Are Telecommunications and Travel Substitutes or Complements? An Empirical Analysis of a Developing City in Nigeria

Findings This study examines the relationship between household travel and telecommunication use with data collected from 498 participants in Akure metropolis, Nigeria and finds that a positive relationship exists between telecommunication and travel. While the substitution effect of telecommunication on the trip was not significant in the study, the result of the multinomial logistic regression revealed telecommunications complements travel more than substitute.


QUESTIONS
The possible relationships between information and communications technology (ICT) (or telecommunication) and physical travel include substitution, complementarity or modification, generation, and neutrality (Mokhtarian 1990; Mokhtarian and Salomon 2002;Nobis and Lenz 2003). Complementarity occurs where there is a modification or change of spatial and temporal characteristics of existing travel patterns by the use of telecommunications, while substitution refers to the number of trips being replaced by ICT (Senbil and Kitamura 2004;Oyesiku 1996). Unlike previous studies (Matous 2017; Olawole 2013) that focused on mobile phones, this study examines the relationship between household travel and telecommunication, which extends beyond call linkages to virtual activities such as e-shopping, e-banking, telework, to mention just a few. The research questions include:

Research Hypothesis
H 0 -There is no significant relationship between telecommunication usage and travel.
H 1 -There is a significant relationship between telecommunication usage and travel.
• What are the nature and characteristics of telecommunication in the study area?
• Do telecommunications substitute for physical travel?
• Is there clear evidence of the relationship between telecommunication and travel in the study?

METHODS
This study uses primary data collected from a survey, which was conducted using trained research assistants and a questionnaire. A probability sampling technique was adopted, and a multi-stage sampling procedure was used. The first stage involved the stratification of residential areas and the selection of political wards. Of 11 political wards, 6 were sampled. Random sampling was adopted in selecting 38 registered streets from a total of 139, representing 20% of the streets in the selected wards.
Further information from the Area Town Planning Area office revealed that there are 5,123 registered buildings in the selected wards, of which 10% were systematically selected. Bookwalter, Fuller, and Dalenberg (2006) and Owoeye, Fadare, and Ojekunle (2018) justified surveying only the household head within the family on the basis that the response of household heads is mainly determined by factors shared by the entire household and not on those experienced primarily by the head. Where the respondent was not available, the next building was selected for the sample. A household head not below the age of 18 years (adult) on the first floor of each selected building was sampled. Based on this, a total of 512 respondents were surveyed. However, a total of 498 questionnaires (97%) were found usable for analysis.
The questionnaire designed for the research had three main sections. Section 1 comprised information relating to the socio-economic characteristics of the respondents. They included gender, age, education, income, marital status, employment status, occupation, and car ownership. Section 2 focuses on the travel characteristics of the people, information such as trip frequency (Average number of round trips), trip purpose (work, shopping, health, recreation), transport mode and travel cost were acquired. The last section comprised information on the telecommunication usage of respondents, questions relating to the type and number of telecommunication own (GSM or PC), the average number of trips complemented/substituted by telecommunication, trip activities/complemented or substituted by telecommunication among others.
Information on the socio-economic characteristics of respondents revealed that the majority were male (54.1%), married (72.8%), had tertiary education (72.1%), 54.8% earned between ₦40,000 -₦60,000. Hence the majority earned above the federal government minimum income wage of ₦20,000. Analysis of trip frequency showed that 50.8% made an average of 4 trips daily trips. Work trips comprised the majority (82.8%) of trips. A smaller majority (50.7%) made use of public transport. The telecommunication usage of households revealed that 98.2% owned a mobile phone while (55.2%) did not make use of virtual activities/platforms such as e-banking, email, and eshopping. Facebook and WhatsApp were the most used smartphone applications, with 40.3% and 28.2% of the respondents, respectively. The socio-economic characteristics of respondents revealed that the majority were educated and employed with the government as civil servants. The number of government-employed respondents is not surprising because an individual's highest level of education determines the kind of occupation such a person can engage in and the income level (Ahn 2001;Badiora 2012;Stead and Marshall 2001).
The result of the average number of self-reported complemented and substituted trips by different ICT means is explained in Table 2. It is evident from the study that the use of a mobile phone (Global System of Mobile Communication (GSM)) for making calls was more predominant in the study area as a larger volume of trips were influenced through that platform, unlike emails, e-shopping, e-banking and social media platforms.

FINDINGS
The result on trip frequency corroborates the findings of (Hine, Barnejee, and Kashyap 2012), which asserted that urban form influences one's travel behaviour, as such, urban dwellers generate more trips than those in remote  locations due to factors ranging from their socio-economic characteristics, access to public transport or private cars among others. As such, the large number of trips generated by respondents may be a function of some of these highlighted factors since most of them had access to either a private car or public transport.
Concerning telecommunication usage, the use of virtual platforms such as ebanking, e-shopping, telework, and e-business to replace physical movement were not significant as respondents rely more on telecommunications for a reorganization of activities or modifications with respect to time and space. It, thus, stipulates that a larger percentage of the population still embarks on physical trips against virtual activities. Overall, 53.4% of trips were complemented by telecommunication. A close reflection showed that 58.5% of total shopping trips were replaced by telecommunication, thus, in contrast Are Telecommunications and Travel Substitutes or Complements? An Empirical Analysis of a Developing City in Nigeria Findings with the findings of Farag et al. (2007) and Douma et al. (2004), which concluded that e-shopping is used as an additional shopping method which does not change trip making behaviour nor replaces in-store shopping but does change shopping behaviour.
The hypothesis testing result revealed that a positive relationship exists between the average number of trips complemented and telecommunication usage (call volume). This implies that the null (H 0 ) hypothesis is rejected, and the alternative (H 1 ) hypothesis is accepted. The correlation coefficient for the relationship between telecommunication usage and complemented trips in the study is 0.535, significant at 0.00.
Multinomial logistic regression (logit) analysis was further carried out using SPSS to determine the impact of telecommunication on travel of households. The result of the likelihood ratio test which explains the overall contribution of each independent variable revealed that household size and gender with P = 0.00 and P = 0.04 respectively were significant variables influencing the complementarity ability of telecommunication on travel. From the model equation, even though the complementarity effect of telecommunication on travel was significant in this study, male household heads were more likely to complement their trips (P = 0.035) than females. For the parameter table, the coefficient represents comparison between telecommunication capacity to complement (1) and telecommunication capacity to substitute (0). In this case, gender (male) and household size were significant. This implies that telecommunication was more likely to complement trips rather than substitute.
Although the study found a substitution effect of ICT on shopping trips, it, however, supports the perspective that the overall trend of the relationship is shifting from substitution to complementarity, which has been widely discussed in previous studies (De Graaff and Rietveld 2007;Wojuade 2014). Findings from the study corroborate the studies of Oyesiku (1996) and Choi, Choo, and Kim (2020) that telecommunication cannot fully compensate for face-to-face contact, for instance, the conveyance of complex, non-structured, or potentially ambiguous information. Also, the social and cultural  background of people in Nigeria, a developing country, is such that the physical presence of friends and relatives is often appreciated; hence telecommunication may not be able to fully substitute travel. Overall, the impact of telecommunication on travel depends largely on individuals' socio-economic attributes such as gender and household size, and activity patterns (shopping, work, banking, etc.) of an individual or household.

Acknowledgement
The