Estimating transportation demand can be conceptualized from three main perspectives that are also models:
- Constant. Transportation demand is proportional to a variable, so an increase in demand is expected with the corresponding growth of the variable. This does not necessarily imply causality, although variables having a strong association tend to be causal. The associations depicted are usually linear, but non-linear relations can also be represented. This is commonly known as the multiplier effect in economic impacts assessments. For instance, additional employment in an area could be associated with (proportional to) a defined number of additional commuting trips. An additional number of tons handled by a port could be associated with a proportional number of truck or rail movements to the hinterland.
- Deterministic. Transportation demand is a direct function of a number of known parameters. Knowing the parameters allows for estimating the demand accurately with a modal such as multiple regression. For instance, knowing supply and demand functions, including their elasticity, allows estimating how many passengers would purchase a ticket at a specific price level. A spatial interaction model is a standard example of deterministic demand, which is a general function of the attributes of at least two locations pondered by a function of their distance.
- Stochastic. Transportation demand cannot be accurately estimated because of the complexity of the parameters and the possibility of random events. In a complex transport market, demand becomes a bounded probability. A certain demand level is assessed to be probable, but another demand level remains possible. For instance, freight demand and rates can have substantial variations, including surges and crashes associated with the volatility of commodity and energy prices, transportation prices, geopolitical events, and currency fluctuations.