At its simplest, retail power is a physical logistics and marketing business. A retail energy provider (REP) contracts with customers to supply power and, in turn, contracts with a generator (or generates power from owned assets) to serve that obligation and lock in a margin. That said, assessing the credit risks associated with the retail power business has turned out to be a challenge. As power industry analysts used to evaluating the economics of the generating asset along the dispatch stack, it's difficult for us to appreciate the long-term sustainability of a business that is, for all practical matters, a marketing business that sees gross customer attrition of over 20% annually even for industry leaders (countered by new customers gained that mitigate net attrition).

S&P Global Ratings has been skeptical about the long-term prospects for this business, not in terms of its viability, but in terms of the sustainability of its margins. While Vistra Energy Corp. and NRG Energy Inc. have pointed at their past performance in dismissing any suggestions that market shares could come under threat, we think that as one the few avenues left for growth, the retail power business will experience increasing competition that could indeed erode margins. At the least, we think residential margins are at such levels that they should witness underpricing attacks. Yet, the retail power business has been an extremely profitable one for power companies. Despite our beliefs, the retail power business continues to thrive and currently underpins the performance of the independent power producers (IPPs).

A recent conversation helped us frame the credit risks of this business from a different perspective and sets the background for this commentary. At a meeting with an investor, one of the authors of this comment was asked whether he liked his cable provider. The analyst responded that he hated it. That response was followed by "how long have you had your cable provider?". The analyst responded that he had had the provider since 2004. "Well, how do you feel about the retail power business now?" came the rejoinder. As economic purists, we've been surprised by the stickiness that residential customers display and have been thinking of a reasonable response to that question.

Here, we take a deeper dive into the retail business. We discuss whether this business has unique attributes that require specific competencies only possessed by a few, or if the first mover advantage--and abnormal returns--is temporary and will ebb over time. We have largely focused on the residential retail business but have provided our views on commercial and industrial (C&I) business where relevant.

What is the retail market opportunity?

On Jan. 1, 2002, Texas (Electric Reliability Council of Texas [ERCOT]) deregulated its retail electricity market. The main goal of deregulating the market was to foster competition and give customers a choice for price, new products, innovative programs, diverse energy sources, and customer service as a differentiator. The market has grown since deregulation and currently comprises about 7 million customers. It is now largely a fully competitive market where 94% of customers have made a choice.

But ERCOT has some unique attributes that make it a front-running retail market. It is the best one for retailers to operate in for many different reasons, including:

• Retail providers, not regulated utilities, hold the relationship with the customer (send bills, collect payments, and manage the customer experience).
• There is a single set of market rules in ERCOT.
• The ERCOT market has been allowed to work unhindered by a regulated "default rate."
• It has a larger average mass customer annual usage in Texas (driven by air conditioner load)

Table 1

In contrast, only about 35%-40% of residential customers have switched from the incumbent utilities to a REP in most northeastern states, suggesting that the opportunity to gain market share in these market is substantial (see table 2). Yet, as we explain later, the ability to engage directly with the customer appears to be a meaningful advantage in the retail power business. However, in these Northeast markets, the electric distribution company owns the billing and servicing and the customer does not have an ongoing relationship with the REP, which is a significant disadvantage when it comes to customer acquisition and retention.

Table 2

Is the retail power business profitable?

The best way to evaluate the performance of a retailer is its unit margins (in $ per megawatt hour [/MWh]). However, that analysis is a bit of a challenge and perhaps not a practical approach in a market like ERCOT because there is no "average price" for comparison. That is, there is no power-to-choose (PTC) benchmark. Given that ERCOT is a completely competitive market, a retailer could have hundreds of different offers for new customers and even more for current customers at any time. Even if we limit our focus to only the nearly 100 different 12-month term offers currently on PTC for a single Houston zip code, the price points (at a fixed 1,000 kilowatt hour [kWh] usage level) range widely. This makes comparison to the "average market price" problematic.

Instead of doing a granular top-down analysis, we approximated a simpler bottoms-up approach. To illustrate the profitability, we used Vistra's 2016 and 2017 publicly disclosed retail business EBITDA. We chose Vistra over NRG only because it offered a relatively cleaner "pure play." Before its acquisition of Dynegy, Vistra's retail business was confined to ERCOT and was largely in the residential segment (see table 3).

Table 3

Our calculation of NRG's margins was also in a similar range, albeit a bit lower. While NRG has been able to increase its net customer counts because of its presence in more markets, we suspect that the non-ERCOT markets are not as profitable because the REP does not have a direct relationship with the customer and needs to advertise more. Also, we note that these calculations are illustrative to present the range that the residential retail business earns. In practice, term contracts will have lower margins while month-to-month contracts will have a higher range. Pricing discounts, concessions, and offers make this complicated, but the bottom line is that the retail business has been profitable for major brands and also generates a significant amount of free cash flow.

So the retail business is currently profitable (at least for the major brands) but is it sustainable?

What kind of attrition does retail power witness?

Even as net attrition levels (current customer count, less losses, plus new customers) have been largely steady for the larger retailers, gross attrition levels (current customer count, less losses) are typically high. As per ERCOT's retail transaction reports, the competitive ERCOT residential market experienced about 1.25 million switch and about 2.9 million-3.0 million customer move-out/move-in transactions in 2018 for a total of 4.15 million-4.25 million transactions on a base of about 7.25-7.5 million residential customer meters.

However, the details provided by ERCOT are not a sufficient source of information to assess what the market or retailers' "real" attrition performance is. In particular, move-outs and move-ins are not necessarily customers leaving a retailer--they could also be transfers of service or people moving in and out of state. Also, the document includes merger and acquisition (M&A) customer book transactions that affects the results as well--we estimate that there were probably as many as 250,000 meters changing hands this way in 2018. However, what this data demonstrates is that most of the population has switched and the large volume of transactions gives a sense of the magnitude of activity in the market and the seasonality that the business experiences. As an example, we can see that most move/switch activity occurs from May to August.

We estimate gross attrition rates for both Vistra and NRG of about 30%-33%, including transfers of service. Excluding the transfers of service, we estimate the gross attrition rate would be lower at about 22%-24%. These levels of portfolio churn appear to be typical in the retail business but are an aspect that is new to power sector analysts. We will monitor gross attrition as we consider any changes in those levels as a lead indicator of a weakening portfolio. However, net attrition rates at the major brands still tend to be lower than 1%. We think the ability to maintain low net churn rates is impressive given there are typically 200 or more offerings within a typical urban zip code, with the ability to switch daily due to smart meters.

The primary driver of attrition is customers moving. In the East, customers who move are required to return to the utility for at least one month. In Texas, if customers are moving from a competitive territory to a regulated territory, they cannot keep service. In addition, in deregulated markets, like Texas, market participants are given plenty of alternatives to choose from. Retailers have to work hard to craft offers that make customers more likely to stay. They often predict churn risk and manage customer lifecycle events by leveraging attributes like rewards (miles) or secondary products (like natural gas delivery or home security) to extend customer lifetimes.

We think the modestly declining net customer count for Vistra can be explained by the fact that the retail coverage of all competitively shopping customers in ERCOT is now very high (94% have exercised a choice). As a result, there are fewer options to expand the retail business within the state. While Vistra's net attrition rate is still fairly low (we estimate 0.3%-0.4%), we think the company has to move out of ERCOT to increase its net residential customer count. In contrast, NRG's net attrition rate is negative (i.e it is seeing overall portfolio growth) largely because it is in markets where competition is still limited, with the incumbent utility serving about 65% of residential load. NRG has been successful in taking market share from the incumbent utility to increase its portfolio. Consequently, we see the need for Vistra to expand into other regional markets not so much as a choice but a necessity (even though rules aren't as clear or efficient as ERCOT). Because of its higher residential retail proportion, its largely ERCOT operations, and fewer brands, we think Vistra's retail margins are higher than NRG's on a per-unit basis. However, its ability to maintain margins when it expands into the eastern markets in a meaningful way remains to be seen.

Can retailers keep customers?

We estimate that the major brands achieved average retail residential margins of nearly $35-$40/MWh in ERCOT between 2016 and 2018. In comparison, mid-tier players made about $15-$25/MWh and flanker/attacker brands made only about $5-$20/MWh. Stated differently, major brands have prices that are about 25% higher than the "average PTC" price in ERCOT (averaging out the prices offered by all participants in a region across similar product offerings). However, the majority of market share is within the top 10 parent companies, with a number of companies having a few brands and multiple offerings within a region.

Unlike the media industry, where streaming content offered by service providers can be highly customized and differentiated, we consider electricity as a commodity. Our view is that customers care most about rates and reliability. As a result, our view is that as long as a competitor offers rates that are meaningfully lower, major brands would lose market share. That has not happened in a significant way and customer accounts have proven to be stickier than we expected. The question remains whether this will continue to be the case.

All customers don't see value the same way. Some customers want airline miles, some want a nest thermostat, and some want no frills electricity or are willing to interact only electronically with their provider. Let's take a very transparent channel and look at the situation. Right now, on PTC in Texas, a number of mid-tier brands have no less than four different term offers with rate of eight cents/kWh or less (at the 1,000 kWh usage level). At any given time, mid-tier and flanker/attacker brands will have attractive acquisition offers available somewhere that yield margins considerably below the major brands overall portfolio. This is the nature of the retail business. However, we discovered that the following behavioral factors also appear to affect customer decisions and are not merely based on prices:

• Nearly 33% of enrollment is now online compared with 5% as recently as 2010.
• Acquisition price points do not reflect average portfolio price points.
• Products, prices, and offers vary by sales channel, so even the acquisition prices seen in one channel are not necessarily what most new customers pay.
• Customers do not buy electricity like a commodity. They make purchase decisions on many different factors. It's the same reason there are five to 10 different brands of bottled water at most stores, each at a different price point. And the prices for the same bottle are different in a vending machine, in a convenience store, at the airport, and at Walmart.

The value propositions for meeting different customer needs can be broadly bucketed into full-service energy provider, low-cost service, and environmentally sustainable service (see table 4).

Table 4

We are somewhat puzzled by this customer behavior but it appears that branding, smart data-driven marketing, and pricing structures help sustain margins. Electricity is a product that typical customers don't want to think about as long as there are no price spikes and the bills stay predictable. So far, major brands have demonstrated that they can construct and present value propositions that customers are willing to pay for, or alternatively, customers do not care to educate themselves enough about better alternatives. This is not to imply that customers are not price sensitive, but rather, they migrate to offers that they value at the time and place of purchase.

This dynamic plays into the hands of the incumbents and major brands. The ability to extend customer terms with add-on products and new offerings is key to a retailer's success. This is why many of them offer a multi-brand strategy. The objective is to occupy "shelf space" in the market and capture customer mindshare along key value dimensions. Smart meters and emerging technologies provide retailers with lead indicators of a change in a customer's monthly electric bill that they can mitigate with promotional offers before a sticker shock is felt by the customer. That's why retailers work hard to present offers with compelling attributes to the customer at the right place with the right price at the right time.

What are S&P Global Ratings' views on the C&I business?

From the perspective of projecting load, we see the C&I business as more predictable because these customers tend to operate at consistently high load factors. Engagement touchpoints with C&I customers also tend to be much lower than those with residential customers, which tend to be numerous. There are also fewer, if any, advocacy or regulatory entities observing the C&I market, in contrast to residential markets, which are under constant scrutiny. As a result, many companies--such as Calpine and Constellation--prefer the C&I retail segment. The downside is that these markets comprise sophisticated buyers who invariably open a request-for-proposals each time a contract comes to term. While previous and existing relationships are important for contract extensions, the C&I business tends to be extremely competitive and has relatively lower margins than the residential segment, typically in the range of 50 cents to $1/MWh (industrial) to $3/MWh-$5/MWh (commercial). However, small-scale commercial can achieve margins of about $8/MWh.

So where do we see risks?

The credit and market risks associated with retail power operations are obvious and relate to the ability of the retailer to accurately model the expected load and hedge its full requirements load following obligation. Many retailers have asserted that their modeling skills of projected load, and their ability to hedge it, have become more sophisticated over time (see table 5). However, our view is that since the projected load depends on a complex set of factors, including customer preferences, structural changes in secular load growth (energy efficiency, behind the meter generation etc.), and the weather, predicting this load, and hedging it, is not always accurate, and often not possible.

Table 5

An unexpectedly high bill is the major reason for customers to switch and retailers proactively try to manage customers' bills. There is no doubt that commodity cost spikes (as expected over the next two summers in ERCOT) could be an issue for retailers. In a rising power price environment, retailers can lose some margin and customer counts as they try to pass through costs. However, we think the retail players without generation will be more vulnerable because the cost of hedging will likely spike. So integrated retailers could still increase customer count and market share (if smaller retailers exit the market), which overall could still have a positive EBITDA impact. In a falling wholesale power price environment, retail margin could rise but attrition risk would increase.

We've seen instances of this on a number of occasions, such as NRG's under-performance in August 2011 during an ERCOT heat wave, Constellation's miss during February 2012 during a cold snap in ERCOT, and, more prominently, when FirstEnergy Solutions significantly underperformed (about $200 million impact) during the polar vortex. Vistra and NRG also reported weaknesses in that year with Vistra's retail margins dropping to about $680 million in 2014 from about $810 million the previous year. Similarly, NRG's unit margins also declined that year, even as aggregate EBITDA rose because of the acquisition of Dominion's 400,000 retail customers.

We note that all major brands maintain a balanced mix of term and month-to-month accounts. While pricing of the variable/month-to-month customers can be changed as the market changes, margins on these accounts are the ones that come under pressure during scarcity periods because wholesale pressure rises (situations like the polar vortex, the bomb cyclone etc.).

What are other major risks for retail providers?

Regulatory and consumer advocacy remains the principle risk for retailers in the residential segment given the high margins earned on customers today. Even in ERCOT, the most supportive market for retailers, we note some lingering discussions on retail markets from the PUC. The New York Public Service Commission (PSC) has been the most vocal in its opposition of mass market retail. In 2017, it opened an investigation (Cases 15-M-0127 et al.) to consider removing mass market retailing altogether in favor of other alternatives.

Briefly, the New York staff's eventual recommendations were that the energy supply companies (or ESCOs) not be allowed to use the utilities' systems to distribute their commodity unless the retailer provided a guarantee that each customer's total electric or gas bills would be lower than, or no greater than, that charged by the utility for delivery and commodity service over the calendar year. That said, a retailer may provide customers energy generated through 100% renewable resources that are delivered to and consumed in New York at a premium price to utility service. Specifically, in order to ensure mass-market customers are receiving value from the retail energy markets, the PSC has ordered that the retailers meet one of the following two provisions:

• Guarantee savings compared with what the customer would have paid as a full service utility customer, and
• Provide at least 30% renewable electricity.

We believe that the risk here is one of transparency and asymmetric information. There's a huge disparity in knowledge, which could make it easy for residential customers to be taken advantage of. The PSC is particularly concerned about high-pressure sales situations, deceptive marketing, slamming, and lack of expected savings. As a result, the PSC effectively placed restrictions on mass retail choice by substituting its judgment for that of consumers in determining which energy products offer value. Given that many major brands are extremely profitable, there is a risk that this view can gain traction.

We also see risks that local communities could band together or formal community choice aggregation could compete for residential customers, eroding the current high margins for mass retailers. For example, in New York, the staff also recommend that aggregation of customers be allowed through either a not for profit, municipal entity, or Community Energy Aggregation (CEA) using a professional energy buyer acting as a fiduciary to the CEA and independent of the retailer.

Are there any tail risks that we are concerned with?

We even think of hypothetical scenarios where Google or Amazon would consider getting into retail power as an extension of their current businesses to provide complete home solutions. Lately though, our view on Google has shifted somewhat--we believe it is not interested in taking oncommodity exposure and likely more interested in enabling the use of its home devices by partnering with retailers, rather than competing with them. However, Amazon poses a totally different risk, in our view. Through its forays into grocery (Whole Foods) and online pharmacy businesses, Amazon has shown a willingness (and ability) to enter mass retail and the residential channel offers it an incremental avenue. On the other hand, rather than building its own retail business, it could also acquire one of the existing businesses, offering IPPs an opportunity to monetize their operations at a premium.

Bringing In The Verdict

From the perspective of debt financing, we're still skeptical of an asset-light retail strategy, likening it to picking up pennies in front of steam-rollers. It is not a question of if, but when, you will get crushed. However, we cannot ignore the fact that the retail power business for major integrated companies has clearly outperformed our expectations. In particular, we view an asset-backed retail business as beneficial to the wholesale generation platform because it offers it a natural hedge (helps insulate market risk for 18-months--the typical length of a retail contract). Both Vistra and NRG estimate that transaction and collateral efficiencies/synergies between the wholesale generation and retail power business is worth about $3/MWh, which can be as much as C&I margins.

Moreover, from a timing perspective, retail tends to exhibit stronger performance in the second and fourth quarters; a seasonal balance to wholesale generation, where best performance generally occurs in the first and third quarters. From a credit perspective, the profitability of the retail power and wholesale generation businesses moves up and down in counter cycles (see chart). Capital charges that typical retail businesses use--including the cost of working capital, credit facilities, and contingent collateral, as well as equity costs required to cover risk capital requirements--increase roughly in proportion to commodity prices. With high power prices, capital charges are also high and cut into gross margins. Yet customers are less inclined to lock in prices at these levels. As a result, at elevated prices, we expect fixed-price sales to fall, reducing total capital requirements and increasing average margins on existing retail volumes. At low power prices, capital charges decline. Although customer migration ensues, gross margins for retail volumes rise due to increasing headroom between locked-in retail prices and wholesale prices. Thus, although the generation business' profitability declines when prices are low, the retail business' profitability improves, and vice versa.

Counter-Cyclicality Impact On C&I Contracts Margins

We note that while no company requires its retail supply business to source its hedges from the company's wholesale business, they are sometimes motivated to do so, and we expect a natural evolution toward wholesale hedges greater than 50%, reducing the need for collateral support for the retail operations. Lenders now also appear to be more comfortable with an "asset-heavy" retail power strategy. Investor appetite for these retail power offsets also appears to be high, so many companies are now seeking to expand margins with further acquisitions.

Up until now, we saw IPPs and diversified power companies expanding retail offerings in an effort to provide complementary revenue streams to their otherwise volatile wholesale businesses. However, in the context of disruptive renewables and storage, the retail power business becomes both a survival strategy and the next battleground. With the grid expected to become much smarter over the next two decades, getting inside a customer's premises could mean opening up an entire avenue of growth through sales of everything from thermostats and smart meters at one end to PV solar systems and eventually battery solutions at the other. In this way, retail business provides a great defense against renewable penetration.

With all this said, the question then becomes what level of debt can retail power, or more specifically, integrated power companies, support? We think that depends on the sustainability of margins and the cash flow conversion of the company. We will respond to that question in a future commentary. Stay tuned.