Electricity is the wheel that drives all facets of our daily lives such as production, consumption, communication, transportation and routine financial transactions. Pakistan’s electricity sector has long been facing unprecedented challenges. It seems to be stuck in a bad equilibrium characterised by costly energy production, overloaded infrastructure, high transmission and distribution losses, extraordinary load shedding, and mounting circular debt. The economy as a whole suffers from unreliable and poor quality of electricity provision.

Despite the attention given to this sector by every government, why have we not been able to emerge from this bad equilibrium? A major reason for the continued woes of the electricity sector is that past reforms have been short-sighted and typically provided band-aid solutions. In 1994, the government allowed private sector entry into generation and subsequently unbundled the monolithic Wapda in order to improve operational efficiency of generation, transmission and distribution. In 2005, electricity provision in Karachi was privatised to bring about a financial turnaround, and a vertically integrated private utility now known as Karachi Electric (KE) was created. Under CPEC agreements, a large amount of generation capacity was added without upgrading transmission and distribution. None of these reforms were successful in bringing about the envisioned improvement in efficiency or service delivery. Rather we continue to experience an ever-rising cost of provision and never ending load shedding.

What do these experiences tell us about the future proposals for reforming the electricity sector? One clear lesson is that unbundling and privatising different parts of the sector will not improve performance, unless we can put in place a competitive regulatory market structure governing sale of electricity by generation companies to distribution companies and from distribution companies to consumers. Under the present regulatory system, generation companies and distribution companies (whether public or private) are compensated on the basis of cost-plus model, with guaranteed margins. This creates a moral hazard problem, that is, these companies do not have sufficient incentives for improving operational efficiency. The government and ultimately the taxpayers bear the brunt of difference in actual cost of provision and recoveries. Future reform efforts must address this by bringing about more competition among generation companies for lowering cost of production, and among distribution companies for lowering the retail cost of provision.

While the challenges in the generation sector are hugely important, in the remaining part of this article, we focus on the distortions in the distribution sector, drawing upon findings from our recent research. We use two features of the distribution system to summarise performance: (i) technical and distribution (T&D) losses which is the fraction of electricity lost due to technical (engineering) reasons and unmetered consumption or theft, and (ii) revenue recovery ratio which is the fraction of bills paid by consumers. Figure 1 and 2 show the trend in these measures over the past eight years (data for these two figures are drawn from Nepra reports).

Except for three out of four distribution companies in Punjab (Fesco, Gepco and Lesco) and for IESCO, T&D losses for all other companies in the country are above 15%. In 2019-20 these losses were highest at 38.9% for Pesco, followed by 36.3% for Sepco. The best performing companies (Fesco and IESCO) have losses below 10%. Revenue recovery rates are above 94% for all Discos in Punjab, but fall significantly as we move to Discos in Sindh, Balochistan and KP. Together, the electricity lost and unpaid bills translate into billions of rupees of losses for the Discos every year. No wonder, they are unable to make a sizable investment in the upgradation of infrastructure.

Furthermore, we observe wide dispersion in T& D losses and revenue recovery rates even within distribution companies. For eexample, using monthly feeder line data from Karachi Electric, we find that areas such as Defence, Clifton, Bahdurabad have T&D losses below 15% and revenue recovery rate of 94%, but areas such as Surjani, Korangi and Landhi lose 30% electricity due to T&D losses and experience revenue recovery rates of 80%. Such wide dispersions in performance cannot be attributed to the poor infrastructure alone. They partly reflect the variations in customer profiles, usage and payment behaviours, and the norms related to the acceptability of using electricity without paying for it.

There is visible evidence of electricity theft in large cities especially in the provinces of Sindh, KP and Balochistan. Although it is difficult to derive a precise estimate of the fraction of electricity lost due to theft, T&D losses above the level of the best performing distribution company provide a reasonable approximation. The usual methods of stealing electricity involve meter tampering and illegal hooking of wires (kundas) to the main line from the pole mounted transformers (PMT) leading to customer meters. Such sights of illegal connections are quite visible in the densely populated areas of the cities with high T&D losses.

What has been done to improve distribution performance? Discos have faced pressure from the regulator to bring down T&D losses and improve recoveries, which is reflected in falling T&D losses and rising recovery ratios for most companies. The segmented loadshedding policy which was introduced first by KE was later adopted by all the Discos, implements increasing load shedding based on past feeder line losses. Although this policy has saved money, it is often perceived as unfair by the customers which creates significant resentment towards utilities. There has also been a push to implement technological solutions such as aerial bundled cables (ABC), smart-phone metering, and smart meters to make it difficult for customers to use electricity illegally. However, these solutions are costly to implement, often require donor assistance, and face resistance both from employees and customers, who quickly find new ways to circumvent these technologies.

However, past reform efforts have not paid attention to understanding customer behaviour. Why are customers unwilling or unable to pay for the electricity they use? In this context, we are evaluating the results of an ongoing programme of Karachi Electric which is underway in the high loss IBCs of Karachi. The high loss areas of KE such as Korangi, Landhi, Surjani and Orangi cover approximately 40% of the metered customer base. Since 2018, these IBCs have an average T& D loss of 30% and revenue recovery rate of 80%.

In November 2019, KE initiated the ambitious Project Sarbulandi in six of the twelve high loss IBCs to improve their performance. In addition to increasing the coverage of ABC cables in areas previously steeped in kunda culture, KE introduced several innovative and complementary programme features, such as counselling of local influentials to gain their support, gaining confidence of people to meter the unmetered, community engagement and facilitation for winning the trust of customers. These initiatives were enabled by empowering IBC management with authority and resources, to determine the best strategies based on local conditions, and then holding them accountable through close monitoring.

Project Sarbulandi delivered phenomenal results in the first four months of its rollout, especially in improving bill recovery. In Figure 3 and 4 we show the T& D loss and recovery ratios of Sarbulandi IBCs relative to the remaining six high loss IBCs. When the Covid-19 pandemic hit in March 2020, T&D losses rose in all high loss IBCs but less so in Sarbulandi IBCs. There were no observable differences in recovery rates, despite Sarbulandi IBCs having brought in a large number of previously non-paying customers into the metered network. These initiatives have now been scaled to remaining high loss IBCs of Karachi. While part of Sarbulandi success appears to be due to improved infrastructure which makes it difficult to steal, it remains to be seen to what extent the programme worked by changing the willingness of customers to pay for electricity. This is the next step in our research, which will involve conducting customer surveys to understand the factors driving illegal electricity usage and bill payment behaviours in these areas.

The experience of Project Sarbulandi illustrates the promise of complementary approaches to dealing with the challenges faced by the electricity distribution sector. It also highlights the need for identifying and addressing demand side distortions which are one of the fundamental causes of high losses in this sector, in a data-driven manner. Customer behaviour will ultimately determine the effectiveness of any reform programme. Unless we find ways to break the culture of theft and non-payment, those who want to steal will find new ways to do so.

The Analytical Angle is a monthly column where top researchers bring rigorous evidence to policy debates in Pakistan. The series is a collaboration between the Centre for Economic Research in Pakistan and Dawn.com. The views expressed are the authors’ alone.