Tuesday, March 10, 2009

Second best models in outsourcing

This post is a note of caution to the gung-ho proponents of outsourcing. I confess to having been a member of this group, till experience chastened me!

The Andhra Pradesh Eastern Power Distribution Company (APEPDCL), one of the four distribution utilities set up in the state after the unbundling, has been trying to outsource loss (both technical and commercial) reduction in some of its high loss urban feeders, to private agencies on an Energy Saving Company (ESCO) model, where the ESCO and the utility would share the reduction in losses. This approach, while under consideration in a few utilities, has not been successfully implemented anywhere (to my knowledge) in the country.

Here is a more general formulation of the task. A public service utility seeks expertise from professionally competent consultants/service providers to lower power distribution losses or water distribution leakages, by taking a power feeder network or a water distribution zone as a management unit. There are two broad approaches

1. Completely outsource the loss/leakage reduction process, including investments (on a BOT or other modes), on a sharing of the benefits model.
2. Procure a consultancy service provider, on a fixed plus success fees model, and then the utility will implement (all or some of) its recommendations.

From the perspective of the utility, the first approach would appear to be a more effective option, given its outcome-based nature. Achieve the specified loss reduction target, take home the payout. On the face of it, the utility appears to bear no risk. The second alternative, even if structured properly, carries inherent implementation risks - wrong or unprioritized recommendations, uncertainty about the implementation costs, transaction costs in implementation etc. In other words, the first approach of outsourcing loss reduction is clearly the best-practice model.

However, a closer analysis of the problem and the market will bear out that the second-best approach is a more effective model for achieving the desired loss reduction objective, both from the cost and efficiency sides. This is especially true of the not-so-large contracts. Here are a few reasons

1. The transaction costs in managing an outsourcing contract - from preparing a contract to monitoring and enforcing the contractual obligations - are substantial and it takes considerable time to finalize such tenders. A consultancy is a far simpler proposition.
2. The market for such outsourcing services lack the required depth to facilitate an accurate price discovery, and the successful bidder is more likely than not to end up with a good deal. This is the typical "first mover disadvantage".
3. In the absence of adequate number of successful demonstration examples and reliable baseline data, the bidders tend to hedge all the possible risks into their bid quotes, thereby boosting the bid prices.
4. Such works have three components - professional survey cum technical analysis, civil/electrical works and Operation & Maintenance. Clearly, the success of loss reduction dependes on the quality of the survey and analysis. Given the absence of adequate market depth (which demands keeping technical qualification norms relatively liberal and flexible), such outsourcing tenders are likely to attract the not-so-competent contractors whose expertise lies not so much in technical analysis but in contract works implementation.
5. Outsourcing contracts raises problems about redeploying the existing employees. This often runs into problems with unions and gets politicized, making the process even more complex.

A consultancy model can avoid many of the pitfalls, but its terms of reference will need to be carefully formulated so as to avoid the usual ambiguity associated with such contracts that renders them ineffective and discredits the model itself. The consultant should be incentivized to make his recommendations based on comprehensive survey and analysis, and his payouts should be linked to the achievement of specific outcomes. A fixed and variable components to the payout can achieve this purpose. In this case, it should consist of location-specific recommendations about the loss/leak reduction potentials of various technical and non-technical interventions.

No comments: