On the Economics of E-pharmacies: Potential Issues for Anti-trust Analysis

E-pharmacies are a relatively new phenomenon and are, in contrast with offline sales of drugs and medicinal formulations, a small segment of the market in most countries.¹ Recently, PharmEasy has acquired Medlife, becoming the largest e-pharmacy in India and showing the tendency towards consolidation in the market (Ahmad 2021). Not only this, but the ongoing pandemic worldwide has also increased the relevance of and consumer dependence on e-pharmacies much more than in the pre-pandemic scenario.² Like most online entities, these e-commerce ventures have become an increasingly important destination for buying palliative medicines as well as nutraceuticals like vitamin supplements.² At the present moment, an anti-trust lens on these e-pharmacies should be of the following kind: what are the likely competition concerns that these entities raise in the near future? Even if e-pharmacies are a marginal segment at present,² given clear indications of surging demand, should the competition regulators take any steps to ensure that competition is not distorted in the future in this market?

To address this question, we need to define the market first. The legal framework in many countries, such as India, prevents sales by e-pharmacies outside the national geographical boundary.³ Within a country, an e-pharmacy’s actions are likely to impact (i) offline pharmacies, (ii) consumers, and (iii) future entrants in the market. Along each of these dimensions, there are potential issues for anti-trust authorities as well as for public health policy. We first define e-pharmacy formally before evaluating the competition issues. 

Typically, the term “e-pharmacy” is ascribed to “the business of distribution or sale, stock, exhibit or offer for sale of drugs through a web portal or any other electronic mode” (Musyuni 2020). In most countries, the legal framework exists to prevent advertisements of medicines, though during the ongoing pandemic, there have been instances of violation of this condition.⁴ By the very definition, e-pharmacies have the capability to stock prescription as well as over-the-counter drugs. The very fact that a consumer does not have to physically identify themself to a pharmacist for online transactions provides perverse incentives to self-medicate, particularly hidden medical conditions that the consumer is embarrassed to reveal to any medical practitioner/pharmacist while ignoring potential side effects of the medicines (Glover-Thomas and Fanning 2010). Any kind of self-medication and self-determined line of treatment (even with vitamins) has its dangers and e-pharmacies facilitate this problem of “medicalisation” (Glover-Thomas and Fanning 2010). 

At the same time, it cannot be denied that these entities introduce an alternative retail channel for selling and distributing drugs. During uncertain times like the COVID-19 pandemic, e-pharmacies provide an alternative channel for consumers, where the purchase can be done from the safe confines of one’s own residence (Mathew 2020; Ahmad 2020). Given this resurgence, we believe that a detailed perusal of potential anti-trust issues with e-pharmacies is the need of the hour. 

Anti-trust Issues with E-pharmacies

When it comes to competition issues, we need to understand the competitive forces that are enjoined upon e-pharmacies at present, and what changes we are likely to see in the near future. Some countries like China have already seen a few dominant players coming up in this space, such as Ali Health, JD Pharmacy and Tmall Pharmacy (Li and Huang 2019). Forecasts for countries like India indicate a sizeable part of the market in the near future with projected share of $3,657 million by 2022 (Frost and Sullivan Report, 2019).⁵

E-pharmacies and Offline Retail Drug Stores

It is a fact that along one dimension, e-pharmacies present an additional competitive constraint on offline drug stores by presenting consumers an alternative purchase destination. A number of anti-trust cases exist in many countries, such as India,⁶ regarding the conduct of offline pharmacies. Collusion in pricing, using their trade associations for pressuring drug companies to price in their favour,⁷ fixing trade margins, constraints on distribution of drugs, prohibition on retailers from giving discounts to customers, constraint on the appointment of stockists or wholesaler, tying arrangement between doctors and pharmaceutical companies, resale price maintenance and exclusive supply agreements are some examples (Jain 2021). Note, however, that allegations of collusion among pharmaceutical companies (the recent case in the United States (US) for generic drugs (Keown 2020, for instance) are matters of anti-trust deliberations, but are not directly relevant for e-pharmacies. Nonetheless, the nature of ownership of e-pharmacies and their financing as well as potential bundling of offline and online services and, therefore, collusion among larger offline retail outlets with online counterparts remain open questions. In order to unbundle these issues, we dwell on each of them in some detail. 

Online Pricing Mechanism by E-pharmacies

This question brings us head-on with the appropriate framework to study pricing by e-commerce firms. The current literature and debates in the anti-trust space use the terminology of “platform pricing” to analyse issues of pricing, in particular, online discounts. The nascent literature on e-pharmacies is no exception: Li and Huang (2019) characterise e-pharmacies as two-sided platforms, which solve the matching problem for online buyers of medicine on the one side and stockists of medicines on the other side. The modern understanding of two-sided platform markets is that they need to differentially price the two sides of the markets (keeping in mind the different elasticities in terms of utilities the two sides derive from platform transactions), which often results in zero or subsidised/discounted prices for online buyers. These discounts solve the problem of network externalities that mark transactions on online platforms and introduce efficiency in transactions. Li and Huang (2019) show that the e-pharmacy (as a profit-maximising platform) will choose to subsidise consumers only if the net income from consumers is lower than the total revenue of drug retailers and platforms in each transaction.

However, this blanket treatment of all e-pharmacies as platforms engaging in discounted pricing ignores the underlying heterogeneity of these entities in terms of their business model and their underlying financing. Some of these pharmacies own/control the stock of medicines in the inventory model, whereas some others are in the marketplace model or are franchisees.⁸ 

Nature of Ownership of E-pharmacies

Some e-pharmacies have deeper pockets as they are financed and aided by established firms in other markets. For example, in August 2019, Reliance Retail (a leading industrialist group in India) had acquired majority stakes in Netmeds which is a leading online pharmacy in India (Sharma and Peermohamed 2020). Following suit, now Amazon and Tata Health are eyeing takeovers Apollo Pharmacy and 1mg, respectively (BusinessToday.In 2020; Peermohamed and Sharma 2020). Parsheera et al (2017) have pointed out towards high concentration of e-pharmacies due to lack of incentives to compete on account of a common set of investors owning significant shares of different online entities. Efficiency in operations for e-tailers, the authors note, is an important criterion and their success should not be driven by “abusive or exploitative practices.” Multiple other competition concerns, such as the exchange of sensitive information, the creation of monopsonies through price-control strategies like joint buying and selling and price parity clauses (Eccles 2015; Colangelo and Zeno-Zencovich 2016), exclusivity and exclusionary practices (see Dontoglou 2002) might become important for anti-trust agencies in the near future.

Potential Bundling of Online and Offline Services

E-pharmacies sell many offline services like diagnostics and add-on online services like doctor’s consultations.⁹ The bundling of these offline diagnostic services with online sale of medicines might lead to the creation of online backyards for dominant offline drug stores which also have a presence in other medical services. By selling a bundled product, with the convenience of online purchases, a dominant offline drugstore can engage in a form of non-linear pricing that challenges the notion of platform-induced discounts and competitive constraints on offline stores. 

Existing E-pharmacies and Potential Entrants 

The direct concern regarding the nature of online pharmacies and potential competition is that of the contestability of this market. While our earlier discussion discusses the overlap between online and offline competition, some competition concerns remain even if we treat these two spaces as two different markets. Fixed costs of operation and operational details of offline and online pharmacies are not similar. Existing laws in most countries enjoin upon online entities multiple restrictions that are absent for other e-commerce entities.¹⁰ However, for the consumer, both are spaces from which to purchase medicines.

To address this, we studied the online market for a limited basket of medicines in India over late 2019 till recently. We also noted the local offline prices for this medicine basket¹¹ in the Delhi region to contrast online prices. There is evidence of significant entry as well as exit in the online market. Between December 2019 and November 2020, six new e-pharmacies entered the market while four stopped their operations. 

For most online markets worldwide, there seems to be a similar pattern.¹² There is some fringe competition from smaller pharmacies, like Publix in the US, whose presence is limited to a few locations alone at present. The deeper the pockets of existing e-pharmacies, the less likely a potential entrant will be able to stay in the market post-entry. This raises the question as to whether contestability is sufficient to address distortion in potential competition due to concentration among existing players in the market. Entry, in such market configurations, lacks the effective monitoring power to bring in competitive forces as simultaneous exit reduces its threat. This also implies that consumers purchasing medicines online will finally be left at the mercy of a handful of players with the potential to abuse their dominant position even within the limited space of only online pharmacies. We now bring up the last dimension of e-pharmacies and consumers in the next section.   

E-pharmacies and Consumers

One major issue in this regard is about transparency of information and over-medicalisation, which might be more of a public health issue rather than one of competition. However, pricing of medicines and their artificial deflation using online discounts, including coupons, cashbacks and loyalty rewards, can potentially be anti-competitive. Many leading e-pharmacies provide a multiplicity of these discounts. Take OptumRx Inc for instance. It has announced on its website: “Save up to 80% on prescriptions. Optum Perks is accepted at over 64,000 pharmacies nationwide. No personal information or insurance required. Get your coupon now.¹³” Similarly, Giant Eagle Inc also in the US offers e-coupons on specific medicines.¹⁴ 

For this purpose, we need estimates of these discounts in relation to the underlying gross price of the medicine. In what follows, we provide a framework to model and empirically estimate the extent of online discounts for e-pharmacies.

Potential Framework to Analyse E-pharmacy Discounts: Application of Hedonic Pricing

The concept of hedonic pricing is based on the theory that the utility a consumer derives from consuming a product is a function of these innate characteristics. Initiated by Rosen (1974), this theory has found multiple applications in the arena of non-market-based valuation models in real estate (Witte et al 1979), environmental goods (Garrod and Willis 1992), consumer goods, such as wines (Combris et al 1997) as well as the pricing of traditional medicine (Saha and Vasuprada 2020). Rather than depending on the forces of demand and supply, the price of the good is estimated from those of its component characteristics. In the case of modern biomedicines sold online, estimating the actual demand and supply is not easy. Many of the e-pharmacies have an internet presence, but are not functional; in the case of our sample based in India, we found only 20 active e-pharmacies from among a total of 50 between 18 to 30 December 2019. Though e-pharmacy markets are nascent worldwide, the characteristics of bio-medicines they sell are well-defined. Most drugs sold commercially are a function of the type of packaging (tablets/syrups/injections), the strength of the dosage of the underlying salt in the medicine (insert footnote as example and the non-linearity in dosage) and specifically for tablets, its type depending on strength.¹⁵  

We propose a hedonic regression of the following form to predict the posted online price of the ith formulation

where log P_i  is the logarithm of online posted price, is a white noise error term, dosage captures the amount of salt present in the formulation, dosageᵢ² captures potential non-linearity in pricing, tabletstrengthdummyᵢ measures whether or not a tablet formulation is of high or low strength, packagingdummyᵢ specifies the form of the medicine: tablet/syrup/injection and onlinecontrolsᵢ include (i) the adjusted discount d_i given by the e-pharmacy defined as the difference between gross and discounted price normalised by the discounted price: d_i/P^d_i=(P_i- P^d_i)/ P^d_i and (ii) a dummy variable to indicate the presence or absence of coupons.  

Why do we not use the discounted online price as the dependent variable? The manner in which coupons are activated and used for a given purchase is not announced for any transaction, hence it is not possible to know the actual price that a consumer pays online. Data is available on the posted price, the presence or absence of coupons and the amount of discount declared by the e-pharmacy per formulation as well as its hedonistic characteristics. By using the undiscounted price as the independent variable, it is possible to check the nature of discounts that are passed on to the consumer. We use a measure of actual discount given by the e-pharmacy (d_i/P^d_i for the ith formulation), but also create a counterfactual discount using the online price as the posted price (for the ith formulation, it is c_i/ P^d_i= (P^o_i-P_i)/ P^d_i), where Poi is the offline undiscounted price of the formulation. 

The extent to which an e-pharmacy engages in platform pricing and provides discounts to consumers can be studied in this framework by comparing different regression specifications: one with d_i/P^d_i  as the discount measure and another with c_i/ P^d_i  as the measure of the discount. A comparison of the coefficients of these discount measures would reveal if the discount passed on to the consumer by the e-pharmacy is only notional and is used to strategically raise the undiscounted price of the medicine online. Any differential impact of these two types of discounts would also indicate the potential for distorted competition between online and offline e-pharmacies, strengthening the concerns that we have raised earlier. 

Summary of Findings

We implement our methodology using data on eight salts (which belong to the NSAID [Non-steroidal Anti-inflammatory Drugs] class) under Sections 2.1 and 2.2 of the National List of Essential Medicines for India (these are mostly analgesics and pain-relief formulations) sold online as well as offline.16 For the 178 formulations that result from these eight salts,¹⁷ not all give discounts. Among the 11 out of the 20 e-pharmacies that do give up-front discounts, the quantum has a large variation.¹⁸ Note that none of formulations we study have any patents associated with them, hence we have not controlled for their patent status in our causal analysis.¹⁹ The hedonic regression indicates that while d_i/ P^d_i  does not have a statistically significant effect on the gross price but the counterfactual discount  c_i/P^d_i  has a positive impact on the undiscounted online price. This is clearly indicative of strategic discounting: higher the offline posted price, in order to show a discount to the online customer, the offline posted price is raised. In fact, we find that either e-pharmacies do not provide any discounts or if they do, then there is significant evidence that the online undiscounted price is higher than the offline posted price. This is the case for one-third of our sample, raising concerns about the true nature of discounts that the e-pharmacies are passing on to consumers. There is minimal evidence of dynamic pricing online as prices online were mostly steady during our survey. We find that our model (all four specifications of the regression model in Table 3 in the Supplementary Material) predicts the online posted prices well. The adjusted R-square of all four specifications are equal to or above 0.9. Further, the plot of residuals against the fitted values of the dependent variable (log of undiscounted prices of medicines) is random.²⁰

Discussion

The online market for e-pharmacies, though nascent in many countries, is highly concentrated. There is evidence of significant entry, but that goes hand in hand with simultaneous exit, leaving only a few entities to serve customers online. Mergers and co-ownership of e-pharmacies by a small number of financiers will limit effective competition. Strategic discounting and the problem of excess medicalisation indicate the need for more transparency in pricing, supply-chain details and regulation of prescription requirements for e-pharmacies. As mentioned earlier, increasing dependence of consumers on e-pharmacies due to the ongoing COVID-19 pandemic indicates this market will become bigger in the near future. Effective regulation for e-pharmacies will need intervention from public health authorities and anti-trust intervention, particularly for mergers.²¹