Brainstorming Session No.7 

Forecasting Motorbike Market in Vietnam

August 10, 2006

Mr. Nguyen Anh Nam and his IPSI team reviewed past data and proposed methodology for forecasting motorbike demand in Vietnam. Statistical data from Vietnam Register Department, Road Transport Police Department, Customs Department and team's survey showed that the share of "Chinese" motorbikes had declined while Japanese and Taiwanese motorbikes had increased sales since 2001. The number of registered motorbikes at the Register Department is much larger than the number of registered motorbikes at the Road Transport Police Department. Domestic sales were projected in two ways: (i) using average motorbikes per person as benchmark; and (ii) using average motorbikes per household with rural and urban areas considered separately. Both approaches predicted that the stock demand for motorbikes in Vietnam would be around 31-32 million units in the year 2020. The team also looked into detailed figures at provincial levels and export figures.
　

Forecast for domestic sales using the average motorbike-per-person approach

Note: Vietnam is assumed to be equal to Thailand's current motorbike per person of 3-3.5 units.

Participants discussed two main issues: (i) reliability of past data; and (ii) forecast methodology.

First, as noted above, past data showed a large gap between the two sources. This may be explained by (i) double-counting in data reported by some producers as they re-assemble unsold units many times; and (ii) motorbikes used in rural areas which are reported by producers but not registered with Road Transport Police. Reliable base is needed for projection, but it is not clear where the real number lies between these reports. Mr. Nam's team proposed a 3-5% markup over the Police data. However, such adjustment needs to be discussed more carefully in the future. Mr. Nam also wanted to know, from producers, the average retirement age of motorbikes that they produced. However, answers would vary depending on product quality, rural vs. urban market, and consumers' preference toward new models. In addition, the official definition of "urban area" covers only five cities, which are Hanoi, Ho Chi Minh City, Haiphong, Danang and Cantho. Prof. Ohno argued that the classification of urban and rural areas should also be carefully examined in projecting demand. He also suggested that age profile estimates of existing motorbike stock was needed to translate stock to flow, and vice versa.

Second, as for forecast method, the IPSI team assumed that Vietnam would gradually approach the maturity state of motorbikes per person, which would be around 3-3.5 units if Thailand was the reference country. Participants generally agreed that using a long-term ratio of motorbikes per person (or household) was acceptable. Forecasted figures were also within the reasonable range. It was suggested that high and low scenarios should be calculated in addition to the benchmark forecast. Demand is affected by various factors, such as GDP, infrastructure, consumer preferences, etc. and their shifts can be incorporated if forecasts have a range. Production capacity of assemblers, which is a supply-side factor, should not be extrapolated to forecast demand. One producer argued that precise forecast was difficult and probably not necessary. But others thought that scientific method behind the forecast must be explained, if the master plan was to convince the Vietnamese government. One participant suggested that all figures and calculation should be presented in the Appendix so that anyone could check and re-calculate forecasts.

Participants agreed that all data presented today should be shared among JWG members as soon as the IPSI team signed official agreements with organizations which provided data.
　

[by Mai The Cuong]