Having just arrived in Accra, Amit and I were trying to wrap our heads around taxi prices. We knew that we had to negotiate with the taxi driver but often the agreed-upon price kinda didn’t make sense…at all.
Our first hypothesis was that the price was completely random! Then we noticed we tended to pay more than others as foreigners (obrunis) AND newcomers, even after negotiation. But we suspected a lot of other factors were coming into play that we hadn’t fully measured the extent of yet: time of day, day of the week, phenotype, gender, where we were going or coming from, etc.
So we wanted to ask around and compare our experiences with others to see if a pattern emerged…
The 10 things we found out
Finding 1: Most people travel short distances
Most respondents don’t travel very far so taxi rides are mainly used for short trips across town. That said, anyone who knows Accra and its layout will also know that short distances mean nothing when traffic is at its peak or even worse, at a standstill. You also know that walking is both uncomfortable (because of the heat) and unsafe (in the absence of pavements). So it makes sense that people use taxis for short trips that would just be too annoying, or logistically impossible to do on foot.
Finding 2: Cost per kilometer is not standardized
This scatter plot illustrates that most respondents traveled short distances and paid reasonable prices, though they varied quite a bit. You see this with the cluster of dots around the low ranges that signal both short distance and low cost per km. There are also a few clear outliers in this data such as people who paid almost nothing for long distances and people who paid a lot for very short distances (we suspect the latter are to do with travel to and from the airport, but more on that later).
Finding 3: Out of all the age groups (disaggregated by gender), the cheapest price goes to men in their 40s!
Men in their 40s only paid 1.01 cedis per kilometer (when disaggregating by age and by gender). But while it’s definitely better, from a cost perspective, to be a man than a woman in your 40s, in your 30s, it’s definitely better to be a woman! As you can see from the figure below, they paid a lot less than men in the same age group.
Finding 4: Young people tend to pay less
First we looked at the average taxi cost per km by age. And it looks like young people between the ages of 20 and 29 paid less (1.74 for women and 1.66 cedis for men per km). The highest amount paid was among 30 to 39 year old men with a whopping 3 cedis per km! However, when we looked at the data more closely, we noticed that a few people in their thirties had to pay hefty amounts to travel extremely short distances to the airport. And of course, any travel to and from the airport automatically costs more…just because.
Finding 5: Men and women kinda pay the same thing, on average
We found that on average women paid 1.9 cedis per km and men paid 2.1 cedis. However, once we start to disaggregate by phenotype, things get interesting…
Then we had a look at cost per km by gender and phenotype. Interestingly, we were wrong to assume that as white obrunis, we were paying more. It turns out Amit and I just suck at negotiating because the average cost per km was pretty low for white respondents! Also, white men and women paid about the same thing. However, out of all our respondents, the prize for best negotiation skills goes to black men who only paid an average of 0.72 per km. This is also very low compared to black women who paid 1.86 on average, about the same as white respondents.
Finding 7: …but Indian respondents get the lowest price of all phenotypes!
Impressively, Indian respondents had the lowest average of all groups (1.4 cedis per km). Here, there seemed to be no significant difference between what women and men paid (1.29 cedis for women, 1.61 for men).
Finding 8: Middle-Eastern, East Asian and “other” respondents paid the highest prices
On average, Middle-Eastern respondents paid 2.7 cedis per km. Women paid 3.1 and men 2.4 cedis per km. That said, one Middle-Eastern respondent was going to the airport, the other to the American Embassy so that may explain why the price got hiked up.
There was only one East Asian respondent and he also paid a high price per km but with a sample size of one it’s hard to deduce anything here.
In the category “other”, the average price was the highest of all! However, we have no idea what phenotype this represents so there’s not really much we can say about this group.
Finding 9: Time of day and day of the week DEFINITELY matter
Using the following bins (Early morning = 0:00-6:59, Morning = 7:00-12:00; Afternoon = 12:01-17:59; Evening = 18:00-24:00), we noticed two interesting things about this data. During the week, the later it is in the day, the more you pay. On the weekend, you pay most in the afternoons. It makes me wonder if people tend to go out more in the afternoons than in the evenings on Saturdays and Sundays, thus making traffic particularly heavy and hiking the price up…
Finding 10: Length of stay doesn’t seem to make a difference
We included “length of stay” in order to see if the longer your stay in Accra, the better your negotiation skills and therefore the lower the price you pay. However, based on the data below, there’s really no clear indication that the length of your stay in Accra has anything to do with the price you pay for your taxi rides.
We sent out a Google Form on two Facebook groups: Ghana Expat and Ghana Intl Expats. On the form, we asked respondents the following:
- The cost of their last taxi ride
- Where they were leaving from and going to
- Time of day
- Day of the week (either Monday to Friday or weekend)
- How long they have been in Accra
Once we had the cost of each taxi ride and the beginning and end point of their journey, we could calculate the cost per km. We used Google Maps to tell us the total number of km for each journey. Once we divided the total cost by total number of kilometers, we had the cost per km.
We used this to analyse the relationship between this cost and the other factors we thought could influence cost: age, gender, phenotype, time of day, and weekday or weekend.
You can find all the raw data and analysis here: accraTaxiCostProject And just in case people keep completing the form, the google spreadsheet is here. Please feel free to play around with the data and tell us what you find!
Here are the highlights but please see figure below for a summary of results.
- Overall, 52 people completed the form, of which 37 were women and 15 were men;
- The vast majority of both female and male respondents were white (68% of all women and 33% of all men); and
- Most of our respondents were young: 44% were 20 to 29 years old and 33% were 30 to 39. That said, all of our respondents are clearly young at heart because they humored us with our fun little data quest 🙂
Limitations of this data
- We are very grateful to those who participated in our mini-study, thank you! That said, we were hoping for a larger sample size in order to get some more conclusive evidence. Given that we have 52 respondents, obviously there are biases in this data that we can’t fully detect or adjust for.
- We only asked for the last trip because we figured that’s probably easiest to remember. However, the last trip may not representative of taxi-riding experiences in Ghana. For example, a respondent could have paid more that day, be less willing to negotiate, or go somewhere (like the airport) which obviously comes with a higher price tag attached.
- There’s a reporting bias in that people who are not proud of their taxi-negotiation skills might feel less comfortable filling in this questionnaire.
Extra bargaining tips and tactics!
- Many people who commented on our study stressed the importance of negotiation! Although this is very difficult to measure in a study like this, people offered advice like “learn to say 4-5 phrases in Twi, it’ll help!”.
- Another respondent suggested: “Saying ‘Please, I beg!’ also helps!”
- A young white female respondent said she traveled for 12 cedis after “extensive bargaining” but that her Ghanaian friends pay 10 cedis for the same trip!
- One person didn’t agree on the price before getting in which is what they normally do, because of the ‘not allowed to park’-situation. I’m not aware of a no parking policy except at the airport… If anyone knows more, please share!
- Many people stress the importance of having a regular taxi driver. One respondent got a great price per kilometer from her regular guy. When you establish a relationship of trust and loyalty, it seems to affect the price. However this is beyond the scope of this study!
- The time of day at which you travel and day of the week have a lot to do with what you pay for your taxi ride. We didn’t ask about the total time it took for the trip but we suspect that as afternoons and evenings are peak traffic periods in general, travel time goes up and therefore cost.
- Your phenotype matters but not in ways we expected! Indians paid the least and Middle-Easterns, East Asians and “other” paid the most. Sandwiched in between, white and black respondents paid about the same thing. That said, black men get some pretty great bargains! Who knows what this is due to. It could be experience, cultural norms or the impression you make on the taxi driver.
So there you have it! Some interesting quantitative and qualitative findings to use in your future taxi ride negotiations!! If the info itself doesn’t help, I’m sure starting to explain the findings of the study to your potential driver will bore him enough to bring the price down, just to get you to stop. 😀
Many, many thanks to Amit Kohli for content and data advice, comments and suggestions for improvement! Please help us further improve our Random Curiosity-Driven Research Projects (RCDRP!) by leaving constructive comments or suggestions for us!