Determinants of Vehicle Sales in South Africa: Evidence from Macroeconomic and Market Dynamics

Abstract

Vehicle sales constitute an important component of household consumption and a key transmission channel of macro-financial conditions in South Africa. This study investigates the macroeconomic determinants of vehicle sales by examining the roles of economic activity, interest rates, and inflation over the period 2000Q1 to 2025Q4. Using quarterly data, the analysis employs the autoregressive distributed lag (ARDL) bounds testing approach to estimate both long-run and short-run relationships, complemented by an error correction model and Granger causality analysis. The results confirm the existence of a stable long-run cointegrating relationship among the variables. In the long run, vehicle sales respond positively to economic growth, while inflation and interest rates are associated with reduced demand. Short-run dynamics indicate that vehicle sales respond positively to economic growth, and negatively to interest rates and inflation, reflecting affordability and credit constraints, alongside rapid adjustment to macroeconomic shocks. The Granger causality results suggest that vehicle demand is largely driven by macro-financial conditions rather than exerting feedback effects on them. Overall, the findings highlight the sensitivity of South Africa’s automotive sector to macroeconomic stability and underscore the importance of prudent monetary policy and price stability in sustaining durable goods demand.

1. Introduction

The automotive industry plays a critical role in the global economy, contributing significantly to employment, innovation and international trade (Organisation for Economic Co-Operation and Development [OECD], 2024), while also exerting wide-ranging effects on production and macroeconomic dynamics (International Monetary Fund [IMF], 2024). Vehicle sales are widely regarded as an important indicator of macroeconomic performance, as demand for automobiles (classified as durable goods) is highly procyclical and responds strongly to changes in income, credit conditions, and broader economic fluctuations (Caballero, 1993; Baghestani & Fatima, 2021; Gavazza & Lanteri, 2021; Bertolotti et al., 2023). Because vehicles involve substantial financial commitments, their demand is particularly sensitive to macroeconomic conditions. Consumption theory suggests that durable goods purchases respond strongly to income and interest rate fluctuations (Blanchard & Johnson, 2013), while empirical evidence shows that automobile expenditures are procyclical and decline sharply during economic downturns (Gavazza & Lanteri, 2021; U.S. Bureau of Labor Statistics, 2024).

This positions the automotive sector as an important transmission channel through which macroeconomic shocks propagate to real economic activity, consistent with evidence on shock amplification in networked production systems and sector-specific supply-chain dynamics (Acemoglu et al., 2016; Flori et al., 2026). International evidence further demonstrates that vehicle demand is strongly influenced by income growth and financial conditions, where rising income and favourable financing stimulate purchases, while monetary tightening and adverse economic conditions lead to declines in sales as households postpone durable goods expenditure (McCarthy, 1996; Baghestani & Fatima, 2021; Gavazza & Lanteri, 2021). Macroeconomic pressures reduce household purchasing power and weaken demand for automobiles, which are high-value and postponable durable goods sensitive to such conditions (Caballero, 1993; Sin et al., 2024).

These dynamics tend to be more pronounced in emerging market economies, where macroeconomic volatility, credit constraints, and structural labour-market challenges amplify the responsiveness of durable goods’ demand to economic shocks. Empirical studies on emerging economies indicate that vehicle sales are particularly sensitive to interest rates and inflation, reflecting the importance of borrowing costs, income dynamics, and broader macroeconomic conditions in shaping automobile demand (Patra & Rao, 2017; Chisasa & Dlamini, 2013; Copeland et al., 2019). Although the financial accelerator framework does not focus specifically on household consumption, it implies that sectors reliant on external finance, such as credit-intensive durable goods, are more sensitive to changes in monetary policy and financial conditions due to amplification through credit market frictions (Bernanke & Gertler, 1995; Bernanke et al., 1999).

As a result, analysing vehicle sales dynamics in these economies requires empirical frameworks capable of capturing both short-run adjustments and long-run structural relationships. South Africa represents a particularly relevant emerging market case, hosting a mature and globally integrated automotive sector that contributes approximately 5.3% to gross domestic product and about 22% of manufacturing output, while supporting substantial employment across extensive value-chain linkages (Lamprecht, 2024; Barnes, 2025). The sector serves a dual role by meeting domestic demand and participating in global export markets, primarily supplying Europe and selected developing regions. Vehicle sales in South Africa reflect household consumption patterns and play a significant role in shaping industrial performance and trade outcomes, while also serving as a key indicator of broader macroeconomic conditions (Trade & Industrial Policy Strategies [TIPS], 2023).

Despite its strategic importance, South Africa’s automotive market operates within a constrained macroeconomic environment characterised by inflationary pressures, elevated interest rates, and persistent structural labour-market challenges. Inflationary pressures, together with elevated interest rates following a period of monetary tightening, have increased the cost of credit and dampened household consumption, while persistent labour-market weaknesses have constrained income and weakened overall demand conditions (South African Reserve Bank [SARB], 2024; Statistics South Africa [Stats SA], 2024; Parliamentary Budget Office [PBO], 2025). Cost pressures within South Africa’s automotive value chain have contributed to weaker affordability conditions in the vehicle market, reinforcing the sensitivity of vehicle demand to macro-financial variables such as interest rates, inflation, and household income. Collectively, these conditions indicate that vehicle sales in South Africa are shaped by the interaction of income dynamics, monetary conditions, price pressures, and ownership costs.

Within this context, the study aims to determine the principal macroeconomic factors influencing vehicle sales in South Africa and to assess their effects on sales behaviour over both the short- and long-term. In particular, the analysis evaluates the impact of income levels, the monetary environment, and price movements on aggregate vehicle sales, employing the autoregressive distributed lag (ARDL) bounds testing technique. Two research questions guide the investigation: Firstly, which macroeconomic indicators have a statistically significant effect on vehicle sales in South Africa? Secondly, do the influences of these variables differ between short-run adjustment dynamics and long-run equilibrium outcomes?

This study makes three distinct contributions to the literature on vehicle demand in emerging economies. Firstly, it provides empirical evidence based on a time period spanning from 2000Q1 to 2025Q4, capturing multiple structural shifts in the South African economy. Secondly, the study jointly examines economic activity, interest rates, and inflation within a unified dynamic framework, allowing for a comprehensive assessment of the interaction between macroeconomic conditions and durable goods demand. Thirdly, by employing the ARDL bounds testing approach, the study distinguishes explicitly between short-run adjustment dynamics and long-run equilibrium relationships, thereby providing policy-relevant evidence on the transmission of macroeconomic shocks to the automotive sector.

The importance of this study lies in its policy, academic, and practical relevance. By identifying the macroeconomic drivers of vehicle sales, the study provides insights into how income dynamics, monetary conditions, and inflationary pressures shape demand for durable goods. These insights are valuable for policymakers seeking to understand the transmission of macroeconomic and monetary policy to broader economic activity and industrial performance. In addition, the findings offer guidance for demand forecasting, credit-risk assessment, and strategic planning by financial institutions and automotive industry participants operating in an environment of heightened economic volatility.

However, despite the importance of the sector, empirical research on the determinants of vehicle sales in South Africa remains relatively limited and fragmented. Existing studies often examine individual determinants, such as exchange rates or fuel prices, or focus on industrial performance, without integrating core macroeconomic fundamentals into a unified dynamic demand framework (Makoni & Chikobvu, 2023; Copeland et al., 2019; Trade & Industrial Policy Strategies [TIPS], 2023). Moreover, relatively few studies explicitly distinguish between short-run shocks and long-run equilibrium relationships, despite the relevance of this distinction for policy formulation within a volatile emerging-market context.

2. Literature Review

The theoretical foundation of this study is anchored in economic theories of consumption, durable goods demand, and financial market transmission mechanisms, providing an integrated framework to analyse the responsiveness of vehicle demand to macroeconomic conditions. At the core lies the standard demand theory, which posits that consumer demand is a function of income, prices, and preferences (Varian, 2014). Vehicles, as high-value durable goods, are particularly sensitive to changes in income and relative prices, making the demand theory central to understanding fluctuations in vehicle sales.

The Keynesian consumption theory further explains short-run variations in vehicle demand by emphasising the role of current income and macroeconomic conditions in shaping household spending decisions. According to Keynes (1936), consumption responds primarily to changes in current income, implying that economic downturns and rising unemployment can rapidly compress aggregate demand. Given the postponable and income-sensitive nature of durable goods, such contractions disproportionately affect demand for items such as motor vehicles. This framework is especially relevant in South Africa, where cyclical economic downturns and labour-market constraints are likely to exert immediate effects on household consumption, in line with standard macroeconomic theory (Blanchard & Johnson, 2013).

By contrast, the permanent income hypothesis adopts a long-term viewpoint, arguing that household consumption behaviour is primarily determined by anticipated lifetime income rather than solely by contemporaneous earnings (Friedman, 1957). Given the substantial financial commitment associated with vehicle purchases, households may postpone buying vehicles when future income prospects are uncertain, even if short-term income conditions improve. In emerging economies characterised by macroeconomic volatility, income uncertainty and credit constraints tend to depress durable-goods consumption during periods of policy uncertainty and structural instability, a mechanism that is particularly relevant for vehicle demand (Deaton, 1992).

Credit market theory provides an additional lens through which to analyse vehicle sales, particularly in economies where vehicle purchases are predominantly credit-financed. The financial accelerator mechanism posits that credit market imperfections amplify economic fluctuations through changes in borrowers’ balance sheets and the external finance premium, thereby affecting borrowing costs and access to credit (Bernanke & Gertler, 1995; Bernanke et al., 1999). Rising interest rates increase the cost of vehicle financing, while tighter credit conditions constrain households’ ability to smooth consumption over time. In South Africa, the period of monetary tightening following inflationary pressures has been associated with tighter credit conditions, which have constrained vehicle affordability and weighed on market demand (Nedbank, 2024). Taken together, these theories suggest that vehicle sales respond to income dynamics, monetary conditions, price developments, and shocks in both the short term and the long term.

Behavioural factors such as consumer confidence and sentiment play an important role in shaping demand for durable goods. Consumer sentiment theory suggests that households’ expectations about future economic conditions significantly influence spending decisions, particularly for high-value and postponable purchases such as motor vehicles (Ludvigson, 2004). During periods of macroeconomic uncertainty, declining consumer confidence may lead households to delay or forgo vehicle purchases, even in the absence of immediate income shocks (U.S. Bureau of Labor Statistics, 2024; Gavazza & Lanteri, 2021). Although consumer sentiment is not explicitly included in the empirical model of this study, due to data limitations, it remains an important underlying factor influencing vehicle demand, particularly within the emerging market context.

Empirical studies consistently confirm the sensitivity of vehicle sales to macroeconomic conditions. International evidence shows that GDP growth, inflation, interest rates, and broader financial conditions are among the most significant determinants of automobile demand. Similar cyclical patterns in vehicle sales are documented in emerging markets, where downturns in income and tighter financial conditions are associated with contractions in vehicle demand (Patra & Rao, 2017). These findings highlight the strongly procyclical nature of vehicle sales and underscore the importance of macroeconomic stability in sustaining demand for durable goods such as automobiles.

Within the South African context, empirical research similarly emphasises the dominant role of macroeconomic fundamentals in shaping vehicle sales. Existing studies show that vehicle demand is closely linked to broader economic conditions, with evidence of a reciprocal relationship between vehicle sales and economic growth (Mothibi et al., 2025). This bidirectional relationship suggests that economic expansion stimulates vehicle demand, while developments in the automotive sector may also influence overall economic performance. In addition, Moodley (2023) found that inflation and other macroeconomic variables, including interest rates, play significant roles in influencing automotive sector performance, reinforcing the importance of the broader macroeconomic environment.

A more detailed examination of specific macroeconomic drivers further reinforces these insights. Economic growth is consistently found to be positively associated with vehicle sales, confirming the procyclical behaviour of automobile demand. For example, Mothibi et al. (2025), Muhammad et al. (2012), and Pehlivanoğlu and Riyanti (2018) all found that GDP exerts a positive effect on vehicle sales across different regions and economic contexts. In contrast, interest rates generally exert a negative effect on vehicle demand by increasing borrowing costs and constraining access to credit. Empirical evidence from Copeland et al. (2019) and Ustabaş and Buyun (2024) supports the view that tighter monetary conditions reduce vehicle sales by limiting affordability and financing options.

Similarly, inflation plays a critical role in shaping vehicle demand by eroding real purchasing power and increasing economic uncertainty. Empirical studies such as Pehlivanoğlu and Riyanti (2018), Muhammad et al. (2012), and Sin et al. (2024) found a negative relationship between inflation and vehicle sales, indicating that rising prices reduce households’ ability and willingness to undertake large durable goods purchases. These findings suggest that inflationary pressures weaken demand not only through reduced affordability but also through increased uncertainty, which may encourage households to postpone vehicle purchases. Despite these contributions, important gaps remain in the empirical literature on vehicle sales in South Africa. Existing studies typically examine individual determinants in isolation and frequently rely on static or single-equation frameworks, limiting their ability to capture dynamic adjustment processes. Furthermore, the existing literature rarely makes an explicit distinction between short-term reactions to macroeconomic shocks and long-term equilibrium linkages. The limited application of integrated dynamic approaches, therefore, provides the rationale for employing the auto-regressive distributed lag (ARDL) methodology in this study, as it allows for the concurrent assessment of the short-run and long-run impacts of key macroeconomic factors on vehicle sales in South Africa.

3. Materials and Methods

The analysis makes use of secondary quarterly time-series data to investigate the macroeconomic factors influencing vehicle sales in South Africa. The dataset covers the period from the first quarter of 2000 through to the fourth quarter of 2025, ensuring uniform quarterly observations for all variables included in the study. This timeframe encompasses several economic phases, notably the global financial crisis and the COVID-19 pandemic, thereby facilitating an assessment of both short-term adjustments and long-run relationships in vehicle sales. The data was sourced from the South African Reserve Bank (SARB) and Quantec EasyData, with a summary presented in

Table 1. Summary of variables.

Variable	Abbreviation	Role in Model	Measurement
Vehicle sales	LVS	Dependent variable	Number of vehicles sold
GDP at constant prices	LGDP	Independent variable	Constant prices
Interest rate	IR	Independent variable	Percentages
Inflation rate	LCPI	Independent variable	Index

Source: Author’s own compilation.

. All variables are expressed in natural logarithmic form, except the interest rate, which is expressed in percentages to enable elasticity-based interpretation of estimated coefficients and to help stabilise variability within the time series.

To examine the relationships among the variables, the empirical investigation is conducted in three main phases. The first phase of the analysis will start with a correlation analysis, and the corresponding correlation matrix will be presented. This analysis is intended to provide preliminary insights into the pairwise associations among the variables. Additionally, to ensure the suitability of the variables for time-series analysis, the stationarity properties of all variables will be examined using standard unit root tests. In particular, the augmented Dickey–Fuller (ADF) test and the Phillips–Perron (PP) test will be applied to determine the order of integration of each variable. These tests are conducted to confirm that none of the variables are integrated of order two, I(2), which is a prerequisite for the application of the ARDL bounds testing approach.

This study adapts and extends the empirical demand frameworks commonly employed in the vehicle demand and durable goods literature. In particular, the model specification draws on the approach adopted by McCarthy (1996), who estimated income and price elasticities of vehicle demand, as well as Patra and Rao (2017), who analysed the impact of macroeconomic variables on automobile demand using time-series techniques within an emerging-market context. Although these studies focus on different country settings, their modelling strategies provide a robust and suitable foundation for analysing vehicle sales dynamics in South Africa.

Accordingly, total vehicle sales are modelled as a function of aggregate economic activity, borrowing costs and inflationary pressures. However, GDP is included as a proxy for overall economic activity and income conditions, reflecting employment dynamics and the broader macroeconomic capacity of the economy. Formally, the empirical relationship between vehicle sales and the selected macroeconomic variables is represented by a reduced-form demand function, which is subsequently expressed as both a long-run equilibrium relationship and a corresponding short-run error-correction specification.

As such, the linear relationship between vehicle sales and its macroeconomic determinants is specified as a reduced-form demand function as follows:

$$\ln \left({VS}_t\right)={\propto }_0+{\propto }_1\ln \left({GDP}_t\right)+{\propto }_2{(IR}_t)+{\propto }_3{(CPI}_t)+{\epsilon }_t$$

(1)

where ${VS}_t$ denotes total vehicle sales at time $t$, followed by ${GDP}_t$, which represents gross domestic product, serving as a proxy for overall economic activity and labour market conditions, reflecting aggregate macroeconomic capacity, followed by ${IR}_t$, which is the interest rate, capturing borrowing and financing costs faced by households, followed by ${CPI}_t$, which denotes the consumer price index, reflecting inflationary pressures that affect purchasing power and economic uncertainty. Furthermore, ${\propto }_0$ denotes the intercept term, while the long-run elasticities are denoted by ${\propto }_1$ to ${\propto }_3$ and the error term is denoted by ${\epsilon }_t$.

To examine both short-term adjustments and long-term relationships between vehicle sales and macroeconomic variables, this study adopts the autoregressive distributed lag (ARDL) bounds testing approach to cointegration, as proposed by Pesaran et al. (2001). The ARDL technique is particularly appropriate for the present analysis for several reasons. Firstly, it enables the estimation of long-run relationships among variables that are integrated at different orders, as long as none are integrated beyond the first difference, I(2) (Pesaran et al., 2001). Secondly, the ARDL framework is well-suited to small sample sizes, as asserted by Pesaran et al. (2001), which makes it appropriate for country-specific time-series studies in emerging economies where data constraints are common. Thirdly, ARDL modelling permits the simultaneous estimation of short-run dynamics and long-run equilibrium relationships (Pesaran et al., 2001), rendering it suitable for the analysis of durable goods demand. In the second phase, the presence of a long-term relationship between vehicle sales and the set of explanatory variables is examined using the ARDL bounds testing approach. Where cointegration is confirmed, the final phase involves estimating the long-run parameters alongside the associated short-run error-correction dynamics within the ARDL framework. Accordingly, to assess the existence of a long-run relationship among the variables, an unrestricted autoregressive distributed lag (ARDL) specification is estimated as follows:

$$\begin{gathered} \ln \left({VS}_t\right)={\beta }_0+{\displaystyle \sum _{i=1}^p}{\beta }_1{∆\mathrm{l}\mathrm{n}\mathrm{V}\mathrm{S}}_{t-i}+{\displaystyle \sum _{J=0}^p}{\beta }_2{∆\mathrm{l}\mathrm{n}\mathrm{G}\mathrm{D}\mathrm{P}}_{t-i}+{\displaystyle \sum _{k=0}^p}{\beta }_3{∆\mathrm{I}\mathrm{R}}_{t-i}+{\displaystyle \sum _{l=0}^p}{\beta }_4{∆\mathrm{l}\mathrm{n}\mathrm{C}\mathrm{P}\mathrm{I}}_{t-i} + \\ {\lambda 1\mathrm{l}\mathrm{n}VS}_{t-1}+{\lambda 2\mathrm{l}\mathrm{n}GDP}_{t-1}+{\lambda 3\mathrm{l}\mathrm{n}\mathrm{I}\mathrm{R}}_{t-1}+{\lambda 4lnCPI}_{t-1}+{\mu }_t \end{gathered}$$

(2)

In this framework, $\Delta$ represents the first difference operator, while $p$ denote the optimal lag lengths selected on the basis of information criteria. The coefficients $\lambda 1$ to $\lambda 4$ capture the underlying long-term relationship among the variables, and ${\mu }_t$ denotes the stochastic disturbance term. The bounds test involves estimating an unrestricted ARDL model and evaluating the joint significance of the lagged level variables through an F-test. The calculated F-statistic is then assessed against the critical value bounds reported by Pesaran et al. (2001). Where the F-statistic exceeds the upper critical bound, the null hypothesis of no cointegration is rejected, indicating the existence of a stable long-run equilibrium relationship. Following confirmation of cointegration, short-run dynamics are analysed using an error-correction representation derived from the ARDL model.

The error-correction term reflects the rate at which deviations from long-run equilibrium are corrected after short-run disturbances. A negative and statistically significant coefficient on this term provides evidence of convergence toward the long-run equilibrium path, with its magnitude indicating the speed at which adjustments occur (Engle & Granger, 1987; Enders, 2014). The short-run model allows vehicle sales to respond contemporaneously and with lags to changes in macroeconomic variables, recognising that households may not adjust durable goods purchases instantly in response to economic shocks. This feature is particularly important within the South African context, given high financing costs and uncertainty, which may delay consumption decisions. Once cointegration is confirmed, the short-run dynamics are estimated using an error correction model (ECM) derived from the ARDL specification, where ${ECT}_{t-1}$ denotes the lagged error correction term derived from the long-term equation, followed by $\varphi$, which represents the speed of adjustment coefficient and all other variables as previously defined, as shown in Equation (3).

$$\Delta \ln {VS}_t={\gamma }_0+{\displaystyle \sum _{i=1}^p}{\gamma }_1{∆\mathrm{l}\mathrm{n}\mathrm{V}\mathrm{S}}_{t-i}+{\displaystyle \sum _{J=0}^{p1}}{\gamma }_2{∆\mathrm{l}\mathrm{n}\mathrm{G}\mathrm{D}\mathrm{P}}_{t-i}+{\displaystyle \sum _{k=0}^{p2}}{\gamma }_3{∆\mathrm{I}\mathrm{R}}_{t-i}+{\displaystyle \sum _{l=0}^{p3}}{\gamma }_4{∆\mathrm{l}\mathrm{n}\mathrm{C}\mathrm{P}\mathrm{I}}_{t-i}+{\displaystyle \sum _{m=0}^{p4}}+\varphi {ECT}_{t-1}$$

(3)

To assess the robustness and reliability of the estimated model, a range of diagnostic and stability procedures is implemented. These include tests for serial correlation, heteroskedasticity, and functional form misspecification using the Ramsey rest test. Parameter stability is further evaluated using the cumulative sum (CUSUM) and cumulative sum of squares (CUSUMSQ) tests. Additionally, the Granger causality test is also conducted to determine the direction of causality among the variables. Evidence of stability over the sample period enhances confidence in the estimated relationships and supports the policy relevance of the findings. Overall, the study employs a dynamic and theoretically grounded econometric framework to analyse the determinants of vehicle sales in South Africa. Through the use of reliable data sources, an established ARDL methodology, and carefully selected variables, the approach facilitates a comprehensive evaluation of both short-run dynamics and long-run structural relationships in vehicle demand. The following section presents and discusses the empirical results obtained from this analysis.

4. Results and Discussion

This section reports and interprets the empirical findings of the study, with particular attention given to the dynamic linkages between vehicle sales and selected macroeconomic indicators in South Africa. The analysis follows a systematic progression, starting with the correlation analysis, and thereafter presenting the results of the augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) unit root test. The discussion then proceeds to the ARDL bounds test for cointegration, followed by estimates of long-run relationships and short-run dynamics, a causality test, and concludes with the outcomes of the diagnostic and stability tests. The empirical results are evaluated in relation to established economic theory and relevant empirical studies, with specific emphasis on their implications for vehicle demand behaviour.

4.1. Correlation Analysis

This subsection presents the correlation analysis, where the Pearson correlation matrix was used to examine the pairwise relationships among the variables. This tests provide complementary insights into the relationships between the explanatory variables and ensure the reliability of the estimated model coefficients.

The results of the correlation analysis are presented in

Table 2. Correlation analysis.

Correlation	LVS	LGDP	IR	LCPI
LVS	1.00000
LGDP	0.5531 ***	1.00000
IR	−0.2874 ***	−0.6285 ***	1.00000
LCPI	0.0561	−0.1027	0.2970 ***	1.00000

*** Denotes significance at 1%. Authors own estimation obtained from EViews 13.

. The findings indicate that the correlations between the explanatory variables are generally moderate. In particular, the correlation between LVS and LGDP is positive and moderate, reflecting the procyclical relationship between vehicle sales and economic growth. Additionally, the correlation between GDP and interest rates is negative and moderately strong, suggesting that periods of higher economic activity are associated with tighter monetary conditions. In contrast, the correlation between LVS and interest rates is relatively weak, indicating that the direct pairwise relationship between vehicle sales and borrowing costs is not strong in isolation. Furthermore, a weak positive association is observed between inflation (LCPI) and interest rates (IR), which likely reflects the monetary policy response to rising price levels, where higher inflation is often accompanied by increases in policy interest rates.

4.2. Unit Root Tests

To assess the stationarity characteristics of the variables, a series of unit root tests is performed, specifically the augmented Dickey–Fuller (ADF) test and the Phillips–Perron (PP) test. These tests are applied to confirm that none of the variables are integrated of order two, I(2), a prerequisite for the appropriate application of the ARDL bounds testing methodology. The outcomes of the unit root tests are reported in

Table 3. Unit root test: Augmented Dickey–Fuller (ADF) test.

Augmented Dickey–Fuller (ADF) Test
Variable	Level		First Difference
	Intercept	Intercept & Trend	Intercept	Intercept & Trend	Order of Integration
LVS	0.0655 **	0.2020	0.0000 ***	0.0000 ***	I(0)
LGDP	0.1613	0.7688	0.0000 ***	0.0000 ***	I(1)
IR	0.0399 *	0.0519 *	0.0000 ***	0.0003 ***	I(0)
LCPI	0.0000	0.0000 ***	0.0000 ***	0.0000 ***	I(0)

* Denotes significance at 5%; ** denotes significance at 10%; *** denotes significance at 1%. Source: Author’s own estimation obtained from EViews 13.

and

Table 4. Unit root test: Phillips–Perron (PP) test.

Phillips–Perron (PP) Test
Variable	Level		First Difference
	Intercept	Intercept & Trend	Intercept	Intercept & Trend	Order of Integration
LVS	0.0026 ***	0.0068 ***	0.0000 ***	0.0000 ***	I(0)
LGDP	0.0039 ***	0.7095	0.0000 ***	0.0000 ***	I(0)
IR	0.1774	0.38873	0.0001 ***	0.0007 ***	I(1)
LCPI	0.0000 ***	0.0000 ***	0.0001 ***	0.0001 ***	I(0)

*** Denotes significance at 1%. Source: Author’s own estimation obtained from EViews 13.

, respectively.

The outcomes of the ADF and PP, as shown by Table 2 and Table 3, reveal that all variables attain stationarity either in levels I(0) or after first differencing I(1). According to the ADF results, LVS, IR, and LCPI are stationary at level I(0), whereas LGDP achieves stationarity after first differencing I(1). By comparison, the PP test indicates that LVS, LGDP, and LCPI are stationary at levels I(0), while IR is stationary at I(1). Overall, the unit root results indicate that all variables are either I(0) or I(1), satisfying the requirements for the ARDL bounds testing approach. This satisfies the necessary conditions for applying the ARDL bounds testing procedure to assess cointegration.

4.3. Bound Test for Cointegration

After establishing that the variables exhibit a mixed order of integration, the ARDL bounds testing procedure is applied to assess the presence of a long-run cointegrating relationship between vehicle sales and its macroeconomic determinants. Model selection is guided by the Akaike information criterion (AIC), which is commonly favoured for identifying parsimonious specifications in small and medium-sized samples. From a theoretical standpoint, the bounds test examines whether the combined effects of income, inflation and interest rates give rise to a stable long-run equilibrium relationship in vehicle demand. Income is proxied by GDP, capturing overall economic activity and purchasing conditions. The results of the ARDL bounds test, summarised in

Table 5. ARDL bounds test.

Test Statistic	Value	Significance	I(0)	I(1)
F-statistic	16.80302	10%	2.37	3.2
K	4	5%	2.79	3.67
		1%	3.65	4.66

Source: Author’s own estimation obtained from EViews 13.

, indicate whether cointegration is present among the variables.

The ARDL bounds test results provide evidence of a long-run cointegrating relationship among the variables. Under the null hypothesis of the bounds testing procedure, it is assumed that no long-run relationship exists between vehicle sales and the selected macroeconomic determinants. As shown in Table 4, the calculated F-statistic of 16.80302 exceeds both the lower and upper critical bounds at standard significance levels. As a result, the null hypothesis of no cointegration is rejected, indicating the presence of a stable long-run equilibrium relationship among the variables. In light of this evidence of cointegration, the analysis advances to the estimation and interpretation of the long-run coefficients in the following section.

4.4. Long-Term Results

Having established the existence of a long-run cointegrating relationship among vehicle sales and the selected macroeconomic variables, this section presents and discusses the estimated long-run coefficients derived from the ARDL model. The long-run estimates capture the equilibrium relationships between vehicle sales and its key determinants, abstracting from short-run fluctuations and transitory shocks. These coefficients provide insight into the magnitude and direction of the long-term effects of income, interest rates and inflation on vehicle demand in South Africa. The long-term equation is formulated as follows:

$$LVS=1.1652+{0.7363}^{***}LGDP-{0.0893}^{**}IR-{0.0658}^{***}LCI$$

(4)

The results indicate a positive and statistically significant relationship between real GDP and vehicle sales, where a 1% increase in real GDP results in approximately a 0.7% increase in vehicle sales in the long run. Within the South African context, this result suggests that sustained economic growth plays a critical role in stimulating vehicle purchases, reflecting improvements in employment prospects and consumer confidence. This finding aligns with McCarthy (1996), who found that vehicle demand exhibits strong income elasticity, particularly over the long term. Similar evidence is reported by Patra and Rao (2017) and Mothibi et al. (2025), who also found a positive relationship between vehicle sales and economic growth, confirming that economic growth is a key driver of automobile demand in emerging markets.

The findings of the study also reveal a negative and significant relationship between interest rates and vehicle sales, where a 1% increase in interest rate results in a 0.08% decrease in vehicle sales. This result is consistent with economic theory, which suggests that increases in borrowing costs reduce consumer demand for high-value durable goods such as motor vehicles. Higher interest rates increase the cost of vehicle financing, thereby constraining affordability and discouraging households from undertaking large credit-financed purchases. This finding aligns with empirical evidence from Copeland et al. (2019) and Chisasa and Dlamini (2013), who demonstrate that rising interest rates exert a contractionary effect on vehicle demand through higher financing costs. Within the South African context, this effect is particularly pronounced due to high levels of household indebtedness and sensitivity to credit conditions, which amplify the impact of interest rate increases on consumption decisions.

The findings of the study also confirm a negative and statistically significant relationship between inflation and vehicle sales, where a 1% increase in inflation results in a 0.06% decrease in vehicle sales in the long-term. The findings suggest that inflation exerts a contractionary effect on vehicle sales, although the magnitude is smaller. Within the South African context, persistent inflationary pressures reduce real disposable income and heighten economic uncertainty, prompting households to postpone or cancel vehicle purchases. This finding is consistent with recent empirical studies by Pehlivanoğlu and Riyanti (2018) and Muhammad et al. (2012), who also found this negative relationship, highlighting that inflation weakens durable goods demand by eroding real purchasing power, thereby increasing economic uncertainty.

Overall, the long-run results highlight the dominant role of broader macroeconomic conditions in shaping vehicle demand in South Africa. Economic growth emerges as the primary driver of vehicle sales, while other variables such as interest rates and inflation exert a negative and statistically significant effect on vehicle demand. These findings indicate that borrowing costs and price pressures impose significant constraints on vehicle purchases. The results suggest that structural constraints related to income, credit access, and cost pressures influence how households translate economic improvements into durable goods purchases. Taken together, the findings point to a demand environment in which economic growth stimulates vehicle sales but is simultaneously moderated by higher financing costs and inflationary pressures that reduce affordability and constrain consumption.

4.5. Error Correction Model and Short-Term Results

After establishing the presence of a long-run cointegrating relationship among the variables, this section examines the short-term dynamics estimated using the error correction model (ECM). The ECM captures both the immediate impact of variations in the explanatory variables on vehicle sales and the rate at which short-run deviations from the long-run equilibrium are corrected. This approach enables an evaluation of the speed with which vehicle sales respond to macroeconomic shocks while ensuring convergence toward the long-run relationship identified through the ARDL framework. The results of the ECM estimation are reported in

Table 6. Error correction model and short-term results.

Variable	Coefficient	Std. Error	T-Statistic	Probability
D(LVS(-1))	−0.568777	0.079764	−7.130780	0.0000 ***
D(LVS(-2))	−0.349329	0.100307	−3.482598	0.0008 ***
D(LVS(-3))	−0.528635	0.086585	−6.105361	0.0000 ***
D(LGDP)	5.572856	0.290451	19.18693	0.0000 ***
D(LGDP(-1))	3.877594	0.612989	6.325720	0.0000 ***
D(LGDP(-2))	2.645583	0.712260	3.714348	0.0004 ***
D(LGDP(-3))	3.327634	0.589084	5.648825	0.0000 ***
D(LGDP(-4))	0.665153	0.282649	2.353281	0.0209 *
D(LCPI)	−0.011393	0.002494	−4.568900	0.0000 ***
D(IR)	−0.027516	0.010400	−2.645709	0.0097 ***
CointEq(-1) *	−0.206585	0.022020	−9.381674	0.0000 ***

* Denotes significance at 5%; *** denotes significance at 1%. Source: Author’s own estimation obtained from EViews 13.

.

The short-run results from the error correction model indicate that vehicle sales in South Africa respond immediately to changes in macroeconomic conditions. In the short-term, the results indicate a positive and statistically significant relationship between real GDP and vehicle sales, consistent with the demand theory. This finding is supported by recent empirical evidence showing that GDP growth significantly drives automobile demand in emerging economies through its effect on purchasing power and consumption dynamics (Johnson & Gopakumar, 2025; Johan, 2019). Within the South African context, this reflects the sensitivity of vehicle purchases to short-term improvements in economic activity, where increases in output quickly translate into higher demand. The results also show a negative and statistically significant relationship between interest rates and vehicle sales in the short run, suggesting that increases in borrowing costs immediately constrain vehicle demand. This contrasts with the long-run findings, where interest rates were found to be insignificant, indicating that their effects are largely short-lived and diminish over time. In the short term, higher interest rates reduce consumers’ ability to access credit and finance vehicle purchases, leading to a contraction in demand for high-value durable goods such as motor vehicles. This finding is consistent with Copeland et al. (2019), who found a negative relationship between interest rates and the demand for vehicles, highlighting the rise in borrowing costs and reduction in consumer demand.

Aligned with the long-run results, inflation also exhibits a negative relationship with vehicle sales, consistent with economic theory, which suggests that rising prices erode real purchasing power and discourage consumption (Mohr, 2004). This is supported by recent empirical findings of Sin et al. (2024), who also found the negative relationship between inflation and vehicle sales.

The error correction term is negative and significant, confirming the presence of a stable adjustment mechanism toward the long-run equilibrium. The magnitude of the coefficient indicates that approximately 20% of deviations from the long-run equilibrium are corrected within one quarter, implying gradual but persistent convergence following short-term shocks. Together, these results suggest that vehicle demand in South Africa is largely driven by broader economic activity, while short-term fluctuations are shaped by affordability and macroeconomic constraints. Over time, these constraints are absorbed as demand gradually adjusts, returning to the underlying long-run relationship defined by economic fundamentals.

4.6. Granger Causality Results

To further explore the dynamic interactions between vehicle sales and the selected macroeconomic variables, the study applies the Granger causality test. This framework evaluates the direction of predictive linkages between variables by examining whether historical values of one variable improve the forecasting ability of another. In contrast to cointegration and ARDL techniques, which focus on long-run and short-run relationships, the Granger causality approach offers additional insights into the direction of influence among the variables, thereby enhancing the interpretation of the empirical findings. The results of the Granger causality analysis are presented in

Table 7. Granger causality test.

Null Hypothesis	Probability	Outcome
LGDP does not Granger-cause LVS	0.0111 *	Bi-directional causality
LVS does not Granger-cause LGDP	0.0564 **
IR does not Granger-cause LVS	0.0176 *	Unidirectional
LVS does not Granger-cause IR	0.9807
LCPI does not Granger-cause LVS	0.1331	No causality
LVS does not Granger-cause LCPI	0.5595
IR does not Granger-cause LGDP	0.0040 *	Unidirectional causality
LGDP does not Granger-cause IR	0.1447
LCPI does not Granger-cause LGDP	0.7690	No causality
LGDP does not Granger-cause LCPI	0.6694
LCPI does not Granger-cause IR	0.0677 *	Bi-directional causality
IR does not Granger-cause LCP	0.0552 *

* Denotes significance at 5%; ** denotes significance at 10%. Authors own estimation obtained from EViews 13.

.

The Granger causality results indicate a complex dynamic relationship between vehicle sales and key macroeconomic variables. The bidirectional causality between GDP and vehicle sales suggests a strong feedback mechanism, where economic growth stimulates vehicle demand while expansion in the automotive sector contributes to broader economic activity. This finding is consistent with the literature, which views durable goods consumption, particularly automobiles, as both a driver and indicator of economic growth (Gavazza & Lanteri, 2021; Mothibi et al., 2025).

In contrast, the results show predominantly unidirectional causality from interest rates and inflation to vehicle sales, indicating that macro-financial conditions influence automobile demand without a reverse feedback effect. This suggests that monetary policy variables act as external constraints on vehicle consumption, primarily through borrowing costs and purchasing power channels. Such findings are consistent with theoretical and empirical literature, which shows that inflation reduces real purchasing power while interest rates influence durable goods’ demand through borrowing costs, with these effects being particularly pronounced in the short run (Mankiw, 2019; Copeland et al., 2019). The absence of causality between certain variables, such as inflation and GDP, further suggests that the transmission of macroeconomic shocks to vehicle sales may be indirect and mediated through other channels, reinforcing the importance of considering both direct and dynamic effects in understanding vehicle demand behaviour.

4.7. Diagnostics and Stability Test Results

To evaluate the robustness and reliability of the estimated ARDL and error-correction models, a set of diagnostic and stability checks is performed. These tests examine whether key model assumptions, including the absence of serial correlation, homoscedastic residuals, and appropriate functional form specification, are satisfied. Furthermore, parameter stability tests are applied to assess whether the estimated coefficients remain stable throughout the sample period, thereby supporting the consistency of the estimated relationships and the policy relevance of the empirical findings. The outcomes of the diagnostic and stability tests are reported in

Table 8. Diagnostics and stability results.

Test	Probability Value	Conclusion
Jarque–Bera normality test	0.6572	Residuals are normally distributed
Breusch–Pagan–Godfrey heteroskedasticity test	0.8734	No heteroscedasticity
Breusch–Godfrey serial correlation LM test	0.1499	No serial correlation
Ramsey RESET test	0.6406	The model is correctly specified
Cumulated sum of square test (CUSUM)		The model is stable
Cumulated sum of square test (CUSUMQ)		The model is stable

Source: Author’s own compilation estimation obtained from EViews 13.

.

The outcomes of the diagnostic and stability tests suggest that the estimated ARDL-ECM is correctly specified and empirically sound. The Jarque–Bera normality test does not reject the null hypothesis of normally distributed residuals, indicating that the error terms conform to the normality assumption. The Breusch–Pagan–Godfrey test provides no evidence of heteroskedasticity, the Breusch–Godfrey LM test for serial correlation confirms the absence of autocorrelation in the residuals, while the Ramsey RESET test confirms that the model is correctly specified. In addition, the CUSUM and CUSUMQ stability tests show that the cumulative sum of recursive residuals remains within the 5 per cent critical bounds over the entire sample period, supporting the stability of the estimated coefficients. Taken together, these findings indicate that the model is not subject to significant econometric violations and that the estimated parameters remain stable over time, thereby strengthening the robustness and credibility of the empirical results and their appropriateness for inference and policy evaluation.

5. Conclusions and Policy Recommendations

This study provides evidence that vehicle sales in South Africa are best understood not solely as a function of macroeconomic output, but as an outcome shaped by macro-financial conditions, household balance-sheet constraints, and monetary policy transmission. The results show that the automotive sector is closely linked to broader demand and credit cycles, responding to variations in interest rates, inflation, and overall economic activity rather than operating as an autonomous growth driver. This reinforces the classification of vehicles as discretionary, credit-dependent durable goods that are particularly sensitive to changes in affordability and macroeconomic expectations.

A key insight from the analysis is the divergence between short-run and long-run dynamics. In the short run, higher interest rates, inflation, and macroeconomic and affordability pressures significantly constrain vehicle demand, reflecting immediate affordability and credit effects. However, these constraints persist into the long run, as both interest rates and inflation continue to exert a negative and statistically significant influence on vehicle demand, and demand gradually adjusts toward broader economic fundamentals. This indicates that while vehicle demand is highly responsive to macro-financial conditions in the short term, it is ultimately anchored by long-run economic activity. In an environment characterised by high household indebtedness and rising living costs, vehicle sales remain closely tied to credit conditions and household financial capacity in South Africa.

From a policy perspective, the findings emphasise the importance of maintaining macroeconomic stability to support sustainable vehicle demand. In particular, price stability and predictable monetary policy are critical for preserving purchasing power and reducing uncertainty. In addition, policies aimed at improving access to affordable credit, especially for lower-income households, could enhance participation in the vehicle market. Strengthening household financial resilience through income support measures and reduced debt vulnerability would further contribute to demand stability, while improved policy communication can help sustain consumer confidence and mitigate precautionary behaviour.

Despite these contributions, certain limitations should be acknowledged. The analysis is constrained by the availability of consistent quarterly data, which restricts the inclusion of variables such as detailed credit conditions and consumer sentiment. Furthermore, the use of aggregate data may mask important heterogeneity across income groups. Future research could address these limitations by incorporating more disaggregated datasets and behavioural indicators, particularly at the household level, to better capture the underlying drivers of vehicle demand in South Africa.