal sector (Dutta et al., 2013; Loayza, 1996; Nagac, 2015; Schneider & Enste, 1998, 2000). These results again confirm our hypotheses. TAXGDP is significant at 1 per cent, while INETNSITY and UNEMP are significant at 5 per cent.

The bounds testing approach for the long-run cointegration is presented in Table A.5. Since the calculated F-statistic of 6.495 is above the upper bound; therefore, cointegration exists among the variables. The long-run normalized equation which will be used for the estimation of the shadow economy is

\begin{matrix} C 2 D D_{t} = - 5.407 + 0.127 T A X G D P + 0.081 U N E M P + 1.014 L n I N T E N S I T Y \\ + 0.318 L n G D P P C + 0.032 I N F L + 0.058 L n H H C O N P C \\ + 0.00003 C D_D I C - 0.684 S B . \end{matrix}

Size of the Shadow Economy of Pakistan

To obtain the estimates of the shadow economy, we use our results of the CD equation. Simulations based on coefficients on the cause variables are used to estimate the size of the shadow economy as a percentage of GDP. The general representation of the CD equation with theoretically accepted variables inducing demand for currency, and cause variables of the shadow economy can be expressed as

C u r r e n c y_{T o t a l} = β_{1} T A X G D P + β_{2} U N E M P + β_{3} L n I N T E N S I T Y + X

(4)

In Equation (4), the subscript ‘Total’ represents the demand for currency in formal and informal sectors combined and ‘X’ is a vector of all theoretically accepted variables inducing official demand for currency. The base period for the shadow economy cause variables is where they are at their minimum. Hence, the assumption—informal activities are relatively lowest in the base period—is used to estimate official demand for currency:

\begin{matrix} C u r r e n c y_{O f f i c i a l} = β_{1} (M i n i m u m_T A X G D P) + β_{2} (M i n i m u m_U N E M P) \\ + β_{3} (M i n i m u m_L n I N T E N S I T Y) + X . \end{matrix}

(5)

In Equation (5), the subscript ‘Official’ stands for the demand for currency when there is no shadow economy. The theoretical official demand for currency is subtracted from the observed (total) demand for currency, leading to the demand for currency generated due to informal economic activities, which is represented by the subscript ‘Informal’:

C u r r e n c y_{I n f o r m a l} = C u r r e n c y_{T o t a l} - C u r r e n c y_{O f f i c i a l} .

(6)

It has to be assumed that money velocity is the same in the formal and informal sectors. When multiplied with the velocity of money, the demand for currency from informal activities gives shadow economy as a percentage of GDP:

\frac{S h a d o w E c o n o m y}{G D P} = \frac{C u r r e n c y_{I n f o r m a l} \times M o n e y V e l o c i t y}{G D P} .

(7)

Using the definition of money velocity from the formal sector, we can rewrite Expression (7) as follows:

\frac{S h a d o w E c o n o m y}{G D P} = \frac{C u r r e n c y_{I n f o r m a l} \times G D P / C u r r e n c y_{O f f i c i a l}}{G D P}

(8)

\frac{S h a d o w E c o n o m y}{G D P} = \frac{C u r r e n c y_{I n f o r m a l}}{C u r r e n c y_{O f f i c i a l}} .

(9)

$C u r r e n c y_{I n f o r m a l}$ from Equation (6) is replaced in Equation (9):

\frac{S h a d o w E c o n o m y}{G D P} = \frac{C u r r e n c y_{T o t a l} - C u r r e n c y_{O f f i c i a l}}{C u r r e n c y_{O f f i c i a l}} = \frac{C u r r e n c y_{T o t a l}}{C u r r e n c y_{O f f i c i a l}} - 1.

(9)

The last expression in Equation (10) is used to estimate the shadow economy as a percentage of GDP, where Equations (4) and (5) can be used for $C u r r e n c y_{T o t a l}$ and $C u r r e n c y_{O f f i c i a l}$ , respectively. It needs to be highlighted that estimates of shadow economy from any technique, at best, give a trend in the shadow sector activities and in no case can be treated as exact, since estimates are subject to change with small changes in parameters. The main reason is the hidden nature of this sector, and even direct estimates are as downwards biased as the indirect estimates. Figure 1 is the graphical representation of the estimated size and development of a shadow economy in the percentage of GDP from all the three models.

Figure 1.

Estimates of Shadow Economy as a Percentage of GDP Using Models 1, 2 and 3

Pakistan has had unstable political history which is evident from its periods of dictatorships and democracies spread over time. This was one of the main reasons for choosing Pakistan, that is, to correctly identify the role of intensity of regulations. In order to elaborate on the yearly changes in estimates, each year in the figure has been coupled with the government regime. The name of the head of state is mentioned against a particular year if he/she was in power for six months or more in a given year. Interestingly, the policies under two regimes might also differ specifically in terms of our new variable, that is, ‘intensity of regulations and control over the economy’. Displaying the estimates of shadow economy across political regimes gives a meaningful understanding to our estimates. The shadow economy in Pakistan has been increasing overall since 1973; however, the increase has been more rapid during the period 1975–1980, which can be seen in all the three models. East Pakistan declared independence in year 1971 and became Bangladesh. A sharp increase in the shadow economy during the period 1975–1980 might be due to the effects of losing a part of the economy, especially when it had shown a better performance in terms of foreign revenue generation (Gull, 2015). Moreover, the democratic government of Zulfiqar Ali Bhutto was replaced by the Dictatorship of General Zia-ul-Haq in 1977, which might be another reason for a sharp increase in shadow economic activities.

Models 1 and 2 show declining shadow economic activities in the latter periods of Zia’s regime which is not clear in Model 3. For the rest of the years, the three models follow a similar pattern. The reversal might be due to more liberal policies along with structural adjustment plans under the International Monetary Fund by Zia in contrast to Bhutto’s period which was covered in the nationalization of many industries and strict control over the economy (Zaidi, 2005). From 1988 to 1990 and 1993 to 1996, Ms Benazir Bhutto was elected as the prime minister of Pakistan, but the government was dissolved in 1990 by President Ghulam Ishaq Khan and in 1996 by President Farooq Ahmed Khan Laghari on charges of nepotism and corruption (Bahadur, 1998). The same took place for Mr Nawaz Sharif’s tenure in 1993 (Bahadur, 1998). But during his later tenure, after 1996, the previous increasing trend seems to have reversed. However, the sharp decline in shadow economy from 1999 onwards represents the period of coupe by General Pervaiz Musharaf, which was again a dictatorship. Immediately after his coupe, some extreme measures were taken to control corruption in the government and an independent organization, National Accountability Bureau (NAB) was established to handle corruption cases. This along with setting up a taskforce on reforms of tax administration (Nazar, 2008) may have resulted in a sharp decline in corruption. However, after 2008–2015, there is again an increasing trend which might be owing to the resignation of President Musharaf.

Conclusion

Large size of an informal sector means that the burden on law-abiding citizens and firms is more than essential. Informal activities might be different from one culture to another and from developed to developing economies, but in general non-payment of taxes and license fees results in a similar effect as cross-subsidization, where the official sector participants bear the burden of informal sector firms and workers. Being a developing country, Pakistan needs to utilize its existing limited resources up to the maximum potential; however, due to such a large unofficial sector, it is losing a considerable amount of tax revenue.

The most important implication from our results is that we highlight a previously ignored aspect of the role of increased intensity of regulations in increasing the informal sector, with size of its coefficient almost matching the one with taxes. New laws and regulations often feed the bureaucracy’s hunger for power rather than making the system beneficial for its users. This aspect needs further attention in informal economy estimates of other countries, especially in developing economies where economic activities are excessively governed and controlled by the public institutions.

Similar to most of the studies on informal sector estimates, our results also show that increased taxation by the government is a major reason behind the growth of a shadow economy. In general, it is important to have policies which attract people towards the official sector by providing incentives. Economic benefits may have better returns in reducing informal activities rather than increasing the intensity of control over the economy. A recent policy initiative⁶ (2014–2015) is linking tax filing with the reduced cost of other documented facilities such as the reduced vehicle registration fee and reduced withholding tax at cash withdrawal from banks. However, the outcomes of these incentives will be visible in near future.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Appendix

Table A.6.

Diagnostic Tests for Model 3

Breusch–Godfrey LM test for autocorrelation H₀: no serial correlation	chi² = 0.135; Prob > chi² = 0.7129
Breusch–Pagan/Cook–Weisberg test for heteroscedasticity H₀: Constant variance	chi² = 2.16; Prob > chi² = 0.1416
Cameron & Trivedi’s decomposition of IM-test
Heteroscedasticity	chi² = 40.00; Prob > chi² = 0.4256
Skewness	chi² = 7.13; Prob > chi² = 0.7134
Kurtosis	chi² = 2.29; Prob > chi² = 0.1299
Total	chi² = 49.42; Prob > chi² = 0.4965

Notes

References

Ahmed

, & Ahmed

Q. M.

(1995). Estimation of the black economy of Pakistan through the monetary approach, The Pakistan Development Review, 34(4), 791–807.

Ahmed

Q. M.

, & Hussain

M. H.

(2008). Estimation of the black economy of through a monetary approach: A case study of Pakistan. Economic Issues, 13(1), 45–60.

Ali Kemal

(2007). A fresh assessment of underground economy and tax evasion in Pakistan: Causes, consequences and linkages with the formal economy [Working Paper No. 13, Pakistan Institute of Development Economics].

Ardizzi

, Petraglia

, Piacenza

, & Turati

(2011). Measuring the shadow economy with the currency demand approach—A reinterpretation of the methodology, with an application to Italy [Working paper No. 2], Department of Economics and Public Finance “G. Prato” University of Torino, Department of Economics and Public Finance G. Prato.

Aslam

(1998). The underground economy and tax evasion in Pakistan: Annual estimates (1960–1998) and the impact of dollarisation of the economy. The Pakistan Development Review, 37(4), 621–631.

Bahadur

(1998). Democracy in Pakistan: Crises and conflicts. Har-Anand Publications.

Bhatti

(2019, March 21). SC accepts Bahria Tows Karachi’s Rs460bn offer, halts NAB references. Dawn. Retrieved from https://www.dawn.com/news/1471002

Buehn

, Lessmann

, & Markwardt

(2013). Decentralization and the shadow economy: Oates meets Allingham–Sandmo. Applied Economics, 45(18), 2567–2578.

Buehn

, & Schneider

F. G.

(2009). Corruption and the shadow economy: A structural equation model approach [IZA Discussion Paper No. 4182]. Available at SSRN: https://ssrn.com/abstract=1409286

10.

Dell’Anno

(2008). What is the relationship between unofficial and official economy? An analysis in Latin American countries. European Journal of Economics, Finance and Administrative Sciences, 12, 185–203.

11.

Dell’Anno

, Gómez-Antonio

, & Pardo

(2007). The shadow economy in three Mediterranean countries: France, Spain and Greece. A MIMIC approach. Empirical Economics, 33(1), 51–84.

12.

Dickey

D. A.

, & Fuller

W. A.

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366a), 427–431.

13.

Dreher

, Kotsogiannis

, & McCorriston

(2009). How do institutions affect corruption and the shadow economy? International Tax and Public Finance, 16(6), 773.

14.

Drummond

, Éthier

, Fougère

, & Rudin

(1994). The underground economy: Moving the myth closer to reality. Canadian Business Economics, 2(4), pp. 3–17.

15.

Dutta

, Kar

, & Roy

(2013). Corruption and persistent informality: An empirical investigation for India. International Review of Economics & Finance, 27, 357–373.

16.

Engle

R. F.

, & Granger

C. W.

(1987). Co-integration and error correction: Representation, estimation, and testing. Econometrica: Journal of the Econometric Society, 251–276.

17.

Feld

, & Schneider

(2010). Survey on the shadow economy and undeclared earnings in OECD countries. German Economic Review, 11(2), 109–149.

18.

Frey

B. S.

(1997). Not just for the money. Edward Elgar.

19.

Frey

B. S.

, & Reto

(2001). Motivation crowding theory, Journal of Economic Surveys, 15(5), 589–611.

20.

Gull

(2015). Development of underdevelopment: The case of east Pakistan 1947–196. Journal of the Punjab University Historical Society, 28(1). Retrieved from http://pu.edu.pk/images/journal/HistoryPStudies/PDF-FILES/Ayyaz%20Gul_v28No1jun2015.pdf

21.

Gulzar

, Junaid

, & Haider

(2010). What is hidden, in the hidden economy of Pakistan? Size, causes, issues, and implications. The Pakistan Development Review, 49(4), 665–704.

22.

Iqbal

(2015, July 5). Tax on cash withdrawals stokes dollar demand. Dawn. Retrieved from https://www.dawn.com/news/1192489

23.

Iqbal

, Qureshi

S. K.

, & Mahmood

(1998). The underground economy and tax evasion in Pakistan [Research Report No. 158, Pakistan Institute of Development Economics].

24.

Johnson

, Kaufmann

, & Shleifer

(1997). The unofficial economy in transition. Brookings Papers on Economic Activity, 1997(4), 159–239.

25.

Kemal Ali

(2003). Underground economy and tax evasion in Pakistan: A critical evaluation [Research Report No. 184., Pakistan Institute of Development Economics].

26.

Khan

, & Mohsin

A. S.

(1993). Taxation system of Pakistan: Structure and trends. Economic Review. FindArticles.com.

27.

Kiani

, Ahmed

, & Zaman

(2015). Combining qualitative and quantitative approaches for measuring underground economy of Pakistan. Quality & Quantity, 49(1), 295–317.

28.

Kwiatkowski

, Phillips

P. C.

, Schmidt

, & Shin

(1992). Testing the null hypothesis of stationarity against the alternative of a unit root: How sure are we that economic time series have a unit root? Journal of Econometrics, 54(1–3), 159–178.

29.

Loayza

N. V.

(1996). The economics of the informal sector: A simple model and some empirical evidence from Latin America. Carnegie-Rochester Conference Series on Public Policy 45, pp. 129–162.

30.

Mauleón

, & Sardà

(2017). Unemployment and the shadow economy. Applied Economics, 49(37), 3729–3740.

31.

Mazhar

(2019). Terrorism and firm performance: Empirical evidence from Pakistan. The BE Journal of Economic Analysis & Policy, 19(1), 20180041. https://doi.org/10.1515/bejeap-2018-0041.

32.

Nagac

(2015). Tax system and informal economy: A cross-country analysis. Applied Economics, 47(17), 1775–1787.

33.

Nazar

(2008, April 28). The fiscal imperative of tax reforms. Dawn. Retrieved from https://www.dawn.com/news/300129

34.

Overland

, Simons

K. L.

, & Spagat

(2005). Political instability and growth in dictatorships. Public Choice, 125(3–4), 445–470.

35.

Papaioannou

K. J.

, & Van Zanden

J. L.

(2015). The dictator effect: How long years in office affect economic development. Journal of Institutional Economics, 11(1), 111–139.

36.

Pardo

(1995). Morals of legitimacy in Naples: Streetwise about legality, semi-legality, and crime. European Journal of Sociology, 36(1), 44–71.

37.

Pesaran

M. H.

, & Shin

(1996). Cointegration and speed of convergence to equilibrium. Journal of Econometrics, 71(1–2), 117–143.

38.

Pesaran

M. H.

, Shin

, & Smith

R. J.

(2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16, 289–326.

39.

Phillips

P. C. B.

, & Perron

(1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346.

40.

Schneider

(2005). Shadow economies around the world: What do we really know? European Journal of Political Economy, 21(3), 598–642.

41.

Schneider

(2006). Shadow economies and corruption all over the world: What do we really know? [CESIFO Working Paper No. 1806].

42.

Schneider

(2009). Size and development of the shadow economy in Germany, Austria and other OECD-countries. Revue économique, 60(5), 1079–1116.

43.

Schneider

(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.

44.

Schneider

, Buehn

, & Montenegro

C. E.

(2010). New estimates for the shadow economies all over the world. International Economic Journal, 24(4), 443–461.

45.

Schneider

, & Dreher

(2006). Corruption and the shadow economy: An empirical analysis [Working Paper No. 0603, Economics Department, Johannes Kepler University].

46.

Schneider

, & Enste

(1998). Increasing shadow economies all over the world—Fiction or reality. A survey of the global evidence, of its size and of its impact from 1970 to 1995 [Working Paper No. 9819, Department of Economics, Johannes Kepler University].

47.

Schneider

, & Enste

(2000). Shadow economies: Size, causes and consequences. Journal of Economic Literature, 38(1), 73–110.

48.

Schneider

, & Halla

(2005). Taxes and benefits: Two distinct options to cheat on the state, Department of Economics, University of Linz, Working Paper No. 0505.

49.

Schneider

, & Hametner

(2007). The shadow economy in Colombia: Size and effects on economic growth, Department of Economics, Johannes Kepler University, Working Paper No. 0703.

50.

Shabsigh

(1995). The underground economy: Estimation, and economic and policy implications—The case of Pakistan, International Monetary Fund, Working Paper No. 95/101, Washington, DC.

51.

Shahbaz

(2017, November 1). Pakistan now ranked 147th in World Bank’s Ease of Doing Business Index. The Express Tribunal. Retrieved from https://tribune.com.pk/story/1546434/2-pakistan-now-ranked-147th-world-banks-ease-business-index/

52.

Spiro

P. S.

(1993). Evidence of a post-GST increase in the underground economy. Canadian Tax Journal, 41(2), 247–258.

53.

Tanzi

(1980). The underground economy in the United States: Estimates and implications. Banca Nazionale del Lavoro, 135(4), 427–453.

54.

Tanzi

(1983). The underground economy in the United States: Annual estimates, 1930–1980. [IMF-Staff Papers, 30(2), 283, H. 305].

55.

Tanzi

(1998). Corruption around the world-causes, consequences, scope and cures. International Monetary Fund Staff Papers, 45, 559–594.

56.

Yasmin

, & Rauf

(2003). Measuring the underground economy and its impact on the economy of Pakistan. The Lahore Journal of Economics, 9(2), 93–103.

57.

Zaidi

S. A.

(2005). Issues in Pakistan’s economy (2nd ed.). Oxford University Press.