Freight Demand Model™ for South Africa

The Freight Demand Model (FDM™) is an instrument that provides intelligence on supply and demand in South Africa’s current and future freight-flow and logistics market, along with trends that shape and impact national freight flows.

The data in the FDM™ provides flow data for four surface transport modes (road, rail, pipeline, and conveyor belt transport). The data (which is comprised of approximately 1.5 million distinct data lines) includes tonnes, tonne-kilometres and costs for each flow associated with 83 commodities and containers between 356 local districts, seven seaports and eight land border posts in South Africa. The FDM™also includes imports and exports of these 83 commodities and containers per port and border post. Each time the model is updated, the annual demand data (base year) is forecasted for five consecutive years (years one to five), along with years ten, 15 and 30, to enable long-term planning. Each of these forecasts includes a likely, low, and high scenario. See Figure 1 below for a visual summary of the FDM™. 

The concept model was developed between 1995 and 1998, and some of its data were used in the Moving South Africa project in 1998 and the National Freight Logistics Strategy in 2004. The full-blown model was developed in 2006 and has been updated annually since. The data has been extensively used as a strategic tool in South Africa, specifically for capital planning related to rail and ports, such as in the NATMAP of 2009 and the Rail White Paper. Various provinces and cities throughout the country have also utilised it.

The FDM™’s core methodology has been described in many published international peer-reviewed academic articles. A representative recent example is the 2018 article on National freight demand modelling by Prof. Jan Havenga and Dr Zane Simpson, which describes how the model’s methodology works and how it is deployed in macro logistics management. Their 2022 article on Macrologistics Instrumentation, co-authored by Ilse Witthöft, describes how freight flows and logistics costs are derived and used (Havenga, Witthöft & Simpson, 2022). Other prominent articles include Simpson, De Bod, Havenga, Van Dyk and Meyer’s 2021 intermodal solutions article, Havenga and Simpson’s 2018 freight sustainability article and their 2012 domestic intermodal article, co-authored by Anneke De Bod, which all use the model’s inputs extensively.

Global freight flow models use one of three approaches in the development of the models, namely surveys (mostly used in developed countries), observations (basically truck counts) and supply and demand modelling (where economic data is disaggregated as far as possible into commodities and districts to determine points of demand and supply to deduce freight flows). All these approaches have challenges; for example, people are not always honest or competent in surveys, observations are often commodity- and origin/destination blind, while demand modelling is expensive and is not able to understand supply chain routes between supply and demand. The FDM™ follows an integrative approach that combines the best aspects of all three approaches (see Figure 2).

The four most important principles of the FDM™ are grounded theory; its hybrid approach; the extension of metrics to include all weight measures, but also logistics costs, externality costs and forecasts (disaggregated to enable trade-off decision-making); and the standardisation of outputs rather than inputs.

Grounded theory

Grounded theory is a guiding principle because, rather than trying to “prove” something, the model continues the discovery process, nearly indefinitely until the data is saturated enough to enable decision-making (see Figure 3). This approach enables both early use and continuous updates.

Hybrid approach

The Hybrid approach to data sources is not standardised (in terms of being absolutely prescribed) but rather a continuous search for new sources to improve the data. In the FDM™, more than 50% of the data is actual data and only the remaining is modelled (according to tried, tested and peer-reviewed methods). The modelled portion is getting smaller every year, with an increasing number of contributing data providers and detail in their respective datasets.

Our motto for the FDM™ is “we don’t model what we know” and is always inclusive of commonly accepted or verified actual data. Many global modellers do not incorporate real data in models, as the ability to do dynamic triangulation (incorporating known data from disparate sources and modelling the remainder) is uncommon.

Many data sources and inputs exist on national, provincial, and regional levels, as well as extensive industry and geography-specific intelligence. All suitable data sources, such as public information and reports, are utilised and included in the Hybrid data inputs (see Figure 4). Disaggregated known data is leveraged to disaggregate modelled data into supply and demand per district. That which is unknown is modelled to origin and destination pairs and collated with known data for total freight flow outputs.

Extended metrics

The FDM™ not only measures rail flows but uses extended metrics for all transport modes, including 30-year forecasts. It also includes, together with tonnes and tonne-kilometres, costs and externality costs (see Figure 5).

The extended metrics allow for trade-offs and measurements that enable decision-making in most policy, infrastructure and spatial planning areas for logistics, and all transport modes (road, rail, sea, etc).

Standard outputs versus standard inputs

The prescribed nature of other models means that many inputs are ignored, and that modelling becomes very expensive and even fails if these inputs cannot be found. Consultants often protect these approaches, and the “black box” nature of the outputs means that “buying” the outputs becomes expensive and models are repeated by many consultants for many projects. By standardising and using hybrid inputs but being prescriptive on outputs it becomes a common good. Users of the data could require more focus in different areas. In this case, the required output is merely added to the standard, meaning that all the data always talk to each other and are in balance. The primary output is to answer the questions of “How much of what was moved from where, to where and how and at what cost?”

This is done by quantifying geographically and sectorally disaggregated supply and demand data and resulting freight flows with the primary parameters of origin, destination, commodity, volume of freight, and transport mode.

Data output metrics

The FDM™ output metrics can be summarised in the following:

• 83 commodities
• Cargo and packaging types
• Bulk and containerised
• 372 geographic origins or destinations (local places, ports and borders)
• Modes of transport – road, rail, pipe, conveyor
• Tonnes per mode
• Tonne-km per mode, and rail equivalent
• Costs per mode – including road cost components
• Externality costs – various
• Freight flows identified as Import, export, domestic
• 30-year forecasts – annually for 5 years, and long-term intervals
• Assigned to corridors, rural, metro and per provinces
• Market segmentation and rail suitability
• Rail branch line classification – actual and market

Visual of overarching data output

A visual representation of all freight flows for South Africa, as well as a 30-year future projection, shows how all freight for the whole country is accounted for in the FDM™ (see Figure 6). On the left are the base year volumes, and on the right are the 30-year future projections, both to the same scale. The complete FDM™ data of approximately 1.5 million unique data lines is collated to generate this visual. While you can see the formation and development of corridors on the future map, the underlying detailed data can provide in-depth details.