Optimizing vehicle procurement for humanitarian organizations

Andrea Masini, Professor of Operations Management and Information Technology - January 19th, 2015
Optimizing vehicle procurement for humanitarian organizations by Andrea Masini ©Fotolia

No NGO would be able to run development programs in disaster-torn countries without heavy logistics, which represent their second largest expense. At a time when donors scrutinize the way each dollar is spent, how can NGOs rein in operational costs? A new model has identified the optimal strategy for vehicle procurement, and it's surprisingly simple. 

Andrea Masini ©HEC Paris

Andrea Masini holds a PhD in Management from INSEAD but he first studied mechanical engineering at La Sapienza University (Rome). He teaches Operations Management and Information (...)

See CV

Whether it's a 4x4 bouncing on a track through Afghan mountains, or a regular sedan crawling through the slums of Port-au-Prince, transportation is vital for humanitarian organizations to run relief operations and development programs. It is also their second largest overhead expense after personnel. HEC researcher Andrea Masini asks whether donors will they think their money is being spent correctly if they see most of it going to support back-office operations rather than to help populations in need. Humanitarian organizations are paying increasing attention to their performance and are eager for management methods that can help them to streamline operations and reduce costs. “There is a boom of research in that area,” notes Andrea Masini. He noticed, however, that academic literature has so far paid less attention to fleet management in development programs than in relief operations, even though development programs make up a sizeable chunk of any humanitarian organization's activity. He and fellow researchers Mahyar Eftekhar, Andreas Robotis and Luk Van Wassenhove thus worked on identifying optimal vehicle procurement policies for organizations running development programs. 

A specific context

The HEC researcher sees this subject as “an incredible opportunity” to develop new operations management models that “improve the quality of living of populations affected by disasters.” Although models imported from commercial supply-chain management sometimes are a good starting point, he says the intuitions developed in that context can't work in a humanitarian context. Organizations operate in particular environments, where poor infrastructure and security problems make vehicle usage patterns very different. For instance, a mechanical breakdown can be dangerous in remote zones, so in certain countries only newer 4x4s are driven on field trips, while older ones, which have crossed the age threshold for dangerous field drives, are used for administrative trips in safer zones. That is how it works in theory, at least. In practice, headquarters are sometimes not aware of practical constraints. “For example, headquarters assume that specially equipped 4x4s will be used off roads, but they might turn out to be too stiff, or else the head of delegation uses one in town because it's more prestigious,” says Andrea Masini. The challenge is compounded by the fact that fleet planning decisions are made by headquarters. Last but not least, NGOs are often tied by commercial agreements with manufacturers. For instance, the ICRC purchases its Land Cruisers at below market prices, but is not allowed to resell them before they are three years old. “It is part of Toyota's commercial strategy to avoid flooding the market with new vehicles,” says Andrea Masini.

Guillemet

The researchers' findings, using both models, were pretty clear, and were also in sharp contrast to the policies usually adopted by humanitarian organizations

Modeling fleet sizes

In order to help humanitarian organizations design cost-effective policies for purchasing vehicles, the researchers gathered data from the International Committee of the Red Cross (ICRC). They identified three countries – Sudan, Ethiopia, Afghanistan – which are representative of dozens of others in terms of some characteristics of their operating environment such as (un)paved roads, level of conflict and safety that affect logistics. They estimated vehicle cost and capacity, “vehicle” here meaning the ubiquitous 4x4 Toyota Land Cruiser, and applied a mathematical model to calculate the optimal fleet size in each of those countries. They validated their model by running it on past data – actual policies implemented by the ICRC. Then, the researchers expanded the scope of the analysis and developed a general mathematical model potentially usable by any organization wishing to curb the cost of purchasing vehicles. The advantage of this model is that its data requirements are compatible with those usually available in humanitarian organization headquarters. “Our model doesn't need very detailed data, it can function well at the aggregate level,” explains Andrea Masini. He believes that's what will make it easy to apply in real-life contexts. “Collecting data at the local level is an issue, it requires resources,” he says, adding that NGOs only need to characterize the demand profile (how many vehicles are needed at a given time). Incidentally, demand remains relatively stable in development programs. 

A level strategy

The researchers' findings, using both models, were pretty clear, and were also in sharp contrast to the policies usually adopted by humanitarian organizations. Andrea Masini and his three co-authors found that the optimal fleet size remained relatively stable, even when demand fluctuated. In other words, NGOs are better off implementing what is known as a “level strategy” (leveling fleet size around the average demand) rather than a “chase strategy” (matching the exact demand). Andrea Masini explains this result by pointing out that there are high costs associated with purchasing and selling vehicles to accommodate demand, such as the costly difficulties in shipping vehicles to distant countries. He says: “If demand doesn't fluctuate much in terms of magnitude, it's better to just keep the vehicles in the fleet, even if they aren't always used. If the fluctuations in demand occur at a high frequency, it doesn't make sense to chase demand either.” The only case where a chase strategy may be relevant is if peaks in demand are few and far between, say if demand over a year is stable from January to November but doubles in December. The researcher is confident that humanitarian organizations are ready to revise their policies. In fact, he is already talking to the ICRC about the second part of his research, which examines vehicle usage.

Based on an interview with Andrea Masini and the study, “Vehicle Procurement Policy for Humanitarian Development Programs” by Mahyar Eftekhar
, Andrea Masini
, Andreas Robotis, and Luk N. Van Wassenhove
 (Production and Operations Management Journal , June 2014).

Applications for managers
Applications for managers

The first and most evident lesson for logisticians is to resist the temptation to chase demand when purchasing vehicles. As Andrea Masini points out, “in three out of four cases, a level strategy is more effective.” When budgets are tight, NGOs can simply revert to an intuitive strategy of purchasing as many vehicles as budget allows for each period, until peak demand is met. Andrea Masini would also suggest that humanitarian organizations revise their contracts with suppliers when particular conditions (such as a no-resale within three years rule) restrain procurement. “It would be better for humanitarian organizations to relax procurement constraints instead of getting bargain prices,” he concludes. 

Methodology
Methodology

The researchers gathered data (age, average distance traveled, value, and so on) about the vehicles used by the International Committee of the Red Cross (ICRC) between 2000 and 2007 in Sudan, Ethiopia and Afghanistan. They complemented the data with interviews with ICRC representatives. Using this field data, they built a linear programming model to calculate optimal vehicle fleet size. They then drew upon its results to build a stylized quadratic control model to characterize the general structure of the optimal procurement policy.