The benefit of certainty

Executive Summary

In- or mid-contract price rises are a common feature of broadband and mobile contracts and often lead to monthly prices increasing by the rate of inflation plus a markup. Since inflation rates have recently been at their highest level for decades, many customers faced increases of up to 17.3% on their monthly bills in April 2023. This has increased attention on these contract terms and they have been the subject of widespread criticism.

Contracts with unpredictable in-contract price rises are likely to be harmful for consumers. They obfuscate prices and undermine price competition, which will lead to consumers paying more overall. They may also unfairly penalise consumers who have lower levels of understanding about inflation and the impact it could have on their future bills, and sharp and unexpected price rises can strain household budgets and cause affordability issues. 

More generally, consumers almost always prefer to have certainty over the price they pay for goods and services, but this is denied to them by in-contract price rises that are inflation-linked or ad hoc. Unusually for consumer markets, these contract terms shift the burden of managing inflation risk away from communications providers and onto their customers.

To understand the consumer harm from unpredictable in-contract price rises, Which? has worked with PJM Economics to estimate the value that consumers place on having price certainty throughout a broadband contract. We use a stated preference survey methodology, involving a large nationally-representative UK survey of broadband bill-payers. The survey was conducted online from 2 June to 20 June 2023, following several phases of development and testing. The sample was sourced from an online panel, and a total of 4,004 fully completed interviews were achieved. 

In the first stage of the survey, participants took part in a choice experiment. They were asked to choose their preferred options when faced with a series of sets of broadband packages. The offered packages varied according to key attributes of average speed, length of contract, upfront set-up costs, monthly price and whether the price was fixed or subject to an inflation-linked or ad hoc price rise. The presentation of the choices was designed to reflect the reality consumers face when selecting a broadband package, which in the case of the contract price variation meant providing a single line of footnoted explanation at the bottom of the screen. The aim of the choice experiments was to estimate the value consumers place on different broadband package attributes, and specifically how much they implicitly value having fixed prices in their contract when faced with complex choices that have multiple attributes. 

The second stage of the survey simplified the consumer decision and used a contingent valuation methodology to more explicitly value having fixed prices in their contract. Respondents were asked to choose between a fixed price broadband package and a broadband package with an unpredictable in-contract price rise (either an ad hoc or inflation-linked price rise). All attributes except monthly price and contract price variation were held constant and clearer, more prominent information was provided about in-contract price rises. We used a double-bounded dichotomous choice to elicit willingness to pay for price certainty.

In addition to these two stages, respondents were also asked questions about their awareness of in-contract price rises, their attitudes towards these, their knowledge and understanding of inflation and their ability to calculate the actual increase in a monthly price when an inflation-linked increase is applied.

Our findings show that these contract terms are unpopular with consumers. More than three-quarters (78%) of respondents believe that increasing the cost of a service during a contract is always unfair.

Despite this, our choice experiments show that when faced with a multi-attribute choice over which broadband package to purchase, people appear to have little regard for whether or not a contract has an in-contract price rise. This is likely to be driven by a lack of saliency within the presentation of the contract terms, combined with an inability amongst consumers to adequately process them. We find, for example, that many consumers struggle to understand price variation terms and their impact on the price they will pay. Only half of respondents (51%) knew that the Consumer Price Index (CPI) or Retail Price Index (RPI) measure the rate of inflation and only a quarter (25%) could calculate by how much a monthly price would be affected by an inflation-linked in-contract price rise. A predicted consequence of these patterns of information processing is that contract terms will not be set competitively, a prediction that is consistent with the evident similarity of these terms across providers.

The willingness to pay estimates from the second stage’s contingent valuation questions, in which contract price variation terms are made more salient, provides more plausible estimates of the value that consumers place on having pricing certainty. The results unequivocally show that people have a preference for certainty in the price they pay for broadband contracts.

Using the willingness to pay estimates from the contingent valuation questions and taking into account the expected value of an in-contract price rise, we can derive the value of the certainty premium for a representative contract. We find that people are prepared to pay £3.54 per month to avoid inflation-linked in-contract price rises and £3.42 per month to avoid ad hoc in-contract price rises. 

Aggregating these certainty premium estimates to an economy-wide level, we estimate the total certainty premium from having fixed prices in domestic broadband contracts to be £710 million, of which £514 million relates to inflation-linked in-contract price rises, which are the most commonly used.

The estimates are robust to changes in sample restrictions, and assumptions about expected inflation and the actual prices paid by broadband customers. At all times, conservative assumptions were made to arrive at the estimates, but it is likely that these estimates will be affected by the research being conducted at a time of relatively high and volatile inflation. This will have influenced consumer sentiment and so the valuations produced by this research are prudently interpreted as a reasonable upper bound for the value that consumers would place on having fixed prices for broadband contracts. We did not include mobile contracts in our research in order to simplify the design and the sample size needed to carry out the experiment, but we would expect the findings of our study would be qualitatively similar across both broadband and mobile markets. 

Overall, the study provides strong evidence that consumers place considerable value on having price certainty for the duration of their telecoms contract. We expect that the evidence in this report will prove valuable to anyone seeking to understand the potential benefits to consumers of a change in market rules that would prohibit or restrict the use of unpredictable in-contract price rises in mobile and broadband markets.

We're demanding big broadband and mobile phone providers drop hikes buried in the terms and conditions. Agree? Sign the petition

1. Introduction

1.1. Background

In 2013, Ofcom revised its guidance to communications providers on price rises within fixed term contracts for landline, broadband and mobile services [1]. It decided that any price increase that was ad hoc would be considered a contract modification and thus, consumers would have the right to exit the contract without penalty. However, if planned price rises were included in the contract’s terms and conditions, it would not be considered harmful to consumers. Many communications providers then adopted Consumer Price Index (CPI) and Retail Price Index (RPI) linked price rises every April which were included in the terms and conditions, meaning consumers could not exit penalty-free. In 2020, providers began to adopt above inflation price rises (eg CPI + 3.9%) which were justified as necessary to cover  their own increased costs and investments in infrastructure.

These in-contract, or mid-contract, price rises have received greater attention as inflation rates have risen.  During the April 2023 price rises, many customers unexpectedly found themselves with increases on their monthly bills of up to 17.3% as the CPI rate announced in January was 10.5% and the RPI rate announced in January was 13.4%. Ofcom’s Affordability Tracker revealed that three in 10 households surveyed in April 2023 experienced affordability issues with their communications services, equivalent to approximately 8 million households in the UK [2].

In the context of this, Ofcom is currently conducting a review into the practice of in-contract price rises to assess its impact on consumers, including their ability to understand broadband deals and choose the right package, as well as how this might impact the effectiveness of a competitive market. 

The present study has been borne out of Which?'s desire to understand how much value consumers place on having certainty in the monthly price they pay for their telecoms contracts in situations of varying transparency and emphasis of contract terms.  

1.2. Objectives

The objective for the research was to obtain, using a stated preference methodology, consumer valuations for having certainty over how much their home broadband would cost per month. This evidence is intended to contribute to the broader evidence base in this area, with a view that it could potentially inform regulatory decisions around potential rules. 

1.3. Report structure

This is our final report for the study. It is as organised as follows.

• Section 2 discusses the context for the study, including the regulatory landscape, the practice of in-contract price rises, and the harm from in-contract price rises.
• Section 3 describes the survey methodology, including the stated preference design, questionnaire structure, pilot testing and main survey administration.
• Section 4 presents findings on consumer understanding of inflation and attitudes towards in-contract price rises.
• Section 5 reports our main findings, including descriptive and econometric analyses of the choices made by participants, and estimates of the willingness to pay values for price certainty in their broadband contracts, disaggregated by segments of the sample.
• Section 6 concludes the study.

Appendix A contains the main survey questionnaire and the discussion guide used in the cognitive interviews; Appendix B contains demographic characteristics of the sample data; Appendix C contains details and results from econometric modelling analyses. These are downloadable with the full report. 

2. Context for the study

2.1. Permitted practices and the regulatory landscape

In 2011-12, mobile operators were increasing the price of their contracts during the minimum contract period. However, they marketed these contracts to consumers as fixed prices throughout the minimum contract period. Thus, when informed of the price rises, consumers wanted to cancel their contracts without an exit fee but found themselves unable to do so. This was due to clauses in the terms and conditions which stated that an operator may increase their prices without allowing customers to leave penalty-free when the change in price does not cause “material detriment” to the customer. Many operators explicitly stated that price rises in line with RPI cannot be said to cause material detriment, thus removing the opportunity for penalty-free exit from customers.

Ofcom launched a review in January 2012 to investigate the fairness of communications providers’ contract terms and identified several issues with General Condition 9.6 (GC9.6) and the Unfair Terms in Consumer Contract Regulations, which providers have to follow when modifying contracts. Ofcom stated that under those rules, terms with price increases that did not give consumers the right to exit penalty-free are “liable to be unfair” [3]. They specifically identified consumer harm arising from in-contract price rises and identified that the rules under GC9.6 did not provide consumers with adequate protection against the effects of price rises. 

Which?’s 2012 campaign, “Fixed Means Fixed” highlighted the unfairness of this situation and the harm it caused to consumers. The campaign called for prices to remain fixed throughout the minimum contract period and a right to exit without penalty if providers failed to uphold this promise, barring any unavoidable tax or regulatory changes. Which? also made a formal complaint to Ofcom on this issue. 

As a result of Ofcom’s initial review and the Which? complaint, Ofcom consulted on whether additional consumer protection from in-contract price rises in landline, broadband and mobile services was needed. The aim was to update guidance on what constituted “material detriment” under the GC9.6 rule. The consultation proposed four options, with the preferred option being the introduction of a right to penalty-free exit if providers increase prices during the minimum period of a contract, which was also supported by Which?.

After communications providers argued against this preferred option, Ofcom introduced a distinction between planned and unplanned in-contract price rises through their guidance. It stated that if planned price rises are included in the terms and conditions of a contract, they would not be considered a “material detriment” and thus customers have no right to exit without penalty [4]. However, if in-contract price rises are unplanned, meaning communication providers are not transparent about the likelihood and nature of increases at the point of sale, then consumers do have a right to cancel their contract without penalty. Following this guidance, it has become common for communications providers to adopt CPI and RPI linked rises every April, specifying this in their contracts.

In 2020, communications providers began implementing above inflation price rises where a percentage between 2.9%-3.9% was applied on top of CPI increases. The justification providers gave for these additional price rises was a need to invest in their services. 

Currently a majority of broadband providers adopt inflation-linked (ie CPI + x%) increases, a couple of major broadband and mobile providers implement ad hoc price rises with the right to exit penalty-free and some smaller broadband providers promise fixed prices during the minimum contract term. These current practices are detailed in Table 1:

Table 1: Current in-contract price rise practices of selected broadband and mobile providers

Recent regulatory efforts have focussed on improving the transparency of price increases. In June 2022, Ofcom introduced new rules so that providers must provide their customers with a one page document explaining the pricing terms and conditions, amongst other things, at the point of signing up for a contract. Ofcom also introduced new guidance on price variation terms so that, where a contract specifies that the advertised monthly price will increase at a certain time by an amount linked to an inflation, providers should include an example estimate of what the customer’s core subscription price would be once the inflation increase has been applied. Additionally in 2022, the Committee of Advertising Practice (CAP) and Broadcast Committee of Advertising Practice (BCAP) proposed new guidance on the presentation of price variation terms in telecoms advertising. This guidance introduces additional transparency requirements on advertisers and applies from December 2023 [5].

However, high levels of inflation have put pressure on Ofcom to do more than simply improve transparency. Since the annual inflation-linked price rises are based on the CPI / RPI rate as published in January 2023, consumers have this year faced the choice of paying price increases up to 17.3% (as much as £113 per year for an average customer [6]) or, if they wanted to leave, accept exit fees of over £200 [7]. As broadband contract lengths are typically either 18 or 24 months and the inflation rate rose dramatically following the invasion of Ukraine, it would have been difficult for consumers to predict the price rises. Ofcom is therefore currently conducting a review into the practice of inflation-linked in-contract price rises to assess the impact of this on consumers, including their ability to understand broadband deals and choose the right package, as well as how this might impact the effectiveness of a competitive market [8].

2.2. The harm from in-contract price rises

There are numerous reasons why inflation-linked in-contract price rises will be detrimental to consumers, with the most significant likely to be: 

• Price obfuscation: in-contract price rises make it harder for consumers to understand the prices they pay and to compare providers which increases consumer search costs and the risk of a consumer making suboptimal choices.
• Undermining price competition and incentives to innovate/invest: since consumers are likely to have limited regard for the size of an in-contract price rise when choosing a deal, these terms are not set competitively. Hence only part of the total cost of the contract is subject to competitive pressure. Firms also face less pressure to become more efficient given they can easily pass costs onto consumers.
• Unfairly penalises some groups of consumers: consumers with lower levels of understanding about inflation and what impact it could have on their future bills may pay more than others. Less engaged consumers will also be less likely to leave a contract (where they are free to do so) in response to a rise in their monthly costs. Since competitive pressure will mean the prices of new contracts rise by less than within or out-of-contract prices then this will exacerbate the ‘loyalty penalty’, in which captive customers pay higher prices.

In addition to these harms, in-contract price rises may be detrimental to consumers because they deny them certainty over the price they will pay for the duration of their contract. Sharp and unplanned price rises can strain household budgets and lead to financial difficulty for some consumers. Ofcom reported that around three in ten (29%) of the households they surveyed in April 2023 had difficulty affording a communication service, which is equivalent to 8 million UK households [9]. For these households it is important they have a clear understanding of the cost of their telecoms for the duration of their contract. More generally, it is well-established that consumers will pay a premium to have a certain outcome [10], hence the existence of markets for insurance, products such as extended warranties, and offers from some companies for consumers to be able to pay a premium to fix prices, such as TalkTalk’s Fixed Price Plus offer on broadband contracts.

3. Methodology

The core objective for the survey design was to understand and estimate the value of consumer preferences for price certainty in their broadband contracts. We expect the findings of our study will be qualitatively similar across both broadband and mobile markets, hence to simplify the survey design and reduce the size of the sample needed to carry out the experiment, we focus solely on the broadband market.

Section 3.1 describes the approach taken in the design of the survey to obtain these estimates. Section 3.2 details the outline of the questionnaire in full, including the additional questions that were included, Section 3.3 describes the detailed development and testing programme carried out to ensure the validity of the main survey instrument, Section 3.4 discusses the main survey administration and Section 3.5 presents diagnostics concerning the sample quality and the sensitivity samples examined in the research.

3.1. Survey design

The survey was split into two stages, with the first being a set of choice experiments aimed at estimating values consumers place on different broadband package attributes, and specifically the value of having fixed prices in their contract. As the first stage is designed to mimic real-world situations, it is unlikely to generate an accurate certainty premium. This is because, with multiple attributes, the decision is relatively complex and the contract terms lack transparency, so that overall the price variation terms are not expected to be salient in the consumer choice.

This necessitates a second stage to assess the impact on consumers’ valuation of fixed prices when information about the price variation terms were made more explicit. The second stage is a contingent valuation design to elicit consumers’ willingness to pay for fixed prices when other broadband package attributes are held constant and the transparency of price variation is improved.

Design of first stage choice experiments

The aim of the first stage choice experiments was to understand how people make choices in real-world situations when in-contract price variation terms are not well displayed, and to develop a personalised choice that can be used in the second stage where we elicit the value consumers place on certainty. 

In order to simulate how consumers would choose a broadband package with multiple attributes and to assess whether they are able to take contract price variation into account, consumers were presented successively with six choice sets. Each choice set had four alternatives: three broadband contracts that varied across the product attributes of speed, contract length, setup cost, monthly price and price variation term, and a fourth option to opt-out of having broadband. 

Although the experiments are a hypothetical environment, the choice sets were designed to reflect the reality that people face when choosing a broadband package as closely as possible. The attributes selected are those typically presented in broadband comparison websites and communications providers’ websites. The specific levels for each attribute, see Table 2, were chosen following desk research on the average speed, contract lengths, upfront set-up costs, and monthly costs that are available in the broadband market. While our primary interest concerns how consumers value fixed prices versus inflation-linked in-contract price increases, we also included ad hoc price increases as a type of contract price variation.

Table 2: Attributes and the levels of these attributes of the broadband packages respondents had to choose from in the choice sets

The presentation of the choice sets were designed to look similar to the packages consumers may see from a broadband comparison website. To avoid complexity, the choices were solely for broadband and did not include bundled services (eg broadband and TV, broadband and mobile, broadband and mobile and TV).

Respondents in the main fieldwork were randomly allocated to one of 16 blocks of six choice sets. We had tested four blocks in the pilot and found that increasing the number of blocks from four to 16 stabilised the results (see Section 3.3 for more detail). The choice sets were generated and divided into blocks using Ngene, with restrictions imposed on the combinations of average speed and monthly price to ensure their believability:

• Contracts with average speeds of 36 Mbps cannot cost more than £30
• Contracts with average speeds of 67 Mbps cannot cost more than £35
• Contracts with average speeds of 500 Mbps cannot cost less than £30
• Contracts with average speeds of 910 Mbps cannot cost less than £35

These restrictions were based on the findings of our desk research on the speeds and pricing of actual broadband packages. 

Another parameter specified in Ngene was that the levels of each attribute had to be balanced, meaning that the distribution of the number of times a level of each attribute appears is roughly equal across the 96 choice sets. 

Figure 1 shows an example of the type of choice set asked in the first stage. Respondents were asked to select a first choice and a second choice among the four alternatives for each of the six choice sets. This was preferable to a best-worst scaling [11] as it was expected that a large majority of respondents would have chosen the “no broadband” option as their worst choice.

Figure 1: Example of a choice set shown to respondents in the first stage

In this stage, it was necessary to emulate how major communication providers typically presented information about in-contract price rises to consumers. Figure 2 presents examples of how that information is displayed on the website of two major broadband providers.  

Figure 2: Examples of how information about in-contract price rises is presented to consumers on broadband providers’ websites

Example 1:

Example 2:

In both examples, the superscript on the prices directs customers to the information displayed at the bottom of the webpage. A single line on the detail of the inflation-linked in-contract price rise is provided in each case with a click-through to further information. The information on price rises presented in stage one of our experiment closely mirrors this with a footnoted single line of explanation on the webpage. However, we do not include a click-through to further information since studies have shown that these are rarely used by consumers [12] and none of the participants in our cognitive testing noticed a click-through when we provided it.

Design of second stage contingent valuation

In the second stage, we wanted to obtain consumers’ valuations of a fixed price broadband contract when the in-contract price rise information was as transparent as possible and all other attributes were held constant, meaning the only variable attributes were monthly price and contract price variation. This would increase the saliency of the price variation terms compared to the first round, allowing us to compare how consumers regard these terms to stage one when the terms were less salient. 

Figures 3 and 4 show an example of the choice question asked in the second stage. Respondents were asked to choose between a fixed price broadband package and a broadband package with an in-contract price rise (either an ad hoc or inflation-linked price rise). The specific levels of the attributes were obtained from the respondent’s most preferred option in the last series of choice sets they had seen in the first stage. This anchored the alternatives in the respondents' own preferences and increased the believability of the choice. If the participant had chosen a contract with either of the in-contract price rises, the alternative choice in the second stage was a fixed price contract. Respondents who had chosen a fixed price contract in the first stage were randomly allocated either an inflation-linked or ad hoc price for their alternative choice, so that overall half the sample faced a choice between a fixed price contract and a contract with an inflation-linked price rise (Figure 3) and the other half chose between contracts with a fixed price and an ad hoc price rise (Figure 4). 

Figure 3: Example of choices between a fixed price contract and a contract with an inflation-linked price rise seen by respondents in stage two

Figure 4: Example of choices between a fixed price contract and a contract with an ad hoc price rise seen by respondents in stage two

The monthly cost of a fixed price package always had to be more expensive than either of the in-contract price rise packages to discern how much more consumers were willing to pay for the certainty. The difference between the two types of contracts were selected at random from the list = {£1.50, £5.00, £10.00}. These monetary values were chosen as they provide a plausible range. They are in line with amounts that could be charged as in-contract price rises and with TalkTalk’s Fixed Price Plus offer that allows customers to pay an additional monthly fee of about £3 to avoid an in-contract price rise. These payment levels were also tested at the pilot stage and were found to be reasonable (see Section 3.3).

Depending on their initial choice, the respondents were then shown the same choice sets but with the price difference either halved if they chose a package with an in-contract price rise (ie £0.75, £2.50, £5.00) or doubled if they chose a package without an in-contract price rise (ie £3.00, £10.00, £20.00).

This approach is known as a double-bounded dichotomous choice (DBDC) format. In comparison to open-ended and payment card questions, participants find it easier to understand the format, are much less likely to give protest responses, and have less (no) incentive to misrepresent their preferences (Carson and Groves, 2007). This format therefore minimises non-responses and avoids outliers.

The transparency of the price variation terms was also improved in the second stage. This was to obtain consumers’ willingness to pay for a fixed price contract when it is made clear what the exact terms will be and whether they will regard them more highly. For the ad hoc price rises and fixed prices, the explanation was made more prominent. For inflation-linked price rises, the prominence was increased and more explanation was provided on CPI and the concept of inflation. A pop-up was placed directly below the bottom of the broadband package option for inflation-linked price rises. The pop-up then generated this note.

 Figure 5: Explanatory note on inflation generated from the pop-up under the broadband package option for inflation-linked in-contract price rises in stage two

The note on inflation was adapted from a report commissioned by Ofwat and CCWater on best practice guidelines in presenting inflation in consumers’ water bills [13]. We added a worked example to the guidelines to illustrate what the potential price rise will mean for a typical broadband contract. 

3.2. Questionnaire structure

The full survey questionnaire was structured as follows: 

• Participant screening to ensure a nationally representative sample of UK adults 18+ (quotas were set on age, gender, region and socio-economic group) at the start of the survey. 
• Whether they were the bill-payer for their mobile and broadband service  – participants who were not the main bill-payers for both services were screened out of the survey at this stage. 
• First-stage choice experiment section.
• Additional survey questions on: 
  • awareness of whether their mobile or broadband provider could increase their monthly cost during their contract; 
  • attitudes towards in-contract price rises; 
  • knowledge and understanding of inflation and; 
  • ability to calculate the impact of an inflation-linked price increase on a monthly price of a mobile or broadband contract.
• Second-stage contingent valuation section.
• Questions about the respondent’s income, education, disability, and current broadband provider.

3.3. Development and testing

Several phases of development and testing research were undertaken before the main survey was implemented.

3.3.1. Cognitive interviews

First, a preliminary phase of qualitative research was undertaken in May 2023, comprising of 10 in-depth interviews with consumers responsible for their broadband bills [14]. The purpose of this phase was to test the full programmed survey, the credibility of the broadband package options and placement of the price variation terms. The interviewer also looked out for any spontaneous responses or instances where participants found it difficult or took longer to answer, and prompted respondents when these happened. The interviews were conducted remotely by Microsoft Teams and lasted between 40 and 60 minutes.

This research found that, in general, the main choice exercise appeared to work well and respondents found the options to be believable. Respondents generally found it easy to pick the options, with many commenting on how the layout was clear across both stages. At this stage of the design, information on the price variation terms in stage one required respondents to click through a “More info” button to access it. All respondents failed to identify the button, thus the click-through was removed and the information was immediately available at the bottom of the screen. 

There was also feedback on the question in which respondents were given the current price of a contract and the CPI rate and then asked to calculate the monthly price after a CPI+3.9% increase. Some respondents felt the question would be too tough and time consuming for the average person. This was difficult to change since this was already a simplification of the real-life experience, but we included a clarification that consumers could use a calculator and changed the contract price to a round number so that the calculation appeared less daunting.

3.3.2. Pilot testing

A pilot survey was conducted online in the second last week of May 2023 using a sample sourced from an online panel. A total of 328 fully completed interviews were achieved for the pilot survey. The mean completion time was 13 minutes 48 seconds. Once the pilot sample was achieved, the data was subject to rigorous quality checks. The findings from the pilot analysis are as follows:

• There were no issues with recruitment, and the rate at which interviews were added to the dataset was as expected. 
• The length of the survey was as expected. 
• There were no routing errors.
• Simple econometric modelling found wide variation in results depending on which block of six choice sets participants had been randomly allocated to. To address this, we increased the number of blocks from four to 16 in the next pilot round to reduce the likelihood of one block dominating the results.
• Simple econometric models were estimated and performed reasonably well in the second stage. 
• The levels of difference in cost in the second stage for both types of in-contract price rises were reasonable with the predicted choice probabilities from the parametric survival analysis (see Figures 6 and 7) being in line with expectations of a smaller percentage of respondents likely to say yes to bigger cost differences and a larger percentage likely to say yes to smaller cost differences. These predicted choice probabilities were derived from modelling described in Appendix C.

Figure 6: Predicted probabilities of choosing a fixed price broadband contract over a contract with an inflation-linked in-contract price rise at various price difference levels

Figure 7: Predicted probabilities of choosing a fixed price broadband contract over a contract with an ad hoc in-contract price rise at various price difference levels

Given the wide variation in results across the blocks in the first stage, we decided to run a second pilot survey to test the survey design with 16 blocks. The second pilot survey was conducted online in the first week of June 2023 using a sample sourced from an online panel. A total of 190 fully completed interviews were achieved for the second pilot survey, which were added on to the main pilot responses. The mean completion time was 13 minutes 7 seconds. Similarly, the data was subject to rigorous quality checks. The findings from the second pilot analysis are as follows:

• Similar to the first pilot survey, there were no issues with recruitment, length of survey, and routing errors. 
• Simple econometric models were estimated and performed well. Coefficients and statistical significance were in line with priors.
• Estimates from the econometric models and predicted choice probabilities from the parametric survival analysis of differences in cost levels (see Figures 8 and 9) were similar to the first pilot round.

Figure 8: Predicted probabilities of choosing a fixed price broadband contract over a contract with an inflation-linked in-contract price rise at various price difference levels

Figure 9: Predicted probabilities of choosing a fixed price broadband contract over a contract with an ad hoc in-contract price rise at various price difference levels

3.4. Main survey administration

The main survey was conducted online from 2 June 2023 to 20 June 2023. The sample was sourced from Yonder’s online panel. We assume that consumers who are responsible for paying household broadband bills are also likely to have experience of searching for and comparing broadband packages and are therefore the most suitable sample. The sample was designed to be representative by selecting respondents who are responsible for paying household broadband bills from a Census-based sampling frame. 7,303 entered the survey, 2,926 participants were screened out of the survey at the scoping stage and 295 participants qualified but dropped out of the survey part-way through. A total of 4,004 fully completed interviews were achieved. 

3.5. Data diagnostics and sensitivity samples

Understanding and believability of the choice experiment options

Participants were asked to rank whether they understood the options shown and how believable they were, using a 5-point Likert scale. These questions were asked after the choice experiment in the first stage and again after the contingent valuation section in the second stage. As shown in Figure 10, 3% and 1% of respondents in the first and second stages, respectively, ‘tend to disagree’ that they understood the options. Only 9% in stage one ‘tend to disagree’ or ‘strongly disagree’ that they found the options believable, with the rate dropping to 2% in stage two. 

Figure 10: Respondents’ understanding and believability of the options shown in stages one and two

The proportion of participants who did not understand the options was similarly low across both stages. However, the proportion of participants who did not find the options believable was higher than those who did not understand the question in both stages. For the first stage, the proportion of those who did not find the options believable was seven percentage points higher than in the second stage. This is not surprising as the options in the second stage were based on the participants’ choice in the first stage.

There is some concern that consumers who do not understand the options may not be able to answer the choice experiment appropriately. Additionally, those who found them unbelievable may give unreliable answers. As these behaviours may impact the results, we examine, within the econometric analysis, how choices made by participants who did not understand the options or found them unbelievable impact our main estimates.

Option order effects

In surveys, there is sometimes a greater tendency for participants to pick the first option on the left rather than the first on the right. To mitigate the effects of this, the order of response options was randomised in the survey questions and the second stage contingent valuation design. For the first stage choice experiments, we screened out the responses of participants who selected the same position for their first choice or second choice in all of the choice sets, which amounted to 84 responses.  

Completion time

While the mean time taken to complete the questionnaire was 16 minutes and 40 seconds and the median was 11 minutes and 15 seconds, around 5% of participants completed the questionnaire in under 5 minutes as shown in Table 3. There is a possibility that participants who completed the questionnaire quickly do not consider the options carefully enough or may even respond randomly to complete quickly. We therefore test the sensitivity of key findings to excluding participants whose completion time was less than the 10th percentile of the distribution of completion times.

Table 3: Distribution of participants’ completion times

4. Attitudes and knowledge on in-contract price rises and inflation 

In addition to the two-stage stated preference experiment, the survey also included other questions related to inflation and in-contract price rises. The primary purpose of these questions is to provide useful context for the analysis of the experiments in Section 5. For example, we elicit sentiment towards in-contract price rises and understanding of inflationary rises which aid the interpretation of our willingness to pay estimates. We also use the data from these questions to segment respondents in our willingness to pay analysis so that we can better understand differences across different groups of consumers.

This section presents findings on the sample of broadband and mobile bill-payers in regard to:

• Awareness of in-contract price rises in their broadband and mobile contracts and attitudes towards in-contract price rises in these contracts (Section 4.1)
• Understanding and knowledge of inflation (Section 4.2)
• Ability to calculate inflationary increases into broadband or mobile contracts (Section 4.3)

4.1. Awareness of and attitudes towards in-contract price rises

Awareness of in-contract price rises

Awareness of in-contract price rises is patchy. Only 60% of participants told us their mobile network and household broadband provider could increase their monthly payment during their contract term. Most people who said that their broadband provider cannot increase their monthly contract payment were wrong; 92% of people who didn’t think their monthly price could increase during their contract term are in fact with a broadband provider that can increase their monthly in-contract payment. 

Attitudes towards in-contract price rises

​A large majority of people expressed discontent about in-contract price rises. Participants were asked to rank how much they agree/disagree with specific sentiments on in-contract price rises, using a 5-point Likert scale. Figure 11 shows that more than three-quarters (78%) of respondents believe that increasing the cost of a service during a contract is always unfair and nearly two-thirds (65%) would prefer to have fixed prices throughout their contract even if it meant paying a higher price initially. Only 13% were unbothered by having the increase during their contract.

Figure 11: Attitudes of participants towards in-contract price raises

4.2. Understanding and knowledge of inflation

Participants were also tested on their understanding of measures of inflation typically used in contracts. They were asked to select which of the following options best describes the concept measured by Consumer Price Index (CPI) or Retail Price Index (RPI):

• The rate of inflation
• The broadband speed
• The rate of interest
• The rate of economic growth
• Don’t know

About half of participants either didn’t know or selected the incorrect answer (49%). This is similar to Ofcom’s research which showed nearly half of all customers did know what CPI and RPI measure [15].

If participants provided the correct response (the rate of inflation), they were then asked what the CPI rate was in January 2023. Four out of 10 participants who were asked selected “don’t know”. Just over a quarter of participants who were asked (27%) provided an answer within ±0.5% of the correct rate (which was 10.1%) and only 3% could provide the exact rate.

Next, participants were asked what they thought the CPI rate would be in January 2024. They were asked to do this as this is the burden that in-contract price rises places on consumers when they are choosing telecommunications packages. Nearly half of participants who were asked selected “don’t know” (48%). Less than a quarter of respondents who were asked (23%) were able to provide an estimate ±2% of Bank of England’s May 2023 prediction of Q1 2024 CPI, 4.4%, which we use as our interpretation of what a reasonable answer is.

4.3. Ability to calculate inflationary increases to monthly contract price

All participants were tested on their ability to calculate the new monthly payments for a £30 broadband or mobile contract if there was an in-contract price increase of CPI + 3.9%. They were given a CPI rate of 5.7% to be used.

There was a random allocation between whether the participant saw a broadband or mobile network contract in the question but this did not affect the answers.

Many participants were unable to correctly estimate new prices. Over a third of respondents selected “don’t know” as a response (37%). Only a fifth of all respondents (21%) answered correctly with the exact figure (which is £32.88) and only a further 4% provided an answer that was reasonably assumed to have been rounded up [16].

1. Understand what an inflationary increase is; (Section 4.2)
2. Have an estimate of what the future rate of inflation will be; (Section 4.2)
3. Be able to calculate the impact this will have on the monthly price. (Section 4.3)

Taken together, this implies that only one in twenty consumers (5.2%) are capable of estimating what impact an inflation-linked in-contract price rise will have on their monthly telecoms bill [17].

5. Choice experiments and contingent valuation analysis and findings

This section presents findings on the choices made by participants in the first stage choice experiments and the second stage contingent valuation. It begins with the main results from the first stage choice experiments in Section 5.1. Analysis of the second stage contingent valuation choices is presented in Sections 5.2 and 5.3, where 5.2 describes the choices made by the participants and 5.3 contains our main econometric analysis that generates estimates of consumers’ willingness to pay to avoid broadband contracts with in-contract price rises. Details of the econometrics, including model specifications, testing procedures and estimation results, are presented in Appendix C.

The willingness to pay estimates from the second stage contingent valuation choices are used to estimate the value that consumers place on having fixed price contracts, ie the certainty premium, in Section 5.4. This is done at an individual level using a representative contract and then aggregated to give an economy-wide value.

5.1. Valuation of price certainty from the first stage choice experiments

A mixed logit model was used to analyse the results of the first stage choice experiment. This was preferred to other models as it allows for random variation in preferences across participants [18]. Table 4 shows the regression coefficients from the mixed logit model and these are consistent with reasonable assumptions of consumer preferences in the broadband market. Sensitivity analyses show that these estimates are robust to a range of scenarios and assumptions relating to different sample exclusions. Our preferred model is the most conservative approach which applies the most restrictive sample requirements. It excludes the 10% of participants who completed the questionnaire fastest and participants who did understand the options or/and find them believable. For brevity, we do not present the other scenarios here, but details can be found in Appendix C. 

The results in Table 4 indicate that, all other things being equal, respondents prefer cheaper contracts, faster speeds and shorter contract lengths. Lower setup costs are preferred, but this only matters for higher setup costs (£30 in our experiment) as the results are not statistically significant for smaller amounts. We also find that respondents prefer fixed price contracts over contracts with either inflation-linked or ad hoc in-contract price rises. Their preference for fixed prices was higher in the case of ad hoc price increases. 

Table 4: Estimates from the mixed logit regression model used on responses in stage one

We are interested in exactly how much consumers prefer fixed price contracts over those with in-contract price rises. To explore this, we calculate willingness to pay estimates by dividing the coefficient of the price increase type over the coefficient of the monthly cost. 

Table 5 shows that respondents have a positive willingness to pay to avoid contracts that have in-contract price rises. On average, they valued fixed prices at £0.42 per month compared to inflation-linked in-contract price increases and at £0.58 per month compared to ad hoc in-contract price increases. 

These willingness to pay estimates appear implausibly low. They suggest that the total price an average customer is prepared to pay for a fixed price contract is less than the expected total price for a contract with an in-contract price rise. This is inconsistent with the survey data analysed in Section 4, which shows, for example, that 78% of people think in-contract price rises are always unfair. The most plausible reason for this inconsistency is because contract price variation terms are attributes that lack saliency. Since respondents need to consider multiple attributes in this choice experiment, the choice is complex and the saliency of an attribute will affect the value placed upon it [19].

Price variation terms have low levels of saliency in our choice experiments in part because the information about in-contract price rises is presented in a way that is typical of current real-world practices. We know these practices lack transparency, hence the forthcoming improved transparency requirements that advertisers will need to adhere to from December 2023. Further, our testing of consumer understanding shows that many consumers struggle with these price variation terms. By contrast, our estimates of willingness to pay for the more salient attributes of speed and contract length are plausible. Table 5 shows that consumers prefer shorter contracts and faster speeds, although there are diminishing returns to speed.

Table 5: Consumer valuation of price certainty, broadband speeds and contract lengths at different levels in stage one

A predicted consequence of the lack of saliency of these contract terms is that they will not not be set competitively, a prediction that is consistent with the evident similarity of these terms across providers. The low estimates for willingness to pay for fixed-price contracts support our methodological design. They demonstrate the need to present as simple and as transparent a choice as possible when eliciting a value from consumers for pricing certainty, as we do in our second-stage contingent valuation survey.

5.2. Descriptive analysis of second stage contingent valuation choices

To elicit more informed willingness to pay estimates we used a contingent valuation method in the second stage of the survey. This was done using a double-bound dichotomous choice (DBDC) format in which respondents each followed one of four possible paths:

• Yes to paying higher prices for a fixed price contract but No when the difference was doubled. For example, yes to £5 per month, but no to £10.
• Yes to paying higher prices for a fixed price contract and Yes even when the difference was doubled. For example, yes to £5 per month and yes to £10.
• No to paying higher prices for a fixed price contract but Yes when the difference was halved. For example, no to £10 per month, but yes to £5.
• No to paying higher prices for a fixed price contract and No even when the difference was halved. For example, no to £10 per month and no to £5.

Table 6 presents the frequency with which participants chose these combinations for both types of in-contract price rises (relative to a fixed price). The frequencies combine across the range of initial cost amounts shown from the list = {£1.50, £5.00, £10.00}. The results show a fairly even split of participants across the different combinations, although the highest prevalence are for those who chose a contract with higher fixed prices and continued to do so even after the price was doubled. This aligns with our main estimate in the section below where we find consumers' valuations for contracts with fixed prices are much higher than those with in-contract price rises.

Table 6: Frequencies of participants selecting a fixed price contract in stage two

Figure 12 shows the percentage of respondents who selected the fixed price broadband contract at the various price difference levels of {£1.50, £5.00, £10.00} in the first part of the second stage contingent valuation design. The graph indicates that as the price difference between the fixed price and the in-contract price rise options increases, a lower proportion of respondents selected the fixed price option. This is in line with our expectations of consumer behaviour.

Figure 12: Percentage of respondents who chose the broadband contract with fixed prices at various price difference levels from the from the first part of the contingent valuation stage

5.3.1. Main estimates

To estimate consumers’ median willingness to pay from the contingent valuation survey we use an interval regression model as the DBDC provided an interval with the lower and upper bound of the outcomes for each respondent but not precise values between those ranges. Using this method allows us to fit a linear model with the outcomes measured as interval data. More detail on the econometric modelling can be found in Appendix C.

Figures 13 and 14 show the predicted choice probabilities with which fixed price contracts are chosen over contracts with in-contract price rises as the difference in the monthly price of the two contracts increases. These predicted choice probabilities were derived from modelling summarised above and described in Appendix C. The downward sloping curves indicate that as the price difference increases participants become less likely to choose the fixed price option. When the cost of fixing is low, such that the fixed price contract is only about £1 more per month, then the fixed price contract is likely to be almost always preferred, but even when the fixed price contract is £20 per month more expensive it is chosen with a probability of 10%. This is the same across both types of contract containing in-contract prices rises. The results are in line with expectations of consumer behaviour and adds confidence to our findings.   

Figure 13: Predicted probabilities of choosing a fixed price broadband contract over a contract with an inflation-linked in-contract price rise at various price difference levels

Figure 14: Predicted probabilities of choosing a fixed price broadband contract over a contract with an ad hoc in-contract price rise at various price difference levels

Table 7 shows the median willingness to pay derived from the interval regression model. We estimated the model five times to test the sensitivity of our estimates to different sample exclusions. The table shows that, despite some of these exclusions being fairly severe, this made no meaningful difference to the willingness to pay estimates. For subsequent analyses we use model 5, which excludes the 10% of respondents who completed the questionnaire in the fastest time and respondents who did not find it the options or/and believable as this is an appropriately conservative approach. 

Table 7: Median value of fixed prices in broadband: Baseline estimates, sensitivity tests, and final estimates (£/month)

We estimate that consumers value contracts with fixed prices at £5.03 per month more than contracts with inflation-linked in-contract price rises and at £4.54 per month more compared to ad hoc price rises. The difference between these estimates and those found in the first-stage choice experiment are stark and demonstrate the impact of providing a clear and understandable choice to consumers to be able to accurately estimate the value of pricing certainty. 

We favour an estimate of median willingness to pay for fixed prices, rather than the mean willingness to pay. On practical grounds, the mean tends to be more difficult to estimate reliably using stated preference methods due to difficulties in accurately measuring the upper tail of the willingness to pay distribution. The median measure, by contrast, can much more reliably be estimated as it does not rely on capturing data on the shape of the distribution; it only requires knowledge of the price point at which 50% would accept and 50% would reject. Conceptually, the median measures the value to the average consumer. The mean, by contrast, tends to be skewed towards a higher value than the majority would hold due to the influence of an upper tail with very high willingness to pay values. A key advantage of the median is that it is not at all susceptible to influence from outliers with extremely high values.

We discuss the interpretation of these willingness to pay estimates in Section 5.4 below, when we use them to estimate the value consumers place on having price certainty, but first we explore how these estimates vary across groups of consumers.

5.3.2. Segment analysis

To examine whether the willingness to pay for certainty might differ across groups of consumers, we ran separate econometric models, using the same model and exclusion restriction as the main estimates, for different segments and complements. Table 8 provides a list of the segments covered in the analysis. Each segment’s mean fixed price was compared against the valuation of the complement segment ‘Other’ (eg, participants aged 18-34 vs participants aged 35 or above and participants aged 35-54 vs participants aged 18-34 and 55 or above), and statistical significance of the differences (at the 5% level of significance) was assessed via t-tests.

Table 8: Variables used for segment analysis

Statistically significant differences in valuations across segments are shown in Table 9, with higher values being highlighted in blue and lower values in yellow. In terms of socioeconomic and demographic characteristics, younger respondents had a higher median willingness to pay. Education was also a differentiating factor as those with no university education have higher valuations.

For other segments, we find that the willingness to pay for a fixed price is higher for people who:

• did not know CPI / RPI measured inflation compared to those who did; 
• have higher monthly contract costs; 
• are currently with a provider who adopts ad hoc price rises compared to those with a provider who adopts inflation-linked ones, and
• were shown 24-month contracts, which is consistent with there being more uncertainty in longer contracts as they are more likely to be subject to two annual in-contract price increases.  

There were no statistically significant differences between income levels, other ages (participants aged 35-54 vs participants aged 18-34 and 55+; participants aged 55+ vs participants below 55), and whether disabilities were present.

The pattern of estimates is consistent when comparing fixed prices with either inflation-linked or ad hoc in-contract price rises. 

Table 9: Segment analysis: Value of consumer preferences for fixed prices versus in-contract price rises

Additional analysis was conducted on whether consumers’ valuation of fixed prices varied with their inflation expectations. Consumers with higher inflation expectations should expect the total cost of a contract with an in-contract price rise to be higher, and so should have a higher willingness to pay for a fixed price contract. The results suggest that inflation expectations have the expected positive impact on willingness to pay; however they are not able to be precisely estimated. This lack of prediction is partly due to the small sample size that provided inflation expectations within the survey [20].

5.4. Valuation of price certainty

The purpose of this study is to find a value to consumers from having certainty over how much their home broadband would cost per month. To derive the value of this certainty premium we need to calculate how much of the willingness to pay estimate represents the smoothing of an in-contract price rise across the whole contract. Subtracting the expected in-contract price rise from the estimated median willingness to pay for a fixed-price contract gives the value of the certainty premium. We do this for an individual consumer and then aggregate this up to get an economy-wide value.  

We use the estimated median willingness to pay to aggregate up to an economy-wide value. Although mean willingness to pay estimates are typically used for this, the median measure was considered to be practically and conceptually more appealing, as discussed above.

5.4.1. The value to individuals

To calculate the certainty premium for a typical individual we take the approach of identifying a representative contract. In doing so we make the following assumptions:

• The contract lasts 18 months, as this is the most common contract length for broadband packages [21]. 
• A single in-contract price rise occurs in month 11 of the contract. This reflects that we conducted the survey in June 2023 and respondents were told the in-contract price rise would happen from April 2024, so 10 of the 18 months are at the initial price.
• The monthly price of a contract is £35, which was the median price of a broadband contract we found at a price comparison website (and hence the median price in our experiment). 

We also need to make an assumption about consumers’ expectations of the in-contract price rise. For the inflation-linked contracts this means we need an estimate of consumer expectations of future inflation and our survey suggests two possible values for this. First, the median expected inflation rate stated by respondents was 7%. This is considerably higher than official forecasts for inflation, no doubt because people had recently experienced higher inflation rates. It does directly represent consumer inflation expectations but only for a minority as only 27% of respondents were able to provide a forecast. Alternatively, we could use the hypothetical CPI rate of 5.7% that was given to respondents for the calculation question just before they answered the contingent valuation questions in stage two. It seems reasonable to assume that at least some respondents will have anchored on this rate when accounting for inflationary increases in stage two. We estimate the certainty premium for both of these, but our preferred estimate is that which uses the more conservative estimate of 5.7%. As a further sensitivity check we also estimate the certainty premium using the Bank of England’s (May 2023) forecast for CPI in January 2024, which was 4.4%.

Using the assumptions above, the certainty premium for the representative consumer when faced with inflation-linked in-contract price rises is calculated by:

(Final estimates of median value *10)+((Final estimates of median value -
((Inflation rate ∈{4.4%, 5.7%,7%}+ 3.9%) * 35)) * 8)

For contracts with ad hoc increases we have no evidence from the survey on what consumers expect this to be, but we can use actual increases as an indicator. Providers who implemented ad hoc price increases in 2023 applied an average in-contract price increase of 10.5%. This is approximately 73%-77% of the increase of providers who adopted inflation-linked price rises in 2023. Thus, we assume that ad hoc providers will adopt an increase in 2024 that is 75% of the increase inflation-linked providers implement.

The certainty premium for the representative consumer when faced with ad hoc in-contract price rises is therefore calculated by:

(Final estimates of median value * 10)+((Final estimates of median value -  

((Inflation rate ∈{4.4%, 5.7%,7%}+ 3.9%) * 0.75 * 35)) * 8)  

These equations generate certainty premia for a whole 18-month contract, and we divide this number by 18 to get a monthly value, as shown in Table 10. Overall, the certainty premium is fairly consistent across the different rates of inflation expectations. The certainty premium falls as inflation expectations increase, since we are attributing a greater share of the willingness to pay estimate to simply compensating for an in-contract price rise. However, the impact of a change in inflation expectations is relatively modest, with a one percentage point increase in inflation expectations only reducing the certainty premium by 16p.

Table 10: Certainty premiums for having fixed prices versus inflation-linked and ad hoc in-contract price increases

More generally, it seems likely that the value of the certainty premium will be affected by the inflation environment at the time at which the survey was conducted as this will have shaped consumer sentiment. In particular, we expect the certainty premium is positively correlated with the degree of inflation volatility recently experienced by respondents, but the extent to which this is the case is unknowable. Given this, if the inflation rate in the UK falls and becomes more stable, as is forecast, it would be prudent to interpret these estimates of the certainty premium as a reasonable upper bound for the value that consumers would place on having fixed prices for broadband contracts.

5.4.2. Economy-wide value

To estimate the total value of the certainty premium consumers place on having fixed prices for broadband for UK households based on our final estimates of the median value (section 5.4) and using the following data:

• According to ONS estimates, there are 28,200,000 UK households [22]
• According to Ofcom, 87% of UK households take a home fixed broadband service [23]
• According to Ofcom, 72% of UK households are in contract [24]

Multiplying these to get the number of UK households which take a fixed broadband service and are in contract gives a relevant population of 17,665,000. 

According to Ofcom’s Switching Tracker 2022, 69% are with a provider who adopts inflation-linked in-contract price rises and 27% are with a provider who adopts ad hoc ones [25, 26]. The remaining 4% either don’t know their broadband provider or are with a fixed price provider or have listed a provider as ‘Other’, meaning we are unable to tell the type of in-contract price rise their provider adopts.

Scaling up the certainty premium per month to a year and multiplying those figures by the relevant population of households and the proportion of consumers with providers who adopt inflation-linked or ad hoc in-contract price rises, yields the following annual estimates.

Table 11: Total annual value of fixed prices to UK households for a representative contract of £35 and a minimum contract period of 18 months, with different inflation expectations

To get an overall, UK-wide annual estimate of the value consumers place on having broadband contracts with fixed prices we need to sum the estimates for contracts that have inflation-linked and ad hoc price increases. For our preferred consumer inflation expectation of 5.7% this gives a total annual certainty premium of £710 million, with confidence intervals of £599 million and £833 million.

Sensitivity analysis

The estimates above are for a representative contract of £35 and 18 months, but we know from our earlier analysis that the willingness to pay estimates vary according to contract length and monthly price. 

People with 24-month contracts have higher average willingness to pay for fixed prices and so if we were able to weight our aggregation by the proportion of people with each contract length, this would give a higher aggregated certainty premium. Unfortunately, we have been unable to find reliable data on the split of contract lengths across households and so we continue with the conservative assumption treating all households as having an 18-month contract. 

In terms of monthly price, we know that individuals with higher monthly prices have, on average, higher willingness to pay to fix prices. We explore the impact of this on the aggregate certainty premium by using data on the actual prices paid by consumers to weight the willingness to pay at each price point [27]. Table 12 presents the breakdown of the aggregate certainty premium across the monthly prices. It shows that weighting the aggregation leads to only a relatively small increase in the overall value of the certainty premium.

Table 12: Breakdown of population level certainty premium estimates across different monthly prices using a representative contract length of 18 months

6. Conclusion

In-contract price rises are a common feature of broadband and mobile contracts in the UK. These contract terms have been widely criticised, with one issue being that they prevent consumers from having certainty over the price they will pay for the duration of their contract. Sharp and unplanned price rises can strain household budgets and lead to financial difficulty, while it is well established that people have a preference for certainty in financial outcomes. 

This study has sought to estimate the value that consumers place on having price certainty for the duration of a broadband contract. We conducted a large-scale survey that included a choice experiment and contingent valuation questions, and these were used to estimate consumers’ willingness to pay for fixed price contracts. The results unequivocally show that consumers have strong preferences for pricing certainty. 

From the contingent valuation, the willingness to pay estimates are £5.03 per consumer per month to avoid inflation-linked price rises, and £4.54 per consumer per month to avoid ad hoc in-contract price rises. Valuations are higher for, among other things, consumers who are younger and with less education.

There are good grounds for considering these valuations to be robust and reliable:

• An extensive development and testing programme found that the core choices were considered plausible by participants, and that participants had no problems answering these questions appropriately.
• Values were obtained from a large national sample of more than 4,000 broadband bill-payers 
• Valuations vary in line with expectations:
  • Valuations were higher when respondents were shown longer contract lengths and higher monthly costs
  • Valuations were higher when respondents were shown contracts with higher monthly costs

Translating the willingness to pay estimates into a certainty premium for a representative contract, in which we take into account expected in-contract price rises, we find that people are prepared to pay £3.54 per month to avoid inflation-linked in-contract price rises and £3.42 per month to avoid ad hoc in-contract price rises. 

Aggregating these estimates for the certainty premium across the economy we estimate the total certainty premium from having fixed prices in domestic broadband contracts to be £710 million, of which £514 million relates to inflation-linked in-contract price rises, which are the most commonly used.

Where possible we have made conservative assumptions in coming to this estimate. For example, choosing a representative contract priced at £35 per month and lasting only 18 months. However, the estimates will inevitably be affected by the research being conducted at a time of relatively high and volatile inflation, meaning people are likely to be more aware of how uncertainty around inflation can unexpectedly cause them to pay higher prices. We would expect that recent experience of more volatile inflation would increase the certainty premium, but the extent to which it might do so is unknowable. If the inflation rate in the UK falls and becomes more stable, as is forecast, we feel it would be prudent to interpret the valuations produced by this research as a reasonable upper bound for the value that consumers would place on having fixed prices for broadband contracts.

Overall, the study provides strong evidence that consumers place considerable value on having prices that are fixed for the duration of a contract. We expect that the evidence in this report will prove valuable to anyone seeking to understand the potential benefits to consumers of a change in market rules that would prohibit or restrict the use of in-contract price rises in mobile and broadband markets.

Footnotes

[1] Ofcom (2013) Guidance on “material detriment” under GC9.6 in relation to price rises and notification of contract modifications  ↑

[2] Ofcom (2023) Affordability of communications services: April 2023 update  ↑

[3] Ofcom (2013) Price rises in fixed term contracts: Options to address consumer harm  ↑

[4] In accordance with this, providers are obligated to inform their customers about the increase and highlight their right to leave, after which customers have one month in which to exercise their right to penalty-free exit  ↑

[5] CAP and BCAP (2012) Mid-contract price increases in telecoms contracts  ↑

[6] Which? (2022) Broadband customers could face bill hikes of £113 in 2023  ↑

[7] Which? (2023) Millions of broadband customers trapped between price hikes and exit fees of over £200  ↑

[8]  Ofcom (2023) Ofcom to review inflation-linked telecoms price rises  ↑

[9] Ofcom (2023) Affordability of communications services: April 2023 update  ↑

[10] As per expected utility theory, risk averse consumers with concave utility functions will have a certainty equivalent value that is less than the expected value of an uncertain outcome  ↑

[11] Best-worst scaling is a survey method for assessing respondents’ ranked preferences by identifying the extremes (eg best and worst options, most and least important aspects)  ↑

[12] Roth, Wänke, & Erev (2016). Click or Skip: The Role of Experience in Easy-Click Checking Decisions. Journal of Consumer Research  ↑

[13] Ofwat (2022) Guidance for water companies: testing customers’ views of the acceptability and affordability of PR24 business plans  ↑

[14] Discussion guide can be found in Appendix A  ↑

[15] Ofcom (2023) Ofcom to review inflation-linked telecoms price rises  ↑

[16] Answers of £32.89, £32.90 and £33.00 were accepted as plausible rounded up responses  ↑

[17] See Which? (2023) Customer knowledge and understanding of mid contract price rises for more details  ↑

[18] Revolt, D. & Train, K. (1998) Mixed logit with repeated choices: Households’ choices of appliance efficiency level, The Review of Economics and Statistics  ↑

[19] Bordalo, P., Gennaioli, N. and Shleifer, A., 2013. Salience and consumer choice. Journal of Political Economy, 121(5), pp.803-843  ↑

[20] Since only 27% of respondents gave an inflation expectation for January 2024, this may not be a representative sample and so we have controlled for correlated variables such as income in the regression model (see Appendix C)  ↑

[21] Three of the four largest broadband providers either only offered 18-month contacts or offered mainly 18-month contracts  ↑

[22] Office for National Statistics (2022) Families and households in the UK: 2022  ↑

[23] Ofcom (2022) UK home broadband performance, measurement period March 2022  ↑

[24] Ofcom (2022) Ofcom Core Switching Tracker Data Tables. Relevant question from Q6K  ↑

[25] Ofcom (2022) Ofcom Core Switching Tracker Data Tables. Relevant question from Q6C  ↑

[26] We include Virgin Media among the providers with inflation-linked increases since this will apply from 2024  ↑

[27] Data taken from a Which? survey of broadband consumers in 2023 that asked c.4,000 consumers how much their monthly broadband bill is  ↑

About

Which? is the UK’s consumer champion, here to make life simpler, fairer and safer for everyone. Our research gets to the heart of consumer issues, our advice is impartial, and our rigorous product tests lead to expert recommendations. We’re the independent consumer voice that works with politicians and lawmakers, investigates, holds businesses to account and makes change happen. As an organisation we’re not for profit and all for making consumers more powerful.

PJM Economics is a small team of microeconomists led by founding director Dr Paul Metcalfe. The company provides bespoke economic research and consultancy services primarily in sectors subject to economic regulation. Core service areas include, economic valuation, appraisal and the modelling and forecasting of microeconomic variables.

Authors

Nicole Chan (Which?), Stephen McDonald (Which?), Paul Metcalfe (PJM Economics)For further information, please contact [email protected].

Citation

If citing this paper in your own work, our preferred citation is: Which? (2023), The benefit of certainty: Valuing consumer preferences for fixed price broadband contracts.