Preserving the old or building the new? Reputation-building through strategic talk and engagement with stakeholder inputs by the European Commission

Figures & data

Table 1. Explaining the probability of an input receiving a public EC reply.

	Model 1	Model 2
Evidence-based comment	1.638*** (0.218)	2.055*** (0.280)
Stakeholder type (Ref. category: Business)
Citizens	−0.572* (0.298)	−0.699** (0.330)
Non-business	−0.310 (0.462)	−0.392 (0.484)
Controls
Comment demand type (Ref. category: Reduce legislation)
Neutral		0.794** (0.402)
Increase		−0.216 (0.289)
Harmonise		−0.388 (0.365)
Negative tone		0.043* (0.025)
Comment text complexity		0.007 (0.010)
Policy topic
Consumer protection/product regulation		1.028 (0.765)
Environment/energy		0.263 (0.767)
Social policy		1.333 (0.818)
Institutional affairs		1.473* (0.796)
Single market		0.759 (0.750)
Transport		1.308 (0.797)
Laws and human rights		0.477 (0.819)
Intercept	−0.757** (0.308)	−1.533** (0.660)
Log-Likelihood	−260.838	−250.334
AIC	529.676	532.668
N	446	446

Note: Logistic regressions. Standard errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.

Table 2. Predicted probabilities of an EC reply in model 2 (95% CIs in parentheses).

Comment	Stakeholder
	Business	Citizen	Non-business
Evidence-based	0.78 (0.65; 0.89)	0.64 (0.50; 0.77)	0.70 (0.50; 0.89)
Opinion-based	0.31 (0.17; 0.48)	0.18 (0.10; 0.28)	0.23 (0.09; 0.45)

Table 3. Distribution of EC reply categories per comment and stakeholder type.

EC reply	Adoption	Competence	Procedural	Policy substantive
Comment
Opinion-based	0 (0.0%)	29 (53.7%)	19 (35.2%)	6 (11.1%)
Evidence-based	16 (13.0%)	50 (43.7%)	32 (26.0%)	25 (20.3%)
Stakeholder
Business	8 (20.0%)	13 (32.5%)	15 (37.5%)	4 (10.0%)
Citizens	6 (5.0%)	56 (46.7%)	33 (27.5%)	25 (20.8%)
Non-business	2 (11.8%)	10 (58.8%)	3 (17.6%)	2 (11.8%)

Table 4. Explaining the probability of receiving an EC reply indicating comment consideration for the REFIT platform’s opinions.

	Model 3	Model 4
Evidence-based comment	2.534* (1.429)	1.437 (1.123)
Stakeholder type (Ref. category: Business)
Citizen	−1.197** (0.566)	0.136 (0.683)
Non-business org.	−0.351 (0.814)	1.127 (0.981)
Controls
Comment demand type (Ref. category: Reduce legislation)
Neutral		−2.325 (1.569)
Increase		−2.018*** (0.746)
Harmonise		−1.358* (0.822)
Negative tone		−0.085 (0.059)
Comment text complexity		−0.033 (0.021)
Policy topic
Consumer protection/product regulation		−0.262 (1.337)
Environment/energy		−0.567 (1.560)
Social policy		−1.647 (2.364)
Institutional affairs		−2.219 (1.691)
Single market		0.476 (1.600)
Transport		−1.457 (1.752)
Laws and human rights		0.532 (1.780)
Intercept	−3.709** (1.458)	−1.923 (1.570)
Log-Likelihood	−45.871	−37.920
AIC	99.743	107.841
N	177	177

Note: Penalised maximum likelihood logistic regressions. Standard errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.

Table A1. Examples of EC reply categories.

Type of reply	EC reply text
Adoption	‘On behalf of the REFIT Platform, we would like to inform you that on 7 June 2017 the REFIT Platform has adopted an opinion on your suggestions regarding “identity and travel documents” (REFIT Platform ref. XIII.3.a). The adopted Opinion has also been published on the REFIT Platform’s website. The follow-up to this adopted opinion will be made in the context of the preparation of the CWP2018 and published on our website thereafter. With many thanks again for your interest in these issues’
Procedural	‘On behalf of the REFIT Platform, thank you for submitting your suggestion on the efficiency, effectiveness or burden reduction of EU regulation or its implementation by Member States. The Platform has concluded that it does not intend to prepare an opinion on the issues raised. This is because the issues are covered by an EU legislative measure which only entered into application less than two years ago and so there is too short a period of experience of the application of the measure for sound conclusions on its performance to be drawn. Your suggestion has been forwarded to the Commission service responsible for the relevant policy area which has produced an explanation of the current context, which we attach for your information. With thanks again for your suggestion which will be retained for future reference’
Competence	‘On behalf of the REFIT Platform, thank you for submitting your suggestion on the efficiency, effectiveness or burden reduction of EU regulation or its implementation by Member States. The Platform has concluded that it does not intend to prepare an opinion on the issues raised. This decision is based on the fact that this submission is a competence of the Member States, and falls beyond the remits of the REFIT Platform. We would also like to provide you with a reference to the Circular Economy Strategy and Plan: http://ec.europa.eu/environment/circular-economy/index_en.htm’
Policy substantive	‘Commercial rates of interest, commissions and other charges are based on many factors. One of the main factors for the rate of interest is the risk that an individual loan might pose to the bank. For instance, a company with a low risk profile will most likely receive an offer with a low rate, while the offer for a company with a higher risk profile will most likely include a higher rate. These market forces are something that the European Union cannot, and should not, influence. The Commission can however implement policies to help reduce asymmetric information exchange between banks and companies, to ensure that there is a clearer link between the cost of the loan and the risk it really entails. The Commission also supports diversity in the sources of financing for companies for instance by supporting credit unions so that companies can be less dependent on bank financing’.

Table A2. Examples of evidence-based and opinion-based stakeholder comments.

Type of comment	Comment text
Evidence-based	‘Please find in the attachment suggestions and activities recognized at working-seminars (2016) for businesses and NGO’s, that could be implemented at the EU level. They cover 7 policy areas: collaborative/ sharing economy, intellectual property framework, public procurement, mutual recognition, prevention of non-compliant products in the EU market, cloud computing and free flow of data, reducing VAT related burdens and obstacles when selling across borders’.
Opinion-based	‘The gun ban that EU is pushing ahead is illegal. Primarily Germany and France, but also other southern EU countries have messed up the whole idea of EU with their sloppy immigration politics and are now trying to make smaller and northern countries suffer because of their stupid & criminal decisions made under influence of blind humanism’.

Table A3. Examples of stakeholder comment demand type.

Category demand	Comment text
Reduce	‘Radiocentre’s concerns on behalf of the UK commercial radio industry relate to Article 4 of Consumer Credit Directive 2008/48/EC (CCD) which requires standard information to be provided in any credit advertising indicating an interest rate or cost of credit. This affects advertising for financial services, retail and motors, typically adding 12–15 s of airtime to a radio advert. Our concerns include: Cost: we estimate that complying with Article 4 of the CCD costs commercial radio stations and advertisers around €95 million a year in the UK alone. Failure to inform and protect consumers: independent research commissioned by Radiocentre shows low levels of recall of key figures – only 4% of listeners recalled the total amount payable Access to market: advertisers have to pay for additional airtime because of the CCD – a burden, which few advertisers, let alone SMEs, can afford Platform neutrality: radio is the only medium where this information has to be read out in full so the CCD has a disproportionate impact on the radio sector Inconsistency with existing EU consumer law: the Unfair Commercial Practices Directive, which pre-dates the CCD, acknowledges that consideration should be given to “limitations of space or time” of the medium used to communicate No impact assessment: no impact assessment was conducted by the Commission prior to the adoption of the Directive – a key component of Better Regulation. Suggestion: We fully support the principles of the CCD, however consumers ignore information which is too complex or difficult to remember and there is a lot of evidence that simpler information with fewer figures is much more effective at landing critical messages at the advertising stage. Radiocentre have invested (and continue to invest) in independent research exploring the effect of different treatments and this has already highlighted that limiting the information can lead to a three-fold increase in recall of critical figures. We would therefore like to be able to work with Members of the REFIT platform, advertisers, regulators and consumer groups to deliver a solution for the benefit of consumers and businesses’.
Increase	‘Dear Comission, Is there any laws in the planning stage that will restrict the use of packaging to a minimum necessary and fine overpackaging? This kind of law is necessary now and will be even more necessary for the future. It will help to educate the market to a more responsable use of resourced and will reduce general trash impact in Europe. Me as a consumer, I feel forced by the industry and market to buy products which are completely over packed, so i am forced to produce an amount of garbage which in my point of view is insane, considering the state of the world. With many products there is no option, as all brands use similair (& unnecessary) packaging. How can one, as a citizen help to forward this kind of law? Thanks very much for your attention and response!’
Neutral	‘I understand that there is to be an on-line consultation regarding REFIT and the Habitats and Birds Directives which is due to take place in April 2015 – so far I have been unable to find it. The on-line consultation for the banning of driftnets passed un-noticed by most organisations and the NFFO would regard a similar exercise for REFIT as unacceptable unless the on-line questionnaire is widely publicised’.
Harmonize	‘EU Driving Licences. It was decided in 2013 I believe that these permits should be uniform across the Member States. The EU did not go far enough as each licence must be changed or renewed in each Member State when exercising freedom of movement. It would be better for all citizens if there was one universal EU driving licence permitted to be used in each State’.

Table A4. Policy topics identified with the help of structural topic models (STMs).

Name of topic	Min	Mean	Max
T1: Consumer protection/product regulation	0.001	0.094	0.986
T2: Environment/energy	0.003	0.181	0.934
T3: Social policy	0.0002	0.091	0.985
T4: Institutional affairs	0.001	0.103	0.947
T5: REACH/chemical regulation	0.0004	0.083	0.984
T6: Single market	0.0006	0.172	0.953
T7: Transportation	0.001	0.119	0.947
T8: Laws and human rights	0.0008	0.157	0.973

Table A5. Most representative words exclusive to each topic (FREX).

Topic	Top words
T1:	vat, return, advertis, insur, consum, incom, sale
T2:	applic, habitat, water, energi, natur, wildlif, climat
T3:	medicin, italian, famili, educ, women, itali, abus
T4:	attach, smoke, search, recognit, engin, petit, manag
T5:	see, label, substanc, hazard, product, clp, classif
T6:	citizen, countri, languag, immigr, translat, movement, sweden
T7:	region, communiti, speed, account, aid, data, secur
T8:	vote, photo, peopl, websit, social, politician, retir

Table A6. Logistic regression explaining the probability of receiving a public EC reply – ENGLISH TEXTS ONLY (robustness check for Table 1 in the Article).

	Model 1	Model 2
Evidence-based comment	1.311*** (0.272)	1.787*** (0.352)
Actor type (ref. category: Business)
Citizens	−0.805** (0.349)	−0.995** (0.405)
Non-business	−0.668 (0.522)	−0.681 (0.552)
Controls
Comment demand type (ref. category: Reduce legislation)
Neutral		0.804 (0.501)
Increase		−0.308 (0.367)
Harmonize		−0.289 (0.443)
Negative tone		0.062* (0.033)
Comment text complexity		0.0001 (0.013)
Policy topic
Consumer protection/product regulation		1.809* (0.995)
Environment/energy		0.347 (0.913)
Social policy		2.226** (1.006)
Institutional affairs		1.847* (1.062)
Single market		1.435 (0.908)
Transport		1.974* (1.066)
Laws and human rights		0.594 (1.003)
Intercept	−0.458 (0.362)	−1.385* (0.813)
Log likelihood	−174.670	−162.951
AIC	357.341	357.902
N	294	294

Note: Logistic regression models. Standard errors in parentheses for regression coefficients. *p < 0.1; **p < 0.05; ***p < 0.01.

Table A7. Distribution of EC reply categories per comment type and stakeholder type – ENGLISH TEXTS ONLY (robustness check for Table 3 in the article).

EC reply	Adoption	Competence	Procedural	Policy substantive
Comment
Opinion-based	0 (0.0%)	23 (59.0%)	12 (31.0%)	4 (10.0%)
Evidence-based	13 (18.1%)	20 (27.8%)	23 (31.9%)	16 (22.2%)
Stakeholder
Business	8 (25.8%)	8 (25.8%)	11 (35.5%)	4 (12.9%)
Citizens	3 (4.0%)	31 (45.0%)	21 (30.0%)	14 (20.0%)
Non-business	2 (18.2%)	4 (36.4%)	3 (27.3%)	2 (18.2%)

Table A8. Explaining the probability of getting an EC reply indicating comment consideration for the REFIT programme (adoption) – ENGLISH TEXTS ONLY (robustness check for Table 4 in main manuscript).

	Model 3	Model 4
Evidence-based comment	2.444* (1.437)	1.438 (1.220)
Stakeholder type (ref. category: Business)
Citizen	−1.413** (0.690)	0.163 (0.940)
Non-business org.	0.089 (0.887)	1.799 (1.161)
Controls
Comment demand type (ref. category: Reduce legislation)
Neutral		−2.870* (0.939)
Increase		−2.246*** (0.939)
Harmonize		−1.118 (0.893)
Negative tone		−0.119 (0.076)
Comment text complexity		−0.022 (0.032)
Policy topic
Consumer protection/product regulation		1.221 (1.991)
Environment/energy		0.781 (2.238)
Social policy		−0.205 (2.840)
Institutional affairs		−0.085 (2.348)
Single market		1.331 (2.125)
Transport		0.318 (2.462)
Laws and human rights		1.373 (2.503)
Intercept	−3.325** (1.471)	−3.128 (2.025)
Log-Likelihood	−31.931	−24.734
AIC	71.862	81.468
N	111	111

Note: Penalised maximum likelihood logistic regressions. Standards errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.