avsolatorio/wb-prwp-covid-sent · Datasets at Hugging Face

page_content stringlengths 1 2.72k	source stringlengths 24 25	span int64 0 251	cite_spans list	ref_spans list	eq_spans list	section stringlengths 0 228	sec_num stringlengths 1 7 ⌀	url_flag bool 2 classes	skip_flag bool 2 classes	sent_idx int64 0 128	num_chars int64 1 2.72k	num_words int64 1 669	num_tokens int64 2 1.2k	has_covid bool 2 classes
The COVID-19 pandemic exposed limitations to economic measurement.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	0	66	8	16	true
Across the world, governments enacted lockdowns and other containment measures to mitigate the public health crisis, with strong implications for economic activity.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	1	164	22	30	false
Rational policy making requires evaluating economic impacts and assessing the economic situation in real-time.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	2	110	14	19	false
However, in many countries, traditional indicators of economic activity do not provide that.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	3	92	13	18	false
First, national accounts estimates are available with substantial lags and often only at the national level.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	4	108	16	21	false
In addition, national accounts estimates in many countries, including in Bangladesh, are only available at annual but not quarterly frequency.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	5	142	20	26	false
Second, data collection through surveys, which are fundamental for the traditional estimation of gross value added, became more difficult during the pandemic due to severe containment measures.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	6	193	27	35	false
In Bangladesh, for example, it took more than a year to finalize the gross value-added estimates for fiscal year 2020 (FY2020) (from July 2019 to June 2020).	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	7	157	27	42	false
In such a situation, high-frequency indicators of economic activity can provide crucial insights.	wb/data/pdfs/WPS10002.pdf	0	[]	[]	[]	Introduction	1.	false	false	8	97	13	19	false
Many new approaches to tracking economic activity have been introduced since the outbreak of the COVID-19 pandemic.	wb/data/pdfs/WPS10002.pdf	1	[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]	[]	[]	Introduction	1.	false	false	0	115	17	25	true
For example, high-frequency indicators measuring mobility based on phone locations provided by Google and Facebook have been used to assess the economic situation in real-time (Spelta and Pagnottoni 2021) .	wb/data/pdfs/WPS10002.pdf	1	[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]	[]	[]	Introduction	1.	false	false	1	206	30	44	false
Beyer, Franco-Bedoya, and Galdo (2021) use daily electricity consumption and monthly nighttime light intensity to analyze subnational economic activity during the pandemic in India.	wb/data/pdfs/WPS10002.pdf	1	[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]	[]	[]	Introduction	1.	false	false	2	181	24	39	false
Few comparable indicators for tracking the economy in real-time are available in Bangladesh.	wb/data/pdfs/WPS10002.pdf	1	[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]	[]	[]	Introduction	1.	false	false	3	92	13	18	false
In this paper, we establish daily electricity consumption as a useful indicator for tracking economic fluctuations in Bangladesh.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	0	129	18	22	false
First, we confirm a meaningful relationship between electricity consumption and short-term economic fluctuations.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	1	113	13	19	false
Second, we estimate a daily electricity consumption model that explains over 90 percent of the daily variation in electricity consumption.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	2	138	20	24	false
Deviations of actual electricity consumption from the model prediction can then be interpreted as a real-time indicator of economic activity and can be used to assess economic shocks.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	3	183	28	34	false
We provide this measure not just at the national level, but also for the country's eight divisions separately.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	4	110	18	24	false
Using this measure, we show that electricity consumption in Dhaka fell over 40 percent compared with normal during the first lockdown in April 2020 and remained below normal until early 2021.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	5	191	31	36	false
In stark contrast, the later lockdowns had smaller impacts, suggesting strong adaptation to the COVID-19 containment measures.	wb/data/pdfs/WPS10002.pdf	2	[]	[]	[]	Introduction	1.	false	false	6	126	17	27	true
The layout of the paper proceeds as follows.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	0	44	8	11	false
Section 2 reviews related literature.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	1	37	5	8	false
Section 3 describes the data sources.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	2	37	6	9	false
Section 4 presents stylized facts on electricity consumption in Bangladesh.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	3	75	10	13	false
Section 5 analyzes the relationship between electricity consumption and economic activity in Bangladesh.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	4	104	13	17	false
Section 6 estimates a daily electricity consumption model.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	5	58	8	11	false
Section 7 analyzes deviations from the model during COVID-19 in Dhaka.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	false	6	70	11	19	true
Section 8 concludes.	wb/data/pdfs/WPS10002.pdf	3	[]	[]	[]	Introduction	1.	false	true	7	20	3	6	false
The long-term relationship between electricity consumption and economic growth has been well researched.	wb/data/pdfs/WPS10002.pdf	4	[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]	[]	[]	Related Literature	2.	false	false	0	104	13	18	false
Apergis and Payne (2011) find a causal relationship between electricity consumption and economic growth for 88 countries over 1990-2006.	wb/data/pdfs/WPS10002.pdf	4	[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]	[]	[]	Related Literature	2.	false	false	1	136	19	28	false
In high-, upper-middle-, and lower-middle-income countries, the causality runs in both directions.	wb/data/pdfs/WPS10002.pdf	4	[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]	[]	[]	Related Literature	2.	false	false	2	98	12	27	false
In low-income countries, they find only a unidirectional causality from electricity consumption to economic growth.	wb/data/pdfs/WPS10002.pdf	4	[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]	[]	[]	Related Literature	2.	false	false	3	115	15	26	false
Due to the strong association of electricity consumption and economic activity, variations in electricity consumption can be used to assess economic fluctuations.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	0	162	22	26	false
Chen et al.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	true	1	11	3	6	false
(2020) find that electricity usage in Europe declined between 10 and 15 percent during the acute phase of the COVID-19 pandemic.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	2	128	21	31	true
Historically, a 1 percent drop in electricity usage in Europe has been associated with a 1.3 to 1.9 percent drop in output.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	3	123	22	30	false
Cicala (2020) observes a one-to-one relationship between the variables of interest in the United States, including during the global financial crisis, and uses the relationship to assess the economic damage from COVID-19.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	4	221	32	47	true
Similarly, Beyer, Franco-Bedoya, and Galdo (2021) show that energy consumption declined strongly in India after a national lockdown was implemented on March 25, 2020, and confirm that electricity consumption can be a useful indicator in emerging market economies.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	5	263	38	54	false
Ai, Zhong, and Zhou (2022) estimate the impact of the COVID-19 pandemic in Hunan province (China) through firm-level electricity consumption data.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	6	146	21	39	true
They employ a difference-indifferences model to show that electricity consumption in Hunan province dropped by 27.8 percent during the early stage of the pandemic.	wb/data/pdfs/WPS10002.pdf	5	[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]	[]	[]	Related Literature	2.	false	false	7	163	24	34	false
For Bangladesh, empirical evidence of the relationship is ambiguous.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	0	68	9	13	false
Yıldırım, Sukruoglu, and Aslan (2014) do not find a link between energy consumption and economic growth in Bangladesh.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	1	118	18	36	false
Mozumder and Marathe (2007) find a unidirectional causality from per capita gross domestic product (GDP) to per capita electricity consumption, using annual data from 1971 to 1999.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	2	180	27	44	false
Along the same line, Asghar (2008) examines the causality between energy consumption and income from 1971 to 2003 and concludes that in Bangladesh, a unidirectional Granger causality runs from GDP to electricity consumption.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	3	224	33	49	false
In contrast, Sarker and Alam (2010) and Hossain and Saeki (2011) find a reverse causal relationship running from electricity generation to economic growth.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	4	155	23	35	false
More recently, Hossen and Hasan (2018) find unidirectional causality running from electricity consumption to GDP in Bangladesh from 1972 to 2011.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	5	145	21	33	false
Finally, Alam et al.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	true	6	20	4	8	false
(2012) and Ahamad and Islam (2011) find bi-directional, long-run causality between Bangladesh's economic activity and energy consumption.	wb/data/pdfs/WPS10002.pdf	6	[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]	[]	[]	Related Literature	2.	false	false	7	137	17	34	false
Our study contributes to the literature in three ways.	wb/data/pdfs/WPS10002.pdf	7	[]	[]	[]	Related Literature	2.	false	false	0	54	9	12	false
First, it reconsiders evidence on the causality between electricity consumption and economic activity in Bangladesh at annual frequency and adds monthly analysis.	wb/data/pdfs/WPS10002.pdf	7	[]	[]	[]	Related Literature	2.	false	false	1	162	22	30	false
Second, this is the first study to analyze electricity consumption as an high-frequency economic indicator for Bangladesh, with lessons for other emerging markets and developing economies.	wb/data/pdfs/WPS10002.pdf	7	[]	[]	[]	Related Literature	2.	false	false	2	188	26	33	false
Third, it uses the indicator to compare the costs of the different lockdowns that were enacted to contain the COVID-19 pandemic, to help assess them.	wb/data/pdfs/WPS10002.pdf	7	[]	[]	[]	Related Literature	2.	false	false	3	149	25	37	true
Monthly electricity consumption in Bangladesh is available from the Bangladesh Bureau of Statistics (BBS) from 1993 until 2013, when the data series was discontinued.	wb/data/pdfs/WPS10002.pdf	8	[]	[]	[]	Electricity Consumption	3.1	false	false	0	166	24	31	false
Daily electricity consumption data can be scraped from the Bangladesh Power Development Board and is available for all eight divisions (Dhaka, Chattogram, Mymensingh, Sylhet, Khulna, Rajshahi, Barisal, and Rangpur) from 2010 until now, with only a one-day lag.	wb/data/pdfs/WPS10002.pdf	8	[]	[]	[]	Electricity Consumption	3.1	false	false	1	260	38	67	false
For this study, daily data from February 2010 to October 2021 has been compiled.	wb/data/pdfs/WPS10002.pdf	8	[]	[]	[]	Electricity Consumption	3.1	false	false	2	80	14	18	false
1 For 2.5 percent of all the days during this period, data are missing and were hence linearly interpolated using the day before and day after the missing one.	wb/data/pdfs/WPS10002.pdf	8	[]	[]	[]	Electricity Consumption	3.1	false	false	3	159	29	38	false
For the analysis of the relationship with economic activity, we use annual data on GDP and industrial production (2000-21) from the National Accounts statistics published in the BBS Bluebook and the monthly Quantum Index of Industrial Production data (July 2012 to April 2021) from the BBS.	wb/data/pdfs/WPS10002.pdf	9	[]	[]	[]	Economic Activity	3.2	false	false	0	290	46	59	false
Monthly export data (2016-20) were taken from the Export Promotion Bureau.	wb/data/pdfs/WPS10002.pdf	9	[]	[]	[]	Economic Activity	3.2	false	false	1	74	11	18	false
We analyze the correlations between electricity consumption per capita and subnational poverty levels, the share of households with access to electricity (from the Bangladesh Household Income and Expenditure Survey), and firm density (from the Economic Census 2013).	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	0	266	37	48	false
To assess the efficacy of using electricity consumption to analyze the impacts of disasters, we extract disaster data (2016-20) from the Center for Excellence in Disaster Management & Humanitarian Assistance.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	1	208	30	38	false
For estimation of the electricity consumption model, we use temperature data (2010-21) from the Bangladesh Meteorological Department and the dates of holidays and Ramadan from timeanddate.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	2	188	26	37	false
2 To measure the stringency of the COVID-19 containment measures Bangladesh adopted and their impact on mobility, we rely on the Oxford COVID-19 Government Response Tracker (OxCGRT) 3 Stringency Index (2020-21) and Google mobility data from the Google Community Mobility Reports.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	3	279	41	62	true
Finally, data on daily COVID-19 cases and division-level data on the number of cases and deaths were retrieved from the Institute of Epidemiology, Disease Control and Research and the Directorate General of Health Services websites, respectively.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	4	246	36	50	true
first, electricity consumption is increasing over time; second, there is seasonality in the data; and third, electricity consumption varies strongly from day to day.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	5	165	24	33	false
Part of the daily variation may be explained by measurement or reporting issues and may not actually reflect variation in electricity consumption.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	6	146	22	25	false
Electricity consumption can hence be at best a noisy measure of economic activity.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	7	82	13	16	false
The variation is much larger in some divisions compared with others.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	8	68	11	14	false
For example, it tends to be lowest in Dhaka and highest in Rajshahi.	wb/data/pdfs/WPS10002.pdf	10	[]	[]	[]	Other Data	3.3	false	false	9	68	13	19	false
Map 1 shows the total electricity consumption per capita in the eight divisions before the COVID-19 pandemic in 2019.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	0	117	19	27	true
There is strong variation in electricity consumption per capita across Bangladesh.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	1	82	11	14	false
Dhaka has the largest per capita consumption, followed by Khulna, Chattogram, and Mymensingh.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	2	93	13	26	false
These divisions are more developed than the others, showing that the level of development is positively correlated with electricity consumption per capita.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	3	155	22	26	false
Figure 2 shows the correlation between per capita electricity consumption and the rate of poverty (panel a), household access to electricity (panel b), and firm density (panel c).	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	4	179	28	39	false
There is a strong negative association between electricity consumption and the poverty rate, implying that income poor divisions are also energy poor.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	5	150	22	26	false
4 Barisal and Sylhet are outliers and have much lower electricity consumption than a linear relationship would suggest, possibly due to lagging power distribution infrastructure and labor for expanding coverage and operations.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	6	226	32	42	false
As expected, per capita electricity consumption strongly increases with the number of people having access to the grid.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	7	119	18	22	false
In addition, it increases with the density of firms, likely reflecting electricity consumption used for economic production.	wb/data/pdfs/WPS10002.pdf	11	[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]	[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]	[]	Electricity Consumption over Time	4.1	false	false	8	124	17	22	false
Despite obvious noise in the data, electricity consumption growth is strongly correlated with other measures of economic activity.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	0	130	18	22	false
Figure 3 shows quarterly changes in electricity consumption from the first quarter of FY2016 to the last quarter of FY2021 and quarterly changes in industrial production and exports.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	1	182	28	37	false
At 0.60 for industrial production and 0.66 for exports, both correlations are strong.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	2	85	13	22	false
The fourth quarter of FY2020, when economic activity was severely disrupted due to the COVID-19 pandemic, is a clear outlier with a strong decline in exports and electricity consumption.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	3	186	29	43	true
Changes in electricity consumption can be used as a proxy to infer insights on the economic impacts of shocks, for example, the COVID-19 pandemic, associated lockdowns, or natural disasters.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	4	190	29	44	true
Figure 4 shows that electricity consumption declined sharply in the fourth quarter of FY2020, right after the first lockdown was imposed in Bangladesh at the end of March 2020.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	5	176	29	37	false
Electricity consumption recovered subsequently, but year-over-year growth remained minimal in the first quarter of FY2021.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	6	122	15	26	false
It took another two quarters before electricity growth was again close to the rates observed prior to the pandemic.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	7	115	19	24	false
By the fourth quarter of FY2021, electricity consumption growth reached the pre-pandemic level.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	8	95	13	24	false
Electricity consumption can also be used to track short-run economic activity during natural disasters.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	9	103	14	19	false
To show this, we average growth in monthly electricity consumption over the following natural disasters: Cyclone Roanu in 2016, the landslides and floods in 2017, the floods in 2018 and 2019, and Cyclone Amphan in 2020.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	10	219	36	47	false
Looking at average growth in electricity consumption (figure 5) during the month of a disaster, there is an average decline of approximately 10 percent.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	11	152	24	30	false
Electricity consumption then picks up gradually and overshoots in the second month after the disaster, presumably due to reconstruction efforts and to make up for some interrupted economic activity during the month of the disaster.	wb/data/pdfs/WPS10002.pdf	12	[]	[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]	[]	Electricity Consumption Growth and Economic Indicators	4.2	false	false	12	231	35	41	false
First, we determine the long-run association between electricity consumption and economic activity based on annual data.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	0	120	16	22	false
We perform the Johansen cointegration test on two bivariate annual models: (i) real GDP and electricity consumption, and (ii) industrial production and electricity consumption.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	1	176	24	38	false
To uncover the long-run equilibrium and causal relationships between the selected variables, data from 2000 to 2020 are analyzed.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	2	129	19	25	false
Each of the estimated models contains one cointegrating vector, suggesting that an equilibrium relationship holds in the long run.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	3	130	19	25	false
The existence of a cointegration relationship also implies that at least one causal relationship exists among the variables.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	4	124	18	23	false
To check pairwise causality, we employ the Toda-Yamamoto (1995) Granger causality test, which is robust to the identified cointegration relationship.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	5	149	20	36	false
5 It uncovers unidirectional causality from GDP to electricity consumption ( 2 = 6.24, p = 0.04) and bidirectional causality between industrial production and electricity consumption.	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	false	6	183	26	45	false
6	wb/data/pdfs/WPS10002.pdf	13	[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]	[]	[]	Cointegration and Causality	5.1	false	true	7	1	1	3	false
Second, we determine the short-run association between electricity consumption and economic activity based on monthly data.	wb/data/pdfs/WPS10002.pdf	14	[]	[]	[]	Cointegration and Causality	5.1	false	false	0	123	16	22	false
Since Bangladesh publishes only annual GDP data, we can only use industrial production.	wb/data/pdfs/WPS10002.pdf	14	[]	[]	[]	Cointegration and Causality	5.1	false	false	1	87	13	17	false

The COVID-19 pandemic exposed limitations to economic measurement.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

0

66

8

16

true

Across the world, governments enacted lockdowns and other containment measures to mitigate the public health crisis, with strong implications for economic activity.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

1

164

22

30

false

Rational policy making requires evaluating economic impacts and assessing the economic situation in real-time.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

2

110

14

19

false

However, in many countries, traditional indicators of economic activity do not provide that.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

3

92

13

18

false

First, national accounts estimates are available with substantial lags and often only at the national level.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

4

108

16

21

false

In addition, national accounts estimates in many countries, including in Bangladesh, are only available at annual but not quarterly frequency.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

5

142

20

26

false

Second, data collection through surveys, which are fundamental for the traditional estimation of gross value added, became more difficult during the pandemic due to severe containment measures.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

6

193

27

35

false

In Bangladesh, for example, it took more than a year to finalize the gross value-added estimates for fiscal year 2020 (FY2020) (from July 2019 to June 2020).

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

7

157

27

42

false

In such a situation, high-frequency indicators of economic activity can provide crucial insights.

wb/data/pdfs/WPS10002.pdf

0

[]

Introduction

1.

false

8

97

13

19

false

Many new approaches to tracking economic activity have been introduced since the outbreak of the COVID-19 pandemic.

wb/data/pdfs/WPS10002.pdf

1

[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]

[]

Introduction

1.

false

0

115

17

25

true

For example, high-frequency indicators measuring mobility based on phone locations provided by Google and Facebook have been used to assess the economic situation in real-time (Spelta and Pagnottoni 2021) .

wb/data/pdfs/WPS10002.pdf

1

[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]

[]

Introduction

1.

false

1

206

30

44

false

Beyer, Franco-Bedoya, and Galdo (2021) use daily electricity consumption and monthly nighttime light intensity to analyze subnational economic activity during the pandemic in India.

wb/data/pdfs/WPS10002.pdf

1

[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]

[]

Introduction

1.

false

2

181

24

39

false

Few comparable indicators for tracking the economy in real-time are available in Bangladesh.

wb/data/pdfs/WPS10002.pdf

1

[ { "end": 320, "ref_id": "BIBREF14", "start": 292, "text": "(Spelta and Pagnottoni 2021)" }, { "end": 361, "ref_id": "BIBREF5", "start": 323, "text": "Beyer, Franco-Bedoya, and Galdo (2021)" } ]

[]

Introduction

1.

false

3

92

13

18

false

In this paper, we establish daily electricity consumption as a useful indicator for tracking economic fluctuations in Bangladesh.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

0

129

18

22

false

First, we confirm a meaningful relationship between electricity consumption and short-term economic fluctuations.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

1

113

13

19

false

Second, we estimate a daily electricity consumption model that explains over 90 percent of the daily variation in electricity consumption.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

2

138

20

24

false

Deviations of actual electricity consumption from the model prediction can then be interpreted as a real-time indicator of economic activity and can be used to assess economic shocks.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

3

183

28

34

false

We provide this measure not just at the national level, but also for the country's eight divisions separately.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

4

110

18

24

false

Using this measure, we show that electricity consumption in Dhaka fell over 40 percent compared with normal during the first lockdown in April 2020 and remained below normal until early 2021.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

5

191

31

36

false

In stark contrast, the later lockdowns had smaller impacts, suggesting strong adaptation to the COVID-19 containment measures.

wb/data/pdfs/WPS10002.pdf

2

[]

Introduction

1.

false

6

126

17

27

true

The layout of the paper proceeds as follows.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

0

44

8

11

false

Section 2 reviews related literature.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

1

37

5

8

false

Section 3 describes the data sources.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

2

37

6

9

false

Section 4 presents stylized facts on electricity consumption in Bangladesh.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

3

75

10

13

false

Section 5 analyzes the relationship between electricity consumption and economic activity in Bangladesh.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

4

104

13

17

false

Section 6 estimates a daily electricity consumption model.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

5

58

8

11

false

Section 7 analyzes deviations from the model during COVID-19 in Dhaka.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

6

70

11

19

true

Section 8 concludes.

wb/data/pdfs/WPS10002.pdf

3

[]

Introduction

1.

false

true

7

20

3

6

false

The long-term relationship between electricity consumption and economic growth has been well researched.

wb/data/pdfs/WPS10002.pdf

4

[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]

[]

Related Literature

2.

false

0

104

13

18

false

Apergis and Payne (2011) find a causal relationship between electricity consumption and economic growth for 88 countries over 1990-2006.

wb/data/pdfs/WPS10002.pdf

4

[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]

[]

Related Literature

2.

false

1

136

19

28

false

In high-, upper-middle-, and lower-middle-income countries, the causality runs in both directions.

wb/data/pdfs/WPS10002.pdf

4

[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]

[]

Related Literature

2.

false

2

98

12

27

false

In low-income countries, they find only a unidirectional causality from electricity consumption to economic growth.

wb/data/pdfs/WPS10002.pdf

4

[ { "end": 241, "ref_id": null, "start": 216, "text": "countries over 1990-2006." } ]

[]

Related Literature

2.

false

3

115

15

26

false

Due to the strong association of electricity consumption and economic activity, variations in electricity consumption can be used to assess economic fluctuations.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

0

162

22

26

false

Chen et al.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

true

1

11

3

6

false

(2020) find that electricity usage in Europe declined between 10 and 15 percent during the acute phase of the COVID-19 pandemic.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

2

128

21

31

true

Historically, a 1 percent drop in electricity usage in Europe has been associated with a 1.3 to 1.9 percent drop in output.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

3

123

22

30

false

Cicala (2020) observes a one-to-one relationship between the variables of interest in the United States, including during the global financial crisis, and uses the relationship to assess the economic damage from COVID-19.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

4

221

32

47

true

Similarly, Beyer, Franco-Bedoya, and Galdo (2021) show that energy consumption declined strongly in India after a national lockdown was implemented on March 25, 2020, and confirm that electricity consumption can be a useful indicator in emerging market economies.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

5

263

38

54

false

Ai, Zhong, and Zhou (2022) estimate the impact of the COVID-19 pandemic in Hunan province (China) through firm-level electricity consumption data.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

6

146

21

39

true

They employ a difference-indifferences model to show that electricity consumption in Hunan province dropped by 27.8 percent during the early stage of the pandemic.

wb/data/pdfs/WPS10002.pdf

5

[ { "end": 181, "ref_id": "BIBREF6", "start": 163, "text": "Chen et al. (2020)" }, { "end": 699, "ref_id": null, "start": 640, "text": "COVID-19. Similarly, Beyer, Franco-Bedoya, and Galdo (2021)" }, { "end": 940, "ref_id": "BIBREF1", "start": 914, "text": "Ai, Zhong, and Zhou (2022)" } ]

[]

Related Literature

2.

false

7

163

24

34

false

For Bangladesh, empirical evidence of the relationship is ambiguous.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

0

68

9

13

false

Yıldırım, Sukruoglu, and Aslan (2014) do not find a link between energy consumption and economic growth in Bangladesh.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

1

118

18

36

false

Mozumder and Marathe (2007) find a unidirectional causality from per capita gross domestic product (GDP) to per capita electricity consumption, using annual data from 1971 to 1999.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

2

180

27

44

false

Along the same line, Asghar (2008) examines the causality between energy consumption and income from 1971 to 2003 and concludes that in Bangladesh, a unidirectional Granger causality runs from GDP to electricity consumption.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

3

224

33

49

false

In contrast, Sarker and Alam (2010) and Hossain and Saeki (2011) find a reverse causal relationship running from electricity generation to economic growth.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

4

155

23

35

false

More recently, Hossen and Hasan (2018) find unidirectional causality running from electricity consumption to GDP in Bangladesh from 1972 to 2011.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

5

145

21

33

false

Finally, Alam et al.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

true

6

20

4

8

false

(2012) and Ahamad and Islam (2011) find bi-directional, long-run causality between Bangladesh's economic activity and energy consumption.

wb/data/pdfs/WPS10002.pdf

6

[ { "end": 215, "ref_id": "BIBREF12", "start": 188, "text": "Mozumder and Marathe (2007)" }, { "end": 403, "ref_id": "BIBREF4", "start": 390, "text": "Asghar (2008)" }, { "end": 629, "ref_id": "BIBREF13", "start": 607, "text": "Sarker and Alam (2010)" }, { "end": 658, "ref_id": "BIBREF9", "start": 634, "text": "Hossain and Saeki (2011)" }, { "end": 923, "ref_id": "BIBREF2", "start": 905, "text": "Alam et al. (2012)" }, { "end": 951, "ref_id": "BIBREF0", "start": 928, "text": "Ahamad and Islam (2011)" } ]

[]

Related Literature

2.

false

7

137

17

34

false

Our study contributes to the literature in three ways.

wb/data/pdfs/WPS10002.pdf

7

[]

Related Literature

2.

false

0

54

9

12

false

First, it reconsiders evidence on the causality between electricity consumption and economic activity in Bangladesh at annual frequency and adds monthly analysis.

wb/data/pdfs/WPS10002.pdf

7

[]

Related Literature

2.

false

1

162

22

30

false

Second, this is the first study to analyze electricity consumption as an high-frequency economic indicator for Bangladesh, with lessons for other emerging markets and developing economies.

wb/data/pdfs/WPS10002.pdf

7

[]

Related Literature

2.

false

2

188

26

33

false

Third, it uses the indicator to compare the costs of the different lockdowns that were enacted to contain the COVID-19 pandemic, to help assess them.

wb/data/pdfs/WPS10002.pdf

7

[]

Related Literature

2.

false

3

149

25

37

true

Monthly electricity consumption in Bangladesh is available from the Bangladesh Bureau of Statistics (BBS) from 1993 until 2013, when the data series was discontinued.

wb/data/pdfs/WPS10002.pdf

8

[]

Electricity Consumption

3.1

false

0

166

24

31

false

Daily electricity consumption data can be scraped from the Bangladesh Power Development Board and is available for all eight divisions (Dhaka, Chattogram, Mymensingh, Sylhet, Khulna, Rajshahi, Barisal, and Rangpur) from 2010 until now, with only a one-day lag.

wb/data/pdfs/WPS10002.pdf

8

[]

Electricity Consumption

3.1

false

1

260

38

67

false

For this study, daily data from February 2010 to October 2021 has been compiled.

wb/data/pdfs/WPS10002.pdf

8

[]

Electricity Consumption

3.1

false

2

80

14

18

false

1 For 2.5 percent of all the days during this period, data are missing and were hence linearly interpolated using the day before and day after the missing one.

wb/data/pdfs/WPS10002.pdf

8

[]

Electricity Consumption

3.1

false

3

159

29

38

false

For the analysis of the relationship with economic activity, we use annual data on GDP and industrial production (2000-21) from the National Accounts statistics published in the BBS Bluebook and the monthly Quantum Index of Industrial Production data (July 2012 to April 2021) from the BBS.

wb/data/pdfs/WPS10002.pdf

9

[]

Economic Activity

3.2

false

0

290

46

59

false

Monthly export data (2016-20) were taken from the Export Promotion Bureau.

wb/data/pdfs/WPS10002.pdf

9

[]

Economic Activity

3.2

false

1

74

11

18

false

We analyze the correlations between electricity consumption per capita and subnational poverty levels, the share of households with access to electricity (from the Bangladesh Household Income and Expenditure Survey), and firm density (from the Economic Census 2013).

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

0

266

37

48

false

To assess the efficacy of using electricity consumption to analyze the impacts of disasters, we extract disaster data (2016-20) from the Center for Excellence in Disaster Management & Humanitarian Assistance.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

1

208

30

38

false

For estimation of the electricity consumption model, we use temperature data (2010-21) from the Bangladesh Meteorological Department and the dates of holidays and Ramadan from timeanddate.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

2

188

26

37

false

2 To measure the stringency of the COVID-19 containment measures Bangladesh adopted and their impact on mobility, we rely on the Oxford COVID-19 Government Response Tracker (OxCGRT) 3 Stringency Index (2020-21) and Google mobility data from the Google Community Mobility Reports.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

3

279

41

62

true

Finally, data on daily COVID-19 cases and division-level data on the number of cases and deaths were retrieved from the Institute of Epidemiology, Disease Control and Research and the Directorate General of Health Services websites, respectively.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

4

246

36

50

true

first, electricity consumption is increasing over time; second, there is seasonality in the data; and third, electricity consumption varies strongly from day to day.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

5

165

24

33

false

Part of the daily variation may be explained by measurement or reporting issues and may not actually reflect variation in electricity consumption.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

6

146

22

25

false

Electricity consumption can hence be at best a noisy measure of economic activity.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

7

82

13

16

false

The variation is much larger in some divisions compared with others.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

8

68

11

14

false

For example, it tends to be lowest in Dhaka and highest in Rajshahi.

wb/data/pdfs/WPS10002.pdf

10

[]

Other Data

3.3

false

9

68

13

19

false

Map 1 shows the total electricity consumption per capita in the eight divisions before the COVID-19 pandemic in 2019.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

0

117

19

27

true

There is strong variation in electricity consumption per capita across Bangladesh.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

1

82

11

14

false

Dhaka has the largest per capita consumption, followed by Khulna, Chattogram, and Mymensingh.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

2

93

13

26

false

These divisions are more developed than the others, showing that the level of development is positively correlated with electricity consumption per capita.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

3

155

22

26

false

Figure 2 shows the correlation between per capita electricity consumption and the rate of poverty (panel a), household access to electricity (panel b), and firm density (panel c).

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

4

179

28

39

false

There is a strong negative association between electricity consumption and the poverty rate, implying that income poor divisions are also energy poor.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

5

150

22

26

false

4 Barisal and Sylhet are outliers and have much lower electricity consumption than a linear relationship would suggest, possibly due to lagging power distribution infrastructure and labor for expanding coverage and operations.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

6

226

32

42

false

As expected, per capita electricity consumption strongly increases with the number of people having access to the grid.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

7

119

18

22

false

In addition, it increases with the density of firms, likely reflecting electricity consumption used for economic production.

wb/data/pdfs/WPS10002.pdf

11

[ { "end": 783, "ref_id": null, "start": 782, "text": "4" } ]

[ { "end": 459, "ref_id": "FIGREF2", "start": 451, "text": "Figure 2" } ]

[]

Electricity Consumption over Time

4.1

false

8

124

17

22

false

Despite obvious noise in the data, electricity consumption growth is strongly correlated with other measures of economic activity.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

0

130

18

22

false

Figure 3 shows quarterly changes in electricity consumption from the first quarter of FY2016 to the last quarter of FY2021 and quarterly changes in industrial production and exports.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

1

182

28

37

false

At 0.60 for industrial production and 0.66 for exports, both correlations are strong.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

2

85

13

22

false

The fourth quarter of FY2020, when economic activity was severely disrupted due to the COVID-19 pandemic, is a clear outlier with a strong decline in exports and electricity consumption.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

3

186

29

43

true

Changes in electricity consumption can be used as a proxy to infer insights on the economic impacts of shocks, for example, the COVID-19 pandemic, associated lockdowns, or natural disasters.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

4

190

29

44

true

Figure 4 shows that electricity consumption declined sharply in the fourth quarter of FY2020, right after the first lockdown was imposed in Bangladesh at the end of March 2020.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

5

176

29

37

false

Electricity consumption recovered subsequently, but year-over-year growth remained minimal in the first quarter of FY2021.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

6

122

15

26

false

It took another two quarters before electricity growth was again close to the rates observed prior to the pandemic.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

7

115

19

24

false

By the fourth quarter of FY2021, electricity consumption growth reached the pre-pandemic level.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

8

95

13

24

false

Electricity consumption can also be used to track short-run economic activity during natural disasters.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

9

103

14

19

false

To show this, we average growth in monthly electricity consumption over the following natural disasters: Cyclone Roanu in 2016, the landslides and floods in 2017, the floods in 2018 and 2019, and Cyclone Amphan in 2020.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

10

219

36

47

false

Looking at average growth in electricity consumption (figure 5) during the month of a disaster, there is an average decline of approximately 10 percent.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

11

152

24

30

false

Electricity consumption then picks up gradually and overshoots in the second month after the disaster, presumably due to reconstruction efforts and to make up for some interrupted economic activity during the month of the disaster.

wb/data/pdfs/WPS10002.pdf

12

[]

[ { "end": 139, "ref_id": "FIGREF3", "start": 131, "text": "Figure 3" }, { "end": 786, "ref_id": "FIGREF5", "start": 778, "text": "Figure 4" } ]

[]

Electricity Consumption Growth and Economic Indicators

4.2

false

12

231

35

41

false

First, we determine the long-run association between electricity consumption and economic activity based on annual data.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

0

120

16

22

false

We perform the Johansen cointegration test on two bivariate annual models: (i) real GDP and electricity consumption, and (ii) industrial production and electricity consumption.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

1

176

24

38

false

To uncover the long-run equilibrium and causal relationships between the selected variables, data from 2000 to 2020 are analyzed.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

2

129

19

25

false

Each of the estimated models contains one cointegrating vector, suggesting that an equilibrium relationship holds in the long run.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

3

130

19

25

false

The existence of a cointegration relationship also implies that at least one causal relationship exists among the variables.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

4

124

18

23

false

To check pairwise causality, we employ the Toda-Yamamoto (1995) Granger causality test, which is robust to the identified cointegration relationship.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

5

149

20

36

false

5 It uncovers unidirectional causality from GDP to electricity consumption ( 2 = 6.24, p = 0.04) and bidirectional causality between industrial production and electricity consumption.

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

6

183

26

45

false

6

wb/data/pdfs/WPS10002.pdf

13

[ { "end": 747, "ref_id": "BIBREF15", "start": 727, "text": "Toda-Yamamoto (1995)" } ]

[]

Cointegration and Causality

5.1

false

true

7

1

3

false

Second, we determine the short-run association between electricity consumption and economic activity based on monthly data.

wb/data/pdfs/WPS10002.pdf

14

[]

Cointegration and Causality

5.1

false

0

123

16

22

false

Since Bangladesh publishes only annual GDP data, we can only use industrial production.

wb/data/pdfs/WPS10002.pdf

14

[]

Cointegration and Causality

5.1

false

1

87

13

17

false