Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions

ELECTR-6200; No of Pages 18

Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions

Fabian B. Lewis is Director of the Research and Analysis Unit at the Ministry of Finance and Planning, Jamaica. He received his Ph.D. in Economics from the University of Manchester, UK, in 2006. His research focuses primarily on applied public finance, empirical international trade, energy supply analysis, the development of labour market information systems and productivity measurement. The author gratefully acknowledges constructive comments by the Editor, two anonymous referees, officials from the Jamaica Public Service Company, Ian M. Scarlett, Clifford G. Williams, and various participants at selected conferences and seminars. Nevertheless, the author is solely responsible for all the results and, consequently, the views are not necessarily those of the institution to which he is affiliated.

August/September 2015,

Vol. 28, Issue 7

An investigation of the link between electricity theft and power disruptions simultaneously provides a rare and contemporary cross-country assessment of electricity theft. The findings suggest that electricity theft in Haiti ranked first overall, and that extensive power theft has risen internationally since 1980. The direct cost of electricity disruptions ranged from US$0.43 to US$9.91 per kWh. Fabian B. Lewis

I. Introduction Rampant electricity theft is a worldwide predicament even among developed countries. Power theft is manifested in several ways, including illegal connections, non-payment of bills for electricity utilized and fraud on the part of some employees of utility companies. Despite the pervasiveness of electricity theft, there is a surprising dearth of empirical studies that provide any comparative assessment of the issue in an international context or

even with emphasis on a particular country. To our knowledge, Smith (2004) is the notable exception. However, while Smith (2004) specifically explored the issue of electricity theft in a global context, the author did not simultaneously estimate the damage caused by power disruptions. Consequently, one of the main contributions of this article is that we attempt to reduce this gap by providing an update and extension to Smith’s (Smith, 2004) rare comparative analysis of power theft globally.1 Additionally, we examine the

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

1

ELECTR-6200; No of Pages 18

evolution of electricity theft internationally while estimating the direct cost of power interruptions to Jamaica and 95 other countries. lectricity is a key input in many production processes and hence its continuous supply is crucial (Jamil, 2013). Theft of power is a major challenge because it affects the distribution of electricity from power companies by overloading/short-circuiting their systems. This often results in a disruption (i.e. a partial or complete loss) in the electricity supply to legitimate customers. As a result, curtailing power theft is an important policy objective for power utility providers. In this article, our general research problem is as follows: Electricity theft partly by ‘‘throw-ups’’2 (also called ‘‘spider webs’’) sporadically cause power systems to overload and lead to power disruptions (‘‘blackouts’’ and ‘‘brownouts’’).3 Such interruptions in turn, result in inter alia a loss of productive output. In other words, the central theme throughout our article is that electricity theft is costly and contributes to power disruptions, which in turn lead to costs for an overall economy and its various sectors. In our research, we use Jamaica as an interesting case study on power theft because its theft level is relatively high compared with other countries and the island is also frequently affected by power interruptions.4 Reliable electricity supply is crucial because the world is becoming increasingly dependent

E

2

on electronic devices connected to the grid (Coll-Mayor et al., 2012). Disruptions in electricity supply (whether planned or unplanned) are undesirable because they lead to inter alia a loss in value that would have otherwise been created. During an electricity disruption, the output of some firms for example, is directly lost when the production process is halted. This value of lost load (VoLL) essentially represents the

These estimates are useful in particular because they can help to approximate the direct losses faced by the respective economies due to electricity disruptions. losses per hour of electricity not supplied (Coll-Mayor et al., 2012) or alternatively, the cost of unserved energy (Bose et al., 2006). Estimating the value of lost load is important because power supply interruptions can have adverse consequences especially to sectors that are extremely reliant on electricity to produce output. Additionally, estimates of the VoLL by sector can assist power companies to determine priority areas for disrupting power when supply shortage arise (de Nooij et al., 2007). he major objectives of this article are to (1) analyze the

T

phenomenon of power theft over an extended period, (2) present a comparative analysis of electricity theft in an international context and (3) estimate the value of lost load due to power disruptions for Jamaica and several other economies internationally. Hitherto, no study, as far as we are aware, simultaneously provides a detailed comparative analysis of electricity theft globally while computing the value of lost load (as a result of power disruptions) for almost 100 countries. As mentioned earlier, these estimates are useful in particular because they can help to approximate the direct losses faced by the respective economies due to electricity disruptions. While de Nooij et al. (2007), Leahy and Tol (2011), Coll-Mayor et al. (2012), and Linares and Rey (2013) all estimate the value of lost load due to power interruptions, they do not specifically explore the issue of electricity theft in an international context. As a result, our research presents a potentially useful contribution by addressing this lacuna. Approximately 97 percent of the Jamaican population has access to electricity. The Jamaica Public Service Company Limited (JPS); Jamaica’s national light and power company, is the only licensed distributor of electricity in the island, though it also engages in the generation and transmission of power.5 It also purchases electricity generated by independent power providers (IPPs) under various power

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

purchase agreements (PPAs). At the end of 2014, the JPS supplied some 594,196 customers, of which 531,363 (or approximately 89.4 percent) were residential consumers and the remaining 62,833 classified as commercial and industrial customers. Before the formation of the JPS in 1923, Jamaica first received electricity in 1892, at a time when the country was supplied with power by the Jamaica Electric Light Company. JPS was subsequently privatized in July 2001 when the government of Jamaica (GOJ) sold its majority holding in the company for some US$201 million. Currently, Marubeni Corporation of Japan and Korea East West Power each own 40.0 percent of the JPS, with the remaining 20.0 percent stake jointly held by the GOJ and a few individual shareholders. JPS is regulated by the Office of Utilities Regulation (OUR) through the Amended and Restated AllIsland Electric Licence (2011). Since 1995, Jamaica’s OUR has been responsible for regulating Jamaica’s electricity sector (including regulation of the JPS

and independent power producers) under the Office of Utilities Regulation Act. The OUR is also charged with setting rates and maintaining guaranteed service standards for the regulated entities and reports to the Jamaican Cabinet. amaican customers currently pay an average of US 25 cents per kilowatt-hour (kWh) for electricity. This relatively high cost of energy adversely impacts the competitiveness of firms operating in Jamaica (especially those in the manufacturing sector) when one considers, for example, that regional competitors such as Suriname and Trinidad and Tobago are producing electricity at roughly US 5 cents per kWh and US 6 cents per kWh, respectively. Overall, total electricity generated in 2014 amounted to some 4,107,457 MWh6 while 3,012,979 MWh were sold by the JPS to customers (Figure 1). Based on data from the JPS, its ‘‘system line losses’’ (as a percentage of net generation and purchases) stood at 26.6 percent as a result of both

J

technical losses and non-technical losses,7 of 9.6 percent and 17.0 percent, respectively. ‘‘Losses’’ simply refer to electricity supplied but was not billed (presumably lost to theft, billing irregularities, and technical losses) (Min and Golden, 2014). In 2003, JPS’ overall system line losses were estimated at approximately 18.6 percent, meaning that almost one-fifth of the total electricity produced/ purchased by the company in that year was lost. By the end of 2014, the proportion of total electricity generated that was not billed rose to 26.6 percent. For the overall period 2003–14, the JPS’ average annual system losses as a proportion of net generation and purchases was roughly 22.8 percent. The article continues as follows: Section II reviews the most widely used methods for assessing electricity theft and measuring the costs of power disruptions. In Section III we present our results relating to electricity theft in a cross-country setting as well as our computations of the direct cost due to power disruptions.

Figure 1: JPS’ Electricity Produced, Electricity Sold and System Losses, 2003–2014 Source: Compiled by author using data obtained from Jamaica Public Service (2008, 2009, 2014a,b, 2015).

August/September 2015,

Vol. 28, Issue 7

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

3

ELECTR-6200; No of Pages 18

Section IV discusses some resulting policy implications for Jamaica while Section V concludes.

reading on an electrical meter (or turn a blind eye to an illegal connection). Unpaid bills occur when customers deliberately refuse to pay amounts owed for electricity previously utilized.

II. Material and Methods A. Manifestation of electricity theft Not all electricity generated is transmitted or distributed to final consumers because some will be lost during the various stages due to technical and non-technical electrical inefficiencies. In other words, some amount of electricity theft is inevitable (Min and Golden, 2014). There are at least four ways in which electricity theft is manifested: fraud, directly stealing power, billing irregularities, and unpaid bills (Smith, 2004; Depuru et al., 2011; Winther, 2012). All four aspects are related in the sense that they all result in revenue loss for the utility companies. Fraud, in this context, virtually involves efforts by consumers to tamper with electrical meters in order to dishonestly lower their power consumption levels, while directly stealing electricity involves, for example, connecting throw-ups on a distribution feeder or bypassing a meter in order to illegally direct electricity to a premises. Billing irregularities often involve collusion between an employee of the utility company and a customer (Depuru et al., 2011; Winther, 2012) to dishonestly report a lower than actual usage 4

B. Consequences of electricity theft Electricity theft has at least four costly consequences for both utility companies and their

Power theft adversely affects the amount of re-investment and employment that can take place in the electricity sector by curtailing potential revenue that can be collected by utility companies. genuine customers. First, it unnecessarily raises the price of power for legitimate consumers because utility companies are generally forced to pass on both the costs of energy lost (due to theft) and the expenses incurred for additional maintenance of their distribution systems. Second, power theft adversely affects the quality of electricity supply by overloading the system, which often leads to intermittent power interruptions (i.e. power outages) for both paying customers and illegal consumers, loss in output and damage to electrical appliances (Tasdoven et al., 2012;

Tishler, 1993; Mwaura, 2012). Power outages caused by electricity theft are especially costly to firms as they often stop vital production from taking place. hird, power theft also adversely affects the amount of re-investment and employment that can take place in the electricity sector by curtailing potential revenue that can be collected by utility companies. Such constraints limit the amount of money available to fund the development/expansion of generating capacity, and therefore contribute to load-shedding i.e. cutting power to selected customers when there is shortage of electricity relative to demand. Consequently, utility companies are often forced to apply for higher electricity tariff8 rates from their regulator in order to maintain viability in supplying power, and this leads to higher electricity prices for legitimate customers (Jamil, 2013; Winther, 2012). Finally, power theft, especially via illegal connections, creates fire hazards and sporadically results in the death of power thieves or even unsuspecting persons who are inadvertently electrocuted after becoming entangled with illegally strung throw-ups.

T

C. Assessing electricity theft and estimating the costs of power disruptions The most widely used technique to gauge electricity theft is to embrace the extent of non-technical energy losses as a

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

rough proxy (Jamil, 2013). According to Joseph (2010) and Depuru et al. (2011), notwithstanding its limitations, this is the traditional and best available method for measuring electricity theft internationally. A shortcoming that is inherent in the technique, for example, arises because aggregated transmission and distribution (T&D) losses (i.e. losses that occur between the point of generation and delivery of electricity to customers) include both technical and nontechnical line losses, while power theft is mainly related to the latter. Nevertheless, there is a consensus that T&D losses are a reasonable proxy of power theft in crosscountry comparisons despite differences in inter alia governance and electricity industry structures. Moreover, data on such losses are generally not disaggregated (into its technical and non-technical components) and technical energy losses tend to be relatively miniscule (Jamil, 2013; Tasdoven et al., 2012). As a result, in exploring the issue of electricity theft across countries, we inter alia compared the extent of T&D losses for Jamaica with the corresponding indicator for 91 other countries worldwide. here are four commonly used methods to measure the costs of electricity interruptions in the literature9: (1) the ‘‘production loss’’ approach, (2) customer surveys (the ‘‘willingness to pay’’ (WTP) or ‘‘willingness to accept’’ (WTA))

T

August/September 2015,

Vol. 28, Issue 7

method, (3) ‘‘captive generation approach,’’ and (4) ‘‘case studies.’’ The production loss approach directly computes the value of production loss for each unit of power outage (Bose et al., 2006). In other words, the technique attempts to quantify the loss in production due to the unavailability of electricity supplied by the grid. According to this method, the value of lost load is typically calculated by dividing

The ‘willingness to accept’ technique ask respondents how much they would be willing to accept as compensation (via lower electricity bills, for example) for suffering more power outages.

gross value added (for the overall economy or a specific sector) by the amount of electricity consumed. Two major advantages of this approach are that it is relatively easy to calculate and the required data are usually readily available. The VoLL formula based on the production function approach is denoted by (1). VoLL ¼

Gross Value Addedi Total Electricity Consumedi

power interruptions to the country’s overall economy or sector i. he customer survey approach using the ‘‘willingness to pay’’ method involves for example, asking consumers how much they would be willing to pay for a reduction in power interruptions (i.e. achieving a more reliable and uninterrupted power supply). On the other hand, the ‘‘willingness to accept’’ technique ask respondents how much they would be willing to accept as compensation (via lower electricity bills, for example) for suffering more power outages (Coll-Mayor et al., 2012; Hensher et al., 2014; Woo et al., 2014). On the other hand, the ‘‘captive generation’’ technique provides information on the cost of alternative power generation (for example, via captive power plants11 or backup generators) that ensures a reliable supply of power. Finally, the ‘‘case study’’ method involves using data on an actual historical power outage to quantify the cost of that interruption. However, a major limitation of this last technique is that it is not typically representative of power disruptions, thus making it very difficult to generalize the results (de Nooij et al., 2007; Linares and Rey, 2013).

T

(1) where VoLL > 0. Note that the higher the VoLL (due to relatively low levels of electricity usage relative to value added,10 for example), the greater the cost of

D. Data and methodology In our cross-country analysis of power theft, we utilized data extracted from the JPS’ Annual

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

5

ELECTR-6200; No of Pages 18

Reports and the World Bank’s (World Bank, 2014) World Development Indicators database. Specific emphasis was placed on examining the extent of electricity theft suffered by the JPS over the period 2003–2013; a period for which rare disaggregated electricity data were available for Jamaica. Similar to some researchers such as Jamil (Jamil, 2013), we utilized the extent of non-technical losses as a proxy for electricity theft. According to the JPS, of the 26.6 percent system losses registered in 2014, approximately 17.0 percentage points represent non-technical losses. Total systems losses time series data were not originally disaggregated but after receiving clarification from JPS officials, we were able to impute the extent of non-technical system losses for the remaining years of the period under consideration. This information was complemented by cross-country data on electric power transmission and distribution losses as a percentage of output extracted from World Bank (2014). Overall, we are able to create a dataset covering some 100 countries.12 s it relates to quantifying the direct cost of power disruptions, we first calculated the value of lost load for Jamaica using detailed gross value added data obtained from the Statistical Institute of Jamaica and aggregate electricity usage13 information sourced from the Jamaica Public Service Company. In the case of the other countries, we estimated

A

6

their respective VoLLs using gross value added and electricity usage data garnered from the World Development Indicators database. The Value of Lost Load was computed as the total value added divided by total electricity consumed as defined by (1).14 Since we want to make a meaningful comparison between Jamaica’s VoLL and the identical indicator for the other countries, we converted all our value of lost load estimates into U.S. dollars using exchange rate data obtained from the Bank of Jamaica. Our choice of the production function method, from the available measures for estimating electricity interruption costs, was based on weighing carefully the various arguments about which technique provides the most accurate insights on the overall consequence of power disruptions for the economy but at the same time is relatively easy to compute, interpret, and compare internationally.15 The willingness-to-pay method based on a survey for example, would pose a challenge in the Jamaican context because of the high cost of conducting such an inquiry as well as the level of subjectivity that would be involved. Moreover, our choice of technique was done with a specific research objective in mind, which requires an empirical estimation of the value of lost load for inter alia Jamaica and its various economic sectors in recent years. espite the relative strengths of the production function

D

approach, it is not without limitations. For example, it generally assumes that during electricity disruptions, production activities must be completely halted and therefore cannot be shifted to alternative time periods or activities (such as construction and agricultural) for which electricity is inessential (Leahy and Tol, 2011). This assumption is not always valid because some firms can adopt the use of backup power sources (such as generators or batteries) and continue production during an electricity interruption, especially if such outage was previously announced by the utility company (i.e. scheduled outage). As a result, the value of lost load estimates might be overstated for some sectors (Linares and Rey, 2013).16 The results of our VoLL calculations and a corresponding analysis of the findings are presented after our general findings on electricity theft.

III. Results A. Electricity theft in Jamaica Kingston (i.e. Jamaica’s capital from its 14 parishes) is the main area where power theft is prevalent (Figure 2). In fact, 35.0 percent of the estimated 13,425 MWh to 15,915 MWh of total power stolen island-wide during May 2014 was in the corporate area i.e. Kingston and St. Andrew

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

Figure 2: Jamaica’s Transmission and Distribution Losses and Election Years, 1971– 2011. Note: General election years Source: Compiled by author using data obtained from World Bank (2014).

(with 25.0 percent accounted for by Kingston alone). The two parishes have a combined population of 666,041 (or 24.6 percent of Jamaica’s overall population of approximately 2.71 million). Most of the electricity theft in Kingston occurs in relatively poor inner-city communities.17 lectricity theft is also high in the parish of St. Catherine (21.0 percent) but the lowest level of electricity theft (i.e. 1.0 percent of the total electricity stolen) was recorded in the parish of Portland. In sum, based on the data, it appears that the magnitude of electricity theft is generally highest in the most populous and urban parishes of Jamaica (Table 1).

countries. The major rationale for this segment is to inter alia gauge Jamaica’s electricity theft relative to other nations to determine whether it is above or below average. Particular emphasis is

This section analyzes electricity theft in Jamaica vis-a`-vis the extent of power abstraction in other August/September 2015,

Vol. 28, Issue 7

B.1. Electricity theft between 1980 and 2011: a cross-country comparison Before directly proceeding to the cross-country comparison, we briefly examined the trend19 in

Table 1: Jamaica’s Electricity Demand–Supply Profile, 2003–2014 Year

Electricity Produced

Electricity Sales

System Losses

Non-Technical Lossesa

(MWh)

(MWh)

(%)

(%)

2003

3,696,005

3,009,461

18.6

8.0

2004 2005

3,717,022 3,877,990

2,999,639 3,055,154

19.3 21.2

9.3 11.2

2006

4,046,428

3,120,669

22.9

12.9

2007 2008

4,078,771 4,123,290

3,131,494 3,179,078

23.2 22.9

13.2 12.9

2009 2010

4,213,980 4,137,350

3,203,878 3,187,488

24.0 23.0

14.0 13.0

2011 2012

4,136,879 4,135,919

3,215,990 3,133,966

22.3 24.2

12.3 14.2

2013

4,141,644

3,069,688

25.9

15.9

2014

4,107,457

3,012,979

26.6

17.0

Mean

4,034,395

3,109,957

22.8

12.8

Median

4,115,374

3,126,082

23.0

12.9

E

B. Jamaica’s electricity theft in an international context

placed on nations for which reliable data are available and those that facilitate a contemporary update and extension to Smith’s (Smith, 2004) research. Smith provided a useful comparative assessment of electricity theft involving a sample of some one-hundred and two countries and data for the years 1980 and 2000 and concluded in general, that power theft had climbed overall in most regions worldwide (Smith, 2004).18

Source: Compiled using data extracted from Jamaica Public Service (2008, 2009, 2014a,b, 2015). Notes: a Imputed for all years except 2003 and 2014.

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

7

ELECTR-6200; No of Pages 18

Jamaica’s electricity theft over an extended period for which independent and reliable World Bank data were available i.e. 1971–2011. Based on Figure 2, the extent of electricity theft in Jamaica rose overall to 18.4 percent in 2011, up from the 7.4 percent that was registered in 1971 (and the 14.7 percent recorded in 1980). he World Bank (World Bank, 2014) data on electricity transmission and distribution losses for some 100 countries for the years 1980 and 2011 are illustrated by Figure 3. It shows that based on the widely adopted proxy for electricity theft, the 18.4 percent in 2011 for Jamaica was above the simple average transmission and distribution losses of 12.2 percent (and a median of 10.4 percent) for all 100 countries in our sample. The extent of power theft ranged from a low of 1.8 percent in the Slovak Republic to a high of 54.6 percent in Haiti (Figure A1 in the Appendix). In 2011, Jamaica ranked 17th out of all the countries in terms of the magnitude of electricity theft. Additionally, the average electricity theft in all the countries increased slightly (i.e. by 0.6 percentage point) in 2011, up from the 11.6 percent registered in 1980. Based on the cross-country results, the extent of power theft in Jamaica is above that of other Latin American and Caribbean countries such as Ecuador (16.6 percent), Mexico (15.4 percent), and Peru (5.8 percent), but is some

T

8

Figure 3: Transmission and Distribution Losses in Selected Countries, 1980 and 2011. Note: Electric power transmission and distribution losses (% of output) include losses in transmission between supply sources and distribution points and in the delivery to customers (including pilferage). 2011 was the latest year for which T&D losses data were available Source: Compiled by author using data obtained from World Bank (2014).

0.3 percentage point below Nicaragua (18.7 percent), 1.3 percentage point lower than Venezuela (19.7 percent), and 1.7 percentage point lower than Honduras (20.1 percent). India, a country notable for rampant20 electricity theft, had relatively high transmission and distribution losses of approximately 21.1 percent in 2011 (i.e. some 2.7 percentage points higher than the level in Jamaica). everal countries experienced an overall increase in the magnitude of electricity theft between 1980 and 2011 (Figure 3). In fact, exactly half of the sample of countries registered increased power theft over the 31-year time span. Such nations include not only the Republic of Congo (up 37.0 percentage points to total 46.0 percent in 2011), Haiti (up 28.8 percentage points to total 54.6

S

percent in 2011), and India (up 3.2 percentage points to total 21.1 percent in 2011) but even a few developed countries such as Singapore (up 0.5 percentage point to total 5.3 percent in 2011) and Switzerland (up 0.5 percentage point to total 7.1 percent in 2011). Increased electricity theft in developed countries is a surprising finding because such a phenomenon is generally prevalent in poorer (less developed) countries.21 Power theft is also typically widespread in countries where electricity prices are relatively high, which motivates some consumers to steal power as the net reward for doing so is greater (Jamil, 2013). On the other hand, power theft declined in countries such as Bangladesh (down 25.0 percentage points to total 10.3 percent in 2011), New Zealand

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

(down 5.6 percentage points to total 7.0 percent in 2011), Sweden (down 1.6 percentage points to total 7.0 percent in 2011), and the Netherlands (down 0.1 percentage point to total 4.1 percent in 2011). here is a general consensus that efficient power systems typically have less than 6.0 percent overall transmission and distribution losses (Smith, 2004). As part of our analysis, we grouped the countries in our sample based on the extent of their transmission and distribution losses in 1980 and 2011. This was deliberately done in order to facilitate a comparison with Smith’s (Smith, 2004) results for the year 1980 and to extend the author’s findings to the latest year for which data were available i.e. 2011. Our results are presented in Table 2. Based on Table 2, the proportion of countries with ‘‘extensive power theft’’ (i.e. T&D losses of at least 16.0 percent) grew to a combined 23.0 percent in 2011, up from 17.0 percent in 1980. Our 2011 finding is significantly less than the 42.1 percent that was registered in 2000 by Smith (2004) who utilized an almost identical similar sample size of 102 countries. Overall, we can interpret our finding22 as suggestive evidence that not only has the incidence of electricity theft has risen23 for many countries over time but extensive power theft has also climbed since 1980. At the end of 2011, approximately 5.0 percent

T

August/September 2015,

Vol. 28, Issue 7

Table 2: Transmission and Distribution Losses for Selected Countries, 1980 and 2011 T&D Losses

No. of Countries

(%)

No. of Countries

(%)

Range (%)

In 1980

In 1980

In 2011

In 2011

1 to less than 4 4 to less than 11

2 55

2.0 55.0

8 45

8.0 45.0

11 to less than 16 16 to less than 20

26 7

26.0 7.0

24 11

24.0 11.0

20 to less than 55

10

10.0

12

12.0

100

100.0

100

100.0

Total

Source: Compiled by author using data obtained from World Bank (2014).

(i.e. comprised by the Dominican Republic, the Republic of Congo, Nepal, Haiti, and Iraq) of all the countries registered ‘‘high electricity theft’’ (i.e. in the magnitude of at least 30.0 percent) compared with the 2.0 percent (accounted for by Bangladesh and Mozambique) registered in 1980. Importantly, all the countries with high levels of electricity theft were relatively poor, which helps to explain their relatively high levels of power theft. It is instrumental to note that countries such as Albania, Myanmar, Nigeria, and Bangladesh, which (Smith, 2004) demonstrated to have high electricity losses in 2000, no longer have such ignominy, based on our overall 2011 results. C. Estimates of the direct cost of power interruptions to Jamaica The second primary research objective we want to fulfil is to quantify the direct economic impact of power disruptions in terms of lost production.24 Specifically, using a production

function approach, we attempt to provide original estimates of the value of lost load for many countries (including Jamaica) due to power disruptions. Our VoLL results for the total Jamaican economy and its various sectors are presented in Table 3 (and Table A2 in the Appendix). For illustrative purposes, we examine the value of lost load results specifically for 2011 in order to facilitate a comparison with other countries globally. Overall, the average estimated cost per unit of electricity not supplied due to a power disruption (i.e. the value of lost load) in Jamaica was US$2.94 per kWh in 2011 (compared with the US$1.30 that was estimated for 1996).25 In other words, the cost to the Jamaican economy of one kWh of electricity not supplied is almost US$3. In terms of the various Jamaican industries, the highest value of lost load in 2011 (i.e. US$23.49 per kWh) was registered in the Construction sector (Table 3). In other words, the Jamaican Construction

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

9

ELECTR-6200; No of Pages 18

Table 3: Jamaica’s Electricity Consumption, Total Value Added and Value of Lost Load by Sector, 2011

Sector Agriculture, Forestry & Fishing Mining & Quarrying Manufacturing Electricity & Water Supply

Electricity Consumed

Sectoral Electricity Use

Total Value Added at Basic

Sectoral Value Added

Value of Lost Load

(GWh)

Share (%)

Prices (J$ M)

Share (%)

(US$ per kWh)

78.6

1.9

70,438

6.6

10.35

63.2 489.9

1.5 11.7

15,487 96,566

1.5 9.1

2.83 2.28

–

35,867

3.4

0.54

760.0

Construction Wholesale & Retail Trade; Repairs;

38.3 517.8

0.9 12.3

77,921 201,491

7.3 18.9

23.49 4.49

Installation of Machinery & Equipment Hotels & Restaurants

756.9

18.1

45,481

4.3

0.69

Transport, Storage & Communic

182.4

4.4

104,330

9.8

6.60

Finance & Insurance Services Real Estate, Renting & Business Activities

144.5 143.0

3.4 3.4

111,869 130,771

10.5 12.3

8.94 10.56

Producers of Government Services Other Services

665.6 352.6

15.9 8.4

152,708 69,803

14.3 6.5

2.65 2.29

Total Economy

2.94

Source: Author’s estimates. Notes: VoLL computed by dividing sectoral Gross Value Added by the electricity consumed in each sector. GWh means Gigawatt hours. The negative sign on the ‘Electricity’ sector is due to the fact that it generates more electricity than it utilizes and consequently, the sector’s electricity demand is negative.

industry created a value of roughly US$23.49 with one kWh of electricity. This was followed by Real Estate, Renting & Business Activities (US$10.56 per kWh), Agriculture, Forestry and Fishing (US$10.35 per kWh), Finance and Insurance Services (US$8.94 per kWh), Transport, Storage and Communication (US$6.60 per kWh) and Wholesale & Retail Trade; Repairs; Installation of Machinery and Equipment (US$4.49 per kWh). The two sectors with the lowest VoLL estimates were Manufacturing and Hotels & Restaurants with US$2.28 per kWh and US$0.69 per kWh, respectively. Based on Table 3, the Jamaican Construction sector utilized only 0.9 percent of total electricity consumed, despite 10

accounting for 7.3 percent of the country’s total value added. On the other hand, while the Hotel & Restaurants sector used 18.1 percent of the total electricity utilized by the country, it only accounted for 4.3 percent of total value added. In general, these observations are in line with expectations as sectors with the highest VoLLs are usually those that consume low levels of electricity relative to the value added they create. Our overall sectoral VoLL estimates are also consistent with findings previously reported by de Nooij et al. (2007) and Linares and Rey (2013), who both reported that construction ranked among the sectors with the highest value of lost load due to

power interruptions in the Netherlands and Spain, respectively. D. Jamaica’s value of lost load in an international context As a final contribution, we attempted to gauge Jamaica’s value of lost load in an international context. Consequently, we extracted data for some ninety-five other countries from the World Bank (World Bank, 2014) dataset and computed their respective VoLLs using the production function method. The extensive results from our estimation are shown in Table A3 of the Appendix. Based on our estimates, relative to other countries, Jamaica’s VoLL of

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

US$2.94 per kWh is below the simple average value of lost load of US$3.66 per kWh (and a median of US$3.44 per kWh) for the overall ninety-six countries for which the calculations were possible. Overall, the cost per unit of electricity not delivered due to a power interruption ranged from a low of US$0.43 per kWh in Tajikistan to a high of US$9.91 per kWh in Switzerland. Specifically, Jamaica ranked fiftyeight out of all the countries in terms of the direct economic damage per unit of electricity not supplied due to power interruptions.

IV. Discussion A. Overall direct cost of power disruptions to the Jamaican economy After estimating the value of lost load, we found it instrumental to determine the total cost to the Jamaican economy as a result of its relatively frequent power disruptions. To do this, we utilized JPS’ T&D System Average Interruption Duration Index (SAIDI). The SAIDI is one of the most commonly used distribution indices that measure the reliability of power systems.26 Overall, the Index measures the total duration of a power interruption for the average customer during a specific time period (such as a year). In other words, the Index indicates the total duration of interruption for the average customer during a August/September 2015,

Vol. 28, Issue 7

pre-defined period of time and is commonly measured in customer minutes (or hours) of interruption (Rajaiah et al., 2013). SAIDI is usually computed by dividing the sum of all customer interruption durations by the total number of customers served by the respective power company. For 2013, Jamaica Public Service (2014a) reported a system average interruption duration index of 1,529 minutes based on 606,654 customers. Using the JPS’ estimate of total energy not supplied due to power disruptions (i.e. 8,400 MWh) in combination with our value of lost load estimate, we imputed the overall direct cost of power disruptions to the Jamaican economy for 2013 to be approximately US$22.84 million or 0.17 percent of total GDP.27 t is also useful to discuss a practical implication resulting directly from our sectoral value of lost load computations. In Jamaica, when there is a generation forced outage28 and power has to be disrupted, based on our latest sectoral VoLL estimates, it should be disconnected first in the Hotels & Restaurants (i.e. tourism) subsector followed by Manufacturing but last in Construction (see Table 3). The rationale for this proposal is based on the fact that the overall direct economic cost to the Jamaican economy due to power disruptions would be lowest if electricity is rationed by first starting with sectors having the lowest VoLLs (Leahy and Tol, 2011; de Nooij et al., 2007). In other words, the JPS should disconnect

I

power in sectors with the smallest VoLLs first and those with the largest VoLLs, last. Overall, the sector with the lowest value of lost load implies that one hour of electricity disruption is not as costly as in the other sectors. he above recommendation based on our sectoral value of lost load findings, run counter to JPS’ generation forced outages practice, which state that power in industrial areas (such as those involved in Manufacturing) should be turned off last (rather than first). JPS’ current practice for rationing electricity in some areas during power outages without much consideration of the direct economic effects is therefore not ideal policy and should be re-visited.

T

B. Recommendation for curtailing electricity theft At the end of 2013, the JPS reportedly lost approximately US$46 million in revenue (or approximately 18.0 percent of its total fuel bill) as a result of electricity theft by an estimated 180,000 unmetered consumers. The cost of such abstraction is typically shared jointly by the utility company and its paying consumers. Consequently, we find it timely to outline a recommendation to the JPS to reduce electricity theft in Jamaica. It is important to state that while complete elimination of illegal power usage is virtually

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

11

ELECTR-6200; No of Pages 18

impossible, at least one major step could be adopted to curtail power theft. Jamaica’s electricity theft might be lowered if the JPS were to embrace a newly available technology proposed by Depuru et al. (2011). According to Depuru et al.’s (Depuru et al., 2011) novel suggestion, a power company could use a combination of inter alia smart meters and a harmonic generator to isolate illegal consumers connected to the national grid (Depuru et al., 2011). Thereafter, utility company officials could remotely send a unique frequency to premises with illegal connections, which would in turn, damage their electrical appliances (Depuru et al., 2011). To permit this, however, legislation would have to be amended and a public education campaign (and a pilot testing phase) conducted to sensitize persons before wide-scale rollout of the harmonic generator system. Despite the fact that the cost of the proposed system might be prohibitive, it could serve as a plausible method for reducing electricity theft not only in Jamaica but internationally.

V. Conclusion Electricity theft by whatever means is costly. In this study, we investigated the nexus among electricity theft, power disruptions and the resulting direct cost to an economy. Despite their simplicity, throw-ups are effective tools in the abstraction of power and their rampant use has adverse 12

consequences. The first contribution of our research is that we provided a rare and detailed exploration of electricity theft in 100 countries globally over an extensive time interval spanning up to 40 years. Our findings show that electricity theft has risen overall not only in many individual countries but has also climbed in most regions globally

between 1971 and 2011. Worryingly, we demonstrated that extensive power theft had increased internationally since 1980. imultaneously, for the first time to our knowledge, using a production function approach and input-output analysis, we estimated the direct cost of an electricity disruption (in terms of lost production time) to inter alia the Jamaican economy. We found that the cost of one kWh of electricity not supplied in Jamaica for example, was approximately US$2.94 in 2011; a 126.2 percent increase over the US$1.30 estimate for 1996. Nonetheless, the damage to the Jamaican economy due to

S

power disruptions is below the corresponding average cost for a sample containing 96 countries. Simultaneously, using input– output analysis in combination with a production function technique, we measured the value of lost load due to power disruptions for the individual Jamaican sectors. Interestingly, between 2007 and 2013, with the exception of Transport, Storage & Communication and Mining & Quarrying, the cost of one kWh of electricity not supplied had increased overall in all sectors. This finding is contrary to that of Linares and Rey (2013) where the cost of electricity not delivered actually declined for most sectors in Spain between 2000 and 2008. Our sectoral estimates on the value of lost load might be useful to the Jamaica Public Service Company because it can potentially assist to inter alia plausibly determine the sectors which should receive priority attention for disconnecting power in cases of generation forced outages. Ideally, JPS should cut power first in the Hotel & Restaurants sector but last in the Construction industry. Despite the usefulness of our study, there are shortcomings with all the available techniques used to measure the cost of electricity disruptions and as such, our findings should be interpreted with caution. Additional research on electricity theft is therefore warranted. Future work could, for example, extend our analysis to ascertain the damage caused by

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

power interruptions to households (in terms of lost leisure time or spoilage of goods), independently corroborate our Jamaica’s value of lost load estimates using an alternative

methodology (such as a survey), and assess the price-elasticity of electricity in the Jamaican context when the required data are available. The latter exercise would be useful, as it would

provide key insights on the demand response of legitimate JPS customers to the volatility of electricity prices, especially in an atmosphere of relatively high levels of electricity theft.

Appendix

Figure A1: Transmission and Distribution Electricity Losses in Selected Countries by Rank, 2011 Source: Compiled by author using data obtained from World Bank (2014).

Figure A2: T&D Losses and GDP Per Capita for Selected Countries, 2011 Source: Compiled by author using data obtained from World Bank (2014).

August/September 2015,

Vol. 28, Issue 7

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

13

ELECTR-6200; No of Pages 18

Figure A3: Jamaica’s Value of Lost Load, 1996–2011 Sources: Compiled by author using data obtained from the Statistical Institute of Jamaica, Bank of Jamaica and World Bank (2014).

Table A1: Evolution of Electricity T&D Losses (%) by Selected Regions and Years Overall Changea

Region

1971

1980

1990

2000

2011

East Asia & Pacific

7.3

6.4

6.3

6.3

5.7

Europe & Central Asia Latin America & Caribbean

7.3 11.8

7.1 12.5

7.4 14.5

9.2 16.0

7.9 14.9

+0.6 +3.1

Middle East & North Africa North America

7.0 8.5

8.9 8.8

9.5 9.0

11.9 6.0

12.0 5.9

+5.0 2.5

South Asia

17.6

19.4

20.0

26.9

20.4

+2.8

Sub-Sahara Africa

7.9

9.1

8.9

11.2

10.8

+2.9

1.7

Source: Compiled by author using data obtained from World Bank (2014). Notes: a Percentage point change between 1971 and 2011. Discrepancies due to rounding.

Table A2: Jamaica’s Value of Lost Load (US$ per kWh) by Sector: 2007, 2011 and 2013 Sector

VoLL for 2007

VoLL for 2011

VoLL for 2013

Agriculture, Forestry & Fishing

5.79

10.35

10.32

Mining & Quarrying Manufacture

5.69 1.54

2.83 2.28

2.31 2.17

0.36 18.55

0.54 23.49

0.50 21.60

Wholesale & Retail Trade; Repairs; Installation of Machinery & Equipment

3.06

4.49

4.25

Hotels & Restaurants Transport, Storage & Communication

0.53 5.50

0.69 6.60

0.66 5.44

Finance & Insurance Services Real Estate, Renting & Business Activities

6.11 6.21

8.94 10.56

8.03 9.58

Producers of Government Services

1.61

2.65

2.49

Other Services Total Economy

1.58 2.04

2.29 2.94

2.16 2.72

Electricity & Water Supply Construction

Sources: Compiled by author using data obtained from the Statistical Institute of Jamaica and World Bank (2014).

14

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

Table A3: Value of Lost Load (US$ per KWh) in Selected Countries by Rank, 2011 Country

VoLL

Country

VoLL

Country

VoLL

Country

VoLL

Switzerland Nigeria

9.91 9.89

Austria Singapore

5.34 5.29

El Salvador Portugal

4.08 4.05

Jamaica Paraguay

2.94 2.92

Sudan

9.23

France

5.23

Ecuador

4.04

Bolivia

2.90

Denmark Ireland

8.41 7.87

Belgium Spain

5.17 5.16

Azerbaijan Norway

3.97 3.82

Nicaragua Bangladesh

2.85 2.73

Luxembourg Netherlands

6.48 6.35

Tanzania Senegal

5.07 5.04

Cuba Chile

3.79 3.78

Trinidad and Tobago Finland

2.70 2.66

United Kingdom

6.34

Cambodia

5.03

Sweden

3.54

Pakistan

2.61

Nepal Italy

6.10 6.00

Peru Kenya

4.50 4.50

Philippines Morocco

3.47 3.46

Estonia Zambia

2.35 2.33

Ethiopia Japan

5.94 5.84

Mexico Brazil

4.48 4.40

Turkey Argentina

3.42 3.35

Lebanon Congo, Dem. Rep.

2.32 2.23

Colombia

5.81

Botswana

4.38

Namibia

3.31

Turkmenistan

2.22

Sri Lanka Dominican Republic

5.78 5.75

Ghana Costa Rica

4.33 4.27

Croatia Tunisia

3.21 3.10

Kazakhstan India

2.17 2.10

Hong Kong SAR, China Guatemala

5.70 5.64

Panama Greece

4.26 4.25

Romania Venezuela, RB

3.01 2.99

Korea, Rep. Jordan

1.99 1.79

Germany

5.60

Uruguay

4.24

Honduras

2.97

Russian Federation

1.74

Australia

5.41

United Arab Emirates

4.12

Czech Republic

2.95

Armenia

1.74

Source: Author’s own calculations using data from World Bank (2014).

Country

VoLL

Country

VoLL

Mongolia Egypt, Arab Rep.

1.71 1.63

Uzbekistan Ukraine

0.88 0.84

South Africa Belarus

1.52 1.50

Zimbabwe Kyrgyz Republic

0.77 0.61

China

1.45

Tajikistan

0.43

Georgia Bulgaria

1.44 1.30

Bosnia and Herzegovina Serbia

1.21 1.18

Macedonia, FYR

1.12

Moldova Mozambique

1.11 1.06

Montenegro Kosovo

1.06 1.04

Vietnam

1.02

Source: Author’s own calculations using data from World Bank (2014).

August/September 2015,

Vol. 28, Issue 7

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

15

ELECTR-6200; No of Pages 18

References Billington, R., Grover, M.S., 1975. Reliability evaluation in distribution and transmission systems. IEE Proc. (5) 517–523. Bose, R.J., Shukla, M., Srivastava, L., Yaron, G., 2006. Cost of unserved power in Karnataka, India. Energy Policy 34, 1434–1447. Coll-Mayor, D., Pardo, J., Perez-Donsion, M., 2012. Methodology based on the value of lost load for evaluating economical losses due to disturbances in the power quality. Energy Policy 50, 407–418. de Nooij, M., Koopmans, C., Bijvoet, C., 2007. The value of supply security: the costs of power interruptions: economic input for damage reduction and investment in networks. Energy Econ. 29 (2) 277–295. Depuru, S.S., Wang, L., Devabhaktuni, V., 2011. Electricity theft: overview, issues, prevention and a smart meter based approach to control theft. Energy Policy 39, 1007–1015. Hensher, D.A., Shore, N., Train, K., 2014. Willingness to pay for residential electricity supply quality and reliability. Appl. Energy 115, 280– 292. Jamaica Productivity Centre, 2010. Generation and Distribution of Electricity in Jamaica: A Regional Comparison of Performance Indicators Kingston, Jamaica. Available at: http://www.jpc.com.jm/docs/ Generation%20and%20Distribution %20of%20Electricity%20in%20 Jamaica.pdf. Jamaica Public Service, 2008. Annual Report 2007 Jamaica. Available at: http://www.myjpsco.com/aboutus/financials/2007-annual-report/. Jamaica Public Service, 2009. Annual Report 2008 Jamaica. Available at: http://www.myjpsco.com/ about-us/financials/2008-annualreport/. Jamaica Public Service, 2014a. Annual Report 2013 Jamaica. Available at: http://www.myjpsco.com/ wp-content/uploads/JPSAnnual-Report-2013.pdf. Jamaica Public Service, 2014, April. JPS Tariff Application 2014–2019: Going for Growth Jamaica. Available at:

16

http://www.myjpsco.com/ wp-content/uploads/jps_20142019_rate_review_application_ 742014-final.pdf. Jamaica Public Service, 2015. Annual Report 2014. Jamaica. Jamil, F., 2013. On the electricity shortage, price and electricity theft nexus. Energy Policy 54, 267–272. Joseph, K.L., 2010. The politics of power: electricity reform in India. Energy Policy 38, 503–511. Leahy, E., Tol, R., 2011. An estimate of the value of lost load for Ireland. Energy Policy 39, 1514–1520. Linares, P., Rey, L., 2013. The costs of electricity interruptions in Spain.

Are we sending the right signals? Energy Policy 61, 751–760. Min, B., Golden, M., 2014. Electoral cycles in electricity losses in India. Energy Policy 65, 619–625. Mwaura, F.M., 2012. Adopting electricity prepayment billing system to reduce non-technical energy losses in Uganda: lesson from Rwanda. Util. Policy 23, 72–79. Neto, E.A., Coelho, J., 2013. Probabilistic methodology for technical and non-technical losses estimation in distribution system. Electric Power Syst. Res. 97, 93–99. Rajaiah, S., Satyanarayana, R.V.S., Srinivas, K., 2013. Evaluation and analysis of customer-specific distribution reliability indices. IUP J. Electr. Electron. Eng. 6 (2) 56–64. Sampson, C., 2006. Challenges in the Electricity Sector: Jamaica at the Crosswords Presentation at the

Geological Society Exhibition, University of the West Indies. Available at http://gsj.monainformatixltd. com/sampson_presentation.pdf. Smith, T.B., 1993. India’s electric power crisis: why do the lights go out? Asian Survey 33 (4) 376–392. Smith, T.B., 2004. Electricity theft: a comparative analysis. Energy Policy 32, 2067–2076. Steadman, K., 2009. Electricity Theft in Jamaica. Available at: http:// www2.binghamton.edu/ economics/graduate/documents/ prospectus-by-k-steadman.pdf. Tasdoven, H., Fielder, B.A., Garayev, V., 2012. Improving electricity efficiency in Turkey by addressing illegal electricity consumption: a governance approach. Energy Policy 43, 226–234. Tishler, A., 1993. Optimal production with uncertain interruptions in the supply of electricity estimation of electricity outage costs. Eur. Econ. Rev. 37, 1259–1274. Winther, T., 2012. Electricity theft as a relational issue: a comparative look at Zanzibar, Tanzania, and the Sunderban Islands, India. Energy Sustain. Dev. 16, 111–119. Woo, C.K., Ho, T., Shiu, A., Cheng, Y.S., Horowitz, I., Wang, J., 2014. Residential outage cost estimation: Hong Kong. Energy Policy 72, 204–210. World Bank, 2014. World Development Indicators. Available at: http:// databank.worldbank.org/data/ download/WDI_excel.zip.

Endnotes: 1. We go further than (Smith, 2004) by inter alia estimating the value of lost load in several countries globally due to electricity interruptions. Additionally, we depart from Jamil (2013), who explored the relationship among electricity theft, power outages, and electricity tariff rates. 2. In the Jamaican context, the phrase ‘‘throw-up’’ is essentially an illegal connection attached to overhead power distribution lines to abstract electricity. During 2014, some 186,961 throw-ups were removed from the

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal

ELECTR-6200; No of Pages 18

country’s grid and several persons arrested and charged with stealing electricity. 3. A ‘‘blackout’’ represents a situation where there is complete loss of power in a region while a ‘‘brownout’’ describes a case where voltage supplied to the system falls below the specified operating range but there is no total loss of power (Leahy and Tol, 2011). Overall, the two terms can be loosely described as ‘‘power disruptions.’’ For summaries of various characteristics of power supply interruptions (such as their duration, sources, types of users affected, etc.), see Bose et al. (2006) and de Nooij et al. (2007). 4. Sampson (2006), Steadman (2009) and Jamaica Productivity Centre (2010) discuss inter alia issues of frequent power outages and relatively high energy costs in the Jamaican context. 5. ‘‘Generation’’ (i.e. the process of producing electric energy), ‘‘transmission’’ (i.e. the movement of electrical energy from one location to another in an electrical power system), and ‘‘distribution’’ (i.e. the final stage in the delivery of electricity to end users) are the three main components of the electricity sector (Tasdoven et al., 2012). 6. One megawatt-hour (MWh) is equivalent to 1,000 kilowatt-hours (kWh) or 0.001 gigawatt-hour (GWh). Kilowatt-hour is conventionally used as the billing unit for electricity supplied by power companies to consumers. Essentially, kWh is a unit of energy equivalent to one kilowatt (kW) of power consumed over a one-hour time period. Total generation of electricity represents the total amount of electricity supplied to the grid and includes JPS’ own net generation (actual production) plus the generation (purchases) of electricity from other non-JPS sources. 7. ‘‘Technical losses’’ are energy losses that occur naturally for example, due to power dissipation in electricity system components such as transmission and distribution lines while ‘‘non-technical

August/September 2015,

Vol. 28, Issue 7

losses’’ result from actions external to the power system, including electricity theft and non-payment by consumers (Depuru et al., 2011; Neto and Coelho, 2013). Tasdoven et al. (2012) describe technical and non-technical losses as ‘‘internal’’ and ‘‘external’’ losses, respectively. 8. In the Jamaican context, the JPS applies to the OUR every five years for a review of its electricity tariff rates after which its customers face a rate adjustment. It is important to however bear in mind that generally with higher electricity tariffs (and hence the cost of power), persons are even more inclined to steal electricity (Tasdoven et al., 2012). For reviews of some key factors that influence the illegal consumption of electricity, see Depuru et al. (2011) and Winther (2012). 9. For a comprehensive review of these techniques, see Bose et al. (2006), de Nooij et al. (2007), Coll-Mayor et al. (2012) and Linares and Rey (2013). 10. According to Linares and Rey, sectors that are relatively more ‘‘energy intensive’’ are those which have less value added per unit of energy and thus a lower VoLL by definition. As a result, an inherent caveat is that the measure can lead to an overestimation of the impact of power disruptions when there is some flexibility in shifting production during an outage (Linares and Rey, 2013). 11. Joseph (2010) for example, utilized 1994–2005 panel data for some 35 states to demonstrate how the issue of electricity theft was associated with an increased development of captive power plants in India. The rationale was that high levels of electricity theft meant that the state could hardly finance the development of the power plant infrastructure and, as a result, industrial consumers in particular (concerned about a reliable power supply) exited the state run system to set up their own captive power plants to run their operations and, subsequently, sold any excess power generated back to the national grid. Bose et al. (2006) also highlights, in passing, a substantial growth of

captive generation in Karnataka, India, due to massive power theft. 12. Following Smith (2004), we excluded a few countries that had unrealistically low levels of power theft (i.e. T&D loss percentages between zero and one) to arrive at our final sample. Their exclusion was based on the rationale that even in the case of very efficient power systems, some amount of electricity is lost during the transmission and distribution stages (Smith, 2004). Additionally, in order to facilitate a comparative and evolutionary analysis of power theft, only countries that had T&D losses data for both 1980 and 2011 were selected. Due to space considerations, the full list of countries and the detailed crosscountry results are not reported here but are available upon request. 13. Required sectoral electricity usage data were however not available. As a result, we utilized Jamaica’s Input– Output table to disaggregate the total electricity consumed in the economy into the respective amounts used by the various sectors. (For more on input–output tables, see de Nooij et al. (2007).) 14. VoLL estimates based on the production function method are also provided by Leahy and Tol (2011) and de Nooij et al. (2007). For a useful summary of some early studies on the costs of electricity disruptions, see Linares and Rey (2013). 15. Similar to Leahy and Tol (2011), we selected the production function approach as it was the only viable option given our data constraints. Leahy and Tol (2011) used the technique to estimate the value of lost load in the Republic of Ireland and Northern Ireland using data for the period 2001–2008 and 2000–2007, respectively. In the case of the Republic of Ireland, the VoLL increased overall between 2001 and 2008 (Leahy and Tol, 2011). 16. The JPS occasionally gives prior notice via the media for scheduled power outages when it is conducting maintenance work for example. However, based on casual observation (and local experience), electricity

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

17

ELECTR-6200; No of Pages 18

disruptions in Jamaica mostly occur on an impromptu basis. 17. Depuru et al. (2011) corroborate the view that electricity theft is not only confined to Jamaican inner city areas but also occurs in more affluent areas, where few cases of meter tampering have been unearthed. In general, the situation of high electricity theft in relatively poor communities is not unique to Jamaica but is common in other countries such as India (Min and Golden, 2014). It is useful to stress that some persons are forced to steal electricity because they either reside in remote communities that have no access to distribution feeders or because they are informal settlers who lack proof of homeownership documents (which are required by the JPS before it formally supplies them with electricity). Additionally, potential customers who live in areas that are at least 100 meters away from distribution lines are usually responsible for setting up the required infrastructure (such as the utility poles, etc.) before the JPS provides them with power. 18. Tasdoven et al. (2012), in passing, compared power theft rates in selected OECD countries with that in Turkey and found that the country’s electricity theft level was more than double the OECD 7.0 percent average in 2002. Depuru et al. (2011) on the other hand, ranked overall transmission and distribution losses for six countries using 2006 data. 19. Although examination of the link between electricity theft and political business cycles is beyond the scope of this research, similar to Min and Golden (2014), this illustration allowed us to briefly explore the relationship between electricity theft in Jamaica and election years. Based on the visual evidence in Figure 3, it does not appear that the country’s general elections typically coincided with time periods in which total transmission and distribution losses peaked. Min and Golden (2014) examined electricity line losses in India’s most populous state (i.e. Uttar Pradesh) during the period 2000–2009

18

and found that the losses were correlated around election years in that line losses were higher in election years than in other periods. Additionally, the authors found empirical evidence that a 10.0 point increase in line losses was associated with a 12.0 percent rise in the proportion of seats retained by the same party in a district (Min and Golden, 2014). For another exposition of how politics can specifically affect the quality of power supplied, see Joseph (2010). 20. For specific discussions on inter alia electricity theft in India, see for example, Smith (1993) and Min and Golden (2014). 21. Using the World Bank (World Bank, 2014) dataset, we attempted to empirically confirm this stylized fact by investigating the relationship between transmission and distribution losses and GDP per capita (PPP) for some 130 countries in 2011. Overall, we found a strong negative relationship (i.e. correlation coefficient of 0.465 and significant at the 1 percent level) between the two variables in the sample, thus tentatively confirming expectations that high levels of electricity theft are generally more associated with countries that have relatively low income levels. The scatterplot is shown as Figure A2 in the Appendix. 22. We also examined data on transmission and distribution losses by region. In 2011, South Asia (20.4 percent), Latin America & the Caribbean (14.9 percent), Middle East & North Africa (12.0 percent) and Sub-Sahara Africa (10.8 percent) had the four highest levels of regional electricity losses (in rank order). Relative to 1971, electricity theft rose overall in the Middle East & North Africa, Latin America & the Caribbean, Sub-Saharan Africa, South Asia and Europe & Central Asia regions in 2011. Power theft however declined overall for North America and the East Asia & Pacific regions between 1971 and 2011. The full results for selected years including 1971 and 2011 are

summarized in Table A1 of the Appendix. 23. Winther (2012) reiterates the growing problem of power theft worldwide. Using two case studies on Tanzania and India, the author demonstrated how the issue of the ‘‘customer–utility provider relationship’’ affects electricity theft. Specifically, Winther (2012) opined that a lack of trust by customers (especially in the utility provider) can negatively influence voluntary compliance and contribute to electricity theft. The author therefore argued that electricity theft can be curtailed by simply improving the relationship between customers and electrical utility companies (Winther, 2012). 24. The estimation of indirect economic impacts (for example, loss of market share) and social impacts (such as loss of leisure time) of power interruptions are outside the scope of this research. For a brief discussion on those impacts, see Linares and Rey (2013) and Woo et al. (2014). 25. An evolution of Jamaica’s VoLL over the period 1996–2011 is illustrated by Figure A3 in the Appendix. 26. For a review of electric system reliability indices, see for example Jamaica Public Service (2014b), Rajaiah et al. (2013) and Billington and Grover (1975). 27. de Nooij et al. (2007) in a similar vein estimated the overall damage of power interruptions for the state of California in 2001 to be US$0.25 billion. Such an estimate was arrived at by multiplying the value of lost load (in US$ per kWh) by the electricity not supplied due to electricity disruptions (de Nooij et al., 2007). 28. A forced outage essentially occurs when a power station, transmission line or distribution line unexpectedly shuts down or when a generating unit is suddenly unavailable to produce power. As a result, such outages exclude inter alia scheduled outages for inspection, refuelling and maintenance.

1040-6190/# 2015 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2015.07.009

Please cite this article in press as: Lewis, F.B., Costly ‘Throw-Ups’: Electricity Theft and Power Disruptions. Electr. J. (2015), http://dx.doi.org/10.1016/j.tej.2015.07.009

The Electricity Journal