Tuesday, February 19, 2008

Corruption in development programs

The former Indian Prime Minister, late Rajiv Gandhi had famously acknowledged that only 15 paise of every rupee spent on development reaches the target beneficiaries. The rest gets dissipated in the bureaucratic maze in which the program is couched, as transaction costs and rents. The Comptroller and Auditor General's (CAG) recent six-month audit report of the government's flagship National Rural Employment Guarantee (NREG) scheme from 513 gram panchayats in 68 districts across 26 states revealed that only 3.2% of the registered households benefited by securing 100 days of employment a year between February 2006 and March 2007. Siphoning off funds through fraudulent cards was the commonest form of corruption.

All this raises important questions. What should be the role of the bureaucracy in administering a development scheme? Is bureaucracy an inevitable accompaniment to development schemes? Is there a trade-off between procedural rigour and achieving specific development outcomes? Is it possible to reduce the bureaucratic red tape and still deliver the welfare and development programs without lowering their effectiveness?

One of the commonest complaints against government welfare programs is about the amount of bureaucracy involved in administering them. It is observed that the larger the program, the more bureaucratic the delivery channels become. There are essentially two main reasons why development and welfare programs get mired in red tape and bureaucracy
1. The target beneficiaries - individuals, groups and areas - are amazingly heterogenous, especially in a complex and diverse society like India. Typical Central and State government development schemes are designed top-down, and attempts to factor in all the possible requirements and contingencies - all categories of beneficiaries, and all regions and areas - and thereby standardize the scheme for the entire country. The resultant labyrinth of procedures and guidelines tend to make the programs complicated and opaque.
2. The need to put in place adequate checks and balances to prevent the scheme being subverted or exploited by vested interests. A multi-layered regulatory architecture develops, that seeks to minimize official discretion and thereby eliminate all opportunities for corrupt practices. We therefore have elaborate tendering processes and procedures, multiple filters for identifying beneficiaries etc. This requirement often becomes an alibi for bringing in multiple layers of bureaucracy, which often end up losing sight of the very objectives of the scheme.

In the effort to include all the requisite needs and demands, while at the same time plugging all avenues of rent seeking, the scheme gets entangled in a complex maze of rules and regulations, procedures and guidelines. The marginal benefit associated with every additional layer of bureaucracy is overwhelmed by the marginal cost incurred by way of transaction rents and delays. All this introduces three major sources of corruption in any Government welfare program.
1. The rent component of the transaction costs associated with administering the program. This involves the rents in moving files, expediting or influencing process level macro-decisions etc.
2. The sanctions or critical decision making in the program. Quite often this involves decisions like selecting beneficiaries, identifying works, finalizing villages or habitations, and the process of allotting work (tenders, nominations etc).
3. Release of money and its distribution/disbursal. In this case, rent seeking occurs mainly in the form of pilferage and siphoning off the final benefits.

There are no simple solutions to addressing these problems. Any effort should focus on containing the amount of corruption, but should not get bogged down in a futile effort to eliminate corruption. Analysis of the aforementioned sources indicates a few prescriptions
1. All the sanctions and critical decision making should be decentralized and done as close to the cutting edge as possible. This in turn should not be an excuse for blind decentralization, that often leaves fledgling institutions, without the requisite expertise and capacity, to administer activities that are much beyond their competence and ability. The remedy then becomes more harmful than the problem itself. The levels of decentralization will vary across states and should be left to the discretion of individual states.
2. Only the objectives and broad contours of the program should be shaped at the level of the Central or State Governments. The details of the processes and the guidelines should preferably be left to the discretion of the competent authorities at the district or even lower level.
3. The entire work flow of a scheme, from conception to delivery, should be made transparent. Computerization can help.
4. Minimize the number of transactions in the work flow and also the number of interfaces for the public with the Government agency. To the extent that any transaction or interface presents opportunities for rent seeking, this will naturally reduce corruption. Single window clearances, use of technology like cell phone, SMS etc can help.
5. Money should be released directly to the competent lowest level agency that is actually administering the activity. This can be facilitated by computerization of internal processes and using the recent advances in banking processes. Preferably, welfare handouts should be directly released to the bank accounts of the beneficiaries.

All this will not eliminate corrpution. It will at best control and contain it significantly and reduce public harassment. Once these prescriptions - transparency, decentralization, simplification - are addressed in a basic form, we should leave it to the system to take care of rent-seeking. These prescriptions will help align the incentives of the administering officials with the desired scheme outcomes. Any effort to consciously address corruption or reduce official discretion, by building in additional checks and balances will only be counter-productive. By introducing multiple filters to screen the selection process and procedural complexity in tendering process, we are only providing enough grist to the rent seekers to feast on. We can do better by acknowledging that it is not possible to completely eliminate rent seeking, and some extent of rent seeking is inevitable and may even be desirable (more on this in a later post).

After a certain basic level of standardization, an inverse relationship kicks in between procedural rigour and bureaucratic overisght on one hand and efficiency and effectiveness on the other. This is especially true of massive Government programs in a diverse context like in India. Some amount of official discretion, with the attendent risk of rent seeking, is inevitable if we are to achieve our desired outcomes. It is better to have a development scheme that achieves at least a portion of its desired objectives while enriching a few, than have a scheme that only enriches the many without achieving any of the objectives. At the risk of sounding cliched, it is better to be not penny-wise and pound-foolish!


Anonymous said...

Private companies are much better at handling resources than the government. Private companies have to respond to pressures of competition whereas there is no competition when you talk about the government.

gulzar said...

private companies are better at handling resources because they are structured (probably because of the competitive environment in which they have to survive) to respond to incentives (say, maximizing profits) and have the operational flexibility to make their responses.

Government agencies have no such incentives (and of lack of competition is surely a contributing factor) and even if the incentives are structured appropriately, there still remains the problem of rigidity and lack of operational flexibility.