Reference pricing in the presence of pseudo-generics

Another strand of the (theoretical) literature focuses on entry-related issues, insofar as the decision to produce pseudo-generics may deter or delay generic entry (Hollis 2003; Kong and Seldon 2004; Reiffen and Ward 2007; Granier and Trinquard 2012). Empirically, the evidence (broadly) appears to confirm these anticompetitive concerns (Hollis 2002, 2003; Hollis and Liang 2007; Aitken et al. 2013; Grootendorst 2007; Berndt et al. 2007a, b; Appelt 2010); Kamien and Zang (1999) are among the few examples of authors finding pseudo-generics to have procompetitive effects. We are particularly interested in understanding competition between producers after generic entry; therefore, our paper explicitly leaves aside the entry decision (pseudo-generic and/or generic) into the market.

Previous literature on reimbursement schemes for health care providers includes, among others, Ellis and McGuire (1986); Rickman and McGuire (1999) and more recently Bardey et al. (2012). In the pharmaceutical market, the classic reference is Mossialos et al. (2004).

Several criteria could be used to cluster pharmaceuticals: chemical, pharmacological or therapeutic (Miraldo 2009).

For a detailed review of reference pricing, see López-Casasnovas and Puig-Junoy (2000).

Note that reference pricing is only different from FPR when generic competition becomes possible, otherwise only one branded pharmaceutical exists and it would automatically be the reference pharmaceutical.

We thank a referee for suggesting this interpretation.

Without a pseudo-generic, such efforts lead to a reduction in the branded price and an increase in the generic price; the latter is the reference price on which reimbursement is based, which thus leads to an increase in government expenditure. With a pseudo-generic, such efforts lead to branded price reductions and the generic price remains unchanged, which thus implies no changes in government expenditure.

Typically, horizontal differentiation is justified because consumers have heterogeneous preferences (different tastes) or, in the standard Hotelling (1929) model, because they are physically located at different distances from the location of the firm they wish to purchase from. In the pharmaceutical market, Brekke et al. (2007) suggest that consumers support ’mismatch costs’ between their ideal treatment—given by their location along the Hotelling interval— and the pharmaceutical variety they actually consume, because, for instance, the latter may have undesirable side-effects or contraindications which reduce a consumer’s utility.

This is plausible, because the branded pharmaceutical, during the patent protection period, will have already established a reputation for its treatment effectiveness.

We assume that the reference pharmaceutical is the generic and that the copayment rate, \(\theta ,\) is similar across schemes. In equilibrium, we find that indeed the generic has the lowest price and would thus be the ideal candidate for the “reference” pharmaceutical.

For instance, under such a more general formulation, Eq. (3) would become \(\hat{p}_{k}^{R}=p_{k}^{R}-\left( 1-\theta \right) \left[ \alpha p_{pg}^{R}+\left( 1-\alpha \right) p_{g}^{R}\right] , k=b,pg,g, \alpha \in \left( 0,1\right) .\)

In particular, under such a general formulation in which the reference price is a weighted average of the pseudo-generic and generic prices, the location of the indifferent consumers does not change (see “The market with no pseudo-generics” and “Reference pricing” sections in Appendix) because the reference price impacts on all effective prices in the same way. That is, given headline prices, a change in the reference price formulation increases or decreases all effective prices (branded, pseudo-generic and generic) in the same amount. Therefore, equilibrium (headline) prices, quantities and firms’ profits are not affected by the reference price formulation (thus, Propositions 1, 2 and 3 hold under this more general formulation). The latter only affects CS and government expenditure. In particular, this more general reference price formulation [(compared to our assumption—see Eq. (3)] would lead to an increase in the reference price, a decrease in all effective prices and, thus, an increase in CS; but this is exactly offset (because quantities are the same) by an increase in government expenditure. Therefore, welfare results are unchanged: both social welfare (SW) and welfare are exactly the same regardless of the reference price formulation (thus, Propositions 4 and 5 also hold under this more general formulation).

Graphically, this is equivalent to saying that the indifferent consumer under RP is located to the right of that represented in Fig. 1.

For a given headline price vector, note that the effective price of the generic under FPR and RP is the same, but the effective price of the branded drug is higher under RP as long as \(p_{b}>p_{g}.\) See Eqs. (2) and (3).

As in the previous section, for both varieties to be sold in equilibrium, restrictions (32) and (33) must be imposed in the parameters. See “The market with no pseudo-generics” section in Appendix for more details.

We assume that this location is fixed, i.e., it is not a decision variable for the BP. Location, in our model is a metaphor for the degree of differentiation perceived by consumers. The pseudo-generic and the generic differ in observable characteristics such as the physical shape of the pharmaceutical, its colour, the package size and shape or the producer identity (often BPs license the production of pseudo-generics to other firms, rather than producing it themselves). Producers can choose these variables, of course (except, maybe, their identity). However, the resulting differentiation effect in consumers’ eyes is difficult, if at all possible, to manipulate strategically with any degree of precision. This is particularly so for the branded producer, who in this type of market is typically a first-mover (pseudo-generics enter the market early, generally before other generics do). Therefore, he has no ability to control the extent to which generics will physically resemble his own product.

This is discussed at more length in Rodrigues et al. (2014).

Equilibrium demand can be seen in Eq. (42).

Again, three conditions must hold for the three varieties of the product to be sold in equilibrium and for these to be equilibrium prices: Eqs. (50), (51) and (52). See “The market with a pseudo-generic” section in Appendix for more details.

We thank an anonymous referee for pointing this out.

We thank an anonymous referee for suggesting this line of analysis.

For a graphical representation of these restrictions, see Rodrigues et al. (2014).

In equilibrium, the marginal consumer is located at \(c_{b.g}^{R}=\frac{t+\alpha -\beta }{3t}.\)

In equilibrium, \(c_{g.pg}^{F}=\frac{f_{BP}}{6}+\frac{1}{3}\) and \(c_{b.pg}^{r,F}=\frac{\left( 1+f_{BP}\right) t-\left( \beta -\gamma \right) }{2t}.\)

All these restrictions are similar to those of Rodrigues et al. (2014), where their graphical representation can also be found.

In equilibrium, \(c_{g.pg}^{R}=\frac{f_{BP}}{6}+\frac{1}{3}\) and \(c_{b.pg}^{r,R}=\frac{\left( 1+f_{BP}\right) t-\left( \beta -\gamma \right) }{2t}.\)