“Intention to provide ridesharing services: Determinants from the perspective of driver-partners in a gig economy”

Research on ridesharing platforms under the gig economy has focused much on the incentives and barriers of users, leaving many gaps in understanding drivers’ intention to provide ridesharing services. This paper aims to explore, from the perspective of driver-partners, motives that encourage them to continue being gig workers. Data for the study are based on a cross-sectional survey of ridesharing drivers in three metropolitan areas in three regions (North, Central, and South) of Vietnam, conducted from June to July 2022. The paper regresses behavioral intention to continue being a gig driver on their demographic characteristics and self-estimation of economic benefit, time preference, and enjoyment of being a gig driver via ordered probit models. For all three regions, the result suggests that economic benefit, time preference, and enjoyment are good predictors of drivers’ intention to provide the services. Specifically, the probability of remaining in gig work among drivers decreases with their educational and economic status. Higher economic benefit does not predict a higher intention of drivers to stay longer in gig work. Similarly, those with higher levels of enjoyment of traveling and vehicles have a lower intention to remain in this sphere. In the North, the interaction terms between time preference and enjoyment level are significant, suggesting that the effect of enjoyment levels becomes less damaging with an increase in time preference. In other words, time preference is vital in keeping gig drivers in this type of work.


INTRODUCTION
Ridesharing platforms or ridesharing services, as part of the gig economy, provide a promising mode of transportation that significantly contributes to countries' economic and sustainable development.This new business mode creates efficiency and convenience for both users and providers, including technology companies and their partners, via smartphone applications.On one hand, technological companies do not have to own devices and vehicles, and partners can work in suitable and convenient ways based on time preferences, economic returns, or just enjoyment.But on the other hand, customers are more likely to use ride-share if they find the service useful and environmentally protective, and vice versa if they think the service is risky (Wang et al., 2020).
Ridesharing's advantages include reduced travel costs, increased trip convenience, better available seat capacity, reduced use of vehicles for personal trips, and significantly reduced gas emissions.Overall, it provides a platform for many parties to connect despite their different oc-cupations, backgrounds, and personal characteristics.For example, using archival data and the difference-in-differences method, Li et al. (2022) empirically found that Uber's ridesharing service contributed to the decline of traffic jams in busy municipalities.In addition, with a time series analysis, Khan et al. (2022) showed a significant negative relationship between the presence of ride-share services and the number of total crashes and injuries, except for severe injuries in a week.It is also found that shared autonomous vehicles from ridesharing transportation services are likely to decrease traffic crashes when it comes to drivers' mistakes.
Several studies have focused on customers' intentions to adopt ridesharing services, including price, safety, convenience, trust, and satisfaction.These factors significantly affect intentions to use this service.Social influence also directly affects customers' behavioral intentions.For instance, Giang et al. (2017) analyzed data from 328 users in Vietnam and reported that customers have positive attitudes toward ride-share services because they find them useful and easy to use.The results also revealed that attitudes, subjective norms, and cognitive-behavioral control are essential predictors of customers' intention to use carpooling apps.For driver-partners, behavioral intention to stay on such platforms may be affected by socio-demographic factors such as education, age, and family economic status.In general, much attention has been paid to exploring multiple factors associated with the behavior of ride-share travelers (Lu & Wang, 2020).
Although there have been many investigations of customers' points of view on intention to use a ridesharing platform, the opinion of driver-partners from the sharing companies has not been widely examined.In addition to socio-demographic characteristics associated with the motive to use ridesharing services, some intrinsic and extrinsic elements, including economic benefit, time preference, and enjoyment of drivers, play a vital role.To explore these elements in the relationship with regard to driverpartners' intention to remain with the platform, it is also necessary to look at the interactions among these factors.

LITERATURE REVIEW AND HYPOTHESES
It is critical to realize the trend of benefits from ridesharing to the partners who play the role of business groups or workers.Usually, any services provided by state or private enterprises in transportation will benefit one or many parties.From the view of ridesharing, economic, social, and environmental benefits, especially personal preferences for partners, are positively recognized from this business model's emergence.First, ridesharing was initially interpreted as carpooling, where the driver and rider share the same journey.Gradually, with the support of technology and the introduction of online applications/platforms, matching drivers and riders has become more accessible.As a result, ridesharing via online platforms has become a popular transport service in many metropolitan areas and is considered an appealing alternative mode of transportation to traditional taxis.Although ridesharing may affect travel demand and urban areas unfavorably, it has been found to contribute to lower travel costs, higher travel flexibility, decreased air pollution, and more sustainable urban environments (Xiao & Goulias, 2022).2019) concluded that users' attitudes were strongly related to their behavioral intention to adopt ridesharing services.Specifically, for share-riders, prejudiced attitudes toward riders are negative predictors of their dissatisfaction with the sharing choice and those of their shared TNC trips.In addition, these attitudes are inversely correlated with sustained and frequent ridesharing.For non-share riders, rider-to-rider biased attitudes negatively affect their readiness to use ridesharing.
In line with the above view, Goel and Haldar (2020) reported that customers volunteer to pay more for sustainable goods because they care more about environmental issues (Chaudhry et al., 2018).In addition, the care for the environment may lead to commuters' consideration of using ridesharing platforms (Wang et al., 2019).However, other authors found weak or insignificant associations between the two ( From travelers' perspective, Yang et al. (2020) showed that the mismatch among travelers in terms of travel distance and start location will heavily affect the management and operation systems of the ridesharing platforms.
Zhang and Leiming (2020) tested the theory of planned behavior on Chinese college students.It was found that subjective norms and perceived behavioral control are direct and positive predictors of their motive to adopt ridesharing.At the same time, concerns for environmental issues are indirectly correlated with the motive.The mechanism for this indirect correlation is that environmental awareness affects the students' motive to use ridesharing via subjective norms and perceived behavioral control.
Most research has identified environmental awareness, ease of use, personal innovation, financial benefit, social influence, and perceived usefulness as predictors of the motive to adopt ridesharing (Ashrafi et al., 2020;Goel & Haldar, 2020;Litman, 2000;Akbari et al., 2021).However, these correlatives are examined from the viewpoint of customers rather than driver-partners.Therefore, this paper aims to determine the relationship between preferential factors and the intention to adopt ridesharing of driver-partners in the sharing platform.Consequently, the following hypotheses are raised: H1: There is a positive relationship between economic benefit and the intention of driver-partners to adopt ridesharing on a sharing platform.
H2: There is a positive relationship between time preference and the intention of driver-partners to adopt ridesharing on a sharing platform.
H3: There is a positive relationship between enjoyment and the intention of driver-partners to adopt ridesharing on a sharing platform.
H4: There is a positive relationship between interaction among preferential factors and the intention of driver-partners to adopt ridesharing on a sharing platform.

DATA DESCRIPTION AND METHODOLOGY
This paper used survey data from residents in three large metropolitan cities in Vietnam to find socio-demographic and preferential determinants, including economic benefit, time management benefit, and enjoyment in the driver-partners' intention to join the gig economy.Data were collected from 996 respondents from several motorbike and/or carsharing companies in Vietnam, including Grab, Gojek, Be, Shopee, Baemin, and Ahamove.The survey was carried out online and offline from June to July 2022 in Hanoi (North), Danang (Central), Hochiminh City (South), and other nearby areas.For Grab, a link to the questionnaire was sent to drivers to collect data.For other companies, drivers were interviewed faceto-face using questionnaire printouts.
The questionnaire link was sent to respondents online since many interviewees could access and answer the survey simultaneously.
Given independent variables such as time preference, economic benefit, or enjoyment, the conditional distribution of the intention to adopt the ridesharing platform can be derived with the standard normal assumption for the error term.Each response probability can be measured as: The coefficients α and β can be estimated by maximum likelihood.For each i, the log-likelihood function is For the ordered probit type, the marginal effect of every variable or interaction of variables can also be measured: The behavioral intention to adopt ridesharing is the dependent variable of the unstandardized ordered probit model.The dependent variable ranges from 1 to 5, where 1 means driver-partners will not continue, and 5 shows that they will remain with this sharing service for 5 years more.In all ordered probit models, socio-demographic variables and self-estimation of economic benefit, time preference, and enjoyment of partners are used as the explanatory factors (Table 3).

RESULTS
Table 4 presents the results of the ordered probit models.Model 1a presents the results for all three regions, 1b for the North and 1c for the South.The number of observations in the Central region was limited, so the regression for this area was not estimated.The results show that most explanatory variables were statistically significant in all models.Socio-demographic variables significantly affected drivers' behavioral intention to adopt ridesharing in the gig economy.Firstly, age had a positive relationship with the ridesharing participation intention of driver-partners within 5 years.The driver-partners' intention to join the gig economy increased with age, but then at a certain age, it began to decrease.In other words, the relationship between age and intention to join the gig economy is quadratic.The maximum ages for driver-partners to leave the gig economy are 48.26,42.56, and 48.82 years nationwide, in the North and the South, respectively.These ages seem logical since, at these ages, drivers need to find other jobs as their health begins to worsen over time.
The models' findings show that driver-partners' educational level was negatively associated with their intention to stay in the work of ridesharing.In addition, drivers' experiences revealed other information about their behavioral intention to provide ridesharing services.For the vehicle variable, the result indicates that car driver-partners tend to stay longer in this work compared to the motorbike group.Note: *, ** and *** denote significant levels of 10%, 5% and 1% in consequence.
The current income from being partners in the gig economy and economic conditions also provides some good information.With significant coefficients, the results indicate that persons with higher current income working as partners with sharing platform companies are more likely to join this work in all three regions and the North.Interestingly, the economic benefits of driver-partners have a negative relationship with the intention to adopt ridesharing in models 1a and 1b, and it has a significant relationship with the dependent variable.
In Figure 1, in all three regions, the marginal effect of enjoyment levels varies and depends on economic preferences to the probability of intention to adopt gig work.In detail, the likelihood of staying in gig work is strongly affected by workers expressing either of two enjoyment levels, including "Agree" and "Strongly agree."This result was similar in the North and the South (Figure 2 and Figure 3), where the marginal effect of enjoyment levels also varied, depending on the respondent's level of agreement with economic preferences for gig work.

Figure 1.
Marginal effects of interaction between economic benefit and enjoyment on the probability of intention to join the gig economy (all three regions) 0 Probability (Intention to adopt)

Strongly agree=5
Predictive Margins with 95% CIs Generally, the hypotheses H1 and H2 are accepted.It means that economic benefit and time preference have positive impacts on the intention of driver-partners to stay more in the new economy model.Similarly, the hypotheses H3 and H4, the interactions among these variables, are also accepted but in some cases only.

DISCUSSION
The coefficients for educational levels in all models are statistically significant, suggesting that the higher the educational level driver-partners have, the less likely they are to stay in this kind of work.
As expected, drivers with more experience will likely stay longer in the gig economy.Experienced driver-partners may indeed operate their vehicles and services better than other partners; consequently, they will stay in this work longer.Perhaps they are familiar with the job, routes, and regular passengers; consequently, they are more willing to stay in the ridesharing market.

Strongly agree=5
Predictive Margins with 95% CIs In developed countries, ridesharing platforms provide mainly automobiles, but in developing ones, motorbikes are more prevalent in this gig economy.
The results show that car driver-partners stay longer in this work than motorbike partners.This result recommends a higher demand for car services in the gig economy, which may reflect the better economic situation in developing countries.In addition, working with automobiles does not include severe health hazards as those experienced by motorbike driver-partners, such as noise, dust, and accidents.Another aspect is that the benefit and promotion policies of ridesharing companies are better for car drivers; thus, more driver-partners want to remain in this work.
Driver-partners' current income and economic status may affect workers' intention to stay longer or leave the gig environment.As expected, the results show that respondents with better financial conditions exhibit reduced intentions to stay in the gig economy.Some recent studies have postulated that ridesharing mobility is cheaper than non-sharing mobility and that financial reward is an essential factor in the use of sharing services, to the extent that sharing can be an excellent alternative to owning a vehicle (Goel & Haldar, 2020).Being a partner in sharing platform companies indeed improves poor individuals' lives.
When this new type of work came to developing nations like Vietnam, it helped many poor individuals earn money for themselves and their families.This result supports the findings of d'Orey and Ferreira (2015) and Li et al. (2018).They noted that ridesharing facilitates income redistribution by helping poor people with access to affordable transport means (as customers) or secondary jobs (as drivers).This finding is a revealing base for authorities and policymakers to understand the working class characteristics of this platform in the economic sector.
The most critical variables worth exploring are gig workers' economic and time preferences, in addition to job enjoyment.Interestingly, higher economic benefits to workers and their families do not guarantee an increased intention to be driver-partners in the future.Economic preference is only one critical factor that motivates gig workers to stay longer.Similarly, respondents with higher levels of enjoyment of traveling and vehicles have a lower intention to remain on this platform.In the ordered probit models in the present study, interaction terms between economic preference, time preference, and enjoyment for driver-partners were estimated to investigate whether the effect of these explanatory variables on the intention to adopt the gig economy changed depending on their values.For example, in the North, Figure 4 shows the marginal effects of the interaction between time preference and enjoyment on the probability of joining the gig economy.The negative interaction term means that time preference of driver-partners strengthens the negative effect of respondents' enjoyment level on intention to join the gig economy.In other words, the impact of enjoyment levels will be less negative with an increase in time preference.Thus, time preference is vital in keeping driver-partners operating in the ridesharing gig economy.

CONCLUSION
Through investigating determinants of behavioral intention to adopt the ride-hailing service, this paper reveals that joining the ride-hailing platform as a driver-partner is a good option for those with low economic status and educational levels in Vietnam.Thus, the factor of economic benefits will determine the intention of these drivers to stay more in the gig economy.In addition, time benefits also play an important role when the driver-partners believe in their answers from the survey.The paper adds to the current research by providing more profound evidence on determinants of drivers' intention to stay in the gig economy.To some extent, income may as well be a good tool to keep the driver-partners working longer in the gig economy.
The findings of the paper help explain the boom of ride-hailing services in Vietnam from the viewpoint of service providers and show factors that policymakers may consider in forecasting the trend of this labour market.Future studies may focus more on drivers' time preferences and enjoyment to better capture factors that strongly determine driver-partners' intention to join and stay in the ridesharing economy.

Figure 2 .Figure 4
Figure 2. Marginal effects of interaction between economic benefit and enjoyment on the probability of intention to join the gig economy (in the North)

Figure 3 .Figure 4 .
Figure 3. Marginal effects of interaction between economic benefit and enjoyment on the probability of intention to join the gig economy (in the South)

Table 2
time or full-time job?" Accordingly, 66.27% of the respondents reported that they would consider this work a full-time job, while 33.73% considered it a part-time job.

Table 2 .
(Wooldridge, 2001)tics for ridesharing adoptionTo understand the effect of social, demographic, and preferential variables on driver-partners' intention to adopt ridesharing, y can be an ordered choice taking the values {0, 1, 2,..., J} for some known integer J.The ordered probit model for y (intention to adopt ridesharing platform), conditional on independent factors x, can be derived from a latent factor model(Wooldridge, 2001).A latent factor y * is identified as in which β is Kx1.Let α 1 < α 2 < ...< α J be threshold parameters and report

Table 3 .
Main independent and control variables AGE The age of respondents in the year of the survey.The model includes AGE 2 to capture the quadratic relationship between age and intention to adopt ridesharing if any EDU Educational level of respondents in which 1 equals primary and 8 equals the graduate level Participating in the gig economy benefits me, my family, and society/economy.Likert scale from 1 to 5 in which 1 is the lowest level of agreement and 5 is the highest level of agreement TIMEBE Being a driver-partner in the gig economy helps me manage my time better, measured using a Likert scale from 1 to 5 in which 1 is the lowest level of agreement while 5 is the highest level of agreement ENJOY I am a driver-partner in the gig economy to satisfy my passion for traveling and moving vehicles.Measured using a Likert scale from 1 to 5 in which 1 is the lowest level of agreement while 5 is the highest level of agreement

Table 4 .
Ordered probit estimation results Kim et al. (2017)hs (2017)r studies with qualitative approaches to determine the latent and factual factors that motivate gig workers.The findings of this paper do not concur with those reported byWang et al. (2019)orHwang and Griffiths (2017).Specifically, Amirkiaee and Evangelopoulos (2018) showed that in case of high anxiety with transportation, the trust in ridesharing stakeholders, together with economic and time benefits, will encourage people to join ridesharing.Kim et al. (2017)identified motivational variables that frame riders' perceptions of and attitudes toward ridesharing services and proposed a research model including motivational factors and those from the technology acceptance model to explain the adoption of car sharing.These authors found that perceived reliability, compatibility, and enjoyment of car-sharing services, as well as users' innovative tendencies, were positively associated with usage intention.