Tourism levy collection for ‘Marketing South Africa’

Abstract

Governments around the world implement tourism taxes as a means of generating revenue to promote their respective destinations. South Africa introduced a voluntary tourism tax to raise additional funds for destination marketing. This paper focuses on the stationary tourism levy collectors’ base, a growing concern for revenue collection and tourism marketing. This study aided previous studies using stepwise logistic regression to unravel the determinants of tourism tax/levy collection in Gauteng-graded accommodation establishments. The descriptive statistics showed that most establishments collecting tourist taxes are from Tshwane. About 55% of these establishments are hotels registered as small and medium enterprises with white ownership. Determinants of tourism levy collection are levy rates, lack of transparency in levy usage, and murky levy collection processes that limit the likelihood of collecting the levy and revenue base. Extensive marketing for levy collection, clarity in levy usage, and training of levy administrators are necessary to increase levy collection.

Keywords:

• Tourism levy
• destination marketing
• South Africa

REVIEWING EDITOR:

• Pier Luigi Sacco, Gabriele d’Annunzio University of Chieti and Pescara, Pescara, Italy

SUBJECTS:

• Tourism
• Cities & the Developing World
• Hospitality

1. Introduction

Developing countries often lack resources for destination marketing. Destination management organisations use tourism levies to supplement government funding. (Reinhold et al., 2019). There are varying opinions on the importance of indirect and direct tourism taxes, user fees, and levies. Some suggest the reduction or removal of tourism taxes or levies (Brida et al., 2019; Cárdenas-García et al., 2022; Fredman & Choong, 2020; Seetaram et al., 2014), while others advocate for the introduction and increase of the rate of existing tourism taxes and levies (Falk & Hagsten, 2019; Gago et al., 2009; Luis et al., 2020; Safarov et al., 2023). Any implemented tourism tax policy can generate revenue; however, it can also deter tourists from visiting a particular destination and affect its destination competitiveness (Adedoyin et al., 2021). The United Nations World Tourism Organization (UNWTO) (2021) Global Code of Ethics for Tourism recommends that taxes and levies that penalise the tourism industry and undermine its competitiveness be gradually phased out or corrected (UNWTO, 2001).

In this paper, the term ‘tourism levy’ has a meaning and implication similar to ‘tourism tax’. The UNWTO (1998, p. 16) defines tourism taxes as a percentage fee charged to tourists and the tourism sector. Such tourism taxes are applied through a tax system or other mechanisms outside the general tax system. Accommodation taxes are commonly used in many countries because they are easy to collect from establishments and reach more tourists (Borges et al., 2020).

It is worth noting that a tourism tax is levied on top of other taxes in South Africa. The tourism levy, commonly known as the Tourism Marketing South Africa (TOMSA) levy, is a 1% levy applied to specific tourism services in South Africa that is passed onto the consumer. The TOMSA levy aims to generate additional funds for South African Tourism (SAT) to market the country as a tourism destination nationally and globally. SAT is the marketing agency for South Africa’s destinations under the Department of Tourism. This tourism tax is charged on invoices from those who purchase tourism products and services from establishments that collect the tax (Tourism Business Council of South Africa (TBCSA), 2020). Tourism establishments collect the tax from local and international tourists and pay the tax to the TOMSA account every month. The TBCSA is the representative council of the players in the private tourism sector in South Africa that manages the TOMSA account. The TBCSA consolidates the collected levies and transfers 90% of the total funds to SAT as added funding for marketing activities. The TOMSA levy is collected in the accommodation sector and by travel agencies, tour operators, car hire, and conference venue subsectors in all nine provinces of South Africa. As in many countries, accommodation is one of the largest subsectors in the tourism value chain, contributing about 70% to the TOMSA revenue base.

This paper answers the following questions.

What are the determinants of tourism levy collection in Gauteng-graded accommodation establishments?

What are the policy recommendations for improving the revenue base of levy collectors to market South Africa?

Gauteng is South Africa’s economic powerhouse, with several tourist resorts and accommodations catering to domestic and international visitors. In 2022, the economy of Gauteng was more significant than the economies of Western Cape and KwaZulu-Natal combined, responsible for a third of the national economic activity. Gauteng also benefited the most from economic growth in 2022 compared to 2021, expanding by 2, 8%. (Statistics South Africa, 2023). The growth trajectory indirectly impacts the accommodation sector, the focus of corporate guests and other visitors. Accommodation comprises hotels, guesthouses, Bed &Breakfast (B&Bs), self-catering caravans and campsites, backpackers, hostels, and game/nature lodges. The accommodation sector is categorised as graded or nongraded based on star rating. The stars range from 1 to 5, with 1 being entry-level, affordable with basic options, and 5 being luxury options, which are more service-oriented and guest-focused.

Since the inception of the tourism levy in 1998, there has been a slow uptake in the collection of levies amongst establishments, which is a problem statement in this paper. First, the national trend analysis from 2011 to 2016 revealed a fluctuation in the number of levy collectors from 491 in 2011, 506 in 2012, 473 in 2013, 524 in 2015 and 546 in 2016 (TBCSA, 2016; TBCSA, 2013). This analysis illustrates an increase of 55 establishments over five years, translating on average to 11 per year. The statistics on levy contributions for 2017 to 2022 are not well documented. None of the available statistics are disaggregated at the provincial and local levels rather than reported nationally. However, in 2023, only 396 establishments collected the levy, which denoted the continuous decline in collectors (TBCSA, 2023). The number of establishments collecting the levy is deficient, considering the period the levy was introduced. Gauteng province has about 800 graded accommodation establishments. It is the financial capital of South Africa, which should benefit from the collection of levies for the development and marketing of tourism infrastructure (TGCSA, 2020).

To encourage levy collection, the TBCSA introduced certain benefits, including marketing opportunities, free Broad-Based Black Economic Empowerment (BBBEE) points, and discounts on grading fees. Unfortunately, the advantages presented by the TBCSA did not persuade more establishments to collect and remit the levy. Instead, the base of the levy collection remains low. Improved tax collection would allow SAT to market the country to domestic and international markets robustly and directly fund other tourism programmes, such as tourism safety initiatives (TBCSA, 2020).

The literature on tax and levy collection in the tourism sector in developing countries remains scarce (Mahangila & Anderson, 2017). Factors influencing voluntary tourism levy collection have been largely overlooked, and this paper hopes to address the gap. Previous studies focused mainly on the effectiveness of various tax administrations in revenue collection (Djulius, 2018). Consequently, other studies focus on the determinants of tourism levy collection without assuming the impact on revenue collection (Assfaw & Sebhat, 2019; Borges et al., 2020; Carrozzo & Rotaris, 2019; Dube, 2014; Goktas & Polat, 2019; Moreno-Rojas et al., 2017; Jiao et al., 2021). Finally, Denford et al. (2019); Mangayi (2014); WTTC (2018) focused more on the levy collection system, the effectiveness of presumptive tax administration, principles for intelligent tourism tax and awareness of tourist tax. Studies excluded the factors that impact the graded accommodation establishments, that is, whether to collect the levy. We aided the previous studies by using stepwise logistic regression to examine the factors influencing voluntary tourism levy collection in Gauteng-graded accommodation establishments. The study contributes both methodologically and empirically to tourism marketing. It prioritises factors to increase tourism levy collection, encouraging accommodations globally to support the sector’s growth through the collected revenue.

The rest of the paper is structured as follows: Introduction, Review of the literature, Materials and Methods, Results, Discussion, and Conclusions.

2. Review of the literature

Various socioeconomic factors can discourage tourism accommodation establishments from paying taxes or levies. In the context of a voluntary levy, as in the case of the tourism levy in South Africa, defiance and lack of buy-in from the levy payers could affect revenue collection (Carrozzo & Rotaris, 2019). The willingness to pay for a tourism levy can depend on whether the generated revenue is utilised directly for tourism to benefit the sector and improve the visitor experience (Carrozzo & Rotaris, 2019). Transparent use of income affects compliance (Assfaw & Sebhat, 2019; Goktas & Polat, 2019). The government’s misuse of collected levies or tourism-related taxes can influence its perception of spending patterns and further dissuade levy payers from complying (Dube, 2014).

Tax revenue collection could be complex and expensive for businesses, resulting in high compliance costs (Afuberoh & Okoye, 2014; Arsika et al., 2020; Corthay & Loeprick, 2010; Mangayi, 2014). The perception that tourism is a highly taxed sector can create a generally negative attitude toward tax-related fees among taxpayers and cause a revolt against the policy regardless of the accrued benefits from the contribution (Kangave, 2004). Factors likely to influence tourism establishments not to collect the tourism levy from the public voluntarily include lack of consultation, levy education and awareness, simplicity, high tax rate, perceptions of tourism tax spending, equity, and compliance costs (Denford et al., 2019; Mahangila & Anderson, 2017; WTTC, 2018). However, previous studies did not explore the influence of levy collection determinants on revenue collection in graded accommodation establishments. This study argues that the pursuit of levy-based growth demonstrates the growing importance of the revenue generated to address the needs of the tourism sector, including the implementation of tourist safety initiatives. Consequently, accommodation establishments in Gauteng province could be encouraged to become voluntary levy collectors.

2.1. Theoretical framework

This study is grounded in stakeholder and social norm theories. The stakeholder theory argues that ‘managers in organisations have a network of relationships to serve stakeholders including shareholders, creditors, employees, managers, customers, suppliers, local communities, advocacy groups and the public at large’ (Abdullah & Valentine, 2009, p. 91). Stakeholder theory, as a descriptive perspective, outlines how concepts correspond to how directors think of the interests of shared groups (Brin & Nehme, 2019; Farmaki, 2019). From a normative perspective, this explains how managers should address concerns raised by stakeholders (Farmaki, 2019). The benefits of good stakeholder relationships include improved coordination, better knowledge sharing, interdependence with stakeholders, shared sense of purpose, value creation, loyalty, business continuity, lower transaction costs, and greater moral motivation (Elena & Mondonedo, 2021).

TBCSA’s stakeholders come from both within and beyond the organisation. Internal stakeholders include administrators, managers, the Chief Executive Officer of the TBCSA, the TBCSA Board, tourist member associations, and classified businesses. External stakeholders include the government, the SAT, and society. The TBCSA board helps parties informally manage the levy. After collecting public levy payments, establishments immediately transmit the monies to the TBCSA, the central levy administrator. Tourism businesses profit from taxes collected to finance marketing campaigns through SAT. Some rated lodging establishments do not collect the tax due to several factors affecting collection. The identified stakeholders collaborate to lobby for their interests in levy management and collection. The stakeholder theory requires the TBCSA to balance the various interests of stakeholders. Establishing a system to involve multiple stakeholders in levy administration and collection concerns is essential for the TBCSA to improve stationary levy base collection and finance additional safety efforts. The mandate of TBCSA is to unify stakeholders in the tourism industry to advocate and campaign on problems that concern them (TBCSA, 2020).

The social norm theory highlights the hospitality establishments’ willingness to collect and remit levies voluntarily based on the norms they follow. A social norm is a belief system people share that determines how one should behave. It encourages certain conduct by imposing social sanctions on those who violate it (Djulius, 2018, p. 61). For the TBCSA and its members, collecting and paying over the gathered levies is the accepted social norm. Therefore, those who manage the levy and fail to remit it might feel inferior, as the social norm dictates. According to Djulius (2018, p. 62), ‘institutional factors such as support for democracy, trust in government and preference for distribution relate to aspects which explain how citizens perceive government spending taxes’. Another institutional factor relates to the public’s perceived higher benefit of public services. The higher society taxed morale, and the establishments’ perception of the help received from the levy may also affect their opinion of other benefits or services derived from levy collection. The institutional factor assumes that the minimal benefits experienced by establishments might discourage them from collecting the levy. Participating in decision-making boosts confidence and morale and encourages people to collect the levy voluntarily (Djulius, 2018). It should be noted that a complex tax/levy system can also contribute to tax/levy noncompliance (Bruce-Twum, 2014).

2.2. Tourism tax: a cross-country comparison

Different countries apply a tourism levy or tourism tax for various reasons, among others, to reduce over-tourism, to ensure environmental protection and conservation, and to ensure the sector’s sustainability (Alfano et al., 2022). However, the tourism levy was introduced in South Africa to raise additional funds for destination marketing. The advantages of implementing a tourism levy can assist in generating revenue to fund, among others, domestic and international marketing activities, environmental and cultural protection and preservation, tourism safety and security initiatives, tourism infrastructure investment and development, address over tourism or attract more affluent tourists as well as contribute towards local community beneficiation particularly when tourist arrivals increase (Arsika et al., 2020; Durbarry, 2008; Gooroochurn & Sinclair, 2005; Ihalanayake, 2007).

Ideally, any revenue generated from the tourism sector through taxes or related fees should benefit the sector, attract more tourists, and grow the sector (Arsika et al., 2020; Barron et al., 2001; Cetin et al., 2017). Goktas and Polat (2019) argue that many countries strictly use tourism-related taxes for tourism purposes. On the contrary, tourism levies are collected and used for non-tourism development, which can dwindle funding for tourism, including marketing (Khanal et al., 2022). The levies collected by TBCSA are used for the common good, to market the country domestically and internationally, and to attract tourists. Tourism is considered a labour-intensive industry and has a multiplier effect, benefiting other sectors of the economy through tourist expenditure (WTTC, 2018). The public should also support the levy because the more tourism activities occur, the more employment is created, and the country’s gross domestic product (GDP) grows.

The disadvantage associated with tourism levy is that tourism is a price-sensitive sector; as such, a small increase in prices causes instability in tourism demand, and tourists are likely to choose cheaper destinations or destinations that do not charge tourism tax or levy (Ihalanayake, 2007; Mahangila & Anderson, 2017; WTO, 1998). Some establishments are more likely to pass on the cost of the levy to customers through increased prices, while others may absorb the levy to improve occupancy rates (European Commission, 2017). Collins and Stephenson (2018, p. 7) examined the effect of the R5 (US$) per night hotel tax and compared 2014 to 2015 in Georgia. The results revealed that the tax decreased the number of rooms occupied by approximately 92,000 per month, translating into a –0.7% price elasticity of demand and a decrease in revenue collection (Collins & Stephenson, 2018, p. 7).

Goktas and Cetin (2023) alluded that implementing a tourism-related tax as a source of revenue can either stimulate or impede tourism activity, depending on its design and effective administration. Attesting to this, Croatia introduced an accommodation tax referred to as the Sojourn Tax for overnight stays. The accommodation tax is between R0.27 and R.094 euros per night, depending on the accommodation type (PWC, 2017, p. 185). The Croatian tax legislation requires the revenue generated from the accommodation tax to be shared among municipal, district, city, and national tourism bodies (PWC, 2017, p. 185).

The province of Alberta in Canada imposed a 4% tourism levy on accommodation establishments (Government of Alberta, 2021, p. 1). The tourism levy is governed by a Tourism Levy Act, which provides for registration, the administrative collection process, exemptions, reporting, and remittance by accommodation establishments and enforcement by the local municipality. Despite the COVID-19 pandemic, the province of Alberta generated a total revenue of R34.6 million US$ during the financial year 2020/21 (Government of Alberta, 2021, p. 1).

The City of Berlin imposed a 5% levy collected by accommodation establishments and paid to the federal state of Berlin for administration (Daley, 2017, p. 23). Guests travelling for business are not charged the levy provided they can produce proof of the business trip (Daley, 2017, p. 23). The levy is added to the guest bill, expected to raise over 240 million euros (Daley, 2017, p. 3). Vera (2022) agreed that the impact of an occupancy tax on demand is uncertain and likely to vary depending on the destination type and the tax rate. The Edinburgh Council has researched this matter, highlighting their taking the proposals seriously. According to one survey, only 2% of respondents stated that they would not have visited the city if they had to pay a £1-2 per night occupancy tax, indicating a low impact on demand.

The introduction of a tourism tax on accommodation of about $2 US dollars in Zanzibar led to an administrative burden partly due to the lack of specific legislation that regulates tourism levies applied to tour operators, restaurants, and hotels (Arsika et al., 2020).

Various municipalities in Portugal implemented the tourist tax in March 2018 (Borges et al., 2020). The tourist tax is levied to guests from the age of thirteen (13) and in two (2) EUROs per person/night in types of tourism accommodation establishment, up to a maximum of seven (7) consecutive nights person/stay to avoid losing competitiveness in the destination instead to create value in the tourist services provided (Borges et al., 2020). However, the impact of the tourist tax and the determinants of its tax collection within establishments were not determined except for the level of awareness among tourists.

Zimbabwe’s 2% tax on hotels, restaurants, camping safaris and other tourist facilities suffered from non-compliance in addition to economic challenges and inefficiencies of the Zimbabwe Tourism Authority (Woyo & Slabbert, 2021). The TBCSA (2020) posits that certain tourism establishments might not perceive the direct link between the levy and tourist arrivals. However, without the tax, establishments could be negatively affected if South Africa cannot market itself and compete with international tourism destinations to boost the country’s economy.

Previous literature and experiences have shown that introducing a tax in the accommodation sector can increase revenue in the tourism sector. However, non-compliance and mismanagement have also been identified. This study examines the factors that could cause noncompliance and how to use this revenue for tourism marketing, especially in South African cases where the levy is only 1% and voluntary.

3. Materials and methods

We use stepwise logistic regression to analyse the determinants of tourism levy collection in Gauteng-graded accommodation establishments. The first subsection presents the population, sampling, and variables for the study, followed by the empirical model.

3.1. Population and sampling

We adopt a proportionately size-adjusted stratified random sampling technique. The strata comprised seven types of accommodation establishments (Backpackers, B&B, caravan and camping, guesthouses, hotels, lodges and self-catering) drawn from a population of 704 Gauteng-graded accommodation establishments. We tried to access the target population of Gauteng’s graded accommodation establishments by contacting TGCSA, the Gauteng Tourism Authority and local tourism associations. However, TGCSA reported that their database was private and confidential, and local tourism associations declined to share it. Ultimately, we manually compiled a sampling frame by capturing the graded establishments listed on the TGCSA website already in the public domain. The council granted permission to telephonically utilise the list on the website. The results represent the Gauteng-graded accommodation establishments that participated in the study.

Table 1 illustrates the selected sample size of the study.

Table 1. Population and sample size.

Strata	Population	Proportional sample
Backpackers	4	1
Bed and breakfast	94	33
Caravan and camping	5	2
Guesthouse	323	114
Hotels	166	59
Lodge	60	21
Self-catering	52	18
Total	704	249

Source: Authors.

The appropriate sample size was determined using a statistical calculator by entering the total population size of 704 Gauteng-graded accommodation establishments. The study was able to achieve a sample size at a confidence level of 95%, a margin of error of 5%, and a 50% response distribution, as suggested by Creswell and Plano Clark (2018) and Creswell (2014). As a result, a recommended sample of 249 was drawn through the sample size calculator. An online self-administered questionnaire was distributed online to Gauteng-graded accommodation establishments. It produced a low response rate, a typical challenge with online questionnaires. This study carried constant reminders, telephonic follow-ups, and physical walk-ins through appointments with establishments during the week to improve the response rate.

3.2. Variable selection

3.2.1. Dependent variable

The study involved the primary dependent variable levy collection in two binary categories: 1 = ‘current collectors’ and 2 = ‘current non-collectors’ of the tourism levy.

3.2.2. Independent variable

The study used constructs such as challenges associated with the collection of the tax (Denford et al., 2019, p. 8; Makuya, 2017,.p. 12) and (Mangayi, 2014, p. 3) awareness of municipal tourist tax (Arsika et al., 2020) as well as factors which influence the collection of the tourism levy (Mangayi, 2014,.p. 10; Denford et al., 2019, p. 13) factors for intelligent tourism taxation (WTTC, 2018,.p. 2) and (Ihalanayake, 2007, p. 18) to measure factors that influence graded accommodation establishments to decide whether to collect the tourism levy. The reviewed literature speculates on independent variables as factors that influence the collection of levies. These include a lack of understanding and awareness of how the levy is collected, complicated levy administration systems, high levy rate charges, lack of transparency of how the levy is utilised, misuse of collected levies, lack of benefits from collected levies, high costs of levy collection, and corruption among levy collectors.

3.3 Model specification

Binary logistic regression is a statistical strategy to predict the connection between dependent and independent variables (Bohanec et al., 2018). Furthermore, according to Bohanec et al. (2018), logistic regression implies a linear connection between qualities and might result in bias if data exhibit a nonlinear relationship between attributes. The logistic regression equation draws on the following formulae: (1) $$\mathrm{\text{log}}\hspace{0.17em}\left(\frac{p}{1-P}\right)=0^{0^{+0}{1}^x{1}^t}\dots +O_n{x}_n$$(1)

Where $p$ represents the probability that an outcome will occur, $o_i$ weight or influence of associated attributes on the outcome. Delibašić et al. (2018) suggested that reducing the penalty function, as shown in Equation (2)(2) $$l\left(0\right)=\sum _{i=1}\left(1+\mathrm{\text{exp}}\hspace{0.17em}\left(-y\right.i\left(\begin{array}{c}T\\ \mathit{\text{xO}}+C\end{array}\right)\right)\mathrm{ }$$(2) , can improve the likelihood of an outcome. (2) $$l\left(0\right)=\sum _{i=1}\left(1+\mathrm{\text{exp}}\hspace{0.17em}\left(-y\right.i\left(\begin{array}{c}T\\ \mathit{\text{xO}}+C\end{array}\right)\right)\mathrm{ }$$(2)

Where $x$ represents the vector of input attributes, $y$ label from the set $\left\{0,1\right\}$ and $C$ random noise, optimisation is performed using the Newton method to acquire the best value for the penalty function (Delibašić et al., 2018, p. 204). The probability is derived using the following equation: (3) $$\rho \left(\overline{y}\right)=\frac{\mathrm{\text{exp}}\hspace{0.17em}\left(O_0+O_1{x}_1\right.\dots +O_n{x}_n}{1+\mathrm{\text{exp}}\hspace{0.17em}\left(O_0+0,x_1+\dots +O_n{x}_n\right)}$$(3)

The probability will have a value between zero and one to determine whether to predict the outcome as planned by deciding on the decision threshold $\tau$. The probability of an outcome occurring or not if it is greater or equal to $\tau$ (Delibašić et al., 2018).

We use logistic regression analysis to build a stepwise model to predict the factors that influence the collection of taxes. Stepwise logistic regression is a statistical analysis tool utilised to predict the relationship between a dependent variable, levy collector and non-levy collector, and an independent variable. Peric and Radic (2015) state that stepwise logistic regression excludes and selects statistically significant or relevant variables to predict a specific relationship between variables to build a model. This study used stepwise logistic regression from previous related studies (see Jiao et al., 2021; Moreno-Rojas et al., 2017).

The stepwise method drawing on binary logistic regression was beneficial in determining the factors that influence the tourism levy collection in the context of the Gauteng-graded accommodation establishment. The procedure progressively adds the set of control variables to ensure that the candidate variable included in each step significantly reduces the residual sum of squares and removes control variables (Jiao et al., 2021). This procedure was suitable because the tourism levy had many potential key collection determinants. Furthermore, the objective was to select a model with only variables (key determinants) that effectively predicted success in collecting the levy.

4. Results

This section presents the findings of the study. The Statistical Package for Social Sciences (SPSS) was used to run descriptive statistics, diagnostic tests, and a stepwise logistic regression analysis model to test the determinants of levy collection in Gauteng-graded accommodation establishments.

4.1 Descriptive statistics

4.1.1. Levy collection and the characteristics of the establishment

Table 2 illustrates the cross-tabulation of the collected tourism levy against the characteristics of the establishment: region of establishment in Gauteng, geographic location (urban or rural), accommodation sub-sector, size of business, and type of ownership.

Table 2. Cross-tabulation between levy collection and business characteristics.

Business characteristic	Have collected the tourism levy before
	Business characteristic	Yes	No
Gauteng Region of establishment	Tshwane	56.5%	43.6%
	Johannesburg	17.4%	82.6%
	Ekurhuleni	4.3%	95.7%
	West Rand	17.4%	82.6%
	Sedibeng	4.3%	95.7%
Geographic location of establishment	Urban	82.6%	17.4%
	Rural	17.4%	82.6%
Accommodation subsector	Hotel	56.5%	43.5%
	Bed and Breakfast	4.3%	95.7%
	Lodge	17.4%	82.6%
	Guest house	17.4%	82.6%
	Self-catering	4.3%	95.7%
	Other	0%	100%
Size of business	Small, Medium and Micro Enterprise (SMME)	73.9%	26.1%
	Large enterprise	26.1%	73.9%
Type of ownership of accommodation establishment	Black-owned	12.5%	87.5%
	White-owned	85.7%	14.3%

Source: Authors’.

The results revealed that in establishments that collect the tourism levy, the majority originate in Tshwane (56.50%); in urban areas (82.60%); hotel establishments (56.50%); SMMEs (73.90%); and white-owned (85.7%).

4.1.2. Chi-square tests

The Chi-square test of independence was used to determine whether there was an association/relationship between the collection of the tourism levy and various business characteristics. The chi-square test was conducted under the null hypothesis: ‘There is an association between the tourism levy collection and the business characteristics. These findings are presented in Table 3, with their statistical significance.

Table 3. Relationship between business attributes and collection of tourism levy.

Business characteristic	Pearson Chi-Square	Df	Significance Level
Location of establishment (rural/urban)	0.505	1	0.477
Region of establishment in Gauteng	4.868	4	0.301
Accommodation subsector of the establishment	28.606***	5	0.000
Size of establishment	18.197***	1	0.000
Type of Ownership	5.014**	1	0.025

Source: Authors’.

An insignificant relationship exists between tourism levy collection and the establishment’s location (rural/urban) and the location (region of establishment).

The relationship between the tourism levy collection and the establishment’s accommodation subsector was statistically significant at 1%—Pearson’s chi-square = 28.606. The relationship between the tourism tax collection and the business size was statistically significant at 1%—Pearson Chi-square = 18.197.

The relationship between the collection of tourism taxes and the type of ownership was statistically significant at 5%, Pearson Chi-square = 5.014.

4.2. Diagnostic tests

We used SPSS to run diagnostic tests, namely the likelihood ratio test, omnibus test, and Hosmer-Lemeshow to determine the model’s fit, which is most suitable to predict factors that influence the collection of the tourism levy which culminated to the final model using stepwise logistic regression. The stepwise logistic regression assumes a linear relationship among attributes and can result in bias if data has a non-linear relationship among attributes.

Diagnostic tests are presented below.

4.2.1. Likelihood ratio test

The value of the Nagelkerke R square reveals the magnitude of the coefficient of determination in the logistic regression model. The coefficient of determination measures the appropriateness of a model to explain the independent variables.

As shown in Table 4, the Nagelkerke R square is 0.701, indicating that the independent variables explained 70.1% of the variability of the dependent variable.

Table 4. Likelihood ratio test.

Step	−2 Log likelihood	Cox & Snell R Square	Nagelkerke R Square
1	25.765a	0.196	0.266
2	19.488b	0.388	0.526
3	14.033c	0.517	0.701

4.2.2. Hosmer-Lemeshow

The Hosmer-Lemeshow goodness of fit test assesses how well a model’s predictions match the observed data. If the test value is over 0.05, there is no significant difference between the predictions and actual values, meaning the model’s estimates are acceptable. The Hosmer-Lemeshow test is statistically significant in evaluating the logistic regression model’s goodness of fit (Table 5).

Table 5. Hosmer-Lemeshow goodness of fit test.

Step	Chi-square	Df	P- value
1	5.775	3	0.123
2	5.635	6	0.465
3	4.613	8	0.798

* p<.05, **p<.01, ***p<.001.

4.3. Logistic regression results

The results are based on our application of the stepwise regression algorithm as a procedure for statistical model selection. We modelled a stepwise logistic regression of the critical determinants of collecting the tourism levy in Gauteng-graded accommodation establishments using the forward approach. The selected model started with a ‘no model’, that is, there were no control variables at all. The procedure added into the set of control variables step-by-step ensured that the candidate variable included in each step produced the greatest reduction of the residual sum of squares, and the control variables already included in the previous steps had not been withdrawn. In this case, this procedure was suitable because there were many potential key collection determinants of the tourism levy. Furthermore, the objective was to select a model with variables only (critical determinants) that were effective to predict success in the collection of the levy.

The dependent variable was binary with categories for collection and non-collection of the levy.

Stepwise selection is a model-building method that removes any insignificant variables before adding a critical variable. The output displays each addition or deletion of a variable as a separate step, and a new model is fitted at each step.

Table 6 shows the results of the stepwise regression. Model 3 is interpreted as revealing statistical significance.

Table 6. Logistic regression model findings.

Step	Determinants	B	SE	Wald	Df	P value	Exp(B)
Step 1a	The Levy rate charge is high	−0.897**	0.461	3.783	1	0.050	0.408
	Constant	3.122	1.495	4.359	1	0.037	22.685
Step 2b	The Levy rate charge is high	−2.439**	0.968	6.357	1	0.012	0.087
	Lack of transparency on the levy is a challenge to levy collection	1.973**	0.93	4.499	1	0.034	7.194
	Constant	3.081	1.665	3.421	1	0.064	21.77
Step 3c	The levy collection process is simple.	1.61**	0.981	2.695	1	0.010	5.005
	The levy rate charge is high.	−3.74**	1.758	4.526	1	0.033	0.024
	Lack of transparency on the levy usage is a challenge to levy collection	3.475**	1.639	4.498	1	0.034	32.305
	Constant	−0.871	2.304	0.143	1	0.705	0.419

^a Variable(s) entered in step 1: The levy rate charge is high.

^b Variable(s) entered in step 2: Levy usage’s lack of transparency challenges levy collection.

^c Variable(s) entered in Step 3: Levy collection process is simple and easy to understand.

* p ≤ 0.10, **p ≤ 0.05, ***p ≤ 0.01.

Table 6 reveals the covariates of tourism levy collection and the critical determinants of such levy collection. Furthermore, the table analyses the odds of which establishments were likely to collect the levy.

Firstly, the ‘high rate levy’ significantly predicted the probability of collecting the tourism levy (β = -3.74, p = 0.033). The β had a negative sign, and the odds ratio was 0.024, which revealed that the ‘high-levy rate charge’ was likely to reduce the collection of tourism levies. Second, the ‘Lack of transparency of how the tax is used as a challenge to its collection’ significantly predicted the probability of collecting the tourism levy (β = 3.475, p = 0.034). Finally, the ‘clean levy collection process’ significantly predicted the probability of collecting the tourism levy (β = 5.005, p = 0.010).

5. Discussion

This section presents a discussion of the study findings. First descriptive followed by empirical.

The descriptive findings showed that the establishments that collect the tourist tax are mostly hotel establishments, SMMEs, white-owned establishments, and those located in urban areas, particularly in Tshwane. Most levy collectors emanate from Tshwane mainly because Tshwane has many graded establishments and was approached personally to increase the response rate. Regions outside Tshwane were followed by phone due to financial constraints. Gauteng province is mainly urban and comprises two metropolitan municipalities considered one of the country’s major economic hubs (Statistics South Africa (Stats SA, 2019, p. 1). The fact that most small businesses still collect the levy indicates small firms’ dominance in the tourism sector. The dominance of white-owned establishments as majority levy collectors confirms the untransformed tourism sector dominated by one race group (Abrahams, 2019, p. 825; DTI, 2015, p. 8). Mangayi (2014, p. 11) emphasised that the type and size of the establishment can be determining factors for establishments to collect the levy or not.

The Chi-Square test was conducted under the null hypothesis: ‘There is an association between the collection of the tourism levy and the characteristics of the business,

The test findings on the relationship between the levy collection and the establishment region in Gauteng revealed no relationship between the location (rural or urban) and the collected tourism tax. The literature on the collection of the levy was limited. This assertion was also acknowledged by the United Kingdom (UK) Local Government Association UK (2020, p. 6), which said there is limited evidence of how the tourism tax works. The results suggested that any establishment registered as a collector can collect the tourism levy from the public regardless of location as long as it receives customers.

There was no relationship between geographical location and levy collection. Borges et al., (2020) point out that in Portugal, municipalities have the authority to charge tourist tax to tourists who pay for overnight stays in accommodation establishments in the city, which can either be fixed or variable, such as exemptions for children and people living with disabilities. The tourist tax applies only when the accommodation capacity or occupancy rate is equal to or greater than 60% (60%) (Borges et al., 2020). The findings revealed that all the collected levy patterns were the same in all regions. Moreover, there was no unique practice among regions to draw lessons to understand what makes them perform better in collecting the levy than other regions. The levy collection was purely at the discretion of the individual establishment.

A positive relationship exists between the establishment’s type of accommodation sub-sector and the tourism levy collection. According to TBCSA (2016, p. 23), hotels were the highest revenue collection sub-sector within lodging establishments, followed by game lodges, guest houses, and bed and breakfasts. However, the levy collection figures disaggregated per accommodation subsector are unavailable in the public domain. This study argued that hotels were mainly large establishments and the major collectors of the tourist tax.

A positive relationship exists between the size of the business and the levy collection. The study argued that some establishments might have considered themselves too small to collect the tourism tax. This study opined that small establishments that collect the tourism tax could struggle to compete with large establishments due to differences in pricing. Large establishments can benefit more from collection than small establishments through SAT’s marketing initiatives. Corthay and Loeprick (2010, p. 1) argued that large establishments should be treated differently from small ones because they are sensitive to compliance costs and can be expensive to collect tax-related fees such as the tourism levy. UK Local Government Association UK (2020, p. 5) argued that a tourism levy burdens establishments to collect and remit revenue to benefit small establishments. However, the burden would likely be low, provided the calculation is simple. Corthay and Loeprick (2010, p. 1) and Afuberoh and Okoye (2014, p. 27) argued that the cost to collect and administer the tax must be as low as possible to achieve the necessary revenue collection.

The results showed a positive relationship between the type of ownership and the collected tourism tax. The lack of transformation in the sector and access to key international source markets experienced by black-owned establishments can discourage them from subscribing as levy collectors because they could perceive the marketing undertaken by SAT as nonbeneficial. These exclusions can impede black-owned tourism establishments from participating meaningfully in the levy collection because most are classified as survivalist tourism SMMEs. The ownership and control of the tourism economy remains in the hands of several white-owned establishments, which does not reflect the demographics of the South African population. DTI (2015, p. 8) acknowledged that black-owned establishments are excluded from the tourism economy due to the slow pace of transformation. Harilal and Nyikana (2019, p. 4) underscore the need for large establishments to create sustainable business links with black-owned establishments to share in the tourism economy. The Department of Tourism (2017, p. 6) maintains the need to ensure inclusive tourism growth as part of its vision for the industry.

The empirical findings from the stepwise logistic regression revealed that the ‘high levy rate charge’ significantly predicted the likelihood of collecting the tourism levy. The ‘high levy rate charge’ was likely to reduce the collection of the tourism levy. This result could be attributed to the fact that the accommodation subsector is dominated by small and medium establishments that cannot compete with large establishments on occupancy rates and financial position (Culture, Arts, Tourism, Hospitality, and Sport Sector Education and Training Authority (CATHSSETA, 2024; Stats SA 2023). Moreso, tourism suffers from seasonality which is difficult to eliminate because in a period where there is limited tourism activities the accommodation establishments might not receive much levies generated from tourists who visit their establishments (Alshuqaiqi & Omar, 2019). Gooroochurn and Sinclair (2005) noted that as taxes and levies keep increasing in most countries, businesses and consumers have demanded that authorities reduce their range and levels. Tourism is a price-sensitive sector (Mahangila & Anderson, 2017). A slight price increase can destabilise tourism demand, and tourists will likely choose cheaper destinations or those that do not charge a tourism tax or levy. The decreased tourist demand for a destination can ultimately affect the profit margins of tourism businesses. Borges et al., (2020) argue that tourism businesses consider tourist taxes or levies as making destinations less competitive. Hence, the reduced demand can further reduce the revenue collected through tourism levies.

Second, the study reveals that the lack of transparency in using the levy significantly predicted the likelihood of collecting the tourism levy. Those who collected the levy were more likely to perceive this as challenging than those who did not. The perceived lack of transparency can be attributed to the establishments that do not derive individual tangible benefits from collecting the levy. Accountability and transparency apply to public institutions, the private sector, and civil society organisations, which must account for the public and stakeholders (Gyong, 2014). Jashari and Pepaj (2018) argue that transparency is compulsory for public institutions to be trustworthy and work with integrity to build public trust. Borges et al., (2020) argue that most countries that apply tourism taxes or levies find it challenging to define the return on investment from collected levies and assess whether that value is competitive and does not affect tourism and tourists.

Barron et al. (2001) argue that establishments lack trust in the government to plough back collected revenue into the tourism sector amid mismanagement of general funds. Assfaw and Sebhat (2019) add that perceptions of how the government utilises taxpayer revenue could influence tax and non-tax revenue compliance. Dube (2014) argues that if the government does not use collected taxes for its intended purpose, such acts could influence the perception of its spending patterns and further dissuade taxpayers from voluntary compliance. As a result, the OECD (2014) reveals that establishments can oppose a tourism tax if there is no clear relationship between tourism revenue and how the government uses it. Moreover, tourists and establishments are more willing to pay and collect a tourist tax or levy when commitments are to improve the tourist experience and invest for tourism purposes (Borges et al., 2020).

According to Cárdenas-García et al. (2022) and Edwards (2009), all interested stakeholders must be aware of the benefits derived from the payment or collection of the fee. In this regard, accommodation establishments and tourists or service users need to know why the tourism tax is required and how it will be used.

Third, the study reveals that a ‘clear levy collection process’ predicted the likelihood of collecting the tourism levy. As such, a ‘clear levy collection process’ was likely to enhance the collection of tourism levy rates by five. It is argued that the tourism levy collection process could be straightforward for well-resourced establishments with adequate technological accounting systems to incorporate the levy. There could be lack of adequate/robust awareness or information among the accommodation establishments on how to register and collect the levy. However, collecting the levies for establishments that use the manual accounting system could be tedious. Mangayi (2014) argues that small establishments may lack proper records and accounting systems to determine the levy collected for remittance. According to Mangayi (2014), obtaining the figures from small establishments is difficult and costly. This could explain the low number of registered tourism levy collectors, as the accommodation sector mainly comprises small and medium enterprises. Implementing tourism taxes or levies can result in extensive administrative burdens for tourism businesses (Arsika et al., 2020).

Levy collection can be complex and expensive for certain establishments, resulting in high compliance costs (Afuberoh & Okoye, 2014; Corthay & Loeprick, 2010). Mahangila and Anderson (2017) reveal that in Zanzibar, no specific legislation governed tourism levies. This exacerbated the situation, and there was much confusion about full compliance. A simplified, easy-to-understand, interactive and user-friendly levy collection system can improve the levy base and revenue collection (Assfaw & Sebhat, 2019). Simplifying a tax or levy system’s operation is critical to creating certainty and enhanced revenue collection (Ihalanayake, 2007). Mangayi (2014) advocates reducing red tape in collection processes.

Subsequently, there should be innovative ways to create awareness of the importance and significance of the levy collection process. Information on how to collect the levy should be easily accessible to both tourists and potential levy collectors. From the information provided on the TBCSA website, the tourism levy collection process appears simple. However, the website does not diffuse information on the levy collection process. The system’s lack of knowledge and complexity influence payers’ decisions of whether to collect, pay over, seek exemptions, or avoid the levy altogether (Mangayi, 2014).

6. Conclusions

Generating levy-related revenue from tourism has become a popular approach amongst world economies. Revenue funds destination marketing and tourism-related infrastructure, maintaining tourist attractions and ensuring tourist safety and security. This article is based on stepwise logistic regression to examine factors influencing voluntary tourism levy collection in 249 Gauteng-grade accommodation establishments selected by stratified random sampling proportionate to size.

The study accentuates that a high levy rate and a lack of transparency regarding using the levy are significant factors that discourage establishments from collecting the money.

Implications for tourism operators – buy-in and support from accommodation establishments:

To successfully collect any voluntary tax, such as the tourism levy in South Africa, it is necessary to have support for the levy from the tourism establishments entrusted with the responsibility of collecting this levy from users of tourist services. Accommodation establishments operating within the Airbnb platform should also contribute towards the tourism levy because they benefit from tourism marketing and service tourists. This will also address freeriding by establishments operating on the Airbnb platform who benefit from tourism marketing by SAT yet not collecting the levy.

Implications for levy administrators and user – transparency on the usage of levies:

As a user of the tourism levy user, SAT should adequately account for the tax collectors in its expenditure or the return on investment. Therefore, transparency must be stimulated by providing the public with accurate, timely, and easy-to-find information about the levy through various media platforms. Clarity about using the tourism levy should show a return on investment for specific provinces and localities. There should be better reporting on how tourism levy revenue has been used by providing the specific initiatives the levy has funded. The positive impact of the tourism levy on the tourism industry in general also needs to be well-reported and communicated to industry stakeholders. This would encourage other accommodation establishments to register to collect the levy. Implications for policy makers – strengthen levy collection system; robust marketing, tangible benefits as well as monitoring and evaluation:

For TBCSA, the Department of Tourism and SAT, there is a need to strengthen the systems and processes of levy collection and balance all stakeholders’ interests to facilitate buy-in and acceptance of the levy system. Simplifying the process of collecting the levy and being transparent about its usage can help to boost the collection base. Currently, establishments that do not pay the levy can still benefit from marketing South Africa as a destination, which can be referred to as freeriding. Freeriding also calls for the introduction of an automated system that when a guest pays the levy, the system must be able to take the 1% and place it in a separate account that is not accessible to the establishment. The sum is paid directly into the TOMSA account monthly. TBCSA, Department of Tourism and SAT need to help establishments register and comply by offering them advice, providing extensive simplified instructions to less-resourced establishments, and committing to tourism associations and platforms to assist.

Marketing efforts should focus on the levy’s significance and tangible benefits derived from its collection for individual establishments. Marketing must ensure that the money is spent wisely because legitimacy is about where the funds are used. Recently, the focus has been on international marketing. However, a focus on domestic marketing is also needed because there are establishments whose target market is local.

For tourism operators that are registered to collect the levy should be offered more tangible benefits or incentives to ensure compliance. Lack of tangible benefits contributes towards unhappiness and resentment from collecting the levy. The Tourism Incentive Programme (TIP) of the Department can be leveraged to include possible tangible benefits. The government can also play a role by introducing regulations that state that public officials and institutions, including departments, must use graded establishments and collect the levy. Proof of collection will need to be produced during the procurement process. As a result, more tourism establishments would contribute to the levy. If establishments are not convinced or do not see value in subscribing to the levy, the initiative will continue to lack much-needed support.

TBCSA’s continuous monitoring and evaluation of the levy administration process, in collaboration with the Department of Tourism, will ensure that the levy meets its primary goals without impacting tourism demand and supply. The levy programme must be evaluated from a marketing perspective and operation process. This will reveal what is working and not working.

The study’s findings can offer insight into the collection and administration of tourism taxes for destination marketing in many developing countries struggling with tourism growth. From a local perspective, the study could help the government and TBCSA prioritise factors that are more influential on levy collection. Additionally, it enables the government and TBCSA to identify areas to enhance those factors that negatively influence the tourism levy collection. Furthermore, accommodation establishments in Gauteng province could be encouraged to become voluntary levy collectors should negative factors be addressed and positive factors strengthened to influence levy collection. Moreover, the study could also allow the government and TBCSA to determine whether the tourism levy is necessary or has contributed to any development in the sector, particularly the development of accommodation establishments in Gauteng Province, through collected revenue.

Authors’ contributions

BBR: Conceptualised the study, developed the methodology and software, validated the study, and drafted the paper. AM: Conceptualised and supervised the study, revising it critically for intellectual content and giving final approval for the version to be published. FGB: provided a formal analysis, reviewed the study, and edited the final manuscript.

Consent for participation

All respondents filled in the consent to participate forms.

Ethics approval

The University of Pretoria EMS Ethics Committee approved this study. Ethics Protocol number EMS009/21

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets generated and/or analysed during the current study are available in the [University of Pretoria] repository, [https://doi.org/10.25403/UPresearchdata.19029737.v2]

Additional information

Notes on contributors

Bernard Bonginkosi Ripinga

Bernard Bonginkosi Ripinga is a research fellow at the School of Public Management and Administration, University of Pretoria, South Africa. He has a Master’s in Public Administration, and his research interests include tourism marketing.

Adrino Mazenda

Adrino Mazenda is an associate professor in the School of Public Management and Administration at the University of Pretoria, South Africa. He has a PhD in Economics and research interests in livelihoods and development. He has several research awards and is an NRF-rated researcher from South Africa’s National Research Foundation.

Felix Gasten Bello

Felix Gasten Bello is an associate professor in the Department of Marketing Management at the University of Pretoria, South Africa. He has a PhD in Economics and research interests in Tourism management. He is also an NRF-rated researcher from South Africa’s National Research Foundation.