Monitoring
Monitoring COVID-19 impacts
on firms in Ethiopia
COVID-19 REPORT
REPORTNO.
NO.11
11
21 / December / 2020
21 / December / 2020
impacts on firms
in Ethiopia
Authors: Girum Abebe, Tom Bundervoet, Christina Wieser
❖ 49% the share of firms affected by the riots
❖ 6% the share of firms affected by internet shutdown
❖ 9% the share of firms that sustained damages to their properties
❖ 123,000 ETB the estimated value of self-reported property damage per firm
❖ 29% the estimated decline in monthly sales revenue
❖ 29% the share of firms that experienced disruptions in their supply chain
It is long recognized that instability is inimical to economic growth. Instability produces
uncertainty, amplifies risks, undercuts high-return investment, and diverts public policy towards
short-term and quick-fix policies. Instability, accompanied by violent protest and riots, not only
impacts current productive assets but also thwarts physical and human capital accumulation,
weakening future growth. In the last two years, non-state conflict, civil unrest, and violent protest
was rife in Ethiopia. While such conflicts are known to be barriers to a peaceful political and
economic transition, little is known about how they have impacted the private sector. In this brief,
we examine the economic impact of one specific event that led to the outburst of violent protests
on firms in Ethiopia.
Following the death of the Ethiopian singer and activist Hachalu Hundessa on June 29th, 2020, a
wave of largely peaceful protests took place in Addis Ababa and Oromia regional state.
Overshadowing the peaceful protest, violent demonstrations targeting businesses and
communities arose in the days following the singer’s death. More than 200 fatalities and property
damages estimated in the hundreds of millions of ETB were reported in Addis Ababa, Adama,
Shashamane, and other towns across Oromia. In an attempt to contain the violence, the
government shut down the internet connection for two weeks.
How did the riots and the internet shut down impact firms in Addis Ababa? To address this
question, we rely on data from a representative sample of firms in Addis Ababa, which was in the
process of being collected when the riots broke out. The high-frequency phone survey of firms
(HFPS-F), which follows the same firms each month for a period of eight months, was in its fourth
wave. The survey was intended to study the impact of the COVID-19 pandemic on firms in Addis
Ababa.1 Early evidence from the survey indicates that the economy is reeling from the pandemic
and firms have gradually started opening when the violence erupted in the early morning of June
30th 2020.
1
Further details about the population frame and sampling strategy is available here.
1
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
Using this dataset, this brief examines the direct impact of the political instability, riots, and
internet outage on 438 firms in Addis Ababa, which were interviewed immediately after the
incidents. We mostly rely on observations from the fifth round (July 23rd to August 15th, 2020) and
the sixth round (August 17th to September 8th, 2020) of the HFPS. While the violent protests in
Addis Ababa lasted only a few days, they seem to have left a scarring effect on firms and their
operations. We, however, recommend caution in interpreting some of these results as they are
mostly based on self-reported data and are not verified by objective indicators or administrative
data.
About 49 percent of firms in Addis Ababa reported to have been affected by the instability,
violence, or internet shut down associated with the riots. About 47 percent of the firms in the
sample reported to be affected directly by the unrest triggered by the riots (equal to 95 percent
of those that report to be affected somehow). About 6 percent of firms in Addis were affected by
the internet shutdown (Figure 1). Less than 1 percent indicated that the mobile phone service
shutdown was a problem during this time.
Firms experienced a significant shock to their sales revenues following the riots and internet
shutdown. Among those who stated that they had been affected by the riots, about 71 percent of
firms reported that their sales were affected negatively by the unrest or violent protest following
the killing of Hachalu Hundessa. Moreover, 29 percent experienced a notable disruption to their
supply chain and about 9 percent suffered direct physical damage to their properties (Figure 1).
Figure 1. Share of firms reporting to be affected by the instability between June and July 2020
Firms were asked whether and how they were impacted by the riots and unrest that ensued the
June 29th event. Figure 2 shows a closer look at the 47 percent of firms that were affected directly
by the unrest. The riots appeared to impact a greater share of larger firms than smaller firms — 55
percent of small, medium, and large (SML) firms reported to be impacted by the unrest. In
contrast, 44 percent of own-account firms reported that they were affected by the unrest during
the same period. Service sector firms are slightly more likely to be impacted by the riots compared
to firms in the industry sector. With 60 percent of firms reporting business disruptions due to the
riots, the 5-9 year age cohort of firms was particularly hard hit by the instability.
Figure 2 also suggests that slightly more male-owned firms than female-owned firms experienced
business disruptions due to the instability associated with the riots. This is partly explained by size
effects, where male-owned and operated businesses are likely to be larger and larger firms are
more likely to be affected by the unrest.
Firms that reported to have been impacted by the instability were on average closed for 6 days.
The number of days of closure does not vary much by sector, firm size, and sex of the business
2
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
owners. There is a slight variation by firm age, where the youngest firms were closed for 10 days
on average, while the rest were closed between 5 and 7 days (see Table 1).
Figure 2. Share of firms affected by the riots
Ethiopia lags far behind African peer countries in internet penetration, speed, and other quality
dimensions of internet usage. In an earlier brief, we report that about 20 percent of firms used
internet or online platforms for any business related activities. We thus do not expect that the
internet shutdown would affect a large number of firms which could trigger a business crisis.
Indeed, as indicated in Figure 1, only about 6.5 percent of firms report that they were mainly
affected by the internet shut down between June 29th and mid-July 2020. Compared to older
firms, a larger share of younger firms reported to have been negatively impacted by the internet
shutdown. Nearly one in eight (12 percent) new firms that started operations between 2019 and
2020 were affected by the internet shutdown compared to about 6 percent of firms with 10 or
more years of operational experience (Figure 3).
Figure 3 also shows that the internet shutdown affected a large share of SML firms than own-
account firms. Not surprisingly perhaps, young and large firms are more likely to run businesses
that depend on an online presence. We also see sector differences in the impact of internet
shutdown on businesses. About 12 percent of firms in the industry sector and 6 percent of firms
in the service sector report to have experienced disruptions to the business due to the internet
shutdown. There is, however, little variation in the impact of internet shutdown on male- and
female-owned firms.
3
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
Figure 3. Share of firms affected by internet disruptions by firm characteristics
Irrespective of the cause, riots can be extremely destructive. About 9 percent of firms reported to
have sustained damages to their properties following the instability that started in the early hours
of June 30th, 2020. Service sector firms were more likely to report physical damage to their
properties compared to firms in the industry sector — one in ten service firms experienced a
damage to their properties compared to only 1.5 percent of firms in the industry (Figure 4).
Anecdotal evidence and public disclosures suggest that the property damages include broken
windows, arson attacks, looting merchandise, and so forth.
While a relatively small number of firms were affected by damages to properties, the estimated
value of the physical damage for affected firms is large — and larger for the industry sector
compared to the service sector. Firms that reported to have experienced damage to properties
incurred an average self-reported loss amounting to 137,000 ETB in the industry and 123,000
ETB in the service sectors (Figure 5 and Table 2).
Larger firms were affected more seriously by the instability and the riots, both in the intensive and
extensive margins. The larger the firm, the higher the likelihood of being exposed to property
damage or vandalism and the larger the estimated value of the damage it sustains. SML firms
were nearly twice as likely to have been hit by property damages than own-account firms.
Similarly, the average estimated value of the property damage in SML firms is eight times larger
than the damage experienced by own-account firms and more than 30 times larger than by micro-
enterprises. Due to their visibility and their physical appearance on the main roads, that were
swarmed by the protests, large firms are often soft targets for attack, vandalism, and looting.
While the correlation between the probability of experiencing property damage and firm age is
less clear, the estimated value of physical damage seems to be largest in the oldest firms in the
sample. For example, about 4 percent of new firms reported to sustain physical damage to their
properties, estimated at around 4,000 ETB per firm. By contrast, of firms that were in operation
for 10 years or more, 9 percent experienced physical damages, estimated at 476,000 ETB per
firm. This is partly due to the seminal correlation between firm age and firm size where larger and
older firms tend to be affected more by violent protest, vandalism, and looting.
A larger share of male entrepreneurs sustained property damages to their businesses compared
to female entrepreneurs. Strikingly, the average male-owned firm reported and estimated
damage amounting to 137,500 ETB compared to a 3,500 ETB reported by female-owned
4
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
businesses. This is despite about 95 percent of women-owned firms operating in the service
sector, which is likely to be more strongly affected by the instability. Indeed, a closer look at the
data indicates that these female-owned firms that report to be impacted physically are own-
account firms that operate in the service sector specifically in the retail trade.
Figure 4. Share of firms that reported to have experienced physical damage to their properties
Figure 5. Estimated value of physical damage (in thousands ETB)
500
Value in thousnands of
400
300
200
100
ETB
0
0 years
1-2 years
3-4 years
5-9 years
Male
Industry
Female
Own-account
Micro (1-5)
SML (>5)
Services
10 years or
more
Sector Firm size Firm age Sex of owner
Even when firms do not directly sustain physical damages to their properties, their business
earnings may be affected by forced closures to prevent possible damages to their firms. Firms
were closed for an average of 6 days (Table 1) but firms that remained open might have scaled
back operations, for example, by closing earlier than usual or due to disruptions to their supply
networks. Moreover, we expect firms to experience a decline in demand as customers choose to
stay indoors and their spending confidence gets eroded due to the instability. Indeed, as shown
in Figure 1, the instability led to a perceptible decline in sales revenue and disruptions to the
business supply chain.
To examine the extent to which firm turnover was directly affected by the riots, we asked
entrepreneurs to indicate the percentage of monthly sales revenue that were lost due to the
inability to sell or produce during this period. Figure 6 shows self-reported revenue impacts
experienced by firms in our sample. About 29 percent of monthly sales revenue—26 percent in
the industry and 29 percent in the service sector—were lost due to the instability and the
5
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
accompanying shock in demand. Strikingly, more than a third of own-account firms’ and a quarter
of micro-enterprises’ sales revenues were wiped out during this period. Larger firms experienced
a 19 percent decline in sales turnover during the same period. Similarly, the youngest firm in the
age-group under the study reported the largest decline in sales revenues followed by those that
are 1-2 years in operation. The reported decline in monthly sales revenues is also comparable for
male- (29 percent) and female-owned firms (27 percent).
Figure 6. Reported monthly sales revenue lost (percent)
41
35
33
31
Sales revenue lost (%)
29
29
27
26
26
22
19
17
Female
Micro (1-5)
SML (>5)
Male
10 years or
0 years
Industry
1-2 years
3-4 years
5-9 years
Services
account
Own-
more
Sector Firm size Firm age Sex of owner
The survey also inquired about potential disruptions to the firms’ supply chain following the
instability and riots. About 29 percent of firms that reported to be impacted by the instability
stated that their supply chains were disrupted due to the riots and unrest following Hachalu
Hundessa’s death (Figure 1). A third of firms in the industry and 28 percent of firms in the service
sector reported that the instability (unrest, internet or mobile service shutdown) lead to significant
supply disruptions (Figure 7 and Table 3). Most of the disruptions are related to the unrest
following the riots and 9 percent in the industry and 5 percent in the service sector identify the
internet shutdown to be responsible for supply chain disruptions.
The supply networks of own-account firms are slightly less likely to be impacted by instability
compared to be micro- and SML enterprises. Smaller share of younger firms compared to older
firms report supply disruptions due to the instability. However, a larger share of these firms were
impacted by internet shutdown. Supply disruptions were also observed across different firm age
cohorts and irrespective of the sex of business owners.
Figure 7. Share of firms reporting supply disruptions (percent)
30.8
33.9
28.3
31.3
28.5
28.9
33.9
29.4
26.5
32
24.9
21.9
SHARE OF FIRMS (%)
MICRO (1-5)
1-2 YEARS
3-4 YEARS
5-9 YEARS
SERVICES
SML (>5)
0 YEARS
FEMALE
INDUSTRY
MALE
OR MORE
10 YEARS
ACCOUNT
OWN-
SECTOR FIRM SIZE FIRM AGE SEX OF OWNER
Table 1. Share of firms affected by sector, size, age and sex of the owner
6
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
Unrest, internet or Mobile
Internet Number of
mobile service Unrest service
shut down days closed
disruptions shutdown
Sector
Industry 48.2 44.6 11.9 0 6
Services 49.5 47.7 5.6 0.7 6
Firm size
Own-account firm 46.3 43.8 5.6 0.5 6
Micro-enterprise (1-5) 51.2 49 7 0.2 6
Small, medium or large enterprise (>5) 54.6 54.6 8.2 1.8 5
Firm age
0 years 37.8 34.5 11.7 2.4 10
1-2 years 49.5 49.5 5.6 0 5
3-4 years 45.4 41.9 4.1 0 7
5-9 years 59.9 59.9 5.1 0 6
10 years or more 47.4 44.1 5.9 1.5 5
Sex of owner
Male 50.5 48 6.4 0.8 6
Female 46.3 45.5 6.8 0 5
Total 49.3 47.3 6.5 0.6 6
Table 2. Reported lost sales and physical damage firms incurred during the unrest
Estimated value of
Incurred physical Monthly Sales Revenue
physical damage (in 000
damage (%) Lost (percent)
ETB)
(1) (2) (3)
Sector
Industry 1.5 137 26
Services 10 122.9 29
Firm size
Own-account firm 7.6 50.7 35
Micro-enterprise (1-5) 8.3 13.8 26
Small, medium or large enterprise (>5) 14.4 440.2 19
Firm age
0 years 4.4 4 41
1-2 years 11.6 23.6 33
3-4 years 8.4 2.6 22
5-9 years 6.7 87.4 31
10 years or more 9 476 17
Sex of owner
Male 10.7 137.5 29
Female 3.6 3.5 27
Total 8.8 123.3 29
7
Monitoring COVID-19 impacts REPORT NO. 11
on firms in Ethiopia 21 / December / 2020
Table 3. Percent of firms that report supply chain disruptions
Unrest, internet or Mobile
Internet
mobile service Unrest service
shut down
disruptions shutdown
(1) (2) (3) (4)
Sector
Industry 33.9 28 8.5 0.7
Services 28.3 26.8 4.8 0.3
Firm size
Own-account firm 26.5 24.5 4.2 0
Micro-enterprise (1-5) 31.3 28.8 7.2 0.2
Small, medium or large
32 30.3 3.5 1.8
enterprise (>5)
Firm age
0 years 21.9 20.2 7.1 0
1-2 years 28.5 28.1 5.6 0
3-4 years 28.9 24.2 4.7 0
5-9 years 33.9 33.9 3.2 0
10 years or more 29.4 26.1 4.8 2.1
Sex of owner
Male 30.8 28.3 5.4 0.5
Female 24.9 23.7 5.4 0
Total 29.1 27.1 5.4 0.3
This summary brief is the fourth in a series of briefs on special topics and the eleventh on those
reporting on the findings of the high-frequency phone survey of firms. Each round’s summary
brief, table of indicators, and microdata are made available via the following website:
https://www.worldbank.org/en/country/ethiopia/brief/phone-survey-data-monitoring-covid-19-
impact-on-firms-and-households-in-ethiopia.
The high-frequency phone monitoring survey monitors the economic impacts of and responses to the COVID-19 pandemic on firms with a
focus on the effects on firm operations, revenues, and jobs. We call a sample of firms every three weeks between mid-April and mid-
September 2020 for a total of eight survey rounds in Addis Ababa and seven survey rounds in other cities. The final dataset will consist of a
panel of approximately 800 firms (500 in Addis Ababa and 300 in four other cities—Adama, Bahir Dar, Hawassa, and Mekelle).
The sampling procedure was undertaken in three steps. First, the team cleaned the list of registered firms in number in Addis Ababa, Adama,
Bahir Dar, Hawassa, and Mekelle, received from the Ministry of Trade and Industry (MoTI), by removing firms with missing or invalid phone
numbers. Second, all phone numbers of the cleaned list of firms were shared with EthioTelecom and only active phone numbers were kept
constituting the sampling frame. Third, two survey domains were selected (Addis Ababa and other cities) and the team drew a random
sample of firms without replacement, stratified by firm size (proxied by capital) and sector (industry and services). The sample size was set by
a tight budget envelope, with a panel of 500 firms in Addis Ababa and 300 firms in other cities after round 8 (7) of the survey operation.
Expecting a high non-response rate, we drew a sample of 1,550 firms for Addis Ababa and 800 (200 in each city) firms in other cities.
8