[Home] [Papers] [Courses] [Contact]

Jamal Munshi, Transparency International, 1999
All rights reserved


Electricity generation and distribution in Bangladesh are carried out primarily by three state-owned enterprises. These are the Bangladesh Power Development Board (PDB), the Dhaka Electric Supply Authority (DESA), and the Rural Electrification Board (REB). All three organizations report to the Ministry of Energy, which controls all energy resources in the country. The Ministry is staffed with political appointments and is overhauled with new personnel each time a new party comes to power. The parliamentary system of democracy in Bangladesh is characterized by intense political rivalries between the major political parties.


The PDB is the primary nationwide electrical power generation and distribution utility. The DESA and REB are distribution and sales entities that purchase power from the PDB at a discount. The DESA provides power to the Dhaka metropolitan area. The REB provides power to rural areas not directly served by the PDB.


The private sector plays a small but growing role. Independent Power Providers (IPP) generate and sell power to the PDB. Their contribution is expected to rise from 3% in 1999 to over 15% in 2000. Large industrial plants are encouraged to use micro-generators to provide internal power needs and to sell off-peak excess power to the PDB grid. Private sector consultants provide technical support to state owned enterprises.


Non-management employees of state owned enterprises are unionized. The unions are very powerful and union leaders are able to exercise political influence at the national level. Managerial ability to make personnel decisions at SOEs is constrained by the power of local union leaders. In 1999 the PDB had over 24,671 employees consisting of 83.71% union members, 2.1% senior managers, 6% middle managers, and 7.6% junior officers.


The SOEs are financed with both equity and debt provided by the state. Donors that include international development banks also provide debt financing. The operations of these enterprises are not profitable and they require continual infusion of capital. The 1999 PDB financial statements show total assets of 178 billion taka supported by 97 billion in debt mostly to the government and the rest in equity held by the state. During that operating year they suffered a net loss of 4.5 billion and positive cash flows from operations of 2.4 billion from revenues of 24 billion. All figures are in Bangladesh Taka (BDT). The exchange rate at this time was approximately 50 BDT per USD.


Bangladesh is energy rich in theory because she possesses substantial natural gas resources. Natural gas fired generators provide more than 85% of the electrical power with the balance being generated by liquid fuel fired generators and hydroelectric sources.


The generation and distribution of power by the SOEs are characterized by inefficiency. In 1999 PDB had an installed capacity was 3611 MW with a production capacity of 3029 MW and a "firm supply" capacity of only 2519 MW or less than 70% of installed capacity. It fell short of the peak demand of 2881 MW. Of a total of 14.6 thousand gigawatt-hours of electricity generated or purchased in 1999, 21.5% was lost and is accounted for as a "system loss". About a third of the system loss is actually dissipated through an inefficient distribution network. The remainder has a corruption interpretation.


Nature of Corruption in the Electrical Utilities


The data for these findings include confidential and sensitive material elicited by TIB from retired public officials and retired managers of state owned enterprises. These so called "diagnostic reports" contain inside information and anecdotal evidence and give us insight into the nature and mechanism of corruption. A survey of consumers carried out by the TIB in 1998 provides corroborating evidence for the types of corruption suggested by the diagnostic reports. Internal reports generated by TIB's investigative research staff also support these findings. Research methods used by the staff include field interviews with public officials and collection and follow-up study of newspaper articles and other publicly available information relevant to corruption in Bangladesh.


In the context of this paper "corruption" is an undesirable alteration of a transaction between the private sector and the public sector that constitutes a breach of the public trust in which a law, policy, or procedure that has been put in place for the public good is subverted for personal gain at the expense of society by the giving and taking of pecuniary or non-pecuniary bribes or favors and/or by the abuse of power vested in a public official.


Using this construct as our operational definition we are able to identify four areas of corruption in the electrical power utilities. They are inferred from operating and financial data and supported by interviews with managers and consumers. They are described as (1) non-technical system loss, (2) accounts receivables collections, (3) new connections, and (4) procurement. We describe each of these irregularities using supply and demand analysis and then present a synthesis using agency theory as a framework for corruption in state owned enterprises.


Non-technical System Loss


We estimate that about 13.5% of the power generated or purchased in 1999 by the PDB was removed from the system without an invoice. The loss amounts to almost 2 thousand gigawatt-hours of electricity or 4 billion BDT ($80 million USD at 50 BDT/USD).  The total system loss is the difference between the power generated and purchased and the power that is sold. Some of this difference, termed "technical system loss" is accounted for by actual line losses in the distribution network. The net non-technical system loss may therefore be stated algebraically as


            NTSL = G + P - S - TSL


where NTSL = non-technical system loss, G = generation, P = purchases, S = sales, and TSL = technical system loss. The existence of a non-zero NTSL is evidence of corruption. It implies that either G or P has been overstated or that S has been understated. Corroborating evidence suggests that understatement of revenues is the primary source of NTSL. We now consider three different mechanisms by which NTSL may occur. These are (1) falsified meter readings or altered invoices, (2) lost product through illegal connections, and (3) overstated purchases.


Falsified meter readings or altered invoices


A fee paid to the meter reader enables the consumer to disable the meter or to otherwise falsify meter readings. Alternately, a fee may be paid to the ledger keeper to reduce the invoice amount. The practice is fairly widespread especially in Dhaka. TIB diagnostic reports testify to the daring and power of unionized meter readers and ledger keepers and to their conspicuous wealth and real estate holdings. The TIB survey of consumers shows that most consumers are aware of these practices.


Illegal connections


It is estimated that about 10% of the estimated 4 million electrical connections nationwide are not officially recorded as connections by any of the electrical utilities. These clandestine connections are able to remove power from the system without a meter. Twenty magistrates have been appointed to hear illegal-connection cases in mobile courts. These magistrates clear thousands of cases a month and order an equal number of disconnections and hundreds of jail sentences. However anecdotal evidence suggests that many of the disconnected lines are re-connected almost overnight.


The difficulty of obtaining legal connections may force many consumers to resort to clandestine connections. It is unclear whether the "illegal-connection" problem represents outright theft or a form of corruption. There is no evidence of the complicity of utility employees in these cases.



Overstated purchases


We anticipate that as power purchases by SOEs from the private sector increases in volume and importance, new opportunities for corruption will arise. Corrupt officials may profit not only by understating revenues as we have shown but also by overstating purchases. The mechanism may involve a bribe paid by the IPP in exchange for the acceptance of inflated invoices by government officials. Although we have no evidence of this practice in 1999, the bribery predisposition revealed by the diagnostic study makes it possible to anticipate these hypothetical scenarios as the industry undergoes structural changes.



Accounts Receivables


In 1999 DESA showed an accounts receivable balance equal to 25 months of sales. At the same time they were generating new invoices at more than twice the taka rate at which they were collecting old ones, i.e., the average collection period was increasing. We conclude from these data and from anecdotal evidence contained in the TIB diagnostic study that the DESA bill collectors are engaged in a corrupt enterprise possibly in collusion with higher officials. The bribery mechanism may involve paying bill collectors not to collect. Like meter readers and ledger keepers, bill collectors too are union employees and therefore difficult to discipline.


The PDB also shows a similar but less severe collection problem with an average collection period of 11 months and increasing due to an unfavorable collection to bill ratio. However, their accounts receivable anomaly is imposed primarily by DESA who have not been paying PDB invoices. DESA is a sales and distribution organization that purchases most of its power from the PDB. These data may imply that the two organizations use different methods for understating sales. DESA employees use mostly accounts receivables while PDB employees use mostly falsified meter readings.


New Connections


It is very difficult to get a new electrical connection without influence or bribery. In theory one needs only to purchase Form 300/1 for 15 taka and to provide the required information and attachments. However the process is complex and requires clearance from fire and health departments and information about property ownership. About a dozen attachments are necessary and the form passes through six different layers of the bureaucracy. At one time more than 30 signatures were needed. The complexity of the procedure makes it easy to find faults with the application.


A well-defined bribery structure has evolved from this situation. Even those without political influence may obtain a fast track connection simply by paying a bribe. The bribery amount varies from 5,000 to 50,000 BDT depending on the amount of power needed. The average rate in 1999 was 1,000 BDT per KW. The amount is higher if there are unresolved problems in the application material.


The complex application process has also spawned a 'dalal' industry. Dalals are agents who know the application procedure well and have connections with utility officials. They offer one-stop-shopping convenience to consumers for a fee. The fee includes a bribery amount for the officials. Consumers' willingness to pay for this convenience has created a supply side pressure that serves to perpetuate this form of corruption.


Persons with political influence, especially well-connected businessmen requiring large industrial connections, are immune to the bribery procedure and are exempt from the Byzantine requirements of Form 300/1. Where payments are made in such cases they are more in keeping with gratuities than bribes.


The PDB acknowledges that the new connection procedure is unnecessarily burdensome for consumers and, as of this writing, is planning to test a simplified "on-the-spot" connection procedure.





The capital budget for the electrical utilities during 1999 was in excess of 27 billion BDT (more than 500 million USD). There is significant evidence of corruption in the acquisition of these capital goods as well as in the purchase of operating materials and in retaining consultants.


Evidence contained in the TIB diagnostic study suggests that bribery in the procurement process is driven by both supply and demand. When available, political influence neutralizes bribery by making it unnecessary for vendors to supply bribes and impossible for utility officials to demand them. Vendors report that both pecuniary and non-pecuniary bribes are demanded from them by utility officials. Non-pecuniary gifts that have been demanded include international vacations for the family and education in overseas universities for the children. Utility officials report that intense salesmanship of competing vendors imposes bribery on them and corrupts the procurement process.


An informal survey of individuals on both sides of the transaction indicates that the pecuniary amount necessary to secure an order ranges from 10% to 30% of the order with the midrange of 15-20% being most common. These figures apply across the scale even to large projects. The bribery amount is hidden in the invoice by padding prices or by other means. The bribe is therefore ultimately paid by the shareholders of the SOE, in this case the Government of Bangladesh and her donors. Currently capital expenditures of up to 100 million BDT may be made by PDB officials without the approval of the Ministry. The bribery potential at this decision level therefore exceeds 25 million BDT or 500,000 USD at 50 BDT per USD.


The existence of corruption in procurement is inferred from documented cases of procedural irregularities. The following examples are cited.


Unproven technology with anticipated failure modes was purchased and then additional units re-ordered even after the anticipated failure occurred in the initial unit.


"Emergency" repeat-orders have been placed for additional units even before the first unit had been delivered; and when the delivery of the first unit was more than a year behind schedule.


Some purchases are made on a fast track apparently for the sake of expediency while others become bogged down in negotiations for years.


"Crash programs" to acquire generating capacity quickly are used to bypass normal procurement procedures.


In some projects a rift between the Ministry and the PDB or among the Cabinet Committee members is apparent with each interested party lobbying for a competing vendor.


Bidding on a project may be closed and then re-opened repeatedly for re-tender until a certain vendor wins the contract.


Some vendors are able to anticipate "crash program" purchases and to prepare their proposals in advance of the enquiry.


Three different bribery mechanisms are identified. These are: (1) accessing inside information, (2) influencing specifications, and (3) purchasing a recommendation. They range from small fees paid to clerical staff to large amounts payable to senior officials. When political influence is used it is usually applied at the highest levels and then allowed to work its way down the chain of command.


Accessing inside information


A small fee is paid to clerical staff to obtain photocopies of confidential documents relating to a project or a purchase. The information may include specifications, budgets, timetables, and data about competitors and competitive bids. This mode involves the lowest levels of risk, staff hierarchy, and bribery amount. Individuals with political influence may have access to the same information at higher official levels without direct bribery of the clerical staff.


Influencing specifications


At a higher managerial and bribery level, payments may be made to alter the specifications on a project to favor a specific vendor. Individuals with political influence may be able to influence specifications by other means.


Purchasing a recommendation


Finally, payments may be made to the ultimate decision makers on a project to secure an order. This corruption mechanism operates at the highest managerial levels and requires the highest bribery amounts. For purchases of less than 100 million BDT the decision makers are managers of the SOE. Decision makers for higher amounts are may be at the Ministry or the Cabinet Committee on Purchases. Individuals with political influence often operate at the ministerial level even for smaller purchases.



An Agency Framework for Corruption in State Owned Enterprises



The agency theory of the firm


Agency theory offers a model of a publicly traded corporation in terms of the conflicts between stakeholders whose relationships involve imperfect agency. The three primary stakeholders are owners, creditors, and managers. Managers are viewed as imperfect agents of owners and owners are viewed as imperfect agents of creditors. In a perfect world, managers always make decisions to maximize the wealth of owners and owners are never inclined to increase their wealth at the expense of creditors. Imperfections in the agency relationships imply that managers may make decisions that increase their own wealth at the expense of shareholders and owners may likewise transfer wealth from creditors to themselves. These imperfections impose "agency costs" that reduce the value of the firm.


Implementing costly "monitoring" mechanisms that enforce discipline may increase the value of such a firm. For example, performance bonuses or stock options may be offered to managers to align their goals with those of owners; and debt contracts may be written to limit the ability of owners to receive dividends or to take on new projects until creditors are repaid. All of these monitoring mechanisms are costly but it is optimal for the corporation to bear them as long as the reduction in agency costs exceeds the cost of monitoring. The value of the optimally balanced imperfect corporation is therefore less than its perfect counterpart by the sum of monitoring costs and residual agency costs but more than the un-monitored or "completely corrupt" firm by the difference between the monitoring costs and the agency costs that were eliminated by the monitoring devices.



Agency theory applied to state-owned-enterprises


SOE managers may be viewed as imperfect agents of the Ministry in charge of their industry. The Ministry itself is an imperfect agent of the legislature and the legislature is an imperfect agent of society at large. Similarly, we may view union workers as imperfect agents of managers. Such a multi-tiered arrangement imposes very high agency costs and makes monitoring difficult and expensive since each layer in the hierarchy must be monitored to achieve system-wide discipline. In this model, union members, managers, the ministry, and even the legislature may impose agency costs.


In the case of an SOE the government who are the owners also holds most of the debt. This unusual ownership structure of the SOE adds even more agency costs. Fortunately a certain amount of discipline enforcement is available from an unexpected source. In Bangladesh, as in other LDCs, much of the capital available to SOEs is provided or guaranteed by donor countries. In this role donors are able to add invisible wealth to an LDC by providing valuable monitoring services. However, their ability to discipline the stakeholders is limited at either end of the agency chain. On the low end they are unable to negotiate with trade unions and on the high end they face sensitive sovereignty issues if they interfere with the legislature. Within these limits donor discipline reduces agency costs and increases the value of state-owned-enterprises.


Corruption as agency cost


In the context of the agency theory corruption is modeled as a mechanism by which the firm incurs agency cost; and anti-corruption measures are modeled as monitoring devices that impose discipline on it.


Non-technical system loss (NTSL) may be related to the agency conflict between managers and trade unions. This corruption mechanism increases the wealth of union members and reduces the value of the firm by the amount of the loss, which is estimated to be 4 billion BDT per year. This figure also serves as the maximum budget for monitoring devices that may be designed to reduce NTSL. As noted donor discipline is ineffective at the trade union level.


Anti-corruption measures may involve monitoring methods employed by managers to discipline workers through union contracts, job training, and incentives. Privatization of the SOEs in itself may not directly affect the manager-worker agency conflict but may make it possible for the enterprise to recruit individuals with better managerial skills. Supply-side anti-corruption methods may be effective in this case.


Although NTSL constitutes a "loss" from the SOE point of view, it is not clear whether the loss is a drain on society at large. NTSL may be viewed as a combination of an energy subsidy and direct taxation. Funds removed from the system by the meter readers is a form of direct taxation while the amount saved by the consumers by paying off the meter readers is a form of subsidy.


Similarly, losses due to poor accounts receivables management may be ascribed to manager-ministry agency conflicts. The ministry attempts to monitor and discipline managers in cases that do not involve political influence but has a limited ability to serve society's interests in this role because it is itself an imperfect agent. Donor discipline is effective in this case because donors can work directly with managers as well as the ministry and there are measurable performance milestones that may serve as pre-conditions for release of funds.


Corruption of the procedures for providing electrical connections to new customers may also be modeled as a manager-ministry agency conflict. Although the ministry uses political influence to facilitate new connections for certain individuals there is evidence that the ministry has also acted to streamline new connection procedures for all customers. Improved training, motivation, and compensation for managers may achieve further discipline. Alternately, the ministry may choose to institutionalize the dalal mechanism that seems to provide the needed convenience to consumers, procedural simplicity to managers, and employment for a large number of dalals.


Procurement problems involve agency conflicts at the highest levels possibly involving the legislature and the Prime Minister's Office in addition to the ministry and SOE managers. The bribery amounts involved are large. Donor discipline and managerial training and compensation may not be effective in these cases. Privatization of the SOE could be used as a method for reducing agency costs by imposing shareholder and creditor discipline on managers.





Privatization of state-owned-enterprises is often proposed as an anti-corruption measure. Although privatization would be expected to reduce agency costs, there are special considerations that limit its applicability in this case.


First, privatization is not expected to significantly alter the agency conflict between managers and union workers that we believe is responsible for NTSL. Second, privatization of a utility monopoly such as a power distribution company poses special regulatory challenges. In a bribery economy, regulation in turn spawns new corruption opportunities.


Out-sourcing and privatization in small chunks also pose special problems and new bribery opportunities. Whenever a portion of the enterprise is privatized the privatized portion must then carry on business transactions with the public portion and these transactions are subject to corruption. For example, if accounts receivable collections are out-sourced then the vendor may be able to under-report collections and if power generation is privatized, as it has been, then power vendors may be able to over-report sales.


The ideal industry architecture would therefore be one that minimizes the role of the public sector without creating a private sector monopoly. It is anticipated that as micro-generation becomes popular most large industrial and commercial consumers will generate their own power. This new technology may encourage the growth of small scale utility cooperatives that will create a new kind of power distribution architecture with little scope for the government to become involved in the electricity business.



The Supply Side of Corruption


The demand side of bribery has received a great deal of attention in corruption studies but the role of the supply side has been largely ignored. We show that the study of the supply side of bribery can lead to new insight into corruption and better design of anti-corruption measures


A transaction between the private sector and the public sector may become corrupted if the public bureaucrat demands a bribe or if the customer offers one even in the absence of a demand. The supplier of a bribe may be motivated by simple expediency. A bribe may also be offered because the transaction is flawed or illegal and would not otherwise be accepted. Finally, a bribe may be offered because the customer cannot bring sufficient political influence to bear on a public bureaucracy that is not likely to act without it.


Such an offer can corrupt an otherwise honest bureaucracy especially when the amount being offered is large relative to the wealth of the bureaucrat. Once corrupted a bureaucrat is more likely to expect or eventually demand similar payments in the future and less likely to carry out the transaction without a reward.


This insight allows us to rationalize how income disparity may be related to corruption. Poor distribution of wealth and large differences in income at all strata of society are characteristic of LDCs. An individual with an income of $50 a month may transact business with customers earning $5,000 a month or more. A single transaction may be worth more to the customer than the government official’s annual salary. The observed correlation between the poverty of nations and the level of corruption is evidence of the role of the supply side in creating and perpetrating a bribery infrastructure. Poverty alleviation and equitable wealth distribution are in themselves anti-corruption devices.


Supply side analysis enables us to design more comprehensive anti-corruption programs. Anti-corruption measures such as ethics training, job training, better compensation, tough enforcement of criminal laws, and building public sentiment against corrupt officials all address only the demand side. These programs may be expanded to include the supply side by providing educational and promotional programs that inhibit individuals from offering bribes. Information must be provided to customers of government services about the nature of the service that they can reasonably expect without the payment of bribes. Enforcement of anti-corruption laws against of consumers and suppliers of corruption must be equally vigorous.