The role of specific macroeconomic factors on tax revenue policy in an oil-dependent economy: evidence from Oman

It has been widely recognised by economists that in times of economic or financial crisis, the government may try to address the issue through monetary or fiscal policy (Stone, 2020; Nils et al., 2019). From the standpoint of fiscal policy, the government can utilise taxation or spending, contingent on the status of the economy and the crises encountered, to accomplish its goals in managing the economy (Besnik, 2017).

According to Piancastelli and Thirlwall (2020), fiscal sustainability is considered a primary goal for any nation and is essential to both macroeconomic stability and sustained growth. It assists in locating risk and susceptibility points in a nation’s macro and fiscal structures and makes recommendations for suitable measures to avert sudden macroeconomic crises (Pradhan, 2019). In a nation where tax revenue from citizens, businesses and investors is crucial to the government’s budget, taxes are seen as one of the major sources of income (Yushko, 2019). Taxes have the ability to increase income for the government, reduce the cost of collection and, most crucially, increase competition with smart policy design (Besim, 2014). In addition, taxes are seen as a key tool for resource allocation, income redistribution, mitigating externalities, promoting economic growth and stability and enhancing governance (Moore, 2015; Prichard, 2016).

The entire economy is subject to a variety of tax consequences, which can either be positive or negative (Gechert and Heimberger, 2021; Kate and Milionis, 2019; Neog and Gaur, 2020). However, this is contingent upon the tax’s nature and rate structure. Certain taxes have a greater impact than others. According to economic theory, an increase in taxes (excluding lump-sum taxes) could have a negative impact on economic growth (Macek, 2015). Considering a simple production function, it can be seen that taxes have an effect on growth through their effects on physical and human capital, which in turn affects total factor productivity and, ultimately, economic growth (Arachi et al., 2015; Kate and Milionis, 2019).

The recent COVID-19 epidemic crisis has recently highlighted how crucial fiscal policy is to reducing the effects of the crisis that led to the severe global recession (World Bank, 2021). According to the World Bank report from 2021, the Omani government, for example, halted municipal taxes, various government fees and rent payments for businesses located in industrial zones for around half a year. Additionally, loan services for borrowers and small and medium-sized businesses have been delayed, and freight prices for ports and planes have been lowered. The government unveiled the Economic Stimulus Plan in March 2021, which consists of a combination of pandemic-related policy support initiatives and structural improvements to encourage economic diversification.

The COVID-19 pandemic’s effect, for instance, contributed to weak economic development, a large budget deficit, rising inflation and a significant current account deficit for Oman [Central Bank of Oman (CBO), 2020]. Another crisis that affects the economy is exclusive to the inflation, which rose from 1.1% in 2013 to 1.6% in 2017 and then fluctuated up to the study period (NCSI, 2020). Comparably, from a surplus of RO 2,001 million in 2013 before the drop in oil prices to a deficit of RO 4,821 million in 2016 and RO 1,648 million in 2018. This influence affected the status of the current account deficit, which has gotten substantially worse (CBO, 2022). These metrics make it abundantly evident how vulnerable the Omani economy is to shocks to oil prices, which made it imperative for the government to revise its fiscal policy of taxation to be an alternative source for funding the deficit.

Oman has the least amount of fiscal room and resources compared to other Gulf Cooperation Council (GCC) countries, so raising taxes and reforming the taxation system is a top objective for the authority. High tax rates and complicated tax laws, however, carry a number of risks that could negatively affect local business, especially in this challenging time of pandemic recovery, hinder attempts to draw in foreign investment to help the economy’s transition away from oil and possibly spark social unrest.

The majority of countries in the GCC, including Oman, mostly rely on the crude oil market. Because of how erratic the global market prices for oil and gas are, economies are vulnerable to shocks related to rising oil prices. The economy of Oman suffered greatly during the economic recession of 2020, which was brought on by the prices’ decline in the world energy market due to the COVID-19 pandemic. As a result, the government in the Sultanate of Oman began looking for revenue streams beyond the energy sector in order to maintain economic expansion. The tax policy was updated by the Omani economy to increase efficiency and revenue collection and fund the budget deficit. Therefore, determining the important elements that affect tax revenues is significantly raised in this study. Specifically, this study contributes to investigating the factors affecting the magnitude of tax revenue policy changes in the Omani economy from 1991 to 2020.

This investigation is the main focus of this study for a number of reasons. First, the Omani government’s tax and fiscal policies are intended to be the primary source of financial stability for the Omani economy. Notably, the Omani national currency, the Riyal, is fixed to the US dollar as a means of avoiding fluctuations in exchange rates, particularly given the nation’s reliance on a single commodity (Alkhater and Basher, 2016). Second, recent spikes in oil prices have caused Oman’s fiscal policy to shift, just like it has in other oil-dependent nations. Thirdly, there are unique characteristics of the Omani economy, chief among them being the nonexistence of personal income tax. Thus, public spending financed by the oil and gas tax is a major component of expansionary fiscal policies, which could not be sustained for long. Fourth, from a macroeconomic-level analysis, no study has verified the effect of the shadow economy (SE) on the tax revenue policy in the economy of Oman, as an oil-dependent country. Thus, this study is the first to consider the relationship between the duo. Lastly, the Oman economy’s reliance on oil and gas revenues – which normally cover roughly 70% of its yearly budget through taxes on oil and gas sold in the global energy market – means that the externalities of imported inflation have a direct impact on the nation’s fiscal base. Therefore, this study determines how several macroeconomic variables influence Oman’s tax revenues.

The rest of this paper is arranged as follows: Section 2 presents the relevant theories in the context of Oman’s tax revenue policy and the relevant empirical literature on the issue shortly. The methodology, estimation techniques and data used are described in Section 3. Section 4 discusses the empirical results, and finally, Section 5 concludes the study with recommendations.

This section provides a brief overview of the tax policy in Oman and presents a concise summary of the literature relevant to the topic.

Between 1991 and 2020, Oman’s tax policy was designed by its hydrocarbon-dependent economy, with oil revenues dominating fiscal inflows and non-oil taxation remaining underdeveloped. During the 1990s, the economy relied heavily on oil and gas returns, which constituted 75–85% of total tax revenue, while non-oil taxes profile, such as corporate income tax (15% on non-hydrocarbon businesses) and customs duties (5% on imports), contributed slightly (Masan, 2016; World Bank, 2021). The absence of personal income tax and value-added tax (VAT) highlighted a regressive tax structure, listing resource rents over equitable revenue streams. This period also saw limited institutional reforms, with tax administration stalled by the informality activities (valued at 30–35% of the official GDP) and reliance on foreign labour (Al Abri et al., 2023; Medina and Schneider, 2018).

The 2000s oil boom (2001–2014) deepened dependence on hydrocarbons, as oil prices averaging 60–100/barrel inflated oil returns to 85% of total income by 2014 (CBO, 2015); Despite Vision 2020 diversification goals, non-oil tax revenues stagnated at 10–15%, reflecting weak enforcement and excessive facilities for foreign investors (Al Abri et al., 2023). Corporate tax collections from non-oil sectors were subsidized to 6% of total revenue by 2015, while customs duties fell just 4% (CBO Reports, 2010-2021). The 2016 oil price reduction ($30/barrel in January 2016) showed systemic liabilities, prompting austerity procedures and postponed fiscal reforms, including excise taxes (2019) on tobacco goods and energy drinks (5–100%) and plans for VAT implementation (CBO, 2015; Krupa et al., 2019).

Post-2016 reforms were designed to mitigate fiscal risks, raising non-oil tax returns to 28% by 2020 through stricter corporate tax compliance and tourism duties (CBO, 2021). However, structural challenges made Oman’s fiscal breakeven oil price rise to $82/barrel in 2023 (CBO, 2023), while tax exceptions for foreign companies and a large informal economy continued to restrict improvement. The late implementation of VAT (April 2021) marked a serious shift, but the 1991–2020 period highlighted Oman’s struggle to poise hydrocarbon needs with sustainable tax diversification. However, diversifying non-oil industries could help mitigate fiscal stress related to taxes. Nevertheless, this policy alone, without addressing the SE, will not be adequate. According to this concise presentation, Table 1 outlines Oman’s tax revenue structure over the period from 1991 to 2020.

Kassa (2021) evaluated the elements that influence Ethiopian taxpayers’ decisions to engage in tax evasion. The study’s sample comprised 370 taxpayers in the city of Woldia. The study findings showed that tax fairness, tax knowledge and moral obligation are the most influential factors in tax evasion (Mukhlis et al., 2015). Taxpayers who are more knowledgeable about taxes are more likely to comply with the tax system, as seen by their timely filing of tax returns, truthful reporting of tax matters and accurate computation of tax liabilities. Tax compliance was found to be positively connected with the function of the tax authority, how it treats taxpayers fairly and the adoption of the self-assessment system (Serem et al., 2017).

Mina (2015) conducted a study on the influence of the SE on tax income in 125 countries between 1990 and 2011. The study’s findings indicated that if the SE’s size increased beyond a certain threshold, it would have a negative impact on indirect tax collection. The report suggested streamlining tax legislation, bolstering tax enforcement and offering incentives to individuals and businesses to make payments through official banking channels in order to reduce the extent of the SE and increase tax collections. The existence of a negative association between the levels of tax collection and the SE was proven by Awasthi and Engelschalk (2018).

By surveying 419 taxpayers, Nivakan et al. (2020) investigated the impact of social determinants on individual taxpayers’ tax compliance behaviour in Malaysia. Regression and correlation analysis were employed in the study. The study’s primary finding was that political affiliation, referral networks, the function of tax authorities and policy changes are the main variables influencing taxpayers’ tax compliance behaviour in Malaysia.

Lapatinas et al. (2019) used fixed-effects two-stage least squares/instrumental variables (FE 2SLS/IV) to examine the relationship between tax structure and economic sophistication in 17 OECD nations covering the years 1970–2001. The study’s key conclusion, which is further corroborated by the degree of economic development, is that capital taxes have a detrimental impact on economic sophistication. Pokharel (2019) examined the impact of tax reforms on economic growth in South Asian Association for Regional Cooperation countries over the period 1990–2012, using panel data regression analysis and fixed effect and random effect model regression. The study revealed the existence of a significant negative relationship between taxation and economic growth in these countries.

Gechert and Heimberger (2022) demonstrated that reductions in cooperative taxes will increase economic growth in order to enhance economic growth. The meta-regression study used a dataset of 441 estimates from 42 primary research studies. A one percentage point decrease in the corporate tax rate would only slightly raise annual GDP growth rates by approximately 0.02% points, based on the unweighted average of all projections in their study dataset. In a separate study, Kate and Milionis (2019) examined 77 countries, including OECD members as well as a number of other Asian, African and Latin American nations. The results indicated a negative correlation between capital taxes and growth in the sample’s low-income nations.

Victor (2018) studied the determinants of tax in 138 countries over the period 1976–2015, using static and dynamic models. The study found that the share of agriculture in gross domestic product, financial intermediation, natural rents, education, population share above 65 years, quality of government and democracy are the most influential factors for tax revenue. Al Jabri et al. (2022) examined the relationship between oil price shocks, fiscal policy and real GDP in Oman, using quarterly data from 1989 to 2016 to provide empirical insights into the economy of this oil-exporting country. By applying a structural vector autoregression model, the study reveals that oil price shocks account for approximately 22% of the variation in aggregate government revenues. Despite the valuable contribution of this study, it did not consider the impact of macroeconomic variables on tax revenues in Oman, as a stand-alone variable was not considered, which justifies the need for the current research.

The recent study by ALshubiri (2024) examines the impact of FDI inflows on tax revenue across 34 developed and developing countries from 2006 to 2020 in the form of panel data analysis, including Oman. The analysis applies a range of advanced econometric methods, including Feasible Generalised Least Squares (FGLS), a two-step system Generalised Method of Moments (GMM) and Pool Mean Group (PMG)-based panel autoregressive distributed lag (ARDL) approaches, to enhance the reliability of the results. The findings stated that in the developed countries, FDI is negatively affecting tax revenues in the long term, while in the short run, the relation is positively observed. On the other hand, results show that the effect from FDI towards tax revenues is significantly positive, in the countries with developing profiles, while in the short run, the relation effect is negative. The important note of this valuable study is that it confirms that the issue of the impact of FDI inflows on the magnitude of tax revenues does not depend on the level of development of the country in terms of whether it is a developed or developing country but rather depends on the optimal rates of the tax burden imposed on taxpayers, which has to be determined taking into consideration many variables that may affect the profile of tax revenues if ignored, such as the SE. Based on the literature presented, it is emphasised that studies on tax revenues and their determinants in the economy of Oman are scarce. Thus, conducting the current study will add value to the beneficiaries.

This section explains the specified econometric model of the tax revenues and the estimation process for the relevant models.

This study on Oman’s tax revenue policy is grounded in the Resource Curse and Dutch Disease theories proposed by Copithorne (1979) and Corden and Neary (1982). These theories suggest that countries heavily reliant on natural resources, such as oil, often ignore the development of a robust tax system due to the availability of easy revenue streams. This can lead to fiscal vulnerability and hinder efforts toward economic diversification. These theories emphasise the importance of fiscal diversification to reinforce the economic sectors and ensure long-run financial stability. Attaining this also involves addressing economic issues such as inflation and the effect of the SE, as highlighted in this study.

The general econometric model capturing the established relationship between the dependent variable and independent variables is specified in its logarithmic form as follows:

where TRC_t is the tax revenue policy change at time t as the dependent variable, SEA_t is the size of economic activity at time t, proxied by the current gross domestic product (GDP), FDI_t is the foreign direct investment inflows at time t, DSP_t is the variable of the diversification support program at time t, proxied by the share of sectors targeted by the government for diversification, which are manufacturing, tourism and the logistic sector, SE_t is the size of the SE in Oman at time t. CPI_t is the inflation rate in the country at time t, measured by the percentage change of the consumer price index. ϑ₀ is the intercept, and σ₁, σ₂, σ₃, σ₄, and σ₅ are parameters to be estimated. ε_t is the stochastic error term with ε_t ∼ iid(0, σ²) and ln refers to the logarithmic.

The explanatory factors counted in the econometric model of tax revenue were chosen based on their direct relevance to the tax revenue policy change in Oman. The GDP variable, which assesses the size of economic activity, is entered as it is an ultimate indicator of overall economic act, directly affecting tax revenue. FDI is accounted for because of its role in motivating economic activity, which can influence both tax revenues and the competitive environment for domestic industries. The variable of DSP captures the Omani government’s efforts to diversify the economy, which is vital for increasing the tax base beyond normal oil revenues, mainly through targeted industries like manufacturing, tourism and logistics. The variable of SE is considered because unrecorded economic activity leads to tax elusion, reducing potential revenue. Lastly, CPI reflects the economic conditions that affect consumer behaviour and tax compliance, influencing tax revenue.

Since the analysis of the study model belongs to time series data, the requirement of the stationarity investigation for the unit root is needed.

Before estimating the specified model of Equation (1) in the form of the ARDL model, the study utilised the unit root tests of the conventional Augmented Dickey–Fuller (ADF) and Zivot and Andrews (1992) in their form of both intercept and trend as presented in Equations (2) and (3):

where in Equation (2), $Δ$ is a first difference operator, $Xt$ is an economic variable under investigation, t is the time trend, α is the intercept, X_t-1 is the lagged variable under consideration and ΔX_t-1 is the first difference lagged variable commonly considered to take care of the serial correlation problem (Dickey and Fuller, 1979). k is the number of lags length selected based on the Akaike Information Criteria (AIC) or Schwarz Bayesian Criteria (SBC), and ε is a white noise error term with a zero mean and a constant variance, i.e. $εt∼iid(0,σ2)$⁠. In this study, the conventional ADF test is used alongside the Zivot and Andrews (1992) test to ensure a robust stationarity analysis by accounting for both standard unit root behaviour and potential structural breaks in the time series data.

However, because of the weakness found with the conventional unit root tests, including the ADF test (Gregory et al., 1996). This study attempts to test for a unit root against the alternative of a one-time structural break, utilising the Model (B) approach developed by Zivot and Andrews (1992); as in Equation (3), which denotes the change series’ slope with the occurrence of a trend at an unknown time break-point $Tb.$ $DTt$ in Equation (3) denotes a dummy variable that corresponds to a variable $t$⁠, which denotes a shift in the trend. Strictly, $DUt=1$ if $t>Tb$⁠, and 0 if $t≤Tb$⁠, while the dummy of $DTt=t−Tb$ if $t>Tb$⁠, and 0 if $t≤Tb$⁠, whereas $Tb$ is the date or point wherein the structural break occurred. The analysis procedure is carried out using the former two tests, as they were in the literature.

The study adopts the ARDL bounds testing approach of cointegration developed by Pesaran et al. (2001) to investigate the relationship between the tax revenue policy changes and proposed macroeconomic variables in Oman. The ARDL model is used in this study due to its suitability for analysing relationships between variables that are integrated at different orders, specifically I(0) and I(1). It also provides reliable long-run and short-run estimates even with small sample sizes, making it ideal for the available data from 1991 to 2020. The estimated model, in the form of the ARDL model technique, can be specified based on Gamal et al. (2019), with some amendments as follows:

where $Δ$ is the first difference operator, $φ0$ symbolises a drift component and D_t is the dummy variable that accounts for the influence of the structural break date. For determining the exact structural date, the study uses the cointegration technique developed by Gregory et al. (1996) to precisely identify the exact structural break that needs to be taken into account in the analysis of ARDL. This is despite the fact that the Gregory–Hansen test is itself a methodology designed to investigate cointegrating long-run relationships while accounting for structural break (Gamal et al., 2020). The dummy variable for structural break is used to account for significant policy changes or externalities that might influence the connection between the regressors and tax revenue policy’s variable. These variables jointly examine both the economic environment and policy actions that form Oman’s tax revenue dynamics. The variable of $Δln⁡TRCt−i$ identifies the changes in the lagged term of the dependent variable. The coefficients, variables and statistical symbols are as defined in equation (1). The F-test statistic examines the postulations of no or absence of cointegration between the variables in a specified model against the presence of a long-run cointegration and is represented as; H₀: α₁ = α₂ = α₃ = α₄ = α₅ = α₆ = 0, the null hypothesis for no cointegration, against; H₁: α₁ ≠ α₂ ≠ α₃ ≠ α₄ ≠ α₅ ≠ α₆ ≠0, an alternative hypothesis for the existence of cointegration among the variables.

The guidelines for the decision on the presence or absence of the cointegrating relationship will be carried out as in the literature of the ARDL specification.

From equation (4), the dynamic short-run ECM model is specified in equation (5):

with the ECM specification, both long-run and short-run information are incorporated together. The ECMt-1 error correction term measures the speed of adjustment of the model. It shows how much of the disequilibrium is being corrected and should have its advantages; significant, less than 1 and negative.

In this study, the estimated results for the specified variables on tax revenue policy in Oman, based on the ARDL methodology, will be verified using the fully modified ordinary least squares (FMOLS) and the canonical cointegration regression (CCR) procedures. FMOLS and CCR are employed to verify the ARDL results as they provide robust long-run estimates by correcting for endogeneity and serial correlation in cointegrated systems.

The study uses annual data spanning from 1991 to 2020. The data for the study are sourced from various databases. Specifically, the data for the variables of SEA, TR, FDI inflows and DSP are sourced from the Omani National Center for Statistics and Information (2020). This includes four major categories: taxes on companies and establishment profits, taxes on property, domestic taxes on goods and services and customs duties.

The data for the existing SE in Oman are extracted from Medina and Schneider (2018). Although the data are available for Oman during the period 1991–2017, missing data are obtained by interpolation methods using moving average procedures. For the inflation rate, which is measured by the consumer price index (CPI), the variable is sourced from the CBO.

This section introduces the results attained from the data analysis using different procedures. Also, it provides the descriptive statistics and correlation analysis of the variables employed in the empirical investigation of Tax revenue determinants in the Omani’s economy. Table 2 presents the results of the descriptive statistics.

The descriptive statistics in Table 2 provide insights into the central tendency, variability and shape of the statistical distributions for each variable. The tax revenue change (LTRC) variable has a mean value of 21.880 and a standard deviation of 0.506, with values ranging from 20.773 to 22.875, indicating limited variation. Its slight negative skew of −0.234 points to a small slope to the left, while the kurtosis of 2.379 shows that the values are fairly evenly distributed. Similarly, the size of economic activity variable (SEA), with a mean value of 24.500 and a standard deviation of 0.520, also shows little variation across values (range: 23.243–25.179), but its skewness of −0.565 reflecting a left-tailed distribution. Its kurtosis of 2.552 hints at a more peaked shape. The variable of the SE has the highest variation, with a mean value of 18.661, a standard deviation of 3.098, and a broader range from 13.300 to 24.740. However, its near-zero skew (−0.231) and kurtosis of 2.087 indicate a roughly symmetric and slightly flatter distribution.

The variable for foreign direct investment (FDI), with a mean value of 19.589 and a standard deviation of 1.772, exhibits moderate variability (range: 15.464–22.505), near-symmetric skewness (0.004) and platykurtic kurtosis (1.878), indicating a distribution close to normal but with lighter tails. In contrast, the diversification support programme (LDSP) has a mean of 21.544 and a standard deviation of 0.861, showing a right-skewed distribution (skewness = 0.493) and low kurtosis (1.654), suggesting a flatter peak and thinner tails. The CPI variable stands out with a mean of 1.944 and a standard deviation of 2.011, reflecting an extreme range (−0.900 to 9.870), strong positive skewness (1.698) and leptokurtic kurtosis (6.792), which indicate the presence of high-value outliers and substantial volatility. All variables share 30 observations, ensuring sample consistency. These findings underscore the need to address non-normality (e.g. CPI’s skewness and kurtosis) and variability (e.g. SE’s dispersion) in subsequent analyses to ensure robust results.

As for correlation matrix analysis between the tax revenue variable and its determinants, Table 3 presents the Spearman’s rank correlation parameters and their corresponding likelihood values for each pair of variables in the model.

The correlation coefficients, sited above the p-values, present the strength and direction (positive or negative) of the relations. The results show that TRC displays significant correlations with most explanatory variables, including SEA, SE, FDI and LDSP. Specifically, LTRC is negatively correlated with SEA (r = −0.357), FDI (r = −0.706) and LDSP (r = −0.515) and positively correlated with SE (r = 0.541), aligning with the findings of Yanikkaya and Turan (2020). These relationships indicate that restrictive tax policies may diminish economic activity and discourage foreign investment, while potentially driving the growth of the SE. Inflation, however, shows a weak and statistically insignificant negative correlation with LTRC (r = −0.247), indicating no meaningful association. The correlation analysis further displays that SEA is abstemiously and negatively associated with SE (r = −0.590), but positively and significantly linked to FDI (r = 0.666).

Overall, inflation displays weak and insignificant associations with all explanatory variables in the model. Particularly, while these pairwise correlations propose potential symmetric associations, they do not recommend causality. The correlations analysis illustrates as an initial diagnostic for multicollinearity, with none of the correlation coefficients surpassing the 0.80 threshold (Kennedy, 2008). As such, despite outstanding correlations among LTRC and main explanatory variables, there is no immediate effect of multicollinearity. Nevertheless, variables with relatively high associations, predominantly FDI and LDSP, may permit alteration to alleviate risks of spurious interpretation in subsequent econometric modeling. The results justify further examination using more robust procedures capable of capturing accurate linear, nonlinear relationships and direction of causation.

Table 4 presents the conventional ADF unit root test using both intercept and trend. The results showed that the SE and CPI variables are stationary at the 1% significance level. However, the rest of the tested variables are stationary after taking their first difference at the 5% significance level. The order of integration showed that the SE and CPI are integrated of order zero, i.e. [I(0)], while the rest are integrated of order one, i.e. [I(1)].

Evidence in the literature has indicated that the presence of structural breaks in the data-generating process of a series will lead to size distortion and eventually a spurious estimate in the ADF model (Zivot and Andrews, 1992; Perron, 1997). Due to this drawback of the ADF unit root test, this study uses the Zivot and Andrews (ZA) unit root test with structural break to further check the unit root properties of the tested variables in the underlying model. The ZA unit root test with structural break is built on the assumption that the true break point is endogenously estimated at an unknown point (Zivot and Andrews, 1992).

Table 5 presents the ZA unit root test results in levels and at first difference using an intercept and trend.

The findings indicate that variables of the LTRC, SE, FDI and CPI are all stationary in level form, while the rest of the variables are stationary at the 5% significance level after taking their first difference with different structural break points. The ZA test results show that the variables are integrated by the I(0) and I(1) processes, confirming the integration-order properties as in the ADF with different structural break points.

After taking into account the model’s order of integration, which satisfied the criteria for estimating an ARDL model, the study looks at the long-term cointegrating relationship between the tax revenue policy changes and its explanatory variables. The Gregory and Hansen cointegration technique with three forms (level shift model GH-1, level shift with trend model GH-2 and regime shift model GH-3) pits an alternate hypothesis of cointegration against the null hypothesis of no cointegration with a one break point, which is endogenously determined.

The Gregory and Hansen results are shown in Table 6. Over the three models, the null hypothesis was rejected at the 5% significant level, implying the existence of the long-run cointegration between tax revenue changes and its factors with a one-break point of 2002. The year 2002, which is the break date in each of the three models, is linked to a significant structural shift in the Omani economy.

Based on this result, the Gregory–Hansen cointegration technique’s break date will be taken into account. The ARDL model is estimated to examine the factors influencing tax revenue behaviour in the economy of Oman with the inclusion of structural breaks. The result is reported in Table 7, indicating that the F-statistic of 7.412 exceeds the critical values of the lower and upper bounds of 2.685 and 3.960 at the 5% level of significance.

The result, confirming that the null hypothesis of no cointegrating relationship between the variables under study is rejected, implies the presence of cointegration. This agrees with the Gregory–Hansen cointegration approach results.

Table 8 displays the long-run coefficient estimates in Panel A and the short-run estimates with a maximum lag length of four in Panel B. The optimal lag duration of the estimated model is found to be (1, 4, 2, 3, 4, 2, 3, 1) using the AIC. The diagnostic tests are presented in Panel C.

The long-run relationship results are depicted in Panel A of Table 8. It can be seen that, at a 5% level of significance, the SEA has a statistically significant positive effect on the LTRC variable. This suggests that LTRC increases by 57% for every 1% rise in the SEA. The results are consistent with recent empirical studies that highlight a significant positive relationship between revenue and economic growth in Oman (see Javed and Husain, 2024; Al-Saadi and Khudari, 2024; Ahmad and Masan, 2015).

Furthermore, the result demonstrates that the SE has a negative effect on the LTRC variable at the 5% significance level. This implies that a 1% increase in LSE activities decreases the LTRC by 31% in the long run. This result aligns with the findings of Bhuana and Wijaya (2024) and Gnangnon (2023), who demonstrated that the SE weakens tax revenue collections in economies dependent on international trade taxes from a single oil-based commodity, such as Oman. The result could be explained by the fact that the tax revenues can be negatively impacted by the SE through a number of different means. First, the predominance of cash transactions and informal economic activities within the SE diminishes the tax base, as such transactions often evade taxation altogether. Additionally, the concealed nature of the SE makes it difficult for tax authorities to monitor and enforce compliance, leading to widespread tax evasion and revenue losses.

Moreover, there is a negative relationship between LTRC and FDI inflows at a 5% significance level. This implies that the LTRC variable is reduced by 25% for every 1% rise in the FDI inflows. This result aligns with the result found by Shani (2025), but it is contrary to the result found by ALshubiri (2024). The argument behind this result is that, in an oil-dependent economy like Oman, FDI inflows can exert a negative impact on tax revenues through several channels. First, multinational corporations operating in the oil sector often negotiate tax incentives and exemptions as part of investment agreements with the host country, leading to reduced tax revenues for the government. Additionally, the presence of different foreign firms in the country may intensify tax competition among industries, prompting the host government to offer preferential tax treatment to attract FDI, further eroding tax revenues. Moreover, FDI inflows may stimulate imports of capital goods and technology, leading to increased tax deductions for depreciation and royalties, thereby reducing taxable profits and overall tax revenues. This result lends credence to the idea that governments discourage tax collection by promoting FDI through incentives and by luring investors with tax breaks and tariff exemptions. This result is consistent with the findings of Gerardo and Diana (2014).

The outcome also demonstrates that, over the long term, the INF variable has a negative impact on the LTRC variable at a 5% significance level. For every one-point increase in the INF within the CPI leads, the LTRC variable decreases by 29%. The erosion of the real value of tax receipts as a result of the delay between the tax due date and the recovery date can be used to measure this negative relationship. The LDSP variable has a positive effect on the LTRC variable and is statistically significant at the 10% significance level. This result suggests that with additional government support for the private sector, the development of fiscal taxes will be higher. This result is in tandem with the findings of Regassa (2017) and in mismatch with the result found by Basheer et al. (2019). Lastly, the structural dummy variable of 2002 is negative and statistically insignificant, which does not make sense.

These results of the long-run effect of the specified variables on tax revenue policy in Oman are confirmed by the FMOLS and the canonical cointegration regression (CRR) procedures.

Results reported in Table 9 demonstrated that the FMOLS and CRR approaches are in tandem with the results of the ARDL model in terms of the significance, magnitude and directions of effect, except for the variable of the time break effect, where it was significant at the 5% significance level.

The short-run relationship estimates reported in Panel B of Table 8 reveal that the revenue from taxes has a statistically insignificant negative response to the SEA variable, except for the third lagged period, where the variable had a statistically negative effect on tax revenues at the 5% significance level. This indicates that a one-unit increase in SEA causes revenue from taxes to decline by roughly 35%. On the other hand, the SE is statistically significant at the 5% level and has a different impact on the total tax revenues. From the result, at a 5% significance level, there is a positive effect of the current and third-lag variables of the SE on the revenues from taxes.

Furthermore, the results show that the SE lagged variable in the first and second periods has a detrimental effect on the tax revenues at the 5% significance levels, respectively. Furthermore, at a 5% significance level, the results show a significant inverse relationship between the current and lagged FDI inflows variables and revenues from taxes. A 1% rise in FDI’s variables reduces the government’s short-term tax revenue collection by 31%, 15% and 8%, respectively.

The initiative to foster diversity by the DSP projects has a favourable and significant effect on the tax revenues in Oman. Tax revenues increase by 57% and 80%, with a 1% increase in the current and third-lagged DSP variables at the 5% significance level. However, at the lagged periods of one and two, respectively, the DSP’s variable was statistically insignificant. From the results, the current variable of inflation and tax revenues’ variable have an inverse relationship that is significant at the 5% level. A 1% increase in inflation reduces Oman’s tax revenues by 16%. Compared, the long-run ARDL estimation’s result showed that the magnitude of the effect of inflation on tax revenues is almost double compared to the short-run result. In contrast, the short-run break date of 2002 is statistically significant in determining the total tax revenues in the Sultanate of Oman.

At the 5% level of significance, the estimated result of the error correction term (ECT-1) is statistically significant, negative and less than its magnitude, which suggests that there may be some convergence towards the long-run equilibrium. To reach the long-term equilibrium position, the error-correcting mechanism may foster cointegration or convergence. From the result, the ECM coefficient indicates that about 83% of the adjustment speed should be rectified once a year. Long-term equilibrium positions are quickly restored.

The outcomes of the diagnostic test show that the serial correlation-free residuals are supported by the probability value of 0.11 for the LM test, which is insignificant. Heteroscedasticity, normality and misspecification errors all have probability values of 0.90, 0.33 and 0.39, which are higher than 5%, respectively. This demonstrates that the model is fit, appropriately specified, homoscedastic and normally distributed. The model’s goodness of fit is confirmed by the corrected R² of 94%, indicating that about 94% of the variation in the LTRC model is explained by its explanatory factors.

The stability investigation result of the coefficients is demonstrated by the CUSUM and CUSUMSQ, as shown in Figure 1. Plots (a) and (b) in CUSUM and CUSUMSQ remain within the crucial 5% upper and lower boundaries over the sample period, indicating the model is stable. This expected result may reflect the moderate fiscal taxation policy in the economy of the Sultanate of Oman.

In this study, an ARDL model was used to investigate the effect of selected macroeconomic variables on the tax revenue changes in Oman from 1991 to 2020. The main findings of the research reveal that tax revenue change is positively influenced by the SEA and DSP projects, while negatively influenced by the FDI inflows, size of the SE and externalities of inflation rate.

The implications of this study suggest that to protect individuals from the negative effects of inflation and its Externalities, it necessitates that stronger social safety nets and income support programs are required to protect the vulnerable population from price volatility and preserve social cohesion.

Results verified that tax revenue collections in Oman may be affected by a sizeable SE. Given that a sizeable percentage of economic activity in the nation takes place in the SE, it is likely that these activities are either underreported or not subject to taxation. This could prevent the government from collecting all taxes due, which would leave a gap in the tax revenue base. This may make it more difficult for the government to fund infrastructure improvements, public services and other crucial initiatives, which, in turn, might increase reliance on oil earnings and impede attempts at economic diversification. Therefore, increasing tax collections and attaining sustainable economic growth may depend on addressing the SE through better regulatory and tax enforcement methods. Addressing a large SE is needed, since it implies a significant portion of the workforce and business activity works informally, without tax compliance. This undermines public trust, increases social inequality among individuals. The incapacity to tax these illegal or informal activities fairly places an unequal tax load on the formal sector, possibly discouraging the private sector and employment. To avoid this, the taxation authority has to introduce simplified tax regimes and registration incentives to encourage informal businesses to join the formal sector.

Given that FDI inflows have a negative impact on tax revenues, it may exacerbate economic inequality and reduce opportunities for domestic entrepreneurship and innovation. Long-term reliance on foreign capital with minimal returns in public revenue diminishes fiscal space for welfare and human capital development programs. Concerning this, the Omani government should cut back on tax incentives and tax exemptions for foreign businesses. By doing so, local businesses will prosper, and the negative competition effect will be reduced, improving government tax revenues.

To increase tax revenues over time and safeguard the economy from fluctuations in the oil price, the government must expedite and speed up the “Tanweea” for taxation resources rather than heavily relying on oil and gas as the only key economic source for the Omani government. Continued reliance on oil and gas as a dominant source of revenue leaves Oman vulnerable to global oil price shocks. Without substantial diversification, public services and social welfare programs may face funding constraints during periods of low oil prices. However, to avoid the harmful effects of competition and to safeguard local enterprises, the relevant authority in the Sultanate of Oman should exercise caution while offering foreign corporations the Tanweea incentive package policy. For doing so, the government has to form policy alignment between FDI promotion and support for local industries to avoid market distortions. This can be done by conducting a cost–benefit analysis of existing FDI incentives to evaluate their real contribution to the Omani economy and by encouraging revenue generation in non-oil sectors through taxation reforms and economic diversification, ultimately helping to mitigate the impact of oil price fluctuations in the global energy market.

The authors acknowledge the Sultan Idris Education University (UPSI), Malaysia, for supporting this research. Any error remains the authors’ responsibility.