Domestic water charges in Ireland - issues and challenges conveyed through social media
Theo Lynn and Martin Quinn*, DCU Business School.
Binesh Nair, Irish Centre for Cloud Computing and Commerce, Dublin City University
Stephen Jollands, University of Exeter.
*Corresponding author: [email protected], DCU Business School, Glasnevin, Dublin 9,
Ireland
Abstract
This paper analyses a Twitter dataset to explore water governance and stakeholder engagement
during the introduction of domestic water charges in Ireland. The results highlight active Twitter
use during the analysis period, reflective of widespread protest centred on a new utility, Irish
Water. The analysis shows protest activities were dispersed and not cohesive, with tweets largely
focused on economic and political issues and not on the provision of a sustainable water supply.
The findings extend our understanding of these events and provide some insights into the role of
social media in water governance and stakeholder engagement issues in an Irish and wider
context.
Keywords: social media, water management, sustainable water supply, stakeholders.
1
1. Introduction
The Republic of Ireland (hereafter Ireland) has a history of a fragmented domestic1 water supply.
Lack of domestic supply charges over time has led to inadequate infrastructure and related issues
(Brady and Gray 2010; Jollands and Quinn 2015). Researchers have suggested such symptoms
are indicative of poor water governance - the policies, procedures and management to ensure a
sustainable supply of fresh water (Cosgrove and Loucks 2015; Jollands and Quinn 2015;
Stefanovic et al. 2015). This poor governance emerged as an issue within the bailout of the Irish
economy, with a condition being the introduction of domestic water charges. While other
austerity measures were implemented in the bailout, none encountered a public reaction like the
introduction of charges.
The formation of Irish Water and the introduction of domestic charges aimed to provide
resources to improve water governance. As Jollands and Quinn (2015) note, despite the appeal of
the Irish Government to accounting concepts – e.g. cost and investment – to legitimise change,
widespread protest resulted. Evidently, the Irish Government did not recognise a critical success
factor would be broad public support (Dimadama & Zikos 2010; Lennox et al. 2011; Walker et
al. 2015). As noted by Cosgrove and Loucks (2015,4836), “the biggest issue or constraint in the
future will remain what it is today: namely the human component of water management, not the
technical one”.
Jollands and Quinn (2015) outline how the creation of Irish Water and introduction of domestic
charges was fraught and stakeholder engagement was relatively ineffective - there was no invited
public participation in the formation of Irish Water or setting of charges. However, their analysis
1
Domestic water supply refers to water supplied through a water main system by a state-owned or private utility.
2
did not examine the use of social media - suggested by Laspidou (2014) as having a role in
stakeholder engagement within water governance. Thus, the objective of this paper is to extend
their work utilising social media data from Twitter, to explore the extent to which stakeholders
are focusing on issues other than the sustaining water supply and, thereby, the effectiveness of
stakeholder engagement. The use of social media is changing the mechanics of information
diffusion (Stieglitz and Dang-Xuan, 2013). Lynn et al. (2015) suggest an integrative view where
social media is defined as comprising both the conduits and the content disseminated through
interactions between individuals and organizations (Kietzmann et al. 2011). In this view, social
media allows users to create (consume) content that can be consumed (created) by others and
enables and facilitates connections (Hoffman and Novak 2012). This ability for the public to
both publish and consume content but also to identify and connect with others of
similar/opposing views is transforming the political and societal landscape (Agrawal et al. 2011),
including stakeholder engagement in water governance (Laspidou 2014).
The remainder of the paper proceeds as follows. The next section provides the historic context of
water governance in Ireland. Then, Section 3 outlines our method, which consists of four stages.
Section 4 presents the results of the Twitter data analysis and finally Section 5 discusses the
results in the context of stakeholder engagement.
2. Background to water governance in Ireland
This section provides background on Irish water governance over time. Domestic water charges
were introduced in Ireland in October 2014, however the history leading to these charges can be
3
divided in pre- and post-2010. The latter period coincides with Irish economic problems and the
formation of Irish Water, the country’s first water utility.
2.1 Water services pre-2010
Despite abundant rainfall, Ireland has had a poor treated water infrastructure. Increased
population and urbanisation (particularly in Dublin), alongside ongoing lack of investment has
put pressure on water distribution systems (Cashman 2011; Irish Water 2015). Until recently,
Ireland was the only OECD country to not levy domestic water charges. The Local Government
(Sanitary Services) Act, 1962, allowed a local authority (a town, city or county council) to
charge business users only, while domestic users paid rates for local authority services. Rates
were abolished in 1978, and replaced by a rates support grant. A new government in 1983
allowed local authorities to levy domestic water charges. To the early 1990’s, few authorities
levied charges but the largest authority, Dublin County Council, did not. To deal with increasing
population (and other strategic matters), in 1993 Dublin was divided into four authority areas. In
early 1994, three of these authorities introduced a flat domestic water charge. Intense protest
followed, and by 1997 a law was passed to forbid domestic water charges. Local authorities thus
depended more on central government funding. However, an economic boom from the late
1990’s saw increased construction activity and a development levy became a major income
stream from 2004 until 2007/8. The Report of the Comptroller and Auditor General (2010) notes
the investment in water services in Ireland was €5.2 billion in the period 2000-2010, of which
€4.2 billion was spent on new and/or upgraded water infrastructure in urban areas. Despite this
investment, unaccounted for water (UFW) averaged 41% in 2008 which were “levels twice the
OECD average” (ibid). Hence, even with large investments made, a sustainable water supply was
elusive. This is illustrated in examples such as an outbreak of waterborne cryptosporidiosis in
4
Galway (RTE, 2007), and areas of Roscommon were subject to boil water notices from the same
cause (Irish Times, 2014a).
2.2 Water services post-2010
In September 2008, an Irish banking crisis and a global economic crisis had serious effects on
government revenues. On November 21st, 2010 the government accepted a bailout package from
the European Union and the International Monetary Fund. This was a key point in a sequence of
events which resulted in domestic water charges and the formation of Irish Water. The 2010
Budget speech (Lenihan 2009) a few weeks later noted:
The Renewed Programme also contains a commitment to introduce a system of water
metering for homes […]. Water charges, when introduced, will be based on consumption
above a free allocation.
A National Recovery Plan 2011-2014 provided details on domestic water charges, making it
clear the government required a revenue stream to “improve the General Government position”
(p.78) and cover the operational and capital costs of providing domestic water. Simultaneously,
the government negotiated a Memorandum of Understanding (MOU) with its bailout partners.
The MOU committed to recovering the cost of water service provision - a water tax for fiscal
rather than environmental reasons (Wu et al. 2011). A new government was elected in 2011.
They introduced a differing view on operationalising the previous commitment on domestic
water charges, forming Irish Water through the Water Services Act 2013 (WA2013). This Act
vests power in Irish Water to install water meters at domestic premises, and installation began in
5
2014. This is in direct contrast to the flat rate charged in the 1990’s, and follows the “polluter
pays” principle set out in the EU Water Framework Directive.
The WA2013 has two provisions on charging 1) it removed previous bans on domestic charges
and, 2) it appointed the Commission for Energy Regulation (CER) as the authority to approve
charges. By September 2014, the CER had levied a charge of €4.88 per 1000 litres - with some
free volume-based allowances - applicable to December 31st 2016. Following the publication of
these charges, a period of nationwide protest followed. The initial response of the government
and Irish Water was the mobilisation of accounting concepts to address protestor concerns
(Jollands and Quinn 2015). Specifically, these attempts used the language of accounting to
demonstrate the need for Irish Water to cover costs of domestic water provision. However, the
concept of cost was contested by protesters, who associated it with the impact on their resources
and lives rather than to the needs of a water utility (ibid). By the third week of November 2014, a
flat annual charge of €260 per family/€160 per single occupancy household was conceded. These
charges over-ruled the CER's recommendation, and were embodied as the Water Services Act,
2014.
While the government gave ground to protesters, shifting from a volumetric basis to a flat
charge, they also simultaneously implemented a strategy of proliferation (Callon and Law 2005).
Proliferation is “the opening up of things with the inclusion of so many entities that interact with
one another in competing ways and thereby the creation of too much entanglement” (Jollands
and Quinn 2015:8). Simply put, Irish Water’s installation of hundreds of thousands of water
meters made chances of reversing the charges unlikely, and leaves as a possibility of billing
6
volumetrically. The very artefact used by Irish Water to implement this strategy of proliferation
became the focus of protests. For example, groups calling themselves ‘fairies’ used social media
(including Twitter) to advertise how water meters could ‘disappear’ (Irish Independent 2015a).
By the third quarter of 2015, data had been collected from installed meters and households were
in their second billing period, with over 50% of users paying bills. Previously, due to lack of
metering, little was known about Ireland’s water usage (Lyon et al 2010). Thus it is not
surprising that the data has been used to identify leaks and Irish Water claim to have saved 3
billion litres in the six months to September 2015 (Irish Independent, 2015b)
3. Methods
We now detail the methodology utilised to obtain and analyse Twitter data for this study. Twitter
is a popular SNS, reporting 320 million active users worldwide in 2016. It is an open social
network where public messages (tweets) can be seen by everyone. It connects strangers with
common interests, identifying messages by hashtags. This combination of an open network, a
large user base, high volumes of messaging and topic identification makes Twitter a rich data
source (Lynn et al. 2015). It has been utilised in studies of marketing (Jansen et al., 2009),
politics (Tumasjan et al. 2010), finance (Bollen et al. 2011), health (Paul et al. 2011), supply
chain management (Chae 2015) and conflict (Siapera et al. 2015).
As noted earlier, social media may have an important role in stakeholder engagement within
water governance (Laspidou 2014). However, using social media data as a research medium in
this realm is scarce. A search in early 2016 of leading water-related journals - Water Resources
Management, Environmental Processes, European Water, Water Utility Journal, Water
7
Resources Research, Water Policy, Water Research, Water Resources and Economics, and
Ecological Economics - reveals mention but not use of social media/Twitter as a data/analytical
source. Given the use of Twitter in several academic fields, we suggest it is a data source to gain
valuable insights on stakeholders in and around water governance, and it is a platform where
water issues are being discussed. Using http://topsy.com2 on October 21st 2015 the term
“california drought” revealed 28,859 tweets over the previous month. These tweets were from
the Washington Post, CBS, National Geographic, politicians, utilities and private citizens and
mentioned
millennia-old
sequoias,
beekeepers,
wildlife,
shortage
of
water
for
lawns/agriculture/brewing, geo-engineering and water conservation. A search for the more
general term “drought” using the same criteria revealed 195,605 tweets. Additionally, as Cabrera
et al. (2013), Cosgrove and Loucks (2015), Hazelton (2013) and Hunt et al. (2013) note, water is
an inherently political issue. However, political interests tend to not participate in in-depth
research. While no substitute, social media often includes comment from political sources in a
less formal sense than official documents, and thus may provide insights into policy issues - see
Marwick and Boyd (2011).
Our method had four stages:
● Data provisioning - data was collected from DataSift, a data aggregation service, for
January 2014 to June 2015. We created filters using appropriate keywords such as “Irish
Water”, “Irishwater” and “water protests”. The time period covers the formation of Irish
Water, the introduction of domestic water charges and subsequent protest. The final
dataset contained approximately 350,000 records.
2
We should note that topsy.com was shut down in December 2015 by its parent Apple.
8
● Data storage - the data acquired was stored in Google BigQuery to enable ad-hoc
querying and access.
● Data extraction - we used Tableau 9.0 - a data visualisation/analysis software - to extract
data and perform basic descriptive analysis.
● Data analysis - we used R (an open source statistical tool) for pre-processing the data and
for complex analysis. Specifically, we used the ‘tm’ library in R for text analysis. We
used Gephi (an open-source graph and network visualisation tool) for complex network
analytics.
4 Results
The results are now presented drawing on the descriptive, content and network categories set out
by Chae (2015).
4.1 Descriptive analytics (DA)
DA communicates insights on underlying data. Insights on the number of tweets and retweets
were obtained through SQL, whereas R was utilised for insights requiring more complex
processing e.g. the average number of hashtags in tweets.
The dataset comprised 354,739 tweets, of which 171,136 (48%) were original tweets and
183,603 (52%) retweets. In the original tweets, there were 11,622 unique hashtags including
‘#right2water’,
‘#irishwater’,
‘#watercharges’,
‘#dec10’,
‘#antiausterityallianceireland’,
‘#gtadublin’, ‘#bullygovernment’, ‘#gardainottheenemy’. The average number of hashtags in a
tweet was 0.617 and 113,939 tweets (67% of original tweets) had no hashtags. Of the latter, 38%
were replies. Each screen-name represents a discrete account, and a user may have multiple
9
accounts3. We identified 33,433 unique screen-names, with an average of 5.12 tweets and 5.49
retweets per screen-name - the most active and visible users are shown in Figure 1. The most
active users were identified by volume of tweets, retweets and replies. Visibility of users was
measured by sum of the number of retweets and the number of replies each user received. The
analysis revealed that the most active users (e.g. @chezmik and @pauljob5) were not the most
visible users. Similarly, some of the most visible users (e.g. @IrishWater and @LegalEagleStar)
were not the most active. Thus, in general active users were not necessarily visible and visible
users were not necessarily active.
URLs (website links) were widely used by active and visible users, with 60% of tweets
containing at least one URL. The most popular links were www.facebook.com (12,530),
www.youtube.com (9,354), www.journal.ie (6,282), www.ffwireland.blogspot.ie (3,462),
www.irishtimes.com (2,921) and www.independent.ie (1,620). These represent two of the most
popular social networks, a popular Irish news website, two Irish broadsheet newspapers and a
blog by an environmental scientist.
(insert Figure 1 here)
4.2 Content analytics (CA)
CA provides insights from unstructured data (Chae 2015). We conducted word, hashtag and
sentiment analyses. First, we identified frequent words in tweets by transforming the data into a
Term-Document-Matrix (TDM). In a TDM, rows represent tweets and columns correspond to
terms. As the resulting matrix will be highly-sparse, sparse terms are removed to condense the
3
There are approximately 660,000 screen names in Ireland.
10
matrix. The sparsity factor was set at 0.99, meaning all terms not appearing in 99 percent of the
tweets were removed.
The most frequently occurring words were ‘water’ (155,567), ‘Irish’ (129,342) and ‘Irishwater’
(55,854). Others included ‘rightwater’ (48,053), ‘protest’ (24,120), ‘charges’ (12,290), ‘people’
(10,891), ‘Dublin’ (6,707), ‘bill’ (6,678), ‘pay’ (6,611), ‘meter’ (6,179), ‘government’ (5,528),
‘dec’ (4,965), ‘public’ (3,477), ‘tax’ (3,460), ‘march’ (3,329), and ‘protesters’ (3,175). We used
the ‘ngram’ library in R to find frequently co-occurring words. ‘Irish water’ was the most
frequently co-occurring words (116,297). Others included ‘water protest’ (14,653), ‘water
charges’ (7,329), ‘irishwater rightwater’ (6,973), and ‘irish water charges’ (3,211).
We further clustered tweets into two themes, political/economic and sustaining water supply.
The themes were identified by keywords, developed from extant literature (Jollands and Quinn
2015; Barnes and Alatout 201; Cashman 2011; Lewis and Russell 2011; Kurland and Zell 2010)
and a word-search from http://topsy.com (see Appendix 1). The political/economic theme had
68,275 tweets, with 7,486 for sustaining water supply. We identified frequent and co-occurring
words for each of these themes, as shown in Table 1.
(insert Table 1 here)
Second, the hashtag analysis revealed many unique hashtags (11,622). We identified the most
popular hashtags on their usage in tweets/retweets. The most popular hashtags included
‘#right2water’, ‘#irishwater’, ‘#vinb’, ‘#watercharges’, ‘#dec10’, ‘#ireland’, ‘#nov1’, ‘#water’,
‘#news’, ‘#rtept’, ‘#wewontpay’, ‘#irish’, ‘#dail’, ‘#not1pipe’, ‘#dublin’ and ‘#waterprotest’.
11
These can largely be classified into three categories: (i) campaign hashtags (e.g. ‘#right2water’,
‘#wewontpay’ and ‘#not1pipe’) and protest dates (e.g. ‘dec10’ and ‘#nov1’), (ii) media coverage
(e.g. ‘#rtept’), and (iii) geographic hashtags (e.g. ‘#dublin’).
Third, sentiment analysis provides insights on the orientation (positive/negative) and intensity
(strong / weak) of opinions (Pang and Lee 2011). This analysis provides insight on public
attitude, which reflect the offline landscape (Bae and Lee 2012). A lexicon-based approach was
used (Bollen et al. 2011; Gilbert and Karahalios 2010; Tumasjan et al. 2010). A sentiment score
per tweet was calculated on -20/+20 scale, where 0 is neutral, 0-20 have positive sentiment, 0 to 20 have negative sentiment. The analysis showed 56% (95,093) of original tweets were neutral,
the remainder being mostly negative (29%) with 16% highly negative (< -4). Positive tweets
represented 15% (26,751) of total original tweets. The top graphic of Figure 2 depicts the
sentiment distribution of the dataset.
(insert Figure 2 here)
We also performed sentiment analysis on clustered tweets. The sentiment distributions in the
clusters have a similar pattern to the total tweets. However, tweets in the political/economic
cluster had a more negative skew than the sustaining water supply cluster (Figure 2, bottom
graphic). Table 2 gives some exemplars of intense sentiments, and highlights difficulties using
sentiment analysis for explanatory value. The two most positive tweets reflect a satirical piece
(in reality negative) and a sarcastic tweet. Exemplar intense sentiment tweets for each cluster are
also shown in Table 2.
12
(insert Table 2 here)
4.3 Network Analytics (NA)
NA interprets social network patterns including the degree of sparsity of a network, the average
distance between any two nodes (or users) and the network density. NA is useful to discover
influential users and the key brokers. The first component of NA is a topological analysis. The
original Irish Water network had 20,141 nodes and 36,814 edges. We constructed a
representation of this using the @reply4 information. Nodes represent the screen-names who
have sent/received an @reply. Edges represent the relationship between the screen-names linked
through the @reply. To focus on influential nodes in the network, we constructed a
representation of a sub-network composed of nodes with a minimum in-degree of 12 (nodes
which have received at least 12 replies) – resulting in a sub-network with 373 nodes and 3,997
edges. The resulting representation demonstrated the underlying network is well-connected. The
average path length - the average distance between any two nodes - in the network is 2.718 i.e.
every user is about three nodes away from each other. A short average path length reflects the
presence of popular brokers or hubs (which lead the users to influencers) in the network. The
average degree of the network5 was 21.432, indicating the presence of highly influential users in
this sub-network i.e. each node in this network has received at least 21 replies from the network.
The diameter, which represents the longest path (between two nodes) in the network, was 6. A
relatively low diameter is an indication of the presence of popular hubs in the network.
4
An @reply is tweet which is a reply to an original tweet. The source of the @reply is the sender and to whom the
reply is intended is the target. @reply thus establishes a link between two users in a network.
5
Average degree of a network represents the average degree of all the users in the network. It considers both the indegree and the out-degree of a user. In-degree determines the number of replies a user receives to tweets and outdegree determines the number of replies a user sends to other users in the network.
13
A second component of NA is centrality analysis. In-degree is a simple measure of a node’s
connectedness with others and can be an indicator of a user’s popularity (Chae 2015). We used
in-degree to identify the most influential screen-names in the network. The top five were
@IrishWater (in-degree of 165), @Revolution_IRL (57), @williamhboney1 (46), @babsbear
(44), and @JFTAXI (43). Another important node-level metric is betweenness centrality, which
identifies key hubs or brokers in the network. These are the nodes which leads the users to the
influencers. Key brokers have a high value for betweenness centrality. In this data the key
brokers are @IrishWater (betweenness centrality score of 31,256.81), @babsbear (5,605.79),
@GallMandy (4,513.59), @countryboy606 (4,046.91), @farrelleye (3,733.49), @Castletonian
(3,722.39) and @Revolution_IRL (3,629.18).
A third component of NA is community analysis, measured using density. The density of a
network graph is the ratio of the actual number of edges and the number of possible edges. A
dense graph will have a ratio closer to 1 (Coleman and More 1983). The density of the Irish
Water graph is 0.029 indicating a strong sparsity in the network i.e. the network is not cohesive,
with many dispersed groups. We used the community detection method (Blondel et al. 2008)
embedded in Gephi to identify communities. The method found 14 communities reflecting the
dispersion indicated by the density. The three largest communities had 66, 65 and 59 members.
Seven communities had one member each. The largest community had an average degree of
14.515 -, less than that of the sub-network - indicating the presence of nodes which engage less.
The diameter was 5 and the average path length was 2.282, which is close to that of the sub-
14
network. The density of the largest community was found to be 0.196, suggesting that the
community is less dense compared the sub-network.
5. Discussion and final comments
We have noted the history of water management in Ireland and detailed our social media
collection and analysis methods within the critical period of January 2014 to June 2015 - a period
when Irish Water and domestic water charges were implemented. We now turn to discuss what
these sections tell us in terms of water management.
Given the volume of tweets posted over the period, it is reasonable to state that social media was
intensively utilised around the establishment of Irish Water and domestic charges.
Deconstructing the tweets into sustainable water supply versus political/economic issues showed
the latter were in the majority - by almost tenfold. Combining this with findings from Jollands
and Quinn (2015) illustrates that tackling socio-economic issues, which are major impediments
to tackling water issues, is difficult. This suggests the first steps to any water management
program should include efforts to address potential socio-economic issues (Cosgrove and Loucks
2015). What is apparent here is that social media may increasingly undertake a role within this
process. Our data set shows Irish Water as a visible Twitter, but not very active (as per the NA).
The discussion on Twitter, and indeed generally, oriented towards issues of a political/economic
nature rather than on sustainable water supply. Irish Water’s engagement in this discussion
actively and constructively would seem to be limited and piecemeal. This may be explained by
the political/economic backdrop and Irish Water’s chosen communication strategy. While
Twitter can provide a vehicle for stakeholder engagement; it can only be considered a voice
15
within this process if listened to Irish Water's engagement on Twitter may be considered
reflective of a wider lack of true engagement resulting in a lack of trust with stakeholders. Such
“piecemeal reactions” may not be enough to achieve a “more sustainable and secure future”
(Cosgrove and Loucks 2015) of water resources in Ireland.
Jollands and Quinn (2015) note the concept of cost was utilised by the government and Irish
Water to introduce domestic water charges, whereas previous attempts had failed dramatically.
They note the importance, through proliferation (Callon and Law 2005) of the water meter in this
process. The measure of success of this strategy can be seen in that, as of January 2016, 61% of
Irelands’ domestic users were paying their bills (Irish Water 2016). Although there were some
large protests with success in gaining a flat rate charge, the protest movement has not been
entirely successful and our tweet analysis extends understanding of reasons for this. The tweets
suggest a sparse and dispersed network of users, indicating a dispersed and broad protest
movement, inhibiting efforts to be cohesive and gain momentum. This reflects elements of the
actual physical protests, namely 1) while on the whole protests were peaceful, they attracted a
cross-section of society, including some violent elements; 2) they became less widespread and
less in number over time, and 3) they became more than about water and lost potential support
(Jollands and Quinn 2015). We can see that, in the sentiment analysis, while the vast majority are
neutral, tweets with negative sentiment are more common than positive sentiment. A visual
inspection of neutral and positive tweets (see Table 2) points to the overall sentiments of the
tweets being more negative than the analysis suggests. This implies that on the whole, the
average Twitter user in this dataset was not satisfied with the ongoing activities of the Irish
Government and Irish Water.
16
Given the large volume of tweets in relation to this issue and their general negative sentiments, it
seems surprising protests did not gain more momentum – at least during our analysis period.
Beyond the dispersed nature of the network of Twitter users reflecting lack of cohesion, part of
the reason may be related to the types of topics contained within these tweets. The introduction
of subjects beyond the introduction of domestic water charges, for example fluoridisation, by
some of the more active Twitter users may have diluted the message around domestic water
billing and alienated many potential supporters of protest. Right2Water, for example, were
responsible for organising many protests on the introduction of the billing regime and the
activities of Irish Water. The results illustrate that Right2Water was not an influential Twitter
user on this issue - while they had a well-used hashtag, they did not feature in the network
analytics as a top broker. This is likely due to their focus moving beyond water issues to other
social issues (Irish Times, 2015). With such drift from the core issue of domestic charges, it is
not surprising that the Twitter network was less cohesive and a significant number of people paid
their bill - rather than non-payment in protest. After our research period, a national election in
February 2016 saw many politicians campaigning on the abolition of water charges. The result
was about 90 (or 56%) of elected members who were ‘anti-water charges’ (Irish Independent,
2016). This popular vote confirms to an extent the negative sentiment of our tweet analysis, and
confirms the focus on water charges as opposed to sustainable water supply.
Cashman (2011 155) suggests that symptoms of poor governance include “…high unaccounted
for water, lack of proper metering, ineffective collection of water revenue, uneconomic tariffs”).
Jollands and Quinn (2015) noted many of these symptoms were present in the water governance
structures in Ireland in the decades before Irish Water. While many of these symptoms persist,
17
the establishment of Irish Water has started to address some (such as unaccounted for water and
metering) but there is still a general perception that many issues persist, as reflected in our
Twitter dataset. As such, Irish Water may need to consider the best media (including traditional
and social media) to engage with stakeholders to change such perceptions. Whether it be
perception or reality, it is important to remember that Irish Water was set up in a relatively short
period of time and it will require significant adjustment over time. Irish Water management may
not have assisted the situation through some actions taken, and this is played out in the analysis
above. For example, the most frequent words and hashtags illustrate a focus on the
political/economic side and less on issues of sustainable water supply. Hence, while the protests
have been less than successful in achieving their aims, they have dragged Irish Water into
engaging with other issues rather than on the technical aspects of water supply.
As there are many challenges facing Irish Water, not least the issues of serving the growing city
of Dublin (Padowski and Gorelick 2014), there is a strong need for them to refocus on the
governance required to provide infrastructure to maintain a sustainable water supply. This may
require the political process to isolate Irish Water more from the socio-economic issues. This
will require a consensus of political will that given the history of water supply in Ireland may be
hard to achieve. With no political will to shield Irish Water, they will need to engage more
clearly to the wider stakeholder community. They need to communicate the need for funding
from domestic charges to build a sustainable water supply. This may entail considering how best
to use Twitter and other social media. In introducing domestic charges and the establishment of
Irish Water, the socio-economics has not been done to any level of proficiency. While there is
enough water on the planet for every person, Cosgrove and Loucks (2015) note that its
18
distribution means an inadequate supply for some regions to meet development and
environmental needs. In Ireland, a lack of water is not one of the issues affecting a sustainable
supply – as noted in Table 2, “all it does is rain in this country”. Despite the abundant rainfall,
the lack of political will, general unrest and governance issues has meant there are still issues in
meeting Ireland's’ water needs. If these issues are so pronounced in a country of abundant
supply, then this could be even more pronounced in a country where there is inadequate supply.
However, this very lack of supply may be the factor that prompts the required political will and
stakeholder support to address such issues, implying that the abundant supply of water is at the
very heart of the issues in Ireland.
A final implication is it is clear that stakeholder engagement/management was not ideal
throughout the analysis period. This is evident in the content analysis and specifically the
sentiment analysis. Tweets with negative sentiments were more common than positive, and on
the whole potentially more negative than the analysis suggests (Table 2). This suggests
stakeholders were not brought along in the process. Stakeholder engagement and management is
a cornerstone of good water governance (Harvey and Schaefer 2001; Kennedy 2011; Lennox et
al. 2011; Lewis and Russell, 2011). This is reinforced by Cosgrove and Loucks (2015 4836))
who note “(m)ost paths to sustainable development are linked to water, but the decisions that
determine how water resources are used or abused are not made by water managers alone”. We
can see from our analysis a failure to implement an appropriate stakeholder engagement and
management strategy, which resulted in socio-economic issues impeding the process of
constructing a sustainable water supply. If we assume what is reflected in Twitter is reflected
across all social media, then these platforms were not used particularly well to assist with this.
19
To sum up, we could question what this reveals for future policy on water management. The
overarching theme is that there is an integral need to bring stakeholders along. Within the
modern technologically connected era, it is clear that social media is an active means by which
stakeholders communicate. This suggests those involved in water management may need to be
more proactive in analysing and using social media for stakeholder engagement/management
purposes. We should note a limitation of our methods here. Sentiment scores are, as noted,
computed using a lexicon based approach and may not reveal the true sentiment of a tweet sarcasm for example. Thus, the tweets are more negative than the sentiment analysis suggests.
However, this limitation does not affect our discussion above. On the contrary, it adds more
weight to our point that stakeholder engagement on the issue of a sustainable water supply could
have been better, as borne out in the February 2016 election.
20
Acknowledgments
This research was supported by the DCU Office of Vice President for Research Business
Innovation Platform.
21
References
Agrawal D, Budak C, El Abbadi A (2011) Information diffusion in social networks: Observing
and influencing societal interests. Proceedings of the VLDB Endowment, 4(12):1-5.
Barnes, J, Alatout, S (2012) Water worlds: Introduction to the special issue of Social Studies of
Science. Social Studies of Science, 42(4):483-488.
Blondel, VD, Guillaume, JL, Lambiotte, R, Lefebvre, E. (2008) Fast unfolding of communities
in large networks, Journal of Statistical Mechanics: Theory and Experiment, 2008(10):
P10008.
Bollen J, Mao, H, Zeng, X (2011) Twitter mood predicts the stock market, Journal of
Computational Science, 2(1):1-8.
Brady, J., Gray, N.F., 2010. Group Water Schemes in Ireland – Their Role within the Irish Water
Sector. European Water 29: 39-58.
Cabrera E, Pardo MA, Cabrera Jr E, Arregui FJ (2013). Tap water costs and service
sustainability, a close relationship. Water Resources Management, 27(1): 239-253.
Callon M, Law, J (2005) On qualculation, agency, and otherness. Environment and Planning D:
Society and Space, 23(5):717-733
Cashman A (2011) ‘Our Water Supply is being Managed like a Rumshop’: Water Governance in
Barbados, Social and Environmental Accountability Journal, 31(2):155-165.
Chae BK (2015) Insights from hashtag# supplychain and Twitter Analytics: Considering Twitter
and Twitter data for supply chain practice and research, International Journal of Production
Economics, 165:247-259.
Coleman T, More, J (1983) Estimation of Sparse Jacobian Matrices and Graph Coloring Blems,
SIAM Journal on Numerical Analysis, 20(1):187-209.
22
Cosgrove, WJ, Loucks, DP (2015) Water management: Current and future challenges and
research directions, Water Resources Research, 51(6):4823-4839.
Dimadama Z, Zikos, D (2010) Social networks as Trojan horses to challenge the dominance of
existing hierarchies: knowledge and learning in the water governance of Volos, Greece.
Water Resources Management, 24(14):3853-3870.
Dodds, PS, Harris, KD, Kloumann, IM, Bliss, CA, Danforth, CM (2011) Temporal patterns of
happiness and information in a global social network: Hedonometrics and Twitter, PloS
One, 3:1-26.
Harvey B, Schaefer A (2001). Managing Relationships with Environmental Stakeholders: A
Study of U.K. Water and Electricity Utilities. Journal of Business Ethics, 30(3):243-260.
Hazelton J. (2013) Accounting as a human right: the case of water information, Accounting,
Auditing & Accountability Journal, 26(2):267-311.
Hoffman DL, Novak TP (2012) Why do people use social media? Empirical findings and a new
theoretical framework for social media goal pursuit. Empirical Findings and a New
Theoretical Framework for Social Media Goal Pursuit (January 17, 2012).
Hunt CJ, Staunton J, Dunstan K (2013) Equity tension and new public management policy
development and implementation in the water industry, Accounting, Auditing &
Accountability Journal, 26(8):1342-1377.
Irish Independent (2016) “'Fair warning' to politicians as water charge row heats up” available at
http://www.independent.ie/breaking-news/irish-news/fair-warning-to-politicians-as-watercharge-row-heats-up-34651262.html accessed 16 May 2016.
Irish Independent (2015a) “Rogue group of ‘fairies’ will come and make your water meter
‘disappear’”, available at http://www.independent.ie/irish-news/water/irish-
23
watercrisis/rogue-group-of-fairies-will-come-and-make-your-water-meter-disappear31271846.html, accessed 16 July 2015.
Irish Independent (2015b), “Irish Water save 3bn litres of water” available at
http://www.independent.ie/irish-news/water/irish-water-crisis/irish-water-saves-3bn-litres-ofwater-31582270.html ), accessed 10 May 2016.
Irish Times (2014a) ‘I will bring clean water to your village . . . It sounded like I was in the
Third World’, available at http://www.irishtimes.com/news/environment/i-will-bring-cleanwater-to-your-village-it-sounded-like-i-was-in-the-third-world-1.1946535).
Irish Times (2015) “‘Right2Water sets out broad policy platform on social issues”.
available at http://www.irishtimes.com/news/social-affairs/right2water-sets-out-broadpolicyplatform-on-social-issues-1.2248903.
Irish Water (2016) “Irish Water confirms 61% of customers are now paying water charges”
available at http://www.water.ie/news/irish-water-confirms-61-o/, accessed 11 May 2016.
Irish Water (2015) Draft Water Services Strategic Plan, available at
https://www.water.ie/docs/WSSP%20-%20Executive%20Summary.pdf, accessed 10 May
2016.
Jansen BJ, Zhang M, Sobel K, Chowdury A (2009) Twitter power: Tweets as electronic word of
mouth. Journal of the American Society for Information Science and Technology, 60(11):
2169-2188.
Jollands S, Quinn M (2015) Politicising the sustaining of water supply in Ireland – the role of
accounting concepts, Accounting, Auditing and Accountability Journal, in press.
24
Kennedy S (2011) Stakeholder Management for Sustainable Development Implementation: The
Case of a Sustainable Urban Drainage System. Social and Environmental Accountability
Journal, 31(2):139-153.
Kietzmann JH, Hermkens K, McCarthy IP, Silvestre BS (2011) Social media? Get serious!
Understanding the functional building blocks of social media. Business Horizons, 54(3):
241-251.
Kurland, NB, Zell D (2010) Water and Business: A Taxonomy and Review of the Research.
Organization & Environment, 23(3):316-353.
Laspidou, CS (2014). ICT and stakeholder participation for improved urban water management
in the cities of the future. Water Utility Journal 8, 79-85.
Lenihan (2009) Financial Statement of the Minister for Finance available at
http://budget.gov.ie/budgets/2010/FinancialStatement.aspx, accessed 10 May 2016.
Lennox J, Proctor W, Russell S (2011) Structuring stakeholder participation in New Zealand's
water resource governance, Ecological economics, 70(7):1381-1394.
Lewis L, Russell S (2011) Permeating Boundaries: Accountability at the Nexus of Water and
Climate Change. Social and Environmental Accountability Journal, 31(2), 117-123.
Lynn T, Kilroy S, van der Werff L, Healy P, Hunt G, Venkatagiri S, Morrison J (2015) Towards
a general research framework for social media research using big data. In Professional
Communication Conference (IPCC), 2015 IEEE International 1-8. IEEE.
Lyons, S., O’Doherty, J., Tol, R.S.J., (2010). Determinants of Water Connection Type and
Ownership of Water-Using Appliances in Ireland. Water Resources Management 24, 28532867.
25
Malhotra A, Malhotra C, See A (2012) How to get your messages retweeted, MIT Sloan
Management Review, 53:61–66.
Marwick, A. E., & Boyd D (2011), I tweet honestly, I tweet passionately: Twitter users, context
collapse, and the imagined audience, New Media & Society, 13(1), 114-133.
National Recovery Plan 2011-2014 (2010) Government of Ireland, Dublin.
Padowski JC, Gorelick SM (2014) Corrigendum: Global analysis of urban surface water supply
vulnerability. Environmental Research Letters, 9(11):119501-119506.
Paul MJ, Dredze M (2011) You are what you Tweet: Analyzing Twitter for public health.
ICWSM, 20: 265-272.
Report of the Comptroller and Auditor General (2010) Government of Ireland, Dublin.
RTE (2007) Galway water now safer than ever – HSE, available at
http://www.rte.ie/news/2007/0820/92592-galway/).
Siapera E, Hunt G & Lynn T (2015) # GazaUnderAttack: Twitter, Palestine and diffused war.
Information, Communication & Society, 18(11):1297-1319.
Stefanovic N, Radojevic I, Ostojic A, Comic, L, Topuzovic M. (2015). Composite web
information system for management of water resources. Water Resources Management,
29(7): 2285-2301.
Stieglitz S, Dang-Xuan L (2013) Social media and political communication: a social media
analytics framework, Social Network Analysis and Mining, 3(4):1277-1291.
Walker, W.E., Loucks, D.P., Carr, G., 2015. Social Responses to Water Management Decisions.
Environmental Processes 2, 485-509.
Wu Z, McKay J, Hemphill, E (2011) Roles of levies for sustainable domestic water
consumption. Water Resources Management, 25(3):929-940.
26
Appendices
Appendix 1 - theme keywords
Political / Economic
Sustaining water supply
Charge, charging
Conservation, conserving, conserve
Protest, protestors
Infrastructure
Violence, Violent
Governance
Cost
Sustainable, sustainability
Bill
Environment
Tax, taxes
Climate
Revenue, income
Drought
Government
Storage
Subsidy, subsidies, subsidise, subsidising
Aquifer
Pay
Supply
Meter
Natural
March
Freshwater
Staff
Limit, limitation
Minister
Restoration, restoring
Garda, Gardai
Rain
Investment, investing
Groundwater
Resource, resourcing
Hydrology, hydrologic
Price, pricing
Footprint
Politics, political, politicising
Scarcity
Economic, economy, economising
Shortage
Troika
Quality
Boycott
Reliable
Department
Protect, protection
Future
Resilience
27
28
29
Political/Economic
Word
Freq
Sustaining Water Supply
Word
Freq
water
57,910
water
7,099
irish
40,384
irish
4,922
protest
18,908
irishwater
2,509
irishwater
16,861
supply
1,545
rightwater
12,773
rightwater
1,012
charges
10,110
protection
952
pay
5,491
data
829
people
4,313
environment
595
government
4,278
Rain
568
Table 1. Word Analysis in Clustered Tweets
30
Exemplar Tweets
Non-clustered
Political/Economic
cluster
Sustaining Water
Supply cluster
Sentiment
Score
#IrishWater URGENT HELP NEEDED! Ashbrook and Elmvale
Estates #Cork. @IrishWater are being violent with residents & in
breach of @TheHSA regs
-3 (negative)
Irish Water are adding a dangerous chemical to your drinking water.
Why is Margaret Rae, scientist and Irish... http://t.co/ADZpmC8MTX
-4 (negative)
Excel article on @IrishWater by @braoinain in SBP. Handing
consumer protection function to economic regulator CER is recipe for
disaster
-7 (negative)
Surely a modest bonus incentive scheme for @IrishWater staff will
help remove inefficiencies and give public better value for money?
3 (positive)
Good article in Irish Times about Enda Kenny's faith in transparency
and Irish Water http://t.co/2JCuF91hlC
5 (positive)
Christ RT "BrendaKelly_IG: Is it really true that Ernest and Young got
€4m to brand water in Ireland came up with "Irish Water" Seriously?"
6 (positive)
Minister Burton Says Irish Water Must Show It Hasn't Spent
Excessively On Consultants http://t.co/eKXmYj2pCC
4 (positive)
Incredible that Irish Water won't meet us until March. I think that says
it all – Brogan
7 (positive)
@FraffieB Livid @ lack of accountability of ministers & their failure 2
account for taxpayers money in both Limerick &amp; Irish Water.
#cynical
-4 (negative)
You've got to hand to Fine Gael again, they send out the Labour to
defend the Irish Water scandal, by no less than foreign affairs
minister
-5 (negative)
.@RTERadio1: customers of @IrishWater who conserve *too much*
water will be charged more. Could this possibly be true? Can't find a
source.
-6 (negative)
The €1.1bn in savings generated by Irish Water will be composed
mainly of control savings, audited by the Department of Social
Protection.
4 (positive)
IRISH WATER, singular state monopoly to supply THE most
important service, it is therefore essential we know absolutely
everything about it.
5 (positive)
Irish Water can increase water charges if people conserve water. Not
a mention at Wat City council last night by a Cllr. Why?? Very poor.
-3 (negative)
FF Environment Spokesperson @CowenBarry is in studio with
-4 (negative)
31
@lstwrd now to discuss the need to lift the Govt's FOI ban on Irish
Water
Bit strange that we have new Irish Water quango and will be charged
for water when all it does is rain in this country!
-6 (negative)
Table 2. Exemplar Tweets with Relatively Intense Sentiment Scores
32