Firms’ business models and cost recovery

Back to top

Firms’ business models and cost recovery

We looked at:

  • the share of revenue from conditional charges (default, cash withdrawal and foreign exchange fees)
  • the relative profitability of different consumer groups
  • the impact of the Interchange Fee Regulation

Conditional charges account for around 5% to 12% of most firms’ revenue, with over 80% of revenue coming from a combination of net interest income, interchange fees and balance transfer fees. The relative importance of these revenue streams varies by product and target consumer group.

We find that:

  • firms tend to compete strongly on certain product features − the features they compete on differ by product, partly because consumer focus on different features when choosing different products
  • while different types of products may have a similar profitability over a five year lifecycle, the profile of that profit varies significantly over time between product types
  • the consumer behaviour underlying products’ profitability also differs by product.

We do not consider that cross-subsidisation in the credit card market materially restricts entry or expansion in the market. We did not find firms targeting particular groups of consumers or behavioural types with a view to cross-subsidising others − depending on the design of the product any consumer behavioural type can be profitable, except those accounts that default.

We estimate the introduction of the Interchange Fee Regulation will reduce firm income by 5% to 10% on average. Firms that have proportionally higher levels of transacting consumers face bigger losses in income. Firms say they will respond by offering more products with a small or increased annual fee or diluting rewards schemes. These developments do not give us direct cause for concern where appropriately communicated to consumers.

This chapter outlines our interim findings in relation to how firms compete in the market, how firms recover their costs across different consumer groups, and what impact this has on the market. The findings are based on our review of firm data and strategy documents submitted by firms, discussions with firms, and our account-level data analysis.

How firms’ business models and cost recovery affect competition

Understanding credit card firms’ commercial proposition (how products are designed, the services they offer and how they generate profit) is critical to understanding the competitive dynamics of the credit card market.

As stated in our terms of reference62, we want to understand the extent to which product design could inhibit competition or cause consumer detriment. Specifically, we want to explore whether:

  • products that create cross-subsidisation between different consumer groups could restrict entry or expansion in the market
  • cross-subsidisation results in certain consumer groups being unduly disadvantaged

To this end, we are interested in understanding the relationship between firms’ commercial position, consumer demand and the competitive environment this creates.

In the first half of the chapter we consider firms’ revenue streams and their relative importance. We then look at how different types of products make money and the main consumer behaviours driving that profitability.

In the second half of the chapter, we then apply this analysis to understand the impact on competition.

How firms recover costs

As discussed in Chapter 3, firms’ strategies in relation to credit cards revolve around effectively managing credit risk and offering products that meet consumer preferences.

Different firms have different risk appetites that determine their willingness to lend to higher-risk consumers. Within their risk appetite, it appears that firms are in general willing to pursue any product (that is, combination of product features) that is expected to meet internal profitability targets.

Product design is intended to match consumers’ preferences while meeting these targets. Typically, a close match with consumer tastes leads to a high-performing product.

Firms that better match their product(s) to what consumers are looking for or which are more efficient can offer more attractive product features, attracting more consumers and reaping higher profits. As discussed in Chapter 4, where switching can take place freely, this allows for competition between firms.

Revenue streams

Credit card firms have six main revenue streams:

  • net interest income
  • interchange income
  • annual (or monthly) fees
  • balance transfer fees
  • default fees
  • other income conditional on consumers using a service, for example cash withdrawal fees and foreign exchange fees.

The relative importance of these revenue streams will vary depending on a particular product’s features and target consumer groups. For example, rewards cards tend to have a higher proportion of revenue from interchange income while credit builder cards are likely to have a higher proportion of income from interest and default fees.

Most large firms that offer a range of products typically earn the bulk of their revenue from interest income, interchange fees and balance transfer fees. Of the 19 firms for whom we have data, 15 out of 19 firms made 80% or more of their income from these sources with seven of these 15 firms earning more than 90% of their income from these revenue streams.

Conditional charges such as default fees, cash withdrawal fees and foreign exchange fees accounted for around 5% to 12% of overall firm revenue for 14 of the 18 firms we have this data for.

Default fees, in most cases, are a material but small proportion of total income (around 2% to 7%). For firms operating in the higher risk segment the proportion is higher at around 9%. We found default fees are currently clustering at £12 following the OFT review of default charges in 2009.

We have found practices by some firms (although not across the market) that lead to default fees being charged several times per default event. We are pursuing these with the individual firms in question.

How different products make money and main behaviours driving profitability

To understand competitive dynamics in the credit card market, it is essential to understand how firms think about their business. In this section we look at how different products make money and the main behaviours driving profitability by using the framework typically used by firms to forecast the profitability of their products.

This shows that:

  • Profitability is achieved by choosing a combination of product features that balance various conflicting priorities. Credit limits need to be set in a way that maximises revenue while minimising the risk and cost of default. The final design must also meet consumer preferences and attract consumers for whom the product was designed.
  • While different types of products may have a similar profitability over a five year lifecycle the profile of that profit over time varies significantly between product types. For example, a rewards product has a fairly stable level of profit over the five years, while a 0% purchase product will typically make a loss during the promotional period and then profit following the end of the promotional period.
  • The consumer behaviour underlying products’ profitability also differs by product. For example, the profitability of a purchase product is primarily driven by consumers continuing to revolve a balance once the promotional period is over, while the profitability of a rewards product is driven by high levels of consumer spend. All the main credit card firms design, monitor and evaluate their products using a series of linked mathematical models designed to forecast the profitability of each product. Firms use historical data to predict likely consumer behaviour adjusted to reflect any changes in a product’s features. Access to this historical data, along with high investment and expertise, represent the three key costs to entry into the market.

Forecasting profitability

Rather than model individual consumer behaviours, firms typically use a framework that estimates the behaviour of a theoretical ‘average’ consumer for each month for five years starting from the time the card is issued. These estimates can be multiplied by the predicted number of consumers to provide total revenue and cost figures and can then be evaluated using a variety of financial performance measures.

The estimates of each aspect of the ‘average’ consumer’s behaviour can be graphed, for example, active balance, revolve rate63, profit. These collections of graphs have various names across the industry but are commonly referred to as ‘base curves’. By looking at base curves for different product types we are able to gain an insight into:

  • how consumers respond to different product structures
  • the main behaviours driving profitability
  • how different product types make money

Below we present the base curves for several types of product to show how different card features and consumer behaviour affect profitability. The base curves are descriptive but their shape is based on the financial forecasting models provided by several firms in response to our information request.

Why use cohort analysis?

Our analysis mirrors the approach used by firms by using cohorts of consumers. A cohort is a group of consumers who have all taken out the same credit card during the same month.

Most products have a lifecycle. Consumer behaviour and financial performance follow predictable patterns over this lifecycle caused by consumers reacting to the incentives created by the design of the product. For example, active balance tends to spike in the first six months of a balance transfer product then start to decline as the balance is paid off. The number of consumers in a cohort also shrinks over time as consumers switch products or default.

When analysing a credit card product over time, data from a snap shot (for example the first three months of 2015) includes dozens or even hundreds of these cohorts mixed together. Looking at cohorts allows us to examine the product’s life cycle, how the product is designed, and how it recovers costs without the distortions created by mixing cohorts at different stages of their lifecycle.

The descriptive base curve analysis in this chapter and our quantitative analysis using the account-level data are based on cohort analysis.

Long-term balance transfer products

We start by analysing a long term balance transfer (BT) product. To show how the forecasting process works, we look at several base curves relating to long term BT products. All the base curves in this section relate to the same cohort except where we examine how base curves change if we further subdivide cohorts according to their risk profile. Each base curve represents a different key variable that impacts profitability.

Later in this chapter we discuss the profitability base curves for other products. Annex 5 provides a complete base curve analysis for all products in line with the analysis of a long term balance transfer product covered in this chapter.

As the base curves will show, balance transfer products, in which a consumer can transfer existing debt from one credit card to another, rely on the initial balance transfer fee and a proportion of consumers revolving their balance once the promotional period ends to maintain profitability.

Active balances

The balance transfer fee and the interest generated at the end of the promotional period are proportional to the value of the balance on long-term BT cards. The higher the outstanding balance the greater the amount of revenue generated. However, higher outstanding balances also increase the expected charge-off cost. Credit card design requires balancing these conflicting factors.

To do this, firms offer different credit limits, with lower risk consumers typically having higher credit limits and therefore higher balances. The base curve for monthly balance on a long-term BT card has a feature shape (see Figure 11 below for an example of a balance transfer product with a 30 month 0% offer).

Figure 11: Example monthly balance for a long-term balance transfer offer

Figure 11: Example monthly balance for a long-term balance transfer offer

The shape of the base curve is due to consumers’ behaviour and their response to the incentives created by the design of the product. For example the precise shape of this curve is dependent on the length of the promotional period. The peak of balances is generally observed in months three to four. This is due to the promotional balance transfer typically needing to be completed in the first 90 days.

The monthly balance in Figure 11 above decreases from months 4 to 30 due to consumers paying down their balance transfer during the promotional period. Pay down begins after the balance is transferred and tends to stabilise at the end of the promotional period. At the end of the promotional period:

  • Some accounts in the cohort will have fully paid off and may become dormant or revert to transacting behaviour.
  • A further proportion will continue to revolve a balance which will tend to remain relatively static on average after the promotional period has ended.

As with the peak balance the inflection of the base curve is due to the length of the promotional period. If the promotional period was altered, the rate of pay down (the gradient of the base curve between ‘peak balances’ and ‘end of promotional period’) and the period in which balances became relatively constant (the inflection point at ‘end of promotional period’) would both change in response, reflecting consumers’ changing behaviours. These changes would impact the forecast interest income and hence profitability. This dynamic response by consumers to changes in product features is common to credit card profitability modelling, and understanding these dynamics is a key determinant of the accuracy of a firm’s models.

Another way in which decisions about the product can affect balances (and ultimately profitability) is risk appetite and credit limit. In Figure 12 below, we consider two different cohorts, one higher risk and one lower risk, and look at their active balance base curves. Both exhibit similar shapes but we note some differences.

The peak balance is lower for the higher risk group. This is caused by firms seeking to limit bad debt costs by imposing lower credit limits on the higher risk group. Limiting active balance lowers charge off amounts but also limits interest income. Firms need to find the right credit limit for each risk group to balance interest income and bad debt cost.

We also note that the slope of the curve is shallower for the higher risk group indicating the pay down rate is typically lower in the higher risk segments due to a greater propensity towards revolving behaviour by higher risk consumers.

Figure 12: Example monthly balance for a long-term balance transfer offer for a lower risk and a higher risk cohort of consumers

Figure 12: Example monthly balance for a long-term balance transfer offer for a lower risk and a higher risk cohort of consumers

Revolve rate

Interest income is dependent on the revolve rate (the proportion of outstanding balance that attracts interest income). The revolve rate base curve is also a product of consumer behaviour and the structure of the product.

Figure 13: Example monthly revolve rate for a long-term balance transfer offer

Figure 13: Example monthly revolve rate for a long-term balance transfer offer

During the promotional period, no interest is charged on the transferred balance although there may be interest charged on new spending (unless a 0% purchase offer is also included as part of the product). Therefore the revolve rate will:

  • Remain low until the end of the promotional period though the proportion of income attracting interest does slowly creep up as accounts make new purchases or promotional withdrawal64 makes some transferred balances liable for interest charges.
  • Spike at the end of the promotional period, as all balances become eligible for interest charges. Typically, those remaining will pay down a proportion of this balance in the six months after the end of the promotional period.
  • Subsequently remain high as those remaining will typically be revolving their balance and hence attracting interest charges as with a standard product.

Profit before tax (PBT)

The distinctive profile of monthly profit before tax on a balance transfer product can be seen in Figure 14 below. Note, this base curve illustrates the average profitability of an account in each month since the product was issued.

PBT is positive in the first months due to the balance transfer fee income. After the initial positive cash flow, cumulative PBT becomes negative, in our illustration around month 13 as losses accumulate. This is due to reduced interest income in the months where there is the 0% promotional offer, and limited spending on the product combined with running costs for the product.65 Income can also be generated by interchange income from purchases and from charges for cash withdrawal, foreign exchange or late payments. However, these are typically small amounts. In our illustration of a 30 month BT product losses made during the promotional period are fully recovered in month 44.

Figure 14: Example monthly profit before tax for a long term balance transfer offer

Figure 14: Example monthly profit before tax for a long term balance transfer offer

Standard products

A standard product offers different product features and attracts different types of consumers. We can observe how these changes affect the product’s profitability profile by looking at the shape of its profitability base curve, which is a reflection of the forecast behaviour of consumers.

Standard products generate revenue from a steady level of borrowing by consumers. Within a lower risk cohort, average borrowing is typically static in standard products, though it should be noted that the individual accounts that are borrowing in each period may be changing month on month. As a result, the average profit before tax base curve is relatively flat as can be seen in Figure 15(a) below. The initial loss is caused by one-off acquisition costs paid to attract the cohort of consumers.

A credit builder product would have a similar shape. Credit limits for the cohort rise over time (as in a ‘low and grow’ card) this will lead to a stepped profile as shown in Figure 15(b) below. As the credit limit rises over time, so does the value of revolving balance and hence interest income and profit rises.

Figure 15(a): Example monthly PBT for a standard low-rate product in lower risk segment

Figure 15(a): Example monthly PBT for a standard low-rate product in lower risk segment

Figure 15(b): Example monthly PBT for a higher risk credit builder product

Figure 15(b): Example monthly PBT for a higher risk credit builder product

Combined purchase and balance transfer products

Combined purchase and balance transfer products offer promotional periods of varying lengths during which new purchases or balances transferred to the card attract 0% interest. BT-led products typically have long BT offers and short purchase offers. Dual products have BT and purchase offers of similar length. Purchase-led products have long purchase offers and short BT periods.

BT-led products have a similar profit profile to long-term BT products (discussed above) with the initial spike driven by BT fees but also by the end of the short purchase offer period. During the promotional period accounts are loss-making on average, due to low interest income levels and funding cost, but return to profitability after the end of the BT promotional period. In this example, profit spikes as the promotional period ends but falls back as consumers switch away to other products or finish paying off the transferred balance.

Figure 16(a): Example monthly PBT for a BT-led product (30 month BT, 6 month purchase offer)

Figure 16(a): Example monthly PBT for a BT-led product (30 month BT, 6 month purchase offer)

In the dual product the spike caused by BT fees is smaller as fewer consumers in the cohort take out the balance transfer. However, the product returns to profitability more quickly than the BT-led product − the promotional period is shorter and profit spikes due to the end of the BT period, then again with the end of the purchase offer. These spikes are driven by increased interest income caused by interest accruing on the outstanding balance from the promotional period, with profit falling back over time due to switching and repayment.

Figure 16(b): Example monthly PBT for a dual BT and purchase product (16 month BT, 12 month purchase offer)

Figure 16(b): Example monthly PBT for a dual BT and purchase product (16 month BT, 12 month purchase offer)

Purchase-led cards have similar profit spikes. The first peak is driven by BT fees. The second peak corresponds to the end of the purchase offer, when interest begins accruing on outstanding purchase balance.

In this case, balances transferred are too small for the increase in profit caused by higher interest income at the end of the BT period to make the cumulative PBT positive at that stage, although the losses made during the promotional period are much lower than for BT-led products.

Figure 16(c): Example monthly PBT for a purchase-led product (6 month BT, 20 month purchase offer)

Figure 16(c): Example monthly PBT for a purchase-led product (6 month BT, 20 month purchase offer)

Reward products

Reward card products typically have a profit base curve similar to either the standard product base curve (pure reward cards) or the purchase-led base curve (reward plus purchase offer) above.

All profit curves are partly driven by revenue derived from interchange. In the case of reward products the relative importance of interchange will be higher. Profit levels per account are also likely to be higher than for other simple products due to typically higher levels of spending.

Conclusion

Forecasting consumers’ response to the incentives created by a product’s design is critical to predicting financial performance and allows firms to accurately set each product feature (for example, interest rates, credit limits, balance transfer duration) so as to achieve a specified expected level of return.

Where consumer behaviour matches expectations, products are typically profitable provided enough accounts are opened. Additionally, consumers that select cards designed for their typical behaviour tend to be better quality from the firm’s point of view, for example, they are more likely to be engaged with the product and are more likely to activate and use the card. While different products focus on different consumer needs, firms can also make profits from other types of consumers using that card as well. For example:

  • a firm can generate profitability on a balance transfer card from additional borrowing
  • a pure transactor with high spend could be profitable on a mixed BT/purchase offer card
  • a reward product may be designed almost entirely to generate additional spend but still makes money from revolving behaviour

Even loss-making propositions in manageable numbers can have financial and strategic benefits for a firm, though accounts that default carry little benefit and firms do everything they can to screen out these accounts at application.

Impact on competition

In this section, we consider the implications for competition of this analysis of the behaviours driving profitability.

Firms wishing to attract consumers have an incentive to compete strongly on those features that consumers focus on, with other features likely to be subject to less competitive pressure. The features on which consumers focus will depend on the type of product they are purchasing and to some extent the features that firms emphasise in their promotions.

While evidence suggests that firms compete strongly, competition is not uniform across all the features of a product. The design of each product provides different product features (other than interest rate) on which firms can compete:

  • long-term BT products and mixed products can compete on the balance transfer (and purchase offer) length and balance transfer fee
  • reward products can compete by offering different reward structures and, where applicable, compete on the annual fee charged
  • credit builder products can compete on risk appetite, offering greater access to credit than other competitors

This can be seen, for example, on long-term balance transfer products and purchase products, where the increase in promotional lengths over recent years (Figure 17(a) has been offset to some extent by increasing post-promotional interest rates (go-to rate) (Figure 17(b)). This is consistent with our findings from the consumer survey (Chapter 4) that consumers tend to focus more on introductory promotional offers and less on post-promotional interest rates.

Figure 17(a): Average (mean) lengths of 0% balance transfers and 0% purchase offers taken out over time

Figure 17(a): Average (mean) lengths of 0% balance transfers and 0% purchase offers taken out over time

Source: FCA estimates based on sample of account-level data
Note: Figures are three month rolling averages (mean)

Figure 17b) Average go-to interest rates on 0% balance transfers and 0% purchase offers taken out over time

Figure 17b) Average go-to interest rates on 0% balance transfers and 0% purchase offers taken out over time

Source: FCA estimates based on sample of account-level data
Note: Figures are three month rolling averages (mean)

The developments in acquisition channels, such as the emergence of price comparisons websites, may have contributed to the increase in competition on these product features. Historically, PCWs only compared cards on a few key metrics. A lead in one or several of these comparison tables led to higher levels of activity and hence more profitable products. Further, a high table placing for a product tended to mean firms attracted better credit risk consumers. For any given risk profile, such an account tended to be more engaged, use the product more heavily, and be less prone to default. Firms may face an incentive to adjust the product features of a proposition so as to be as competitive as possible on features measured by PCWs so long as forecast profits were still met.

As we found in Chapter 4 the features consumers focus on varies between different segments of the market. The consumer survey, found that for consumers in the ‘low and grow’66 segment, the likelihood of acceptance was the most commonly cited feature, followed by APR or interest rate.

Higher risk consumers also face a more limited selection of products as fewer firms operate within this segment than elsewhere in the market. Some firms have suggested that they choose not to operate in the higher risk for strategic or brand reasons. Firms also noted that the score card needed to correctly estimate default rates in the higher risk segment differs from score cards used for other products, increasing the cost of expanding into this market.

Profitability in higher risk segment

The issues discussed above suggest there may be less competitive pressure on firms in the higher risk segment than in the lower risk segment. We therefore look at relative rates of return between products from both higher and lower risk segments.

To look at the relative rates of return on lending for products offered by different credit card firms, we use our account-level data set. We calculate a return on lending measure by dividing the total profits made by groups of accounts over five years by the outstanding balances of that cohort over five years to produce a monthly rate of return for an average account, which we then annualise.

We found that, between 2010 and 2014, returns were on average higher for higher risk products typically by around 6 percentage points. Returns on some lower risk products were as high as those in the higher risk segment.

Annex 6 lays out how these calculations were performed in detail and demonstrates how altering our methodology did not affect the relative profitability of the two groups. The products used in our analysis were selected to give a representative sample, while also being based on data that was as comparable as possible.

Despite this we are cautious about placing too much weight in this result since:

  • Despite an attempt to standardise the account-level data between firms, we are conscious that differences in firm data, driven by differences in firms’ structure, accounting systems and data submissions, make these results illustrative rather than declarative.
  • Methodologically our approach is similar to one performance metric used by the industry although we acknowledge several performance measures are more typically used to give a complete view of profitability. Our analysis focuses primarily on marginal return from products and makes no attempt to factor in economic or regulatory capital costs, which could alter relative profitability.
  • We analysed profitability over five years, in line with standard practice in the industry, however our analysis does not estimate terminal values on accounts and so may be omitting the effect that account longevity can have on firms’ incentives.

Cross-subsidisation

As set out in our terms of reference, we wanted to investigate the existence and impact, if any, of cross-subsidisation in the credit card market.

Cross-subsidisation is difficult to define and measure. Two common definitions we considered as a lens for our preliminary analysis were:

  • The sale of one set of goods or services at a loss enables or increases the sale of another good or service.
  • Profits from one group of consumers are used to cover the losses from another.

Cross-subsidisation is not necessarily anti-competitive or detrimental. We are only concerned about cross-subsidisation where it impacts competition or is actively detrimental to consumer welfare. This includes where it:

  • Creates a barrier to entry and expansion
  • Results in certain consumer groups being unduly disadvantaged.

Based on our analysis, we do not consider that cross-subsidisation in the credit card market materially restricts entry or expansion in the market. We did not find firms targeting particular groups of consumers or behavioural types with a view to cross-subsidising others - depending on the design of the product, any consumer behavioural type can be profitable, except those accounts that default. Within a given product, we do not consider that groups of consumers whose behaviour is profitable for the firm are significantly less restricted in their ability to switch than those whose behaviour is loss-making.

Firms’ strategies and cross-subsidisation

Firms’ submissions and our own analysis of the models provided by firms, make it clear that cohort analysis (discussed above) is the primary way in which the financial performance of products is managed and forecast. Firms’ strategies are focused on creating profitable cohorts of consumers.67 In general, we find that each credit card product is designed to be profitable in its own right rather than one credit card product being designed to cross-subsidise another.

This approach accepts that some consumers within a cohort will behave in a way that creates a loss on individual accounts. This could include very low level transacting behaviour, switching at the end of the promotional period or defaulting. Other consumers will behave in a way that generates sufficient profit such that the expected performance of the cohort as a group is overall profitable. Where firms can screen materially loss-making consumers, such as those likely to default, they do so. In other cases, firms may have strategic reasons for retaining loss-making consumers, for example linked to their retail banking offering.

When looking at actual financial performance we note that individual accounts (or even whole cohorts) can become loss-making if consumer behaviour shifts away from the assumptions made when designing the product. Firms can also make mistakes regarding their forecasting and set prices/credit limits at the wrong level, leading to loss-making accounts.

While firms are aware of where losses are being made on their products, we do not consider that firms are targeting particular groups of consumers or behavioural types with a view to cross-subsidising others.

Impact of cross-subsidisation on competition

Barrier to entry and expansion

Cross-subsidisation can be detrimental when it creates a barrier to entry or expansion.

Products with long promotional periods can involve cross-subsidisation between those accounts that switch at the end of the promotional period and other accounts, as well as cross-subsidisation over time with consumers in general being loss-making during the promotional period and profitable thereafter.

We find that this can strengthen barriers to entry and expansion that already exist within the market.

As discussed in Chapter 368, an entrant or expanding firm in the credit card market would require a large amount of investment capital (along with considerable expertise and access to a great deal of consumer data). The prevalence of products with a long promotional period that are expected to make a loss in the first few years with profits not expected until the last couple of years of a five year cycle means an entrant will also require capital to cover the three years of loss-making in order to compete in the market, thus increasing the costs of entry.

At this stage we do not consider that these costs are significant enough to constitute a material additional barrier:

  • Any entrant into the market would require a large amount of investment capital regardless of what product(s) they offered. While products with long promotional periods require higher levels of working capital than other products it seems unlikely that the level of this additional capital would be such that it would prevent the entry of a firm able to raise the capital to enter the credit card market.
  • There are various entry strategies into the credit card market.69 Firms can enter and gain market share with a number of different products, including those that do not have long promotional periods.

There are also benefits to competition arising from products with attractive promotional offers such as encouraging switching and increasing competition for existing consumers.

We therefore do not consider that cross-subsidisation materially restricts entry or expansion in the market, and do not propose to take any further action in relation to how firms recover costs or cross-subsidisation.

Certain consumer groups unduly disadvantaged

Cross-subsidisation can also be detrimental if it results in certain consumer groups being unduly disadvantaged.

As discussed above, firms’ cohort analysis centres on analysis based on averages, with less consideration given to the distribution of the income generated. In principle, this could mean that a group of consumers within a cohort are making a disproportionate contribution to firm revenue, and be suffering harm as a result.

In this chapter we have shown that different behaviour patterns can be more or less profitable to firms and more or less beneficial to consumers depending on the products being offered.

Within a given product, we do not consider that groups of consumers whose behaviour is profitable for the firm are significantly less restricted in their ability to switch than those whose behaviour is loss-making.

Taken together, we conclude that a strategy that relies on differentiating price to create market conditions where only non-switching consumers are profitable is unlikely and impractical.

In Chapter 6 we use the account-level data to look at the distribution of firm income across consumers. In particular, whether firms’ returns on products is due to a small number of very profitable consumers and whether those consumers who may be struggling with problem credit card debt are the most profitable accounts.

Cross-subsidisation between transactors and revolvers

The terms of reference set out our intention to examine in more detail the suggestion that transactors (consumers who typically pay off their balance every month) are cross-subsidised by revolvers (consumers who typically revolve their balance from month to month).

We found that a consumer who revolves their balance is typically more profitable than the same consumer who pays off their balance every month due to the net interest income they generate. However, transactors are typically profitable rather than loss-making and firm business models are not predicated on turning transactors into revolvers over time.

As transactors typically pay off their balance every month they tend not to attract interest income on their balance. These consumers generate revenue from interchange fees. These fees are derived from the value of transactions made using the card, hence greater spending on the card generates higher revenue. In addition, fees for cash withdrawals, overseas spending etc. are also a source of income from these consumers.

The value of spending is the critical factor that determines if transactors are profitable as cost varies less with spending than revenue. Firms that attract transactors willing and able to spend more will find transactors more profitable. Firms that attract low spending transactors will typically find them less profitable, and in some cases loss-making.

Firms’ ability to attract high spending transactors is in part dependent on their strategy. The types of products offered by a firm (and to an extent the marketing and branding of that product including any co-brand partnerships) will determine what types of transactors they attract. If the design of these products is successful, transactors will find that they tend to benefit most from the products offered by these firms and switch to them. Firms that have designed their products to attract transactors will have a higher proportion of naturally high-spending transactors than other firms and would tend to find transactors are more profitable. These products are also more likely to encourage engagement and spending on the card and hence maximise revenue from each transactor in a way that other cards may not.

In addition, to transactors, accounts that end up in default and accounts that quickly switch at the end of promotional periods could also be considered loss-making if considered in isolation and hence be considered subsidised by other consumers.

A consumer can also cross-subsidise themselves over time. This means that while over the lifetime of having the product the consumer is profitable there may be periods when they are loss-making for the firm and other periods when they are profitable. Cross-subsidisation over time is inherent in the design of products with promotional offers. Consumers in general are loss-making during the promotional period and profitable thereafter. Consumers exhibiting transactional behaviour may also cross-subsidise themselves over time, for example they may be loss-making during periods when they spend very little but profitable during periods when they spend more or revolve a balance for a couple of months.

As discussed above we are only concerned with cross-subsidisation where it impacts competition or is actively detrimental to consumer welfare. We do not consider that this cross-subsidisation gives rise to these concerns.

Potential impact of the Interchange Fee Regulation

The Interchange Fee Regulation70 caps consumer credit card interchange fees at 30 bps (0.3%) of transaction value. This means credit card firms’ revenue from interchange will fall from around 80bps71 to 30bps. This is a gross reduction of around 60% in interchange revenue across the industry, and equates to around a 5% to 10% reduction in overall revenue.

The firm by firm impact varies significantly, based on a firm’s reliance on interchange income. Firms that have a larger proportion of transacting-type consumers facing proportionally bigger losses in income and these consumers becoming less profitable or even loss-making.

The impact on the higher risk segment is likely to be minimal, reflecting the fact that interchange income accounts for a much smaller proportion of overall income in the higher risk segment. Indeed the changes may lead to more firms entering this segment as a way of diversifying their income streams and attracting more revolving behaviour.

Firm responses

Firms’ responses to the Interchange Fee Regulation can be broadly summarised as:

  • trying to reduce the number of transactors/increase the number of revolvers they attract, and/or
  • adjusting their commercial proposition to make transactors more profitable.

As set out in Chapter 3, to achieve either of these, firms can alter one or more of the product features:

  • Rewards − the reward segment is likely to be hardest hit, given that reward cards attract transacting-type behaviour and rewards have typically been funded by sharing the interchange income with the consumer.
    • Some firms have already intimated that they are likely to dilute or remove reward propositions with others likely to follow.
    • Firms with co-brand partners are likely to look for the co-brand partner to fund a larger proportion of the rewards.
    • Firms are also likely to explore other ways of rewarding consumers such as merchant funded offers, for example, statement credits when spending a certain amount at a particular retailer.
  • Annual fee A number of firms are considering introducing or increasing annual fees. In some cases, fees may be waived for consumers who reach certain spending thresholds or are paying a large amount of interest, or in the first year to attract new consumers.
  • Annual interest rate − The annual interest rate could be altered to increase revenue. A higher annual interest rate could compensate for lower interchange revenue. This potentially creates a mismatch in the product offering, with those exhibiting profitable behaviour finding the price going up. Some firms therefore are considering decreasing their interest rates to encourage revolving behaviour.
  • Other − Other potential changes include entry into unexplored market segments, for example the higher risk segment, which relies far less on interchange fees, or reducing the number of interest-free days.

Based on the discussions we have had with firms to date, they say they will respond by offering more products with a small or increased annual fee or diluting rewards schemes.

These potential developments do not give us direct cause for concern where appropriately communicated to consumers (whereas we would be concerned, for example, if the loss of interchange revenue led to increases in conditional charges which are less susceptible to competitive pressures). However, in a market where there are competitive pressures on firms we would expect to see continued attempts to gain business and minimise the impact of such changes on consumers. We will consider any changes to firms’ response to the cap on Interchange Fees as we progress this study.


62. FCA, Credit card market study terms of reference (2014) MS14/6.1

63. The proportion of outstanding balance that attracts interest income.

64. See paragraph 3.34.

65. Base curves can also be plotted to show how cost evolves over time. Typically, the acquisition cost for a cohort is incurred only in the first month. This cost is why PBT base curves typically start out negative since acquisition costs are incurred before revenue is generated. This can be seen in Figure 13 above and in the PBT base curves of other product types below. All other main sources of cost; default, funding and operational costs, are typically forecast as a constant cost over time. Whilst accurate costing is needed to correctly estimate profit levels costs estimated as constant over time do not affect the shape of the PBT base curve, though they do shift the curve up or down, which clearly affects overall product profitability.

66. As mentioned in Chapter 4, in the consumer survey we asked respondents whether they had a credit card that was designed for someone with ‘no credit or poor credit history’ which may or may not be a ‘low and grow’ card. Given that most cards of this type are ‘low and grow’ we refer to them as such throughout when referring to the consumer survey for simplicity.

67. While firms can accurately predict the behaviour of a cohort, it is far more difficult for a firm to predict an individual consumer’s behaviour. However, most firms are aware which of their products contain an element of cross-subsidisation.

68. See paragraph 3.39-3.44.

69. Ibid.

70. European Parliament & European Council, Regulation (EU) 2015/751 on interchange fees for card-based payment transactions (2015)

71. As noted in the Payment Systems Regulator’s CP14/1 (November 2014), interchange fees for consumer credit card transactions have ranged from 0.65% to 1.85% of transaction value, depending on the card payment system (MasterCard or Visa), the card type (e.g. standard or premium) and the transaction type (e.g. Chip & PIN, card not present, etc.). The average interchange fee on such cards has been around 0.8% of transaction value (80bps).