Determinants of South Africa’s Wine Exports to Selected East African Markets

Abstract

South Africa’s wine industry has traditionally concentrated on developed nations as its principal export markets. In recent years, regional markets within East Africa have emerged as promising areas for growth. However, these markets have not been examined, with existing research predominantly centred on the export of unprocessed agricultural commodities. This study investigates the factors influencing South Africa’s wine exports to Kenya, Tanzania, and Mauritius, considering wine as a final, value-added agricultural product. A trend analysis was conducted to examine export performance from 2010 to 2022, and an augmented gravity model was employed to identify key factors. The results show a steady increase in wine exports to these markets, with strong demand in Kenya and Tanzania. The gravity model demonstrates that higher production capacity in South Africa, larger populations in the importing countries, advantageous import duty structures, and a depreciated exchange rate positively influence exports, whereas high inflation rates significantly constrain export volumes. Consequently, strategies focused on increasing production, maintaining favourable trade conditions, and leveraging market opportunities could enhance export performance. Additionally, mitigating the effects of inflation through strategic pricing policies and industry collaboration could further strengthen South Africa’s wine export position within East African markets.

1. Introduction

South Africa’s wine industry plays a crucial role in the nation’s agricultural export sector, consistently making significant contributions to both domestic and international markets. In 2022, wine exports totalled 386.5 million litres, valued at R10 billion, which was 22% more than the previous year. Domestic consumption was even higher at 452.2 million litres, showing strong local demand. The sector has also made a significant impact on the economy. In 2019, it contributed R55 billion to the economy (SAWIS, 2021), and by 2024, this figure had risen to R56.5 billion, representing 0.9% of South Africa’s gross domestic product (GDP) (SAWIS, 2024). It supported approximately 270,364 employment opportunities (1.8% of national employment), generated R18.8 billion in labour income, and added R30.3 billion to household incomes. Furthermore, the industry accounted for R80.5 billion in net capital formation and utilisation, alongside R19.3 billion in tax revenue, constituting 1.2% of the total national tax income (SAWIS, 2024). Globally, South Africa ranks as the eighth-largest wine producer, contributing approximately 4% of the world’s wine output (SAWIS, 2022). It also ranks as Africa’s leading wine producer and exporter, surpassing other nations such as Algeria, Morocco, and Tunisia (Meloni & Swinnen, 2020; Abidellaoui, 2023; Faouzi, 2024), thereby confirming its strategic position within both regional and international trade.

Given the industry’s substantial economic contributions and its strategic importance in trade and employment, this study examines the factors influencing South Africa’s wine exports to selected East African countries, where demand is increasing yet remains under-researched (WESGRO, 2021; Kenya Tourism, 2023). This research is supported by South Africa’s comparative advantages, including a Mediterranean climate, diverse grape varietals, and adherence to international standards, which facilitate the production of high-quality grapes and enhance its competitiveness in high-value markets (Vinpro, 2021; Wine Tourism, 2023). These conditions have positioned South Africa as the leading wine producer in Africa and as a competitive alternative to established exporters such as France, Italy, and Chile. In addition, favourable production factors have facilitated the development of robust export relationships with developed countries, including the United Kingdom, the Netherlands, Germany, and Sweden. Furthermore, China has become an important market due to geopolitical developments, notably trade tensions with Australia (WESGRO, 2021), which initially created new opportunities for South African exporters. However, recent reports highlight a sustained decline in overall wine demand in China. For instance, wine consumption dropped by 9.3% in 2024, according to the OIV (2025), cautioning against relying on the Chinese market as a stable growth driver for South African wine exports. In contrast, regional markets in East Africa remain underexplored, despite evidence suggesting an increasing demand for wine.

Between 2012 and 2021, numerous challenges, including the COVID-19 pandemic, disrupted the production and trade of wine. Specifically, wine sales and exports were prohibited during South Africa’s 2020 lockdown (SAWIS, 2021). Nevertheless, the East African markets demonstrated a swift recovery following the lifting of restrictions. Countries such as Kenya, Tanzania, and Mauritius contributed to the recovery by substantially increasing imports of South African wine (WESGRO, 2021). For instance, Kenya imported 3.8 million litres, followed by Tanzania with 1.8 million litres and Mauritius with 1.2 million litres (WESGRO, 2021). Additionally, tourism plays a pivotal role in these nations, substantially contributing to both Gross Domestic Product and employment (Kenya Tourism, 2023; WESGRO, 2023), which may, in turn, stimulate demand for wines (Wani, 2024). Consequently, South African wines are strategically positioned to satisfy this potential demand, particularly within the luxury hospitality sector.

Wine production in South Africa is predominantly concentrated in three provinces: the Northern Cape, Western Cape, and Free State, with the Western Cape in particular leading the way due to its Mediterranean climate. Moreover, the climate of the Western Cape, coupled with well-developed infrastructure and a longstanding tradition of winemaking, establishes it as the epicentre of South Africa’s wine industry and a significant contributor to both national supply and export activities (Agribook, 2024; SAWIS, 2024; Vinerra, 2024). Turning to the importing region, Kenya and Tanzania are among the thirteen countries located on the mainland of East Africa (African Development Bank, 2024). By contrast, Mauritius, although not part of mainland East Africa, is one of six island nations in the East African region (World Population Review, 2024). Nevertheless, it is frequently incorporated into regional economic analyses of Eastern and Southern Africa due to its membership in both the Common Market for Eastern and Southern Africa (COMESA) and the Southern African Development Community (SADC) (SADC, 2022; African Regional Integration Index, 2024).

Additionally, the trade relationships between these countries and South Africa are strengthened through their shared membership in regional economic communities. For example, Kenya and Mauritius participate in COMESA, whereas Tanzania and Mauritius are members of SADC, along with South Africa. Furthermore, all three nations, including South Africa, are signatories to the African Continental Free Trade Area (AfCFTA) (COMESA, 2024; SADC, 2024; African Union Commission, 2025). These trade blocs facilitate intra-African commerce, bolster economic collaboration, and underscore the importance of examining wine export dynamics within this regional framework. The choice of Kenya, Tanzania, and Mauritius is therefore justified, as not only have these countries demonstrated strong recovery in South African wine imports following COVID-19 disruptions, with Kenya (3.8 million litres), Tanzania (1.8 million litres), and Mauritius (1.2 million) emerging as top regional importers (WESGRO, 2021), but their tourism-driven economies also generate structural demand for high-value hedonic goods such as wine (Kenya Tourism, 2023; WESGRO, 2023). Taken together, their role as leading importers, coupled with their integration into major regional trade blocs, strengthens the rationale for focusing on these markets to better understand the dynamics of South Africa’s wine exports within the broader East African context.

In the literature, however, prominent research has predominantly concentrated on South Africa’s sugar exports within the Tripartite Free Trade Area (TFTA) (Mamashila, 2024a), overall exports within the Southern African Development Community (SADC) (Mosikari et al., 2016), and fruit exports to West Africa (Phaleng, 2020). Similarly, other scholarly works have examined broader drivers of agricultural export growth and the associated challenges (Potelwa et al., 2016; Seti, 2023; Mamashila, 2024b). Thus, existing research does not sufficiently explore the specific dynamics that influence the export of value-added products, such as wine, which are affected by factors beyond standard economic indicators. Indeed, in markets such as Kenya, Tanzania, and Mauritius, the demand for wine is closely tied to tourism and shifting consumer preferences, which differ markedly from the drivers of raw agricultural exports. These unique market conditions necessitate a targeted analysis of wine exports, as traditional models for raw commodities may not fully account for the demand for processed goods. Accordingly, the study is guided by the following research question: What are the key determinants influencing South Africa’s wine exports to emerging East African markets such as Kenya, Tanzania, and Mauritius? This study addresses this question by examining the trends and determinants of South Africa’s wine exports to selected East African countries, using trend analysis and the gravity model framework, initially developed by Tinbergen (1962) and later advanced by Anderson and van Wincoop (2003). Guided by the literature, it is hypothesised that importer GDP, population size, FDI inflows, trade openness, and production capacity positively influence South Africa’s wine exports to East African markets.

In contrast, higher import duties, inflation rates, exchange rate appreciation, distance and higher wine prices are expected to inhibit export performance. The gravity model results serve to empirically test these expectations and determine the relative strength of each driver and inhibitor. In doing so, this study generates evidence-based insights that advance academic understanding of value-added agricultural trade in emerging East African markets while also providing practical recommendations to enhance the competitiveness of South African wine in these markets.

The remainder of this paper is structured as follows. Section 2 provides a literature overview, drawing on both international and South African studies to identify the knowledge gap and to highlight the novelty and originality of the study. Section 3 outlines the methodology, detailing the data sources, preliminary analyses, and the specification of the gravity model. Section 4 presents empirical results, including the trend analysis, descriptive statistics, correlation and multicollinearity tests, model selection outcomes, gravity model estimates, and robustness checks. Section 5 interprets these results, situating them within the context of previous findings and discussing their implications for the dynamics of wine exports. Finally, Section 6 concludes the paper by highlighting the study’s contributions, practical and policy implications, intended beneficiaries, limitations, and avenues for future research.

2. Literature Overview

2.1. Review of International and National Studies

Both international and South African analyses concerning agricultural export performance have consistently underscored the significance of macroeconomic conditions, production capacities, trade policies, and structural determinants in influencing export outcomes across nations. Regarding global studies, Iskandar et al. (2012) investigated Indonesia’s cinnamon exports to the United States from 1998 to 2010, discovering that export volumes were highly responsive to domestic prices, lagged export values, and fluctuations in the exchange rate, thereby highlighting the crucial role of currency valuation and pricing strategies in shaping trade flows. Similarly, Boansi et al. (2014) examined Ghana’s fresh pineapple exports, demonstrating that although export values increased in response to higher prices, export volumes exhibited less responsiveness. Their research emphasised the necessity of improving production conditions and enhancing competitiveness rather than relying solely on price incentives. Furthermore, Tang (2018) illustrated, in the context of Mauritius, that export sophistication, preferential trade agreements, and historical connections have a positive influence on bilateral trade, indicating that both structural and relational factors significantly contribute to export performance.

Additional evidence from Eshetu and Mehare (2020) regarding Ethiopia suggests that increases in Gross Domestic Product (GDP), improvements in road infrastructure, and enhanced domestic savings all contribute positively to export performance. Conversely, foreign direct investment (FDI) and growth in the labour force exert negative effects, underscoring the complex influence of investment and demographic factors on agricultural exports. Parallel findings were documented by Gbetnkom and Khan (2020) regarding Cameroon’s exports of cocoa, coffee, and bananas, where infrastructure development and access to credit supported trade activities; however, limited price responsiveness persisted due to constraints within international markets. In Kenya, Sato (2020) identified that factors such as economic size, exchange rates, and shared borders with trading partners contribute to bolstering tea exports. Conversely, geographical distance and higher per capita gross domestic product (GDP) in importing countries tended to diminish demand. More recently, Reaz et al. (2020) demonstrated that agricultural exports fostered improved firm performance in Malaysia, while Wani (2024) and Wanzala et al. (2024) emphasised the significance of production capacity, trade liberalisation, and exchange rate volatility in influencing agricultural exports within South Asia and East Africa, respectively.

Studies conducted in South Africa offer additional insights into regional export determinants. For instance, Idsardi (2013) identified that gross domestic product (GDP), production capacity, and importer demand are primary drivers of the expansion of high-value agricultural exports. Mosikari et al. (2016) similarly affirmed that economic growth within importing Southern African Development Community (SADC) nations has positively impacted South Africa’s agricultural trade. Potelwa et al. (2016) demonstrated that GDP and population size in both South Africa and its trading partners exert a positive influence on export performance, while trade agreements such as the African Growth and Opportunity Act (AGOA) and the Tripartite Free Trade Area (TFTA) have strengthened this performance. Mamashila (2017) underscored the impact of tariffs, infrastructure, and governance on South Africa’s sugar exports within the context of the Tripartite Free Trade Area. Additionally, Phaleng (2020) highlighted the significance of tariffs, production levels, and GDP in shaping fruit exports to West Africa. More contemporary research, including studies by Seti (2023) and Mamashila (2024a), has explored South Africa’s agricultural trade under the African Continental Free Trade Area (AfCFTA) and revealed that structural factors such as infrastructure, labour costs, transportation expenses, and participation in regional blocs remain critical determinants influencing export growth.

Overall, the literature reveals several consistent determinants of agricultural exports across international and South African contexts. Export prices, exchange rates, GDP, population size, production capacity, infrastructure, and trade openness have emerged as recurring factors in the literature. At the same time, studies also pointed to challenges such as policy barriers, limited competitiveness, and exchange rate volatility, which constrain export potential. Notably, however, most existing research has focused on raw agricultural commodities, including fruits, sugar, tea, coffee, and grains. Few studies have examined value-added products, such as wine, whose unique demand drivers, including tourism, consumer preferences, and branding, may influence. This gap is particularly evident in the case of South Africa, where the wine industry has a strong export orientation but remains under-researched in relation to emerging markets in East Africa. The demand for wine in Kenya, Tanzania, and Mauritius, for example, is closely tied to tourism and evolving consumption trends, which differ from the determinants typically associated with bulk agricultural commodities.

While the literature offers valuable insights into the determinants of agricultural exports, it has mainly concentrated on raw commodities rather than value-added products. As a result, there is a lack of empirical evidence on the export performance of processed agricultural goods such as wine. This gap is especially evident in South Africa, where the wine industry is internationally competitive but under-researched in terms of the factors influencing wine exports to emerging East African markets, such as Kenya, Tanzania, and Mauritius. These markets are driven more by tourism and lifestyle trends than by traditional drivers of raw agricultural exports. This study addresses this gap by analysing the trends and determinants of South Africa’s wine exports to selected East African countries, employing trend analysis and the gravity model. By focusing on a final, value-added agricultural product, it contributes to the existing body of knowledge and highlights its novelty and originality. Specifically, the study enhances understanding of the drivers and inhibitors of South Africa’s wine export performance in emerging East African markets. It also provides evidence-based insights to mitigate constraining factors and bolster the competitiveness of South African wine in the region.

2.2. Review of Key Factors Affecting Agricultural Exports

The justification for including each explanatory variable in the gravity model, as well as the expected sign of its coefficient, is presented in this section. The rationale is grounded in both theoretical perspectives and empirical evidence from prior international and South African studies. By clearly stating these expectations, the analysis enhances the rigour of the empirical strategy and provides a clearer foundation for interpreting the results. Specifically, the explanatory variables capture supply-side factors (production capacity), macroeconomic conditions (GDP, population, FDI inflows, inflation, and exchange rates), trade-related determinants (import duties, trade openness, and wine prices), and structural determinants (distance), which uniquely shape value-added agricultural exports such as wine. These factors are employed in this study as the key determinants of South Africa’s wine exports to East African markets.

2.2.1. Exporter’s Production Capacity (PC)

Production capacity refers to a country’s ability to produce sufficient quantities of a commodity for both domestic consumption and export, which in agriculture depends on land availability, technology, and infrastructure (Idsardi, 2013). Empirical studies, including those by Idsardi (2013), Phaleng (2020), and Wani (2024), have demonstrated that higher production levels are strongly associated with export growth, as surplus output can be channelled into international markets. In this study, production capacity is expected to have a positive effect on South Africa’s wine exports to Kenya, Tanzania, and Mauritius, as greater supply availability enables the industry to meet external demand and remain competitive.

2.2.2. Importers’ Gross Domestic Product (GDP)

GDP represents the total monetary value of all goods and services produced within a country over a specified period, typically measured annually or quarterly (Mankiw, 2020). Prior studies, such as Phaleng (2020), Mosikari et al. (2016), and Wani (2024), have demonstrated a positive relationship between GDP and agricultural exports. Therefore, in this study, higher GDP in Kenya, Tanzania, and Mauritius is expected to have a positive influence on wine imports, as greater prosperity expands consumer demand for value-added goods, such as wine.

2.2.3. Importers’ Population (POP)

Population size refers to the total number of people in an importing country and directly affects the size of the consumer base for imports (Matyas, 1997). Empirical evidence from Thuong (2018), Mamashila (2017), Idsardi (2013), Potelwa et al. (2016), and Wani (2024) showed that larger populations are positively associated with agricultural imports. Accordingly, in this study, population growth in Kenya, Tanzania, and Mauritius is expected to positively influence South African wine exports by expanding consumption potential.

2.2.4. Importers’ Foreign Direct Investment (FDI)

Foreign Direct Investment (FDI) refers to cross-border investments made to establish lasting economic interest in another country, often linked to infrastructure and market development (OECD, 2008). Empirical studies, such as those by Keho (2020), Mukhtarov et al. (2019), and Karimov (2020), have confirmed that higher FDI inflows stimulate import demand, particularly in cases where domestic production is limited. In this study, FDI is expected to have a positive impact on South Africa’s wine exports, as increased investment in Kenya, Tanzania, and Mauritius fosters greater integration with global markets and expands retail and hospitality demand for imported wine.

2.2.5. Import Duties (IDU)

Import duties are taxes imposed on the value of goods entering a country, raising the landed cost of imports and often reducing competitiveness (SARS, 2021). Previous studies, such as Mamashila (2017) and Phaleng (2020), have shown that tariffs negatively affect agricultural trade flows by making imports less attractive. Hence, in this study, import duties are expected to exert a negative effect on South African wine exports to Kenya, Tanzania, and Mauritius, as they increase overall costs for importers.

2.2.6. Importers’ Trade Openness (TO)

Trade openness measures the extent to which a country’s economy allows free flows of goods and services, often expressed as the ratio of total trade (exports + imports) to GDP (IMF, 2024). Research by Azhar (2014) and Yang et al. (2022) demonstrated that trade openness positively influences export performance by reducing barriers and creating favourable conditions for foreign goods. Thus, in this study, greater trade openness in Kenya, Tanzania, and Mauritius is expected to have a positive influence on South African wine exports by facilitating easier market access.

2.2.7. Importers’ Inflation Rate (IR)

The inflation rate measures the pace at which the general level of prices for goods and services rises in an economy, typically expressed as an annual percentage using the Consumer Price Index (IMF, 2024). Empirical studies, such as those by Ball (2005), Herrera and Baleix (2010), and Okpe and Ikpesu (2021), have indicated that higher inflation in importing countries can stimulate imports if domestic goods become relatively more expensive. Accordingly, in this study, inflation in Kenya, Tanzania, and Mauritius is expected to have a positive impact on South Africa’s wine exports, as consumers may substitute relatively affordable imported wine for more expensive local products.

2.2.8. Exchange Rate (EX)

The exchange rate reflects the value of one country’s currency relative to another and directly affects the price competitiveness of exports (IMF, 2009). Evidence from Idsardi (2013) and Phaleng (2020) suggests that currency depreciation enhances agricultural exports by making them more affordable abroad, although demand for food and agricultural products tends to be relatively inelastic. Therefore, in this study, the exchange rate is expected to have a negative sign, as appreciation of the South African rand would make wine more expensive in Kenya, Tanzania, and Mauritius, while depreciation would improve competitiveness.

2.2.9. Price of Wine (PW)

The price of wine reflects its average export value and serves as both a determinant of competitiveness and a proxy for quality in consumer perception (Mosikari et al., 2016). Studies by Ogede et al. (2020) and standard demand theory suggest a negative relationship between price and demand, though in premium segments, higher prices may signal superior quality and increase demand. In this study, the expected sign of the wine price is negative overall, but potentially ambiguous due to the hedonic nature of wine as a value-added product.

2.2.10. Distance (DST)

Distance represents the geographical separation between trading partners and is commonly used in gravity models as a proxy for transport costs, logistical efficiency, and market accessibility (Anderson & van Wincoop, 2003). Empirical studies, such as those by Tang (2018) and Sato (2020), confirm that greater distance generally reduces trade flows by increasing costs and limiting market integration. In this study, distance is expected to exert a negative effect on South Africa’s wine exports to Kenya, Tanzania, and Mauritius, as higher transport and distribution costs may reduce competitiveness in more distant markets.

Overall, the reviewed factors demonstrate that agricultural export performance is shaped by supply-side capacity (production capacity), macroeconomic conditions (GDP, population, FDI inflows, inflation, and exchange rates), trade-related determinants (import duties, trade openness, and wine prices), and structural factors such as distance. While the expected signs of these variables are broadly consistent with economic theory and prior empirical evidence, their influence may vary across products and markets. By making these expectations explicit, the study strengthens the empirical strategy and provides a solid foundation for interpreting the determinants of South Africa’s wine exports to East African markets.

3. Methodology

3.1. Dataset and Sources

This study utilises secondary time-series data spanning from 2011 to 2022 to examine the factors influencing South Africa’s wine exports to Kenya, Tanzania, and Mauritius. The chosen period encapsulates significant developments within the South African wine industry, including the effects of COVID-19 and the subsequent recovery phase, particularly in regional markets where demand exhibited a strong rebound (SAWIS, 2021). The dataset comprises annual observations for each country, resulting in a balanced panel of 36 observations. The dataset meets the minimum prerequisites for time-series analysis, which ensures robustness and statistical validity (Gujarati & Porter, 2009).

Data were obtained from established international and national institutions to ensure accuracy and comprehensive coverage. Wine export values were sourced from the International Trade Centre (ITC) and cross-verified with UN Comtrade (United Nations, 2024). They were then converted to ZAR millions, representing the monetary value of South African wine exported to Kenya, Tanzania, and Mauritius. Information on wine prices and production capacity was sourced from the South African Wine Industry Information and Systems (SAWIS). Wine prices reflect the average bulk selling price of bottled still and sparkling grape wines (HS 220421) in ZAR. At the same time, production capacity measures the total annual South African wine output in billions of litres. Wine prices reflect the average bulk selling price of bottled still and sparkling grape wines (HS 220421) in ZAR. At the same time, production capacity measures the total annual South African wine output in billions of litres. Import duty rates were obtained from the International Trade Administration (ITA) and WTO databases, recorded as percentages of import value. Kenya applied a fixed 25% tariff, whereas Mauritius and Tanzania applied 0%, which is consistent with the SADC/COMESA trade agreements.

Macroeconomic indicators, including foreign direct investment (FDI), population (POP), gross domestic product (GDP), inflation rate (IR), and trade openness (TO), were obtained from the World Bank’s World Development Indicators (WDI). Foreign Direct Investment (FDI) was measured as net inflows expressed as a percentage of GDP, population (POP) as absolute numbers in millions serving as a proxy for market size, and Gross Domestic Product (GDP), converted to ZAR billions to ensure comparability across countries. The inflation rate (IR) was calculated as the annual percentage change in the Consumer Price Index (CPI), while trade openness (TO) was defined as the ratio of total trade (imports plus exports) to GDP. Finally, exchange rate (XR) data were drawn from IMF publications (IMF, 2009, 2024), defined as the annual average value of the South African rand relative to each importing country’s currency, with all figures converted into ZAR to standardise units across countries.

3.2. Preliminary Data Analysis

To strengthen the empirical model, the analysis began with four steps: descriptive statistics, correlation assessment, multicollinearity diagnostics, and unit root testing. Descriptive statistics summarised the main features of key variables, including wine exports, gross domestic product, population, trade openness, and inflation. This provided an overview of how these variables were distributed in the panel dataset (Gujarati & Porter, 2009; Wooldridge, 2013). Next, correlation analysis examined the relationships between wine exports and the explanatory variables, as well as among the explanatory variables themselves. This helped identify possible multicollinearity, with the strength and direction of correlations interpreted using the thresholds from Cohen (1988) and Evans (1996). To confirm these findings, multicollinearity was also checked using the Variance Inflation Factor (VIF). VIF values below 5 were considered acceptable, whereas values exceeding 10 indicated significant concerns about multicollinearity (Shrestha, 2020; Kyriazos & Poga, 2023). This step was undertaken to ensure the reliability and interpretability of the model. Finally, stationarity was verified using the Im-Pesaran-Shin (IPS) unit root test, which is suitable for small panel datasets that demonstrate cross-sectional heterogeneity (Im et al., 2003). This measure ensured the stability of the time-series properties, thereby preventing spurious regression outcomes.

3.3. Empirical Analysis: Gravity Model: Description, Estimation, and Validation

The determinants of South Africa’s wine exports to Kenya, Tanzania, and Mauritius from 2011 to 2022 were analysed using an augmented gravity model. The gravity model, initially developed by Tinbergen in 1962 and refined by Anderson and van Wincoop in 2003, predicts bilateral trade flows based on economic size and geographical distance. This model has become a widely utilised instrument in empirical trade research (Hatab et al., 2010). The fundamental formula of the gravity model is as follows (Tinbergen, 1962):

$$Trad{e}_{ij}=G\times {\displaystyle \frac{Y_i{Y}_j}{{D^{\alpha }}_{ij}}}$$

(1)

where $Trad{e}_{ij}$ is the trade flow between countries $i$ and $j$, $Y_i,Y_j$ represent the GDP of countries $i$ and $j$ (economic mass), $D_{ij}$ indicates the distance between countries $i$ and $j$ (proxy for trade costs), $G$ denotes gravitational constant, and $\alpha$ represents the elasticity of distance. Equation (1) was linearised using natural logarithms to facilitate econometric analysis and improve interpretability, as shown below (Anderson & van Wincoop, 2003):

$$\mathrm{I}\mathrm{n}{X}_{ij} ={\alpha }_0+{\alpha }_1\mathrm{l}\mathrm{n}{Y}_i+{\alpha }_2\mathrm{l}\mathrm{n}{Y}_j-{\alpha }_3\mathrm{l}\mathrm{n}{D}_{ij}+{\mu }_{ij}$$

(2)

Equation (2) was augmented with trade-relevant variables and applies log-linearisation to improve explanatory power and interpretability, as presented in Equation (3):

$$\begin{gathered} \mathrm{I}\mathrm{n}{X}_{ij} ={\beta }_0+{\beta }_1\mathrm{l}\mathrm{n}GD{P}_{jt}+{\beta }_2\mathrm{l}\mathrm{n}PO{P}_{jt}+{\beta }_3\mathrm{l}\mathrm{n}FD{I}_{jt}+{\beta }_4\mathrm{l}\mathrm{n}T{O}_{jt}+{\beta }_5\mathrm{l}\mathrm{n}ID{U}_{jt}+{ \beta }_6\mathrm{l}\mathrm{n}P{C}_{it} \\ +{\beta }_7\mathrm{l}\mathrm{n}I{R}_{jt}+{\beta }_8\mathrm{l}\mathrm{n}X{R}_{ijt}+{\beta }_9\mathrm{l}\mathrm{n}P{W}_{it} +{\beta }_{10}\mathrm{l}\mathrm{n}DS{T}_{ij}+{\epsilon }_{ijt} \end{gathered}$$

(3)

where ${\beta }_0$ is the intercept, ${\beta }_1$–${\beta }_{10}$ are the regression coefficients for each independent variable, $\mathrm{I}\mathrm{n}$ represents the natural logarithm, $t$ indicates the time period (2011–2022), $j$ denotes the importing East African countries, $i$ represents South Africa as the exporting country, and ${\epsilon }_{ijt}$ is the error term capturing unobserved factors.

Table 1. Description of Variables/Parameters and Their Expected Signs.

Variable/Parameter	Description	Expected Sign
Exports ($X_{ij}$)	Value of wine exported from South Africa to East African countries (in ZAR)	n/a
Common intercept (β0)	The constant term captures baseline trade levels when all explanatory variables are zero	n/a
GDP of the importing countries (GDPjt)	Gross Domestic Product of East African countries (j) at time t (in ZAR)	(+)
Population of the importing countries (POPjt)	Population size of East African countries (j) at time t (in millions)	(+)
FDI of the importing countries (FDIjt)	Net foreign direct investment inflows as % of GDP for country j at time t	(+)
Trade openness of the importing countries (TOjt)	Ratio of total trade (exports + imports) to GDP for country j at time t (percentage)	(+)
Import duty (IDUjt)	Import duty imposed on South African wine in country j at time t (percentage)	(−)
Production Capacity of the exporting country (PCit)	Total volume of wine produced in South Africa (i) at time t (in gross litres)	(+)
Inflation Rate (IRjt)	Annual inflation rate in country j at time t (percentage)	(+)
Exchange Rate (XRijt)	Exchange rate of ZAR against the currency of country j at time t (average annual)	(−)
Price of wine ($P{W}_{it}$)	Average price of wine sold in bulk in ZAR	(−)
Distance ($DS{T}_{ij}$)	Geographical distance (km) between South Africa (i) and country j (time-invariant)	(−)
Error Term (εiit)	Captures unobserved factors affecting wine exports to country j at time t	n/a

Source: Adapted from Phaleng (2020). The (+) and (−) symbols in the table indicate the expected direction of the relationship between each explanatory variable and South Africa’s wine exports, based on economic theory and prior empirical studies.

summarises the variables, their descriptions, and expected influence on South Africa’s wine exports to the selected East African countries, based on economic theory and evidence from the literature, as discussed in Section 2.2.

In the estimation process, models incorporating fixed effects, random effects, and pooled OLS were evaluated. Fixed effects were favoured to account for unobserved heterogeneity among countries. The Hausman test was employed to verify the appropriateness of fixed effects, while the Redundant Fixed Effects Likelihood Ratio (LR) test justified its selection over pooled OLS (Wooldridge, 2010; Greene, 2012; Feenstra, 2015). The Dumitrescu and Hurlin (2012) Panel Granger Causality Test was applied to assess the directional influence between explanatory variables and wine exports. Further diagnostic tests evaluated heteroscedasticity, serial correlation, model stability, and the appropriateness of the functional form to confirm the robustness and reliability of the results.

4. Results

4.1. Global Overview and Trend Analysis of South Africa’s Wine Exports to Africa

To establish context, an overview of South Africa’s wine exports to key global markets, including Europe, the United States, Canada, and China, is presented first. These markets have historically dominated South Africa’s wine trade, reflecting the industry’s reliance on established, high-value destinations outside the continent of Africa. A trend analysis was then conducted to examine South Africa’s wine export patterns to Africa, providing a comparative perspective. The analysis begins with aggregate export trends for the African continent, followed by a breakdown of regional patterns, and concludes with a focus on the three selected East African markets (Kenya, Tanzania, and Mauritius). Overall, the trend analysis provides valuable insights into South Africa’s export performance and highlights the role of the selected East African countries as emerging destinations with strong growth potential.

South Africa’s wine exports between 2010 and 2022 demonstrated a general increase in value, despite fluctuations in volumes. Europe, particularly the United Kingdom, the Netherlands, and Germany, remained the largest markets, with the United Kingdom maintaining a strong role in shaping perceptions of quality and sustaining volume growth through higher-priced bulk exports (USDA FAS, 2024). The United States also represented a significant market. However, competitiveness has been undermined by the imposition of a 30% tariff in 2022, which placed South African wines at a disadvantage compared to those of EU and Chilean producers (Financial Times, 2022; USDA FAS, 2024). Canada emerged as a stable and growing importer, while other African destinations showed promising potential (USDA FAS, 2024).

Furthermore, China has become an important market due to geopolitical developments, notably trade tensions with Australia (WESGRO, 2023), which initially created new opportunities for South African exporters. However, recent reports highlight a sustained decline in overall wine demand in China, with the OIV (2025) noting a 9.3% drop in consumption in 2024, cautioning against relying on the Chinese market as a stable growth driver for South African wine exports. In contrast, regional markets in East Africa remain underexplored, despite evidence of increasing demand, positioning them as key opportunities for diversification. Figure 1 illustrates the overall trend in South Africa’s wine exports to Africa from 2011 to 2022.

South Africa’s wine exports to other African countries generally rose, especially between 2013 and 2015, driven by growing demand and favourable trade conditions. The steep decline in 2020 is attributed to disruptions caused by COVID-19, whereas the decrease between 2017 and 2018 is probably the result of policy or economic changes. The 2021 and 2022 recoveries show that the market was active again after the lockdown. Figure 2 breaks down exports by African region in more detail.

The findings show that Southern and Eastern Africa comprise the largest wine export markets in Africa, with steady growth and occasional peaks. Interestingly, exports to South Africa peaked in 2020, while exports to East Africa performed well in 2015 and 2022. On the other hand, South African wine’s market potential in West and Middle Africa appears to be limited, as evidenced by the region’s consistently low export values over the period. Given the impressive performance of wine exports to East African markets, Figure 3 illustrates the patterns in South Africa’s wine exports to the top importing nations, offering justification for prioritising Kenya, Tanzania, and Mauritius.

Import levels were moderate for Mauritius, whereas Kenya and Tanzania showed strong and stable demand for South African wine, with clear peaks around 2021. On the other hand, countries such as Malawi, Uganda, Seychelles, and Rwanda exhibited limited and irregular import activity, while Zimbabwe and Zambia displayed erratic import trends. Kenya, Tanzania, and Mauritius were crucial to the recovery of South African wine exports to East Africa after the COVID-19 pandemic, as shown by the significant increase in export volumes in 2021 and 2022.

Overall, the trend analysis confirms that Kenya, Tanzania, and Mauritius are the most promising East African markets for South African wine exports, justifying their selection for this study. Their steady demand is underpinned by economic growth and expanding tourism sectors, which make a significant contribution to GDP and employment. These factors may have helped drive sustained demand for wine, particularly within luxury hospitality markets. This economic backdrop positions these three countries as key growth markets with substantial opportunities for sustained export growth, while other East African countries remain niche or secondary markets.

4.2. Preliminary Analysis Results

To enhance the robustness of the gravity model, four essential pre-estimation analyses were performed: descriptive statistics, correlation analysis, multicollinearity testing, and unit root testing. Descriptive statistics were employed to summarise the principal characteristics of variables affecting wine exports to Mauritius, Kenya, and Tanzania, as illustrated in

Table 2. Descriptive Statistics for the Selected East African Countries.

Variable	Mean	Median	Maximum	Minimum	Std. Dev.
PW	345.04	340.67	437.75	212.03	65.50
PC	1.08	1.09	1.18	0.960	0.07
DST	2861.66	2925.00	3070	2590	295
Mauritius
	Mean	Median	Maximum	Minimum	Std. Dev.
EX	8.29	9.25	11.08	1.12	3.78
IDU	0.00	0.00	0.00	0.00	0.00
FDI	0.78	0.71	1.55	0.14	0.46
POP	1.26	1.26	1.27	1.25	0.02
GDP	157.22	170.65	211.50	83.51	39.23
IR	3.67	3.38	10.77	0.41	2.76
TO	106.77	101.31	119.50	85.83	10.69
XR	0.32	0.32	0.33	0.30	0.01
Kenya
	Mean	Median	Maximum	Minimum	Std. Dev.
EX	9.15	9.56	11.42	5.80	3.43
IDU	25.00	25.00	25.00	25.00	0.00
FDI	1.17	0.86	3.10	−0.01	0.86
POP	48.60	48.60	54.25	42.75	3.70
GDP	1120.00	1100.00	1870.00	340.00	487.13
IR	7.17	6.44	14.02	4.69	2.53
TO	40.43	36.88	58.40	27.24	9.48
XR	0.08	0.08	0.13	0.06	0.02
Tanzania
	Mean	Median	Maximum	Minimum	Std. Dev.
EX	8.68	9.56	11.42	0.63	4.13
IDU	0.00	0.00	0.00	0.00	0.00
FDI	2.57	1.95	4.57	1.43	1.09
POP	54.93	54.73	64.71	46.03	5.94
GDP	744.21	721.90	1240.00	251.26	309.87
IR	6.42	5.25	16.00	3.29	4.00
TO	39.37	35.21	56.17	27.96	8.82
XR	0.05	0.05	0.05	0.04	0.00

.

This study quantifies wine exports (EX) as the monetary value of South African wine exported to Kenya, Tanzania, and Mauritius (United Nations, 2024), expressed in ZAR millions. The highest average export value was observed for Kenya at ZAR 9.15 million, followed by Tanzania at ZAR 8.68 million, and Mauritius at ZAR 8.29 million.

Wine price (PW) denotes the average selling price of South African wine in bulk, expressed in ZAR. Beyond reflecting market competitiveness and consumer willingness to pay (Mosikari et al., 2016), price is also widely recognised in both trade and market analysis as a proxy for product quality and brand positioning. This research focused on South African fresh grape wines, specifically bottled still and sparkling varieties (≤2 L, >0.5% alcohol; HS 220421). The mean bulk sale price across three countries was ZAR 345.04, demonstrating relatively stable pricing with minimal variation. Production capacity (PC) signifies the total volume of wine produced in South Africa, measured in billions of litres. It serves as an indicator of South Africa’s export potential as a nation (SAWIS, 2023). The average production was 1.08 billion litres, indicating a stable supply over time. Import duty (IDU) refers to the tariff imposed on imported wine, expressed as a percentage of import value. It influences the final consumer price and may serve as a trade barrier (World Trade Organization, 2023). Kenya maintained a fixed 25% import duty, whereas Mauritius and Tanzania applied a 0% rate, consistent with SADC and COMESA protocols (SARS, 2021). Distance (DST) is defined as the great-circle distance in kilometres between South Africa (exporter) and the capitals of Kenya, Tanzania, and Mauritius (Mayer & Zignago, 2011). The longest average distance was observed for Mauritius at 3065 km, followed by Kenya at 2920 km, and Tanzania at 2475 km. The overall mean distance across the three destinations was 2820 km, with a minimum of 2475 km, a maximum of 3065 km, and a standard deviation of 295 km, indicating relatively low variation in geographic proximity among the selected East African markets.

Foreign direct investment (FDI) is defined as the net inflows of investment into an importing country, expressed as a percentage of the country’s GDP, reflecting economic integration and investment-driven demand (OECD, 2008; World Bank, 2024). Tanzania had the highest average FDI at 2.57%, followed by Kenya at 1.17% and Mauritius at 0.78%. Population (POP) is measured as the total number of residents in each country, expressed in millions. It serves as a proxy for market size and consumption potential (Matyas, 1997; Wani, 2024). Tanzania had the highest population at 54.93 million, followed by Kenya at 48.60 million, and Mauritius at 1.26 million. Gross Domestic Product (GDP) represents the total monetary value of all goods and services produced in a country, measured in billions of ZAR. It is a key indicator of purchasing power and demand potential (Mankiw, 2020). Kenya recorded the highest average GDP at ZAR 1120 billion, followed by Tanzania at ZAR 744.21 billion, and Mauritius at ZAR 157.22 billion.

The inflation rate (IR) is quantified as the annual percentage change in the Consumer Price Index (CPI), representing the rate at which prices of goods and services increase (IMF, 2024). Kenya exhibited the highest average inflation rate at 7.17%, followed by Tanzania at 6.42% and Mauritius at 3.67%. Trade openness (TO) is calculated as the ratio of total trade (exports plus imports) to Gross Domestic Product (GDP), expressed as a percentage, and indicates the degree of economic integration with global markets (Azhar, 2014; IMF, 2024). Mauritius demonstrated the highest trade openness at 106.77%, followed by Kenya at 40.43% and Tanzania at 39.37%. The exchange rate (XR) denotes the annual average value of the South African rand relative to the currency of each importing country, expressed in ZAR. It impacts the competitiveness of exports and is essential in international trade pricing (IMF, 2009). Mauritius recorded the highest exchange rate at 0.32 ZAR, trailed by Kenya at 0.08 ZAR and Tanzania at 0.05 ZAR.

Table 3. Correlation Coefficients.

	EX	TO	GDP	POP	IDU	FDI	IR	PC	XR	PW	DST
EX	1.0
TO	−0.5	1.0
GDP	0.54	0.44	1.0
POP	0.89	0.70	0.05	1.0
IDU	0.12	0.15	0.21	0.13	1.0
FDI	−0.16	0.19	0.27	0.55	0.28	1.0
IR	0.24	0.26	0.12	0.25	0.18	0.24	1.0
PC	0.61	0.30	0.20	0.48	0.70	0.14	0.18	1.0
XR	−0.44	0.10	0.03	0.34	0.08	0.06	0.20	0.15	1.0
PW	−0.12	−0.18	0.04	0.09	0.04	0.06	0.12	0.19	0.20	1.0
DST	−0.42	−0.15	−0.08	−0.20	0.05	0.02	0.10	−0.12	0.07	0.03	1.0

displays the correlation matrix, which assesses the strength and direction of relationships between wine exports and key explanatory variables for Mauritius, Kenya, and Tanzania.

Correlation coefficients exceeding 0.7 indicate strong relationships, values between 0.3 and 0.7 suggest moderate correlations, and those below 0.3 are considered weak (Cohen, 1988; Evans, 1996). A strong positive correlation was found between wine exports and population size (0.89), highlighting the importance of market size. Moderate positive relationships with Gross Domestic Product (GDP) (0.54) and South Africa’s production capacity (0.61) suggest that economic growth and stable supply support export activities. Conversely, trade openness showed a moderate negative correlation (−0.50), possibly due to the presence of protective policies in less open economies. Distance also demonstrated a moderate negative correlation with exports (−0.47), consistent with gravity model expectations that greater geographical separation decreases trade flows. Other variables, including import duty (0.12), foreign direct investment (FDI) (−0.16), inflation (0.24), exchange rate (−0.44), and wine prices (−0.12), exhibited weak associations. Despite these weaker correlations, including all variables remains justified, as they offer additional insights into the complex factors affecting wine exports to selected East African markets.

Table 4. Multicollinearity Results.

Variables	VIF	Tolerance
POP	1.226	0.816
GDP	1.6553	0.604
IR	3.334	0.300
TO	2.360	0.424
PW	4.064	0.246
PC	1.607	0.622
IDU	1.4978	0.668
FDI	1.467	0.682
XR	3.296	0.303
DST	2.87	0.348

presents the Variance Inflation Factor (VIF) results, which were used to assess multicollinearity among the explanatory variables and to support the initial findings from the correlation analysis.

Values under 5 suggest multicollinearity is not an issue, while values between 5 and 10 indicate moderate, but generally acceptable, multicollinearity. Values above 10, however, signal serious multicollinearity that may compromise the stability of the estimates (Kutner et al., 2005; Shrestha, 2020). In this study, all VIF values are below the critical threshold of 10, with the highest being 4.06 for the price of wine (PW). Furthermore, all tolerance values exceed the minimum cut-off of 0.1, further confirming that no variable shows problematic collinearity (Gujarati & Porter, 2009; Wooldridge, 2013). These results support the reliability and interpretability of the estimated gravity model, confirming that all variables can be included without risking inflated standard errors or distorted parameter estimates.

Table 5. IPS Test Results (Overall Panel).

Statistic	Value
W-statistic	−4.01201
p-value	0.000 ***
E(t)	−1.488
E(Var)	1.255

Note: *** Indicates significance at the 1% level.

presents the results of the IPS panel unit root test, which was used to assess the stationarity of the dataset.

The IPS test was employed to evaluate whether each panel series contains a unit root, with the null hypothesis indicating non-stationarity. Stationarity guarantees that the time-series properties of the variables are stable over time, thereby minimising the risk of spurious regression results (Gujarati & Porter, 2009; Wooldridge, 2013). A p-value less than 0.05 usually results in the rejection of the null hypothesis, thereby confirming stationarity (Baltagi, 2008). As demonstrated in Table 5, the W-statistic of −4.01201 produces a p-value of 0.000, which is statistically significant at the 1% level, indicating strong evidence against the null hypothesis. This leads to the rejection of the null hypothesis of a unit root, confirming that the panel data series are stationary. Consequently, the stationarity of the data affirms the validity of time-dependent inferences within the gravity model framework.

4.3. Model Test Results

The gravity model was estimated using three econometric specifications suitable for panel data: Pooled Ordinary Least Squares (OLS), Random Effects (RE), and Fixed Effects (FE) models. To select the most appropriate estimation method, two model selection tests were conducted. First, the Redundant Fixed Effects Likelihood Ratio (LR) test was used to determine whether the fixed effects model offers a statistically significant improvement over the pooled OLS model in capturing unobserved heterogeneity across countries. The LR test results are presented in

Table 6. LR Test Results.

	Statistic	Prob
Cross-section F	8.569113	0.0015 ***
Cross-section chi-square	20.738059	0.0000 ***

Note: *** Indicates significance at the 1% level.

.

The results allow a rejection of the null hypothesis in favour of the fixed effects model. This suggests the existence of country-specific effects not adequately captured by the pooled model, thereby confirming that the fixed effects model is more suitable for analysing the determinants of South Africa’s wine exports to the selected East African nations. Subsequently, the Hausman test was performed to determine whether to adopt the fixed or random effects models, by evaluating whether the random effects (RE) model provides consistent estimates, which depends on the presence of correlation between the unobserved effects and the regressors. The findings of the Hausman test are presented in

Table 7. Hausman Test Results.

Statistic	Value	p-Value
Chi-square test statistic	21.200480	0.0000 ***

Note: *** Indicates significance at the 1% level.

.

The results indicate that the test yielded a statistically significant outcome (p < 0.05), thereby leading to the rejection of the null hypothesis. This substantiates the preference for the fixed effects model over the random effects specification. This finding suggests that unobserved country-specific effects are correlated with the explanatory variables and must be controlled through the fixed effects estimator. Collectively, the likelihood ratio and Hausman test results affirm that the fixed effects model is the most suitable specification, as it accounts for unobserved country-specific effects and produces consistent, unbiased estimates for analysing the determinants of South Africa’s wine exports.

4.4. Gravity Model Estimation Results

Following the results of the model selection tests, which confirmed the fixed effects model as the most appropriate specification, the estimated coefficients from the gravity model employing pooled OLS, random effects, and fixed effects methods are displayed in

Table 8. Empirical Gravity Model Results.

Variables	Pooled Model	Random Effects Model	Fixed Effects Model
PC	14.016 (0.028) **	14.016 (0.050) *	10.051 (0.027) **
PW	21.406 (0.7003)	21.406 (0.7003)	16.129 (0.7523)
TO	−44.246 (0.3323)	−16.220 (0.6437)	−33.617 (0.3307)
IR	−24.97 (0.1309)	−17.28 (0.115)	−22.94 (0.0478) **
POP	0.005 (0.010) **	0.002 (0.100)	0.3770 (0.020) **
GDP	20.742 (0.0249) *	41.887 (0.080) *	41.975 (0.1101)
FDI	−14.129 (0.005) ***	−8.280 (0.007) ***	−8.353(0.0156) **
IDU	21.235 (0.010) **	47.358 (0.6484)	23.959 (0.020) **
XR	63.203 (0.001) ***	23.284 (0.006) ***	30.1229 (0.030) **
DST	−12.547 (0.092) *	−8.764 (0.101)	−10.231 (0.080) *
Number of observation	36	36	36
R-Squared	0.7614	0.7210	0.855
Adjusted R-squared	0.7210	0.6034	0.821
F-statistic	67.7701	52.987	42.8964
Durbin-Watson stat	2.3052	2. 2806	2.0598

Notes: * = 10%, ** = 5%, and *** = 1% significance levels.

. Including all three specifications in Table 8 facilitates comparison and transparency; the subsequent discussion focuses on interpreting the fixed effects results, as justified by the model selection tests.

Alongside the model selection tests, the overall explanatory power of the estimated gravity models was evaluated using the R-squared, Adjusted R-squared, and Durbin-Watson statistics. As shown in Table 8, the fixed effects model performs better, with an R-squared value of 0.855 and an adjusted R-squared value of 0.821, indicating a strong explanatory ability. The Durbin-Watson statistic (2.0598) is near the ideal value of 2, indicating no evidence of autocorrelation in the residuals. Based on both the model selection tests and the model fit statistics, the fixed effects model is deemed the most suitable for analysing the determinants of South Africa’s wine exports to the chosen East African countries. Consequently, the following discussion concentrates on the estimated coefficients from the fixed effects model.

The fixed effects model identified several statistically significant determinants of South African wine exports to Kenya, Tanzania, and Mauritius. Production Capacity (PC) exhibited a positive and statistically significant coefficient of 10.051 at the 5% significance level. This suggests that a 1% increase in production capacity is associated with a 10.05% increase in wine export values, assuming all other factors remain constant. The Inflation Rate (IR) demonstrated a negative and statistically significant coefficient of −22.94 at the 5% significance level, suggesting that a 1% rise in inflation in importing countries is associated with a 22.94% decline in South African wine exports. Population (POP) was found to have a positive and statistically significant impact, with a coefficient of 0.37 at the 5% significance level. This implies that a 1% increase in the population of importing countries is associated with a 0.37% increase in export volumes. Foreign Direct Investment (FDI) showed a statistically significant negative relationship with exports, with a coefficient of −8.35 at the 5% level. Hence, a 1% increase in FDI inflows corresponds to an 8.35% decrease in wine exports. Import Duty (IDU) was also significant at the 5% level, with a positive coefficient of 23.96. This suggests that a 1% increase in import duty is associated with a 23.96% increase in wine exports. Finally, the Exchange Rate (XR) displayed a positive and statistically significant coefficient of 30.12 at the 5% level, implying that a 1% depreciation of the South African rand (ZAR) is associated with an approximate 30.12% increase in South African wine exports.

Conversely, Wine Price (PW), Trade Openness (TO), and Importers’ GDP were determined to be statistically insignificant within the fixed effects model. Although these variables were incorporated to consider theoretical expectations and previous empirical findings, their coefficients did not achieve conventional significance levels (p > 0.10). This suggests that, in the context of the three selected East African nations, fluctuations in these factors do not significantly influence changes in wine export values throughout the study period.

Taken together, the results underscore the significance of production capacity, macroeconomic stability, and demographic factors in influencing export performance, while also emphasising the context-dependent nature of trade determinants. The insignificance of GDP and trade openness highlights the complexity of export dynamics, indicating that broader economic size and liberal trade policies alone may not be sufficient to stimulate demand for specific export commodities, such as wine. These insights are instrumental in formulating targeted trade strategies and enhancing export competitiveness within niche markets across the African continent.

4.5. Validation of Fixed Effects Results

While the fixed effects model revealed statistically and economically significant factors influencing wine exports to East African countries, it is essential to validate these findings through robustness and specification checks. Accordingly, the Dumitrescu and Hurlin panel Granger causality test was employed to assess the direction of causality between wine exports and the explanatory variables included in the fixed effects model.

Table 9. Dumitrescu and Hurlin Test Results.

Null Hypotheses	p-Value	W-Stat	Zbar-Stat	Causality
FDI → Export	0.5898	0.8387	3.6798	No
Export → FDI	0.0235 **	2.4905	0.0245	Yes
GDP → Export	0.5880	3.2850	1.6248	No
Export → GDP	2.0500	3.0457	1.8726	No
POP → Export	0.4100	2.1373	0.7256	No
Export → POP	0.5110	3.3974	1.5600	No
IR → Export	0.7979	3.2566	1.5878	No
Export → IR	0.0400 **	4.5129	3.1793	Yes
PC → Export	0.7000	2.6289	0.6248	No
Export → PC	0.0100 **	3.8640	2.3025	Yes
PW → Export	0.5800	3.2189	1.5582	No
Export → PW	0.0100 **	3.8694	2.4049	Yes
XR → Export	0.8800	2.6209	0.2498	No
Export → XR	0.7100	3.3154	1.9500	No
IDU → Export	0.5200	3.2848	1.4908	No
Export → IDU	0.7900	3.2590	1.4859	No
TO → Export	0.5100	0.5400	0.6400	No
Export → TO	0.1600	2.0200	1.3900	No
DST → Export	0.3200 *	2.4100	1.1200	No
Export → DST	0.2100	2.5600	1.3500	No

Notes: ** and * indicate significance at the 5% and 10% levels, respectively.

presents the test results, which include the direction of causality, W-statistics, Z-bar statistics, p-values, and indicate whether significant Granger causality exists between exports and each variable.

The test demonstrated unidirectional causality emanating from wine exports towards four variables: Foreign Direct Investment (FDI), Inflation Rate (IR), Production Capacity (PC), and Price of Wine (PW), all of which are statistically significant at the 5% level. Conversely, no evidence of reverse causality was found, suggesting that these variables do not Granger-cause exports. Furthermore, no causal relationship was observed, regardless of direction, between wine exports and Gross Domestic Product (GDP), population, trade openness, exchange rate, or import duties. These findings suggest that, although the fixed effects model detects correlations, only a limited number of the explanatory variables are causally affected by exports, rather than vice versa. To ensure the robustness of the fixed effects model, four diagnostic tests were conducted, as detailed in

Table 10. Summary of Diagnostic Test Results.

Test Type	Test Statistic	Value	p-Value	Decision (at 5% Level)	Conclusion
Normality Test	Jarque–Bera	2.6931	0.2374	Do not reject H0	Residuals are normally distributed.
	Skewness	–	–	–	Approximately symmetrical.
	Kurtosis	–	–	–	Approximately mesokurtic.
Serial Correlation	F-statistic	2.2603	0.1520	Do not reject H0	No evidence of serial correlation.
	Obs*R-squared	5.9135	0.1270	Do not reject H0
Heteroscedasticity	F-statistic	0.3001	0.9744	Do not reject H0	No evidence of heteroscedasticity.
	Obs*R-squared	3.8583	0.9535	Do not reject H0
	Scaled Explained SS	12.0533	0.2815	Do not reject H0
Functional Form	T-statistic (RESET)	0.3785	0.6988	Do not reject H0	The model is correctly specified.
	F-statistic (RESET)	0.8156	0.6699	Do not reject H0
	Likelihood Ratio (RESET)	0.0846	0.6396	Do not reject H0

Obs*R-squared is the test statistic from the Breusch-Godfrey test for serial correlation and the Breusch-Pagan test for heteroscedasticity.

.

The Normality Test (Jarque–Bera) demonstrated that the residuals conformed to a normal distribution, with a p-value of 0.2374, suggesting that the assumption of normality was not violated. The Serial Correlation Test (Breusch-Godfrey) indicated no significant serial correlation in the residuals, with a p-value of 0.1520, thus confirming the independence over time. The Heteroscedasticity Test (Breusch-Pagan-Godfrey) evidenced constant variance in the error terms, with a p-value of 0.9744, thereby satisfying the homoscedasticity assumption. Lastly, the Functional Form Test (Ramsey RESET) revealed no signs of omitted variable bias or misspecification of the functional form, with a p-value of 0.6699, confirming the correct specification of the model. In summary, the results obtained from both the Granger causality test and the diagnostic assessments underscore the robustness, reliability, and proper specification of the fixed effects model, thus reinforcing confidence in the empirical findings concerning the determinants of South Africa’s wine exports to East African nations.

5. Discussion

The results derived from the gravity model, as elaborated on in Table 8, illustrate the influence of various economic and trade-related factors on South Africa’s wine exports to selected East African nations. Of the nine explanatory variables examined, five demonstrated statistical significance: production capacity, population size, exchange rate, import duties, and inflation rate. Conversely, the remaining four variables, GDP, foreign direct investment (FDI), trade openness, and wine prices, did not exhibit statistical significance. This discussion aims to interpret these findings in accordance with theoretical expectations outlined in Table 1 and relevant prior research. Additionally, the discussion considers the particular trade dynamics of South African wine exports to East Africa, which are primarily driven by tourism and focused on value-added trade.

5.1. Effect of Production Capacity (PC) on South Africa’s Wine Exports

Production capacity was expected to positively affect wine exports, in line with Phaleng (2020), who showed that greater domestic output strengthens South Africa’s ability to meet external demand. Consistent with this expectation, the coefficient was positive and statistically significant at the 5% level, underscoring the supply-driven nature of South African wine exports. This is particularly pertinent in a tourism-oriented, value-added trade environment where stable and expanding production capacity is critical for sustaining export growth, especially in the aftermath of COVID-19 disruptions. These findings align with prior research (Idsardi, 2013; Phaleng, 2020; Wani, 2024), which stressed the importance of sufficient domestic production in supporting export performance. From a structural perspective, South Africa possesses well-established viticultural regions, modern winery infrastructure, and specialised expertise that underpin its competitiveness in global markets. However, the wine industry also faces constraints, including limited land suitability for wine grapes, periodic droughts, and increasing risks associated with climate change, which may restrict future expansion and long-term sustainability. In this context, although a comprehensive agronomic assessment falls beyond the scope of this paper, these structural realities provide important context for the empirical finding that production capacity has a strong positive influence on export performance.

5.2. Effect of Price of Wine (PW) on South Africa’s Wine Exports

The price of wine was anticipated to have a negative effect on wine exports, as earlier studies found that higher prices generally deter demand (Mosikari et al., 2016; Ogede et al., 2020). Nevertheless, the estimated coefficient was negative yet statistically insignificant, indicating that price variations exerted minimal influence on South African wine exports to Kenya, Tanzania, and Mauritius. Specifically, the price variable used in the model reflects the average bulk export price (in ZAR) rather than final consumer retail prices, as retail channels often absorb or stabilise fluctuations, which helps explain the insignificance. Moreover, in wine markets, price usually serves as a proxy for quality, where instability can undermine a brand’s reputation. In some cases, higher prices can even stimulate demand for premium products. This interpretation is consistent with the specialised and relatively inelastic demand observed in these markets, where consumption is shaped more by tourism, hospitality procurement, and brand positioning than by direct price sensitivity.

5.3. Effect of Importers’ Trade Openness (TO) on South Africa’s Wine Exports

Trade openness was anticipated to exert a positive influence on wine exports, consistent with research showing that trade policies characterised by minimal restrictions generally facilitate a more seamless flow of goods from exporting nations (Azhar, 2014; Yang et al., 2022). Nonetheless, the estimated coefficient was negative and lacked statistical significance, thereby contradicting theoretical expectations. The insignificance of trade openness in promoting South African wine exports suggests that, despite relatively liberal trade policies, other factors, such as regulatory obstacles, regional taste preferences, or well-established local competitors, may constrain growth. Although trade openness typically simplifies the import process (Azhar, 2014; Yang et al., 2022), wine appears to encounter unique market barriers that diminish its responsiveness to broader trade liberalisation. Consequently, trade openness does not necessarily result in increased demand for South African wine within these East African markets.

5.4. Effect of Importers’ Inflation Rate (IR) on South Africa’s Wine Exports

Inflation was anticipated to exert a positive influence on wine exports, as previous research has shown that increasing domestic prices can render foreign goods relatively more affordable and appealing, thereby encouraging consumers and businesses to seek imports to circumvent elevated local costs (Ball, 2005; Herrera & Baleix, 2010; Okpe & Ikpesu, 2021). Nonetheless, within this study, the estimated coefficient for inflation was found to be negative and statistically significant at the 5% level. This indicates that escalating inflation in Kenya, Tanzania, and Mauritius diminishes consumers’ purchasing power and decreases demand for non-essential luxury goods such as South African wine. The observed negative effect reflects the discretionary nature of wine consumption, which is more sensitive to fluctuations in household income and tourism flows than to general price inflation. Consequently, contrary to some prior research suggesting that inflation fuels import demand, in this context, inflationary pressures seem to constrain affordability and reduce imports of South African wine within these East African markets.

5.5. Effect of Importers’ Population (POP) on South Africa’s Wine Exports

The coefficient for population was expected to be positive, as earlier studies found that a larger market size may increase export volumes (Eshetu & Mehare, 2020; Wani, 2024). The estimated coefficient was indeed positive and significant at the 1% level, confirming that population growth in Kenya, Tanzania, and Mauritius expands the potential consumer base for imported wines. This result aligns with previous studies (Matyas, 1997; Potelwa et al., 2016; Eshetu & Mehare, 2020; Wani, 2024), which found that larger populations enhance export potential. In the context of South African wine exports, this suggests that population growth in the selected East African markets broadens the target market for imported wines, particularly in urban areas and the hospitality sector, which caters to middle- and high-income consumers.

5.6. Effect of Importers’ GDP on South Africa’s Wine Exports

The coefficient for Gross Domestic Product (GDP) was expected to be positive, as prior studies had shown that elevated economic activity within importing nations could increase import demand (Phaleng, 2020; Eshetu & Mehare, 2020; Wani, 2024). However, the estimated coefficient for GDP was positive yet statistically insignificant, suggesting that economic growth in Kenya, Tanzania, and Mauritius does not inherently lead to increased imports of South African wine. This observation contrasts with the traditional expectations of the gravity model and prior research (Phaleng, 2020; Eshetu & Mehare, 2020; Wani, 2024), which identified GDP as a key determinant of export performance. The lack of statistical significance for GDP may therefore be attributable to the non-essential nature of wine, which is more sensitive to travel and lifestyle trends than to overall economic growth.

5.7. Effect of Foreign Direct Investment (FDI) on South Africa’s Wine Exports

Foreign Direct Investment (FDI) was expected to have a positive impact on wine exports, based on previous research indicating that increased foreign investment generally signifies greater economic integration and potential improvements in purchasing power (OECD, 2008; Keho, 2020). However, the estimated coefficient was negative and showed no statistical significance, suggesting that FDI inflows into Kenya, Tanzania, and Mauritius do not directly lead to an increase in South African wine imports. This likely reflects the sectoral focus of FDI, which primarily targets infrastructure, construction, and extractive industries, with limited spillover effects into consumer-oriented markets. These findings contrast with previous research (Mukhtarov et al., 2019; Karimov, 2020; Keho, 2020), which found positive relationships between FDI and import demand, especially for manufactured or specialised goods. The difference can be explained by the fact that wine is a final, value-added hedonic product, whose import demand is more influenced by tourism-related consumption and the disposable income of niche consumers rather than by broad-based, investment-driven economic activity.

5.8. Effect of Import Duty (IDU) on South Africa’s Wine Exports

Import duty was anticipated to exert a negative influence on exports, as previous research found that higher tariffs generally elevate costs for importers and may diminish demand (Idsardi, 2013; Wani, 2024). However, in contrast to these expectations, the estimated coefficient for import duty was positive and statistically significant at the 5% level, suggesting that nominal tariff rates have exerted a limited restrictive effect on South African wine exports. This counterintuitive outcome can be elucidated by the fixed and highly disparate tariff structure across the three markets. For instance, Kenya consistently imposed a 25% import duty, whereas Mauritius and Tanzania maintained a 0% duty throughout the study period. This disparity reflects Mauritius and Tanzania’s membership in the Southern African Development Community (SADC) Free Trade Area (FTA), which exempts most intra-regional trade from customs duties. It is essential to note, however, that excise duties and Value Added Tax (VAT) are levied on imported wine at varying rates across these countries, which may impact the final cost and competitiveness of South African wine in each market. Since the gravity model aggregates exports to all three countries, the zero-duty environment in Mauritius and Tanzania likely offsets the negative impact associated with Kenya’s tariffs, resulting in an overall positive coefficient. Essentially, the observed positive elasticity embodies a combination of Kenya’s market strength and the zero-duty conditions in Mauritius and Tanzania, rather than a direct causal benefit derived from imposing tariffs on wine imports.

5.9. Effect of Exchange Rate (ER) on South Africa’s Wine Exports

The exchange rate was anticipated to exert a negative impact on exports, in line with studies showing that a depreciation of the rand renders South African products more affordable internationally (Idsardi, 2013; Phaleng, 2020). The coefficient obtained was negative and statistically significant at the 5% level, corroborating that rand depreciation facilitates an increase in wine exports to East Africa. This observation is consistent with the Mundell-Fleming framework and previous scholarly research (Idsardi, 2013; Phaleng, 2020), emphasising currency competitiveness as a fundamental factor in agricultural export dynamics. Consequently, in the targeted East African wine markets, a weaker rand has contributed to maintaining the price competitiveness of South African wine in both the hospitality and retail sectors, thereby fostering export expansion.

5.10. Effect of Distance (DST) on South Africa’s Wine Exports

Distance was expected to negatively affect wine exports, consistent with gravity model theory, which posits that greater geographical separation reduces trade flows by increasing transport costs and limiting market integration (Anderson & van Wincoop, 2003; Tang, 2018; Sato, 2020). The estimated coefficient for distance was negative across all models, aligning with theoretical expectations, although it was only marginally significant at the 10% level. This suggests that while distance does increase logistical costs and potentially limits accessibility, its overall effect on South Africa’s wine exports to Kenya, Tanzania, and Mauritius is relatively modest. One possible explanation is that regional integration and established shipping routes within Africa may partially offset the effects of pure distance. Nonetheless, the negative sign underscores the importance of addressing transportation and distribution bottlenecks if South African wine is to enhance its competitiveness in East African markets.

Overall, the results from the fixed effects analysis of South African wine exports to the selected East African countries share similarities with those of previous studies. For instance, as with earlier research, this study emphasises the importance of production capacity, importers’ GDP, population size, exchange rates, and trade conditions (such as import duties) on export values. The significant positive impact of production capacity aligns well with findings by Idsardi (2013), Phaleng (2020), and Wani (2024), indicating that increased production boosts export performance, a recurring theme across the reviewed agricultural export literature.

Nonetheless, certain findings deviated from conventional expectations, with both significant and insignificant results offering insights into trade dynamics specific to wine exports. Regarding significant findings, while Foreign Direct Investment (FDI) commonly exerts a positive influence on import demand, as documented by Keho (2020), Mukhtarov et al. (2019), and Karimov (2020), this study observed a pronounced negative correlation between FDI and wine exports. This suggests that increased investment may be concentrated on local industries, thereby diminishing reliance on imports and potentially indicating an import substitution trend. Furthermore, the observed significant negative impact of importers’ inflation contrasts with the conclusions of studies by Ball (2005) and Okpe and Ikpesu (2021), which reported that inflation tends to enhance demand for imports due to increased cost competitiveness. In the current context, inflation appears to suppress wine imports, likely reflecting the non-essential nature of wine and the constrained budgets prevalent in East African markets.

Regarding the insignificant results, while it is often observed that importers’ GDP positively influences import demand, as highlighted in previous studies (Mosikari et al., 2016; Eshetu & Mehare, 2020; Phaleng, 2020; Wani, 2024), this study finds that importers’ GDP did not exhibit a significant effect on South African wine exports. This discrepancy may be attributed to particular preferences or limited demand for non-essential imports such as wine, which may not directly react to overall economic growth.

Similarly, trade openness has often been associated with increased import flows in studies conducted by Azhar (2014) and Yang et al. (2022). However, within this study, trade openness did not exhibit a significant impact on wine exports, potentially owing to product-specific barriers such as local preferences or non-tariff restrictions, which may constrain the anticipated influence of trade openness on wine imports.

Lastly, the price of wine was also insignificant, diverging from standard economic theory, which anticipates that price influences demand. This phenomenon likely reflects a fixed and inelastic pricing strategy adopted during the study period, as well as the niche, tourism-driven nature of demand within these markets.

In summary, although the findings reaffirm established patterns relating to production capacity, population size, and economic conditions in alignment with existing scholarly literature, the distinctive adverse effects of Foreign Direct Investment (FDI) and inflationary pressures offer important new insights into the trade dynamics of wine exports in East Africa. This highlights the necessity of formulating customised trade strategies that account for both universal economic factors and the unique contextual elements influencing niche product markets within the East African region.

6. Conclusions

This study investigated the determinants of South Africa’s wine exports to Kenya, Tanzania, and Mauritius using trend analysis (2011–2022) and an augmented gravity model. Guided by the literature, it was hypothesised that GDP, population, FDI, trade openness, and production capacity would promote exports, while import duties, inflation, exchange rates, and wine prices would constrain them. The results partially confirmed these expectations. Specifically, production capacity, population, exchange rates, import duties, inflation, and distance emerged as significant determinants, while GDP, FDI, trade openness, and wine prices showed no statistical significance. The insignificance of these variables suggests that broader economic growth and investment have not translated into higher wine demand, likely due to consumer preferences, limited market penetration, competing beverages, and niche consumption patterns where cultural and social factors outweigh price sensitivity.

The significant results imply that the exportation of South African wine to Kenya, Tanzania, and Mauritius is primarily shaped by supply-side capacity (production capacity), macroeconomic stability (inflation and exchange rates), and trade conditions (import duties), with population size further reinforcing demand in importing countries. Together, these determinants highlight the interplay of production, economic, and policy factors in influencing South Africa’s wine exports to East Africa.

For producers and industry stakeholders, the findings suggest that scaling up production to meet external demand, coupled with investment in quality enhancement and brand differentiation, will be critical in sustaining wine export growth. Maintaining stable but competitive pricing strategies may also help absorb the negative effects of inflation in importing countries, particularly in tourism-driven segments of the market. For policymakers and trade regulators, the results highlight the importance of targeted interventions, such as negotiating favourable tariff regimes, ensuring exchange rate stability, and developing risk-mitigation tools to cushion inflationary shocks. Furthermore, harmonising non-tariff measures and facilitating cross-border logistics could address product-specific trade barriers that persist even in liberalised trade environments.

By clarifying both the drivers and constraints of wine export performance, this study makes a novel contribution to the literature on value-added agricultural trade in emerging East African markets. Unlike raw commodities, wine exports are often associated with lifestyle and tourism dynamics, as highlighted in the literature, which suggests that they may require a comprehensive approach from both exporters and regulators.

Ultimately, while the study offers new insights, it is not without its limitations. The analysis was confined to three East African countries and a relatively short period (2010–2022). Future research should broaden the scope to include additional emerging African markets such as Uganda, Rwanda, and Ethiopia, which are experiencing rapid growth in tourism and urban middle-class consumption, and cover other product categories within the wine industry beyond bottled still and sparkling grape wines (HS 220421), such as bulk and fortified wines, to capture a fuller picture of export dynamics. It should also extend the timeframe beyond 2022 and adopt a longer historical perspective to capture evolving trade patterns and structural shocks while also integrating data from the International Organisation of Vine and Wine (OIV) to broaden coverage and incorporate the most recent years, thereby enhancing robustness and timeliness. Moreover, tourism plays a role in stimulating wine exports, both directly through tourist consumption and indirectly by influencing demand in tourists’ home countries. Still, its omission here reflects concerns with endogeneity and scope, making it a valuable avenue for future research. Similarly, cultural and religious factors, such as the proportion of the Muslim population, can significantly influence alcohol consumption patterns and merit consideration in extended gravity model applications, thereby enriching the understanding of wine trade dynamics in diverse markets. Additionally, unit values offer valuable insights into export characteristics and market positioning; however, since this study focused on aggregate trade values and volumes, future research could analyse unit value dynamics to better capture product differences and consumer preferences across markets.