Delhi Reforms

Round Table

Power Sector Reforms in India: Distribution Reforms

D

istribution reforms in India began in the mid90s, and followed the World Bank’s model of privatisation. This goal required intermediate steps of unbundling and corporatisation of vertically integrated state utilities. So far, Orissa and Delhi have already privatised their distribution companies, while Karnataka and Andhra Pradesh have undertaken aggressive steps in the same direction (Exhibit 1).


To distance the power utility from government control with the objective of creating autonomy
To attract large amounts of private finance

The Orissa Experience


To introduce competition.

D V Ramana Orissa was the first state to initiate the reform pro-

Exhibit 1: Distribution Reforms: An Overview Orissa

Delhi

AP

Karnataka

Yes

Yes

Yes

Yes

Corporatisation Yes

Yes

Yes

Yes

Privatisation of Distribution

Yes

No

No

Unbundling

cess. In 1993 the Government of Orissa commenced an extensive reform programme of the electricity sector, to improve the quality of electricity supply and stimulate economic growth in the region. The broad objectives of the reform programme were as follows:

Yes

IIMB Management Review, March 2004

To achieve the above objectives, the Orissa Power Sector Restructure Project was designed to undertake the following:
To establish a separate act i.e. the Orissa Electricity Reform Act
To un-bundle and corporatise the Orissa State Electricity Board (OSEB)
To develop an autonomous power sector regulatory body
To privatise the generation business and the distribution business.

71

Process The reform process began with the enactment of the Orissa Electricity Reform Act in 1995. The Act was designed to address the fundamental issues responsible for the poor performance of the industry in the state. The new legislation was enacted for the purpose of restructuring the electricity industry, for taking measures conducive to increasing the efficiency of each activity i.e. generation, transmission, and distribution of electricity, for opening avenues for private participation and for establishing the Regulatory Commission. The Act allowed for the transfer of the assets, liabilities, staff and statutory obligations of the OSEB to the successor companies.

Unbundling and Corporatisation Section 23(I) of the act prescribed the method by which the power sector would be restructured. As a consequence, the Grid Corporation of Orissa Limited (Gridco) and Orissa Hydro Power Corporation (OHPC) were registered as government companies. Accordingly, the following steps were taken:

furcate the field level activities into engineering and commercial. The commercial function comprised revenue management, accounting, customer management and the engineering function comprised system operation, repairs and maintenance, and other technical activities. Coupled with the bifurcation, commercial procedures were implemented. This small intervention improved the business environment at the division level. Introduction of Profit Centre Accounting: Under this, Gridco was divided into Cost Centres and Profit Centres. Transmission divisions were treated as cost centres and the distribution divisions were treated as profit centres. Introduction of Revenue Improvement Action Programme (RIAP): The programme focussed on the reduction of non-technical losses, thereby improving the financial position of Gridco. One of the major contributions of RIAP was the quantification of the actual T&D losses. The energy lost and energy billed in the year 1996-97 worked out to billed: 54%, non-technical losses 28% and technical losses 18%.


The electricity business was divided into three separate activities viz. generation, transmission, and distribution

The finding of high T&D losses against the reported 22% in various forums and published sources was an eye opener to both the shareholders and the management.


The business of generating of hydel power was transferred to OHPC

Privatisation of Distribution Business


The thermal power stations were transferred to the existing Orissa Power Generation Corporation (OPGC)
The transmission and distribution business was transferred to Gridco.

Since the weakest link in the industry value chain is the distribution of power to the ultimate consumers, the government decided to transfer it to private players. Some of the indicators of performance of the distribution business just before the privatisation are given in Exhibit 2.

Prelude to Privatisation Exhibit 2: Selected Performance Indicators Three important steps taken under before privatising the distribution business: Introduction of Commercial Function at the Field Level: OSEB was managed principally by engineers, therefore, the management had a strong technical bias and it was standard practice for engineers to undertake commercial, administration and other non-engineering jobs. One of the major organisational changes introduced during the ISP was to bi-

72

LT/HT Billing as % of Input (units)

51 to 60%

Collection as % of billing (amount) LT consumers

48 to 54%

HT consumers

75 to 86%

EHT consumers

91 to 99%

Effective % of input collected

43.3%

Outstanding Debtors (Sales days) LT consumers

499

HT consumers

121

EHT consumers

64

Source: Gridco’s Management Accounts

Power Sector Reforms in India: Distribution Reforms

Experiment With Management Contract In October 1996, Gridco entered into a management contract, a ‘Distribution Operations Agreement’, with BSES, one of the leading private players of the industry. Under the contract, BSES was given the management of one of the four zones viz. Central Zone in consideration of a fee plus incentives based upon measurable improvements in collections and metering. However, the experiment failed and Gridco revoked the contract in May 1997. This experiment demonstrated that it is not possible to improve the performance of an organisation just by asking private entrepreneurs to run the day-to-day operations without giving overall control. But, on the positive side, Gridco was determined to transfer control to private hands which forced the government of Orissa to take the bold step of divesting its share in all the four distribution zones.

P K Kukde The state of Orissa was the first to proceed with a reform programme in response to the World Bank’s offer. While its estimated percentage of T & D losses was high and collection of bills was low, the state had the advantage of a relatively small agricultural load and hence it needed a lower agricultural subsidy. Electricity sector reform in Orissa was termed as the ‘World Bank model’. It was believed that private finance would develop new generation and enhance availability of existing capacity. Private participation in distribution was expected to improve service quality and increase financial performance. To make distribution more attractive 75% of the shared financial liabilities were transferred to the publicly held transmission sector. Generation was made more attractive by increasing the price charged to Transco, which however was not allowed to pass on the price increase to the distribution companies. Thus Transco built up enormous liabilities that undermined its long term viability.

However in all fairness, it must be recorded that the Orissa experiment provided a powerful demonstration effect. The lessons from Orissa were thus very clear:
There was a need for full and sustained political, administrative and financial support to the distribution companies in their efforts to improve and run the electricity distribution business.
This support would enable them to disconnect illegal consumers, reduce theft and improve collection efficiency.
Baseline data was required to be reasonably correct.
Multi-year regulation was called for to reduce regulatory risk and uncertainty.

Following the lead of Orissa, six other states embarked on the process of reforms, through models largely resembling the Orissa The Orissa model. The lessons learned in Orissa experiment provided a and elsewhere in the country have been constructively applied in the powerful Delhi model.

demonstration effect. The lessons from Orissa were — the need for sustained political, administrative and financial support to the distribution companies, correct baseline data and multi-year regulation.

Revenues from privatisation were not ploughed back into the sector but absorbed into the government budget for other purposes. Subsequently substantial tariff increases were imposed on the public but with few improvements in service, leading to growing public discontent. Provisions were not made

IIMB Management Review, March 2004

for the transitional period. Assets were valued higher, losses and industrial consumption were overstated. The Kanungo report was an eye opener on many of these aspects. There was inadequate government support and this distancing resulted in the failure of this model.

D V Ramana Cost of Service-based Tariff: Will it Help in Rationalising the Tariff CESCO is a distribution company of Orissa, not doing well, but is an interesting case where there is a possibility of forcing the government to show its support for the sector. We are of the view that the government cannot wash its hands of the power sector.

Despite the privatisation and so called reforms the overall performance of the company has been far from satisfactory. The company continues to incur losses and the net worth got eroded over the period. One of the reasons for the dismal financial performance is the gap between the actual and approved

73

Exhibit 3: Sales Mix of CESCO Sales Mix Segment

Units sold (MU)

FY 99 FY 00 FY 01 FY 02

FY 99 FY 00 FY 01 FY 02

LT

59%

69%

63%

71%

1173

1332

1379

1552

HT

19%

18%

17%

18%

367

338

387

394

EHT

22%

13%

20%

12%

435

320

453

266

100% 100% 100% 100%

1975

1990

2219

2212

Total

the exercise raises several questions about the present tariff setting practice and provides some directions in tariff management. It also makes a case for the state government to provide subsidy to the sector.

Power Sector Reforms in Delhi Jagdish Sagar

T&D losses. The actual losses are in the range of 50% to 43%, whereas the Commission approved 34% to 35%. A large chunk of such loss is commercial and revenue related. Another reason for the increased financial losses is the change in the sales mix (Exhibit 3). From the consumption pattern of CESCO we could see that the large consumers were withdrawing from the EHT segment (the most profitable one), both in terms of absolute and relative consumption. The change in the sales mix impacts the loss levels. Decrease in EHT/ HT consumption and increase in LT consumption negatively affects all the performance indicators, viz. loss percentage, collection percentage, and overall financial performance due. The issue was whether the tariff design was responsible for the change and whether Cost of Service (COS) based tariff could address this lopsided consumption pattern. Based on the data of one year (Exhibit 2) and on the basis of the model given by the National Association of Regulatory Utility Commissioners (NARUC) we calculated some tentative numbers and wanted to throw it open to the public at large in Orissa and the stakeholders. Exhibit 4 shows the COS and the gap between the tariff and COS. The actual numbers for the year 2002 of approved tariff shows a the gap between the realisation and the COS for the LT category to the extent of 291 crores (2.91 bn) and for EHT category consumers the gap is around Rs 1.13 crores (11.3 mn). The problem that we encountered while calculating the COS was the quality of the data. At the consumer level or a category level the distribution company does not have the maximum demand data. As a result we had to compute COS on the basis of certain assumptions based on macro data at the CESCO level, the maximum demand consumption per month. Though the basis of computing the COS can be questioned,

74

Background The Delhi Vidyut Board (DVB) was set up in 1997 as an SEB replacing the Delhi Electric Supply Undertaking (DESU). It is one of the largest urban utilities and a purely distribution one. Some features of Delhi’s power supply situation are:
Very high and rapidly growing per capita consumption — 1382 kwh per capita per annum in 2001-02
A high rate of load and consumption growth
Sharp variation between peak and off-peak load, with sharp seasonal variation, and consequent peaking shortages during peak seasons
Limited local generation
The advantage of negligible agricultural load is offset by the presence of a large population of unauthorised colonies and jhuggies; an unstable cityscape and a constantly changing urban scene. From 1959 to1999 the utility was restrained from supplying power to unauthorised areas, which compelled the growing population of such areas to steal it. In 2001, DVB estimated that about 14% of Delhi’s power

Exhibit 4: Cost of Service and Average Realisation Total COS (Rs cr)

COS Reali- Gap per sation per unit per unit unit

Total Gap (Rs cr)

LT Total

503

3.46

2.54

-0.92

-291.21

HT Total

112

2.99

3.91

0.92

38.12

EHT Total 151

3.13

3.1

-0.02

-1.13

Total

3.3

2.87

-0.44

-254.21

766

Power Sector Reforms in India: Distribution Reforms

consumption was going unbilled to the unelectrified or other external agencies. This urgency was an addicolonies and jhuggie basties (squatter settlements). tional reason for adopting the business valuation methIn addition to this, DESU and later DVB were notoriodology. The revenue potential of each of the discoms ously subject to the usual factors, attributable to powas gauged on certain assumptions about reasonable litical interference and the low priority accorded to tariff increases; targetted efficiency improvements (for commercial performance, which affected other SEBs. the business to become self-sustaining within five During recent years, DESU/DVB suffered from a poor years); costs; and the transitional assistance commitpublic image, for its quality of service, consumer related by GNCTD. Such a methodology is in any case tions and commercial performance more appropriate to the purpose and alike. The best known thing about DVB also obviates the impact on retail tarDelhi Vidyut Board was its T&D losses. However, I think iff that overvaluation can result in. our statistics were simply more honhas suffered from a Meeting investor concerns: The est than those of other SEBs. Moredearth of interested investors was over there has been too much simplispoor public image. compounded by the Orissa experience. tic thinking about this problem. It is viewed as a moral issue rather than the The best known thing Unless the concerns of potential investors were addressed, there would managerial problem that it really is. But about DVB was its be little chance of success. even when (down to the early 1990s) DESU’s T&D losses were at relatively T&D losses. There has The experience in Orissa (1999) highacceptable levels, the retail tariff had been too much lighted the importance of reliable openbeen insufficient to cover its costs. ing data, as the level of T&D losses simplistic thinking Delhi’s T&D losses were 12% in 1976 (difference between units supplied – they increased sharply between 1976 about this problem. It and units for which bills are raised) and 1980, but were still in the low twenties, and went up very sharply from is viewed as a moral there turned out to be vastly different from what had been stated, with di1993-94 – and that was the year that issue rather than a sastrous consequences for the invesDelhi first got an elected assembly. Before that Delhi had a different admanagerial problem. tors. Fortunately in Delhi, with metered billing of consumers, the basic ministrative structure. There is a corproblem, arising from the unreliability relation there, in my opinion. DVB of billing figures, was not nearly as serious as in SEBs had become a good boy in the last few years, paying with extensive flat-rate billing; nevertheless there was 95% or so of its power use altogether. some margin for error. Hence the adoption of Aggregate Technical and Commercial (AT&C) losses (differThe Reform Process ence between units supplied and units for which payThe new Government of the National Capital Territory ment is realised), rather than T&D losses, as the criteof Delhi (GNCTD) elected in December 1998 brought rion of commercial efficiency. out a Strategy Paper in February, 1999, the main feaRegulatory uncertainty: The mere establishment of tures of which were the setting up of a regulatory coman independent Commission was initially perceived as mission; unbundling (starting with the establishment an adequate solution for the problem of the political of a generating company); disinvestment of distribuexecutive taking regulatory decisions, but the develtion; interim measures to improve DVB’s performance; opments in Orissa made it apparent that the question and a commitment to protect the interests of staff. of realistic loss reduction targets would have to be However, the reform package as finally implemented in addressed if investor interest was to be sustained. fact evolved empirically with the process over the next three years and did not strictly follow the sequencing originally envisaged.

Key Issues The key issues were as follows: Time frame: Since the government was understandably anxious to show results of its reform programme before the end of its term of office, the resultant time frame was not long enough to involve the World Bank

IIMB Management Review, March 2004

Multi-year tariff packages have commonly been an ingredient of successful privatisation elsewhere but our laws in India do not envisage it and require annual filing. This being so the solution initially proposed was multi-year tariff setting principles premised upon pre-determined T&D loss reduction targets. (The AT&C loss concept had already been conceived by the consultants but at this stage the T&D loss concept was applied to be consistent with the Commission’s own formats for tariff filing.) Along with

75

Exhibit 5: Delhi: Unbundling and Privatisation Process Legend Assets Liabilities

DVB 1. All the assets and liabilities of DVB are acquired by GNCTD

3. All the assets are transferred from GNCTD to successor entities.Assets will be assigned a value equal to serviceable liabilities. Genco

2. All the assets and liabilities of DVB are transferred to Holding Company, entire Equity of Holding Company is issued to GNCTD Holding Co.

GNCTD

Transco

D1

D2

4. Equity and Debt in the successor entities, equal to the value of serviceable liabilities is issued in favour of the Holding Company. D3

the Aggregate Revenue Requirement (ARR) and tariff filing for 2001-02, DVB proposed such five-year T&D loss reduction targets to Delhi Electricity Regulatory Commission (DERC), but there was little appreciation of the need for a multi-year package and the DERC was unable to accept the proposal. Hence the solution was adopted, whereby efficiency improvement targets were pre-set for five years, and simultaneously legitimised, by making them (rather than a premium on the price of equity) the bidding criterion, and building them into the package by means of policy directions with incentivisation for over-achievement (as failure to achieve the target would entail substantial loss). The essence of the policy directions is, thus, simply to bind the Regulatory Commission to the bid levels accepted by the Government.

down. Moreover, there is little scope for generation competition. Such assistance was also a device for ensuring a common retail tariff throughout Delhi for this five-year period (which was necessary keeping in view the disparities in opening loss levels of different companies). Further, those areas with the high losses are also where the poorer people live, and you would necessarily have a much higher tariff in one of the three companies because of the much higher level of the losses there. This would, very understandably, be politically unacceptable. The total assistance was projected at Rs 3450 crores (34.50 bn) for five years; the annual quantum diminishing each year.

Transitional assistance: Transitional assistance was a key issue. It was necessary in this period to balance three things – realistic expected efficiency gains, possible tariff increases, and transitional assistance until the benefits of greater efficiency were realised. DVB’s losses in the last financial year of its existence amounted to nearly Rs 1200 crore (12 bn). Since efficiency improvements would make the sector viable only at the end of a five-year period, it would be necessary either to raise the tariff sharply or to provide interim government assistance for the transitional period. It made sense for the government to consider substantial assistance for the interim period to maintain tariff increases at reasonable levels. By retaining the singlebuyer model for five years, the government would assist Transco to supply power to the discoms at lower rates, the assistance being returned subsequently through loan repayments after the losses had come

Choice of objectives: privatise or corporatise? It is technically necessary to corporatise an SEB before it can be disinvested. The issue is the place in the reform scheme of this intermediary stage (of having government-owned companies) and consequently the time frame for disinvestment in relation to corporatisation. Should we try to create working, viable companies first, to establish the ground for a successful disinvestment? The experience so far does not appear to warrant it. In Delhi, the government was clear in its mind about the objective of privatisation, and it was possible to skip the intermediate stage of having government-owned distribution companies. Mere corporatisation was thought unlikely to address the real problem. Moreover, there was the danger of the political will relapsing with partial achievement. Shell companies had been created in advance, but the transfer scheme, whereby DVB’s assets and liabilities were

76

Reform structure: The process of unbundling and privatisation is compendiously depicted in Exhibit 5.

Power Sector Reforms in India: Distribution Reforms

formally vested in the government and then transferred to the successor companies, was operationalised simultaneously with disinvestment (from July, 2002).

Personnel Issues A major consideration in the early stages was the possible consequences of employee unrest, particularly during the summer months. DVB had a history of very strong trade unionism. It was a major advantage that union leadership and most employees came to understand that reform was inevitable, and also popular with the general public, and therefore opposing it would neither halt the process nor win any public sympathy. Rather, the leadership was able to see that cooperating with the reform process would help them secure their future interests. However, this atmosphere of cooperation could not be maintained without the assurance of job security. The Tripartite Agreement protected the existing employees from the threat of retrenchment and secured their existing conditions of service and committed GNCTD to establish a Trust for their retirement benefits; GNCTD contributed Rs 860 crores (8.6 bn) to make up the initial funding requirement of the Trust of Rs 1329 crores (13.29 bn), and guaranteed the Trust’s discharge of its obligations. In conclusion, the main lesson is that tariff increases, loss reduction through expected improvements in commercial efficiency, and transitional financial assistance, are the three crucial inter-related issues in privatising distribution from a situation of high opening losses. It is absolutely necessary to evolve a viable package that balances these considerations in a manner that is perceived as acceptable to all the stakeholders concerned, in the particular local situation. Secondly, reform cannot be left to regulatory institutions without sufficient government guidance and involvement for which policy directions existing in all Indian reform laws appear to be an appropriate instrument. Another lesson that might be drawn from the Delhi experience is that flexibility during the process may be necessary – assuming that the objective of privatisation is not lost sight of. And finally, the high level of public expectation has already become apparent and poses a major challenge to the new companies. What has been our experience since July 2002 when we handed over? One is, there has to be a focus on system change rather than on fire fighting, which we in the government are always doing. This means that it is going to take time to achieve results and for benefits to be tangible to the public. We have had our share of teething troubles with the introduction of new systems.

IIMB Management Review, March 2004

We also saw some residual issues, partly because of the simultaneous unbundling plus privatisation. There are contractual issues – no matter how carefully you draw up your documents, there are always points and issues that come up later that need to be resolved. If we had to do this over again, then we would probably draft our documents a little differently.

P K Kukde The key issues in the Delhi model were:
Government commitment and sustained support
Improved base-line data on loss levels.
A limited and circumscribed level of past period receivables
Full and continuing involvement of the Regulatory Agency
Rational and fair valuation of assets – valuation based on revenue stream
Clean balance-sheets assured to discoms
Material stocks, loans to personnel to be based on actuals after audit
16% ROE assured if agreed loss reduction benchmarks achieved. 17% reduction in the T&D losses are expected phase-wise in five years.
Continuing but reducing subsidy (via BST) on the scale of past losses incurred. The government has agreed to subsidise the generating company to the extent of Rs 3450 crores (34.5 bn) in five years.
Capital support from government via Accelerated Power Development and Reform Programme (APDRP), and from Power Finance Corporation (PFC).
Model structure assures returns which in turn makes entity creditworthy.

Post Privatisation Approach The approach adopted in Delhi has been multipronged and may be better understood against the time frame i.e. short term, medium term and long term: Short term goals: Consumer friendly measures, consumer grievance redressal, consumer education. Medium term goals: Revamp the billing and collection process to make it fast, efficient and consumer friendly; ensure that all meters work and put in place a credible

77

energy audit system and MIS; identify network weaknesses and implement a reliability improvement plan including refurbishing of test and repair facilities; rationalise the human resources by training and redeployment bringing in needed skills; and define and plan implementation of long term goals Long term goals: Harnessing technology to catch up with global standards – Web enabled consumer services for applications, billing, payment and information, major network replacements including conversion to LT less HV distribution system, Supervisory Control and Data Acquisition System (SCADA) and Geographical Information Systems (GIS) for improved reliability and up-time, shaping the human resources to be productive, honest, consumer oriented and rightskilled.

N M Rothschild & Sons, and the model we are working with is a hybrid model that has many features specifically tailored to Karnataka’s and India’s situation. Over four-fifths of our consulting effort is by Indian consultants. The design includes features of management contracts, concessions and full-scale unbundling of wires companies into a UK-type privatisation. The main thing in the distribution business is to tackle the risk. One of the things that I do as a transaction designer is to ask myself, ‘Would I put a quarter of my pension into this deal?’ And if the answer is ‘no’ then I know that something is wrong.

If you look at ground realities, the operating losses are growing. Tariff structures are not aligned with cost and tariffs are approximately 30% below cost recovery in Karnataka. There is uncertainty about the evoluSince the take-over of North Delhi by the Tata Power tion of tariffs. In the beginning it was suggested that Company, the short term goals listed above have been rural tariffs could be adjusted to reach a reasonable achieved. We have started revamping cost recovery basis, but now it would the substations and putting in mini- It was suggested that appear that the rural areas are not gomum standards of safety. We have ing to face continuing increases torural tariffs could be ward real full cost recovery. This noticed a change in the attitude of our consumers – they are becoming would be a matter of concern to invesadjusted to reach a friendlier. Energy audit systems have tors in the private sector. For the disreasonable cost been implemented. We are doing our tribution business as a whole, the conrecovery basis, but best to control theft, which is of a very dition of assets and the customer base profound nature in Delhi. The T&D is uncertain. In Karnataka, the govnow it appears that losses were 47.1% the day the reforms ernment has tried a number of pilot rural areas are not were put into place, and we have to initiatives and experiments during the reduce it by 17% in five years. The two years but none of them have sucgoing to face AT&C loss was 53.5% in Delhi. ceeded to expectation in enumerating continuing increases customers, in registering unregistered Our Capex plan for six years is eartoward real full cost connections and so forth. An asset marked at Rs 1249 crores (12.49 bn). analysis showed that the recovery — a matter condition Initially about 50% of the amount will condition of assets in the rural areas be spent in the first two years, and the of concern to of the state is much worse than in the remaining 50% will be phased out in urban areas. Most investors would be investors in the the last four years. With this capital concerned about this fact because it investment, Delhi hopefully should private sector. means you know relatively little about have a beautiful power supply. rural distribution assets but you will need to invest heavily in a market that may not supAllen Eisendrath port this investment. Customer resistance and theft are major problems. And we know that the experience The Karnataka Distribution Privatisation in other countries suggests that when you force cusDesign tomers to pay their bills, when you have universal metering, when collection effort is strong, people genI am going to go through the background and the builderally cut their consumption by one third. There could ing blocks of the privatisation transaction design that actually be a big shift in demand. Industry could increase we worked on with the government of Karnataka. The and rural areas might constrain their consumption. initial proposal to unbundle the urban areas separately from the rural areas was rejected by the government, Just to put a finer point on the idea that distribution is so we came back with a design that would fit an urbana risky business, when you see the figures for the cost rural combined zone approach. The consortium conto deliver in Karnataka, in the year 2001 (Exhibit 6) sists of Deloitte, CMS Cameron McKenna, IDFC and only 2% of the total cost is the profit associated with

78

Power Sector Reforms in India: Distribution Reforms

the distribution wires business. The rest, 23%, is the distribution cost, which you control if you are a distributor, but 70% generation and 5% transmission cost is not under your control. The point is that you do not control the majority of costs in the business. IDFC did a quick analysis using the financial restructuring plan and they found that if you change the assumptions in the business plans slightly (due to unexpected events such as droughts and so on), you quickly wipe out the return on capital base. If you start at a 16% return, and reduce the tariff growth by 1% per year due to decisions by the regulator, you have a negative 17% return on equity. You can wipe out your full equity in one single year if your loss level is somewhat higher at the start than was expected. As was noted in the Delhi case, privatisation cannot fix the sector immediately. It takes years to get tariffs to balanced levels. There are major information and service gaps. Even in Delhi, which had reliable metering, there are still some issues related to the quality of information. In Karnataka, only 37% of the load is metered, which means that you have a lot of uncertainty, particularly with agricultural consumption. Further, there is a need to reduce the uncertainty associated with the condition of assets, to find out who your customers are and whether your employees are actually going to work with you or against you. In many other privatisations that have been carried out, for example in the UK or Argentina, policy makers had raised tariff levels to a cost recovery point before the transaction occurred. This is not possible in Karnataka, which means that the investors are buying a business that has cash losses from the day they set foot in it. So there needs to be a transition period, as there was in Delhi.

Key Features of Transaction

Broadly, we are proposing isolating the wires business from the supply business as was done in the UK, Australia, Singapore and other places. This is a very conventional way to step into a restructuring programme. We are giving a reasonable transition period and setting the cost to enter the market at a minimum. We don’t want investors to have to pay a lot of money upfront because the investor is buying a business that, as it is today, has a negative value. Fundamentally, the profitability of the utility business is determined by the regulators and government policy. We want there to be a high probability of recovering the operating and capital costs in the transition years. And then we want it to automatically transition into a conventional incentive-based cost of service regulatory approach. The key elements are that the private operator owns and runs a distribution business and gets to earn a distribution margin during a five-year transition period. The distribution margin consists of two streams of earnings. A Base Revenue is an annual payment for each year of the transition. It will equal approximately 100% of the new operator’s operating cost, capital expenditure and minimum equity rate of return. The Base Revenue includes a set of rules and indices that ensure that the operator faces strong pressure to control operating costs.

Exhibit 6: Breakdown of Karnataka’s Electricity Cost

120% Percentage of Total Cost to Deliver

The overall design objectives of the Karnataka transaction are to move to a viable electricity sector as quickly as possible, allow for a reasonable transition period, and to give as strong incentives as possible to the private operators to reduce technical and commercial losses.

Giving strong incentives to reverse losses while improving the quality of service is the cornerstone of our design. We do not want them to lose a lot of money in the first years because the bankers and investors will not tolerate it. We want realistic rates of return and we want a fair risk allocation – an allocation of risks to those who can control and afford them.

100% 80% 60% 40% 20% 0%

IIMB Management Review, March 2004

123456789012345678901234567 123456789012345678901234567 12345678901234567890123 2%

12345678901234567890123 12345678901234567890123 12345678901234567890123 23% 12345678901234567890123 12345678901234567890123 12345678901234567890123 12345678901234567890123 5%

123456789012 123456789012 123456789012 123456789012 123456789012 12345678901270% 123456789012 123456789012 123456789012 123456789012

123 123 123

Distribution Profit

123 123 123 Distribution Cost 123 123 123 Transmission Cost

12 12 Power Purchase Cost

Type of Cost

79

Exhibit 7: Karnataka Transaction Summary: Estimating the Required Transitional Tariff Rupees Per Unit 3.07

Transition Date

Average Cost of Service Loss or Subsidy

2.03

AverageTariff

2003

2004

2005

2006

Then there will be an Incentive Charge, which is the paise per rupee billed and collected above a minimum collection requirement. In order to limit government subsidies, we advised the government to set a minimum collection requirement and to fix the transitional tariff so that they could estimate the gross revenues from the unit charge. The minimum collection requirement will be a percentage of the previous year’s collection, somewhere between 100% and 150%. Its specific level depends on how potential investors react to the design. Operators will have to meet specific service standards and provide specific information to the government and the Karnataka Electricity Regulatory Commission (KERC). This approach is different from the Delhi approach because given the rural area, with 40% of the load being agricultural and unmetered, the information is not nearly as good as it is in Delhi. The service standards are set at higher levels for each year of the transition period. There will also be penalties if the operators don’t meet their service standards and information requirements. In order to protect rural consumers, there will be wholesale and retail allocation rules which will allocate wholesale electricity between discoms, specify minimum supply to consumer categories and allocate electricity shortfalls and overages so that both the government and the operator are compensated for additional supply going to subsidised customers. Transition tariffs will be rebalanced and increased to cost recovery. After the five-year transition period , an incentivebased cost of service regulatory system will apply, using a multi-year tariff and independent setting of efficiency improvement requirements.

80

2007

2008

While it may sound draconian to say that supply to below average cost customers or rural areas will be limited and even cut off if necessary, we realised that we would just be continuing what the government is already doing to meet its financial constraints. Currently the Government of Karnataka (GoK) manages its rural feeders in order to keep its cash balances in order. It makes perfect sense to apply a similar set of rules under a privatised condition.

Exhibit 7 shows the process that Delhi is going through, that we are proposing for Karnataka. You start out with an average cost reserve of around Rs 3.07 and the average tariff in Karnataka is around Rs 2.03. Over a period of time you expect that tariffs will increase on the average in real terms, and you also expect that the operator will get more efficient and at some point those lines cross. The slope of those two lines is determined by two different factors – the slope of the tariff increases is essentially policy judgment based on what people can bear and what you think is a reasonable approach to tariff increases. The slope of the Average Cost of Service line is a function of the incentives of the private operator – mainly the incentives that they have to reduce their combined losses. The GoK will sell 51% of the equity share capital of the discom to the private sector. Equity will be offered at par value and it will be sold to the bidder who offers the lowest incentive charge. What does the government need to do to make this structure biddable? They need to set Base Revenue and minimum collection charge at levels that attract strong competition, fix retail tariffs and cap the return on equity. We are going to work with KERC to establish regulatory rules and procedures, both for the distribution margin (DM) period, and the post DM period, and we are also going to design a cap on physical supply to low tariff customers. We are going to advise the Government on these decisions and we will continue to help them run scenarios and build a comprehensive financial model of the transaction and its risks, but the government is the seller and will decide final offer terms. There are a number of challenging design issues that we have still to solve. One is, in setting base revenue, what are the reasonable levels of Capex and Opex?

Power Sector Reforms in India: Distribution Reforms

How well can the consultants and the government estimate the required Capex? One possibility would be to make Capex a regulated pass through. It is done worldwide. As to the setting of transitional tariffs, will politics or economics be the guide? We hope there will be some kind of a balance. Average rates will have to be on an upward slope.

the Karnataka Power Transmission Company Ltd. (KPTCL). KPTCL will set up a unit in the transmission company, and then be the single buyer for what is generated. Q: That means the distribution company has absolutely no risks. If they fail to collect, the government will give them the base revenue also.

Base Revenue can be set between 0% and 120% of AE: This brings me back to a fundamentally important estimated Opex and Capex. Its specific level depends point. We have been falsely accused of pushing all of on what investors require to enter the market. This the risk on to the government. Howdepends on perceived and actual risks ever I am trying to make a point here in the market. Obviously, it should be When the government that when the government sets the set at an ‘optimal’ level, one that attracts multiple qualified bidders, but sets the base revenue base revenue and the minimum collection charge, it actually determines the not so high as to eliminate incentives and minimum risk allocation. If the government sets to reduce losses and increase revthe base revenue at zero, the operacollection charge, it enues. tors would have to earn all of their aldetermines the risk One of the problems with our design is lowable charges out of their incentive that the longer the transition period, allocation. If the base charge. If that were the case, then the the more difficult it becomes to propoperators would have 100% of the colrevenue were zero, erly predict what base revenue should lection risk. On the other hand, if the be. Based on extensive discussion operators would have investors refuse to bid under those with potential bidders, the GoK and our to earn their allowable conditions, then the government may consortium members, we have conneed to set it at 90% of expected Opex charges out of their and Capex, or even higher than a 100 cluded that a five-year transition period is required. – allowing a small equity return in the incentive charge. base revenue. Then the operators baIt is important to have an effective in- Then they would have sically share their risks fairly evenly centive-based regulatory methodology 100% of the collection with the government. and procedure for the post-DM perisk. riod. The ‘X’ setting procedure is used Q: Where is the incentive for the disby the regulator to set the annual effitribution company to invest in infrastructure in this model? My question is not about imciency gains that are built into the multi-year tariff. proving the management of collection. A certain Orissa had major problems with that. In the Delhi situamount of political will to increase collections should ation the operators are probably concerned that the see us meeting the required targets. What is needed regulator will come out with difficult efficiency targets is a component where the distribution company is after the transitional period. This is a matter that needs forced to invest in distribution infrastructure where to be solved by the regulator. there is a probably not that significant a margin to improve. Discussion Q: You mentioned about the equity….Opex and Capex…the incentive and the cap on the profits. The operators actually collect a lot of revenue from the consumer, what happens to this money? AE: That is a good question. One of the good things about this model, from the point of view of private operators is that they are the ones who go out and collect the revenues. They have it in their own accounts and have significant control over them. There are clear rules as to how much they are allowed to retain. Every month they take out what they are allowed, and the rest is remitted to the single buyer –

IIMB Management Review, March 2004

AE: If you look at the Capex plans of the Delhi privatised operators, they emphasise two principal objectives. One is better service to customers, on the theory that happy customers pay their bills. And the second is improving technical efficiency, by replacing inefficient equipment, restructuring the network, putting in smart meters and so forth. Our view is that the government should get out of managing the electricity business and should let the private sector make its own judgments as to whether they want to put more money into commercial efficiency, or into technical efficiency.

81

And what we think is going to happen is that the private operator is going to optimise his total plan, not just the Capex investment, but the Opex also. That idea is fundamentally underpinning successful reforms that have occurred in the UK, Australia, New Zealand, Singapore and so forth. The idea is to let good management manage, give them overall efficiency targets and let them decide whether they want new transformers or more fault detection bands, or a better commercial billing system. Since the presentation was made, the design was completed and the Government of Karnataka is currently holding discussions on how to proceed. As we completed the design, several investors expressed strong interest in bidding on the distribution companies using the DM design. In addition, a series of modifications to the design have been completed to ensure that it is consistent with the requirements of the Electricity Act 2003 and to respond to concerns and suggestions raised by KERC.

T V S N Prasad

meter manually and record the reading in a meter reading book (MRB). There are thousands of such books, entries and meter readers. You don’t have access to or control over what he is recording. The data is then taken into a computer database (by a different agency), and from there the bills are printed and distributed. The whole process takes about 45 to 90 days depending on the efficiency of the system. In remote rural areas it takes more than 90 days. When your own process takes 45 to 90 days, you cannot pay the distribution costs on time. We introduced ‘electronic spot billing’ in the whole state of Andhra Pradesh – every consumer, rural or urban, gets his bill on the spot. We have introduced a small hand-held machine called the Common Meter Reading Instrument, which is 2½ kgs in weight and costs around Rs 11,000. It has an attached printer and a GSM modem – you can connect to the central server where the database is located, on the spot. It supports around 2000 records at one go. This in itself compresses the 90 day transaction to one day and the cash flow improves tremendously.

Andhra Pradesh: Interventions in Billing and Collection

With the GSM modem, we can see what the meter reader is doing at any point of time, as his entries get upThere are four power distribution companies in Andhra dated in the central database almost on a real time baPradesh — Central, Northern, Eastern and Southern. sis. A meter reader does 80 houses a day, on an averThe Central Power Company has an annual turnover age, so if 80 people get their bill on one day, even the of Rs 3600 cr (36 bn)and serves around due date gets staggered, cash collec5 million consumers. The geographition becomes much easier as there are In Andhra Pradesh cal area spread is around 90,000 no long queues for payment. square km with around 1.5 lakh km of every consumer, rural the GSM modem, we have also LT line, and 71,000 km of 11KV line. or urban, gets his bill With started on the spot payment on a pilot I would like to speak on a few interon the spot. A small basis, which hooks to the server of the ventions that we made in billing and bank, checks the balance of hand-held machine relevant collection. We focussed on the techthe consumer and payment can be nology and the process. Transmiscalled the Common made within 30-40 seconds. sion or distribution companies are not Meter Reading Once you have a database, other comgiven power on credit by the generamercial services like issuing new contor — most of the PPAs stipulate that Instrument nections, change of title and so on bedistribution companies get a rebate compresses the 90 come easy. The first step in the reform only if they pay within three days. After three days, the penalty sets in. day transaction to one scheme is to have a database, which means networking. Updating through But the practice in our country is to day and the cash flow spot billing machines and maintenance grant two months’ credit to domestic of the database on the net is imporconsumers. The cash flow has not improves tant. All our consumer records are on been designed to make prompt paytremendously. the Internet – all 5 million consumers. ments to generators. There is an Data can be accessed any time and gets inbuilt inefficiency in the billing proupdated automatically. Such transparency brings in cess itself. However, once you have started unbunaccountability and is also a helpful management tool, dling and start focussing on issues, certain intervenhelping both the consumer and the administering autions are possible. thority. Many people are now actually paying their The general practice is for the meter reader to read the bills through the Internet. It is very easy because we

82

Power Sector Reforms in India: Distribution Reforms

have tied up with almost all the banks in Andhra Pradesh. We have taken cognisance of possible privacy issues, and have given separate passwords and log ins to consumers, users and employees. We have tied up with our Commissioner, Commercial Taxes and devised electricity indices for different products and industries. Such integration between government departments helps immensely and everybody benefits. We have a very good intranet as well, where we have a host of other data. We encourage our employees to log in and fill in performance reports online. We use our IT facilities to deal with service issues such as meter change, door lock, meter not existing and so on and drill things down the company, to the circle, the division, the sub-division and then to the crew.

enabled and the chief minister too keeps watching this website and coming back to us with questions. We had also teamed up with ‘e-seva’, an integrated citizen service centre — a single stop solution for all facilities — people can pay their water bills there, buy cinema tickets, furnish income tax returns, and renew their driving licenses. There are about 54 of them in Hyderabad now. Since outsourcing was proving very expensive, we have developed in house training in our call centres. Our call centres service rural areas as well, and people call them using a cellular number. Asset tracking is one area where we have been poor. So we have designed two software models, the Transformer Information Management System and the Meter Information Management System, both intranet based software models to keep track of assets. All transactions are recorded on it and with the latter system, every meter is kept track of. Side by side we must carry out reforms like changing to electronic meters from mechanical ones.

Coming to the details of our energy audit, by using the RS 232 port and ordinary telephone modems, we can access the meter boards of a certain industry, a sub-station or a trans- Through Geographical former. We read the sub-station meter Information System and the consumer meter together, and superimpose the gaps. Since it records (GIS) mapping, all of data in such detail such as, ‘Power Hyderabad has been failed on 29th December 2002 at 1646 — At 1714 it was restored’, we can sat- digitised, every single isfy complaints and rectify problems. transformer, every The RS 232 port can be used for peak load management, to regulate the single pole has been phases of supply, and in customer diligence. All the consumers are automati- coded. Trouble calls cally graded by the database into vari- are now attended using ous slots or quadrants — the amounts the GIS technology. outstanding, the defaulting consumers, the defective meters are classified. Fuse off calls have

We have an accounting software, which helped us get our trial balance on the 2nd of every month, for every unit, whether at the section officer level or village level. In order to improve benchmarking and deal with other problems including that of fictitious billing, we have divided domestic consumption into slabs. Last year 60% of the domestic consumers fell in the first slab of 0-50 units, now the percentage has come down to 52. Above 300 units, we had only 2% of 5 million and now it has grown to 7%. This speaks of better metering. When you don’t have metering for agriculture, then you have to benchmark.

A distribution company can harness been replaced with the power of IT very well. Through equipped vans. Geographical Information System (GIS) Finally, our system of performance mapping, all of Hyderabad has been monitoring. Every employee has to fill a form recorddigitised, every single transformer, every single pole has been coded. Trouble calls are now attended using ing his/her performance before being paid his/her salthe GIS technology. Fuse off calls have been replaced ary. It is not about achieving targets but keeping a with equipped vans. Hyderabad for example, has 47 record of performance. Moreover, the system checks such vans and has been divided into 47 zones. Using all claims registered. We grade every employee as per cellular phones, we are able to instruct the vans about prior weightage. Again, IT can be used very well for which pole or structure they have to attend to. Thus, performance tracking of large utilities. So many IT we can and do address people very quickly. Our call interventions were possible because we decided to centre is also on the Web, so people can actually see have our own mail server and our own intranet. I would how many calls were received. strongly recommend that reform should start with having your own IT backbone. Our customer service centres in Hyderabad are web

IIMB Management Review, March 2004

83

How do you solve the issue of power supply to agriculture? By evolving the right policy, through political will, consensus, metering, but also the upgrading of technology. Our regulator has started giving incentives to people who use demand side management — people who go for frictionless foot-valve, PVC pipes, ISI mark pump sets and so on. Metering the pumpsets will not only bring the consumption down but also help determine the technical loss accurately. Lower capacity transformers will also help bring down losses. Side by side, as part of the effort to upgrade technology we have patented the ‘sub-station logic controller’, a small machine which works on a microprocessor. This machine trips the feeder automatically once the established duration for power supply is over. It functions like an electronic logbook and affords us complete supervisory control — you can switch off 11 KV feeders from the central level corporate office, and use it as a load dispatch centre when you have grid shortages. Further, we have teamed up with International Water Management Institute and the International Crop Research Institute (ICRISAT), in Hyderabad, and are working out packages for small areas. We are trying to ask people to shift to dry crops and commercial crops like Bengal gram, and groundnut, instead of paddy. There are several such possibilities which could be explored. High voltage distribution systems reduce current consumption and loading, improve voltage, and eliminate theft. This is the basis along which the developed world has designed its distribution system. Hyderabad is being converted to high voltage distribution system. With the very useful Supervisory Control and Data Acquisition System (SCADA) system in Hyderabad, we have remote control over every inch of our electricity network in Hyderabad. Cities with overhead

84

lines can convert the existing LT line into a HT line, at a low cost. The High Voltage Distribution System (HVDS) will start paying back in 3 months. We have also come up with a very down to earth commercial parameter to measure loss since the argument is that with electricity, nobody knows what loss is. We calculate the AT & C loss. Right now, including the subsidy, we are breaking even, for some months we did have cash profits. But, come drought, the calculation would change and we would have to go for extra subsidies. This is a continuous process. Two reasons for our high technical losses are: high ratio (1:6) of HT to LT lines and, higher capacity (100 and 63 KVA) transformers instead of smaller capacity transformers. The size has resulted in very long LT lines leading to bad voltage and subsequent high losses. Allen Eisendrath is with Deloitte Emerging Markets. [email protected] P K Kukde is Executive Director, Tata Power Company. [email protected] TVSN Prasad was Chairman and Managing Director, Andhra Pradesh Central Power Distribution Company when the Workshop was held. [email protected] D V Ramana is Professor, Xavier Institute of Management, Bhubaneshwar. [email protected] Jagdish Sagar is Chairman and Managing Director, Delhi Transco. [email protected]

Reprint No 04105 c

Power Sector Reforms in India: Distribution Reforms