In order to operationalise these aims, it is necessary to decide how these groups, stages and services should be targeted. These possibilities may be categorised in terms of delivering the subsidy, through:
• the tariff, to the consumer;
• subsidies to franchises or operators; or
• the regulatory framework.17
Each approach will result in different outcomes and different impacts on the poor. While it is clear that no solution is ideal - and any intervention that dabbles in redistribution will have unintended social consequences - some familiarity with the relative benefits of different systems is necessary to approach the outcomes required.
The degree to which subsidies can assist the poorest and most marginalised households and communities, however, is often misunderstood because the poor are often represented by those who have some (albeit limited) income, some capacity and some level of service. The very poor, however, may not reap any benefits from subsidies. They would need to be formal consumers. The benefits of cross-subsidies are therefore limited to existing consumers and exclude those marginalised from the service in the first place - normally the poorest households in the society.
At the same time, the tendency for the poorest households to occupy marginal unserviced land is also a key factor. In the rural context this might refer to inaccessible areas, but in the urban context it may refer to illegal settlements, low-lying swamplands, steep hillsides and/or peri-urban areas beyond the existing network. Efforts to provide direct subsidies that target these non-consumers (through, for instance, subsidies for infrastructure and connection fees) are thus an important aspect of a pro-poor approach to subsidies.
Criteria for evaluating the most appropriate delivery mechanism for subsidies have been devised for circumstances in transitional economies.18 They include, effective targeting, predictability, minimisation of distortion, and associated costs. In the context of developing countries where the problem is not one of maintaining connections so much as expanding the number of customers with access to network or non-network services, these criteria need to be extended to reflect the specific objectives of different urban service networks, poverty and capacity. Central to this must be the extent to which the vehicle adopted is able to bring about sustainable change in the existing patterns of marginalisation. This will not just require consideration of the avenue for delivering the subsidy, but the capacity of the actors involved in the execution. Another criteria is the relative number of people who have an increased access to affordable service. While they may be second-best solutions from an economic perspective, subsidies often play an important role in reform processes, and it is generally necessary to look at how adverse impacts (e.g., long-term price distortion) can be minimised and how benefits can be maximised.
Despite the general understanding that subsidies are a key way of supporting the poor during reform (and that the injection of funds will support sectoral reform), in practice, one finds many municipalities that cross-subsidise other social and welfare functions. Despite service deficiencies, municipal funds often flow away from, not to, a revenue- earning sector. The widespread occurrence of this informal approach to balancing the books - even where separate accounts are required and the legislative framework guards against such practice - begs the question as to what will happen to other municipal services when these finances are ringfenced, and such cross-subsidisation can no longer occur. Informal practices such as these draw attention to the need for policy-makers to focus on the realities of implementation and include implementers in policy-making discussions.