Rural roads, sometimes called low-volume roads because of the lack of motor traffic, are part of a linked hierarchy of networks. These networks begin at the village level as tracks or very simple earth roads mostly used by people on foot or on bicycles. They must be used with care and can be difficult in bad weather. The next level shades into rural roads. These serve the needs of small numbers of conventional vehicles, as well as non-motorized traffic, but at low speeds and in very moderate comfort but with a reasonable certainty of getting to their destination, albeit with occasional stoppages during the rainy season. At higher levels, they shade into conventionally engineered roads, linking cities and towns and carrying growing numbers of increasingly heavy vehicles at normal speeds and without hindrance.
Roads selected, no matter what hierarchy they belong to, should also form a coherent and sustainable core network, that is one which provides accessibility to higher level networks and the services located on them for as many people as possible. Limits are inevitably imposed by the funds available for maintenance. Roads already in maintainable condition should be a priority, as the marginal cost of their deterioration is far greater than the cost of maintenance.
When needs are great and resources limited, screening methods are necessary to make sure that only the roads with the greatest impact are built. They must not be too costly to apply in proportion to the investment involved, nor should they be so complicated that no one understands how they work.
Selection methods range from a single-criterion index of benefit, preferably monetary, which can be directly compared with the investment, to those which attempt to copy the complexity of real-life situations by incorporating many indices of need and potential impact, such as poverty reduction indices. Afterwards they may sometimes be collapsed into a single one using a weighting system reflecting the values of the investors and more importantly the beneficiaries. The multiplication of variables, although seductive, makes the selection procedure opaque and difficult to understand. An alternative is provided by the Basic Access approach, which attacks isolation explicitly by defining a minimal level of access to motor transport to which all inhabitants have at least a theoretical right.
Selection procedures differ also as to the level at which they are defined and applied. It may be nationally, to ensure a consistent and equitable allocation of funds among regions, locally, in collaboration with road users and beneficiaries, or at both levels, as in the basic access approach where a national criterion for accessibility can be defined.
Robust methods have been developed, such as HDM-4, and RED to evaluate major road programs where savings to road users are high in proportion to the investment and their future behavior faced with an upgraded road reasonably predictable. Others have become available recently, which are more applicable to low-volume roads. However, the impact of rural roads on local mobilty is difficult not only to foresee but also to express in monetary terms that can be directly compared to cost. In such cases, it is better to forget often futile attempts at quantification and rely on local stakeholder participation, not only for information, but to play the primary role in programming. This does not exclude either the use of objective criteria or the participation of higher levels of government, particularly if they are investors. However, it is important that choices be made within an egalitarian process of negotiation where all participants agree upon the selection criteria and understand and accept the consequences of applying them.