Ghana's transport data problem

The most authoritative public figure on how long it takes to get to work in Accra comes from a survey conducted in 2012. It reached most people through a research paper published in 2017. Nothing better has replaced it. A city of roughly four million people is running on transport data old enough to be in secondary school.

I found this out by trying to answer a simple question: what does the average commute in Accra actually cost, in money and in time?

I expected this to take an afternoon. It did not.

What is actually available?

The best single source is a 2017 study by the International Growth Centre, Unequal commutes: Job accessibility and employment in Accra. It is careful work and I have leaned on it heavily. Its headline commuting figure is drawn from Ghana's National Transport Survey 2012.

That figure: the average commute in the Greater Accra Metropolitan Area was 110 minutes, for an average distance of 3.28 kilometres.

Take that seriously for a moment. As a one-way trip it implies an average speed below 2 km/h — slower than walking, which cannot be right. It is far more likely to represent total daily travel across several legs, including waiting and transfers. But the survey does not say, and the paper does not clarify, and so the most-cited transport statistic about Ghana's capital is one whose unit of measurement is genuinely ambiguous.

That is not a criticism of the researchers. They cited the best available source. That is the best available source.

What about fares?

Worse. Tro-tro fares are set through negotiation between the Ghana Private Road Transport Union and government, adjusted when fuel prices move, and then applied across thousands of routes by individual mates and drivers. There is no authoritative published table. There is no API. There is no dataset.

When fares change — which happens several times a year and is national news each time — the change is reported as a percentage. "Fares up 15%." Up 15% from what, on which route, to what new absolute figure, is left as an exercise for 4 million commuters.

You can find crowd-sourced fare sites. You can find news reports quoting a few routes. You cannot find a current, complete, citable source. I looked properly, and I could not find one.

Why does this matter beyond inconvenience?

Three reasons, in ascending order of seriousness.

It makes policy hard to evaluate. If you cannot measure journey times before an intervention, you cannot measure them after. Every road project, every bus lane, every fare adjustment is assessed on anecdote and traffic-count snapshots rather than on what actually changed for commuters.

It makes the commuter's own decisions worse. A household choosing where to live is making a multi-year financial commitment based on a guess about a recurring cost. Housing that looks cheaper can quietly be more expensive once transport is counted, and there is no easy way to check before committing.

It hides distributional effects. This is the one that bothers me most. The IGC work found that a working-age person in Accra can reach, by tro-tro:

• 4% of formal jobs within 45 minutes
• 32% within 60 minutes
• 62% within 90 minutes

Fifteen extra minutes multiplies reachable employment eightfold. Which means where you live, and how reliable your route is, functions as a filter on which jobs are available to you at all. That is a serious statement about economic opportunity, and it rests on a single study working from decade-old survey data.

By comparison, the same research notes that Nairobi matatu users reach about 10% and 20% of jobs in the first two time bands. So Accra is not uniquely badly served. It is just uniquely badly measured — and so is much of the region.

Is this a Ghana problem?

Not really. It is a problem with how informal transport is treated everywhere it dominates.

Roughly 75% of Accra commuters use tro-tro, about 13% walk, and around 12% use private cars. So more than 80% of the city's commuting happens on modes that produce no data at all. There is no ticketing system, no telemetry, no central operator. The network exists entirely in the working knowledge of its drivers and passengers.

Formal transport systems generate data as a by-product of operating. Informal systems do not. That is not a flaw in the informal system — the tro-tro network moves millions of people daily and is, in its own terms, a considerable achievement. But it means the mode carrying most of the city is the mode we know least about.

What would fix it?

Not a study. Studies expire, as this one demonstrates. What is missing is something maintained.

Three things would change the situation materially:

A current fare record. A published, dated, per-route record of what journeys actually cost, updated when fares change rather than every five years. Boring to maintain, and it would immediately become the most-cited transport resource in the country.

Journey times by hour. An average commute time is nearly useless. The same corridor at 6am and 8am is effectively two different roads. Time-of-day data is where the decision-relevant information lives, for commuters and planners alike.

An open stop dataset. Names, coordinates, and which community each stop belongs to. This is public-good infrastructure. Every transport app, every researcher, and every planning exercise in the city currently rebuilds it from scratch.

Who should do it?

Whoever is already collecting it as a by-product of operating.

I should be straightforward about my position here: I am building a commuter transport company, which means I will end up holding fare and journey-time data across Accra corridors. So when I argue this data should be published, I am arguing for something I am unusually well placed to supply. Read the argument with that in mind.

But the point stands independently of who acts on it. A city cannot improve what it does not measure, and Accra currently measures its dominant transport mode not at all. The gap between 4% and 32% job access is fifteen minutes wide. Nobody is tracking whether that gap is getting better or worse.

I intend to publish what I gather, including where it is uncertain and where estimates are estimates. That is worth stating plainly now, before there is anything to publish, so it can be held against me later.

Sources: commute time, modal share and job-accessibility figures from Chen et al. (2017), "Unequal commutes: Job accessibility and employment in Accra", International Growth Centre, drawing on Ghana's National Transport Survey 2012.