The very concept of rational framing effect deserves a thorough analysis. The global rationality of economic man as Simon (Reference Simon1955) puts it (or absolute, Olympian, economic rationality) is definitely not the one contained in the “rational framing effect” phrase used by Bermúdez. The type of rationality in Bermúdez's paper is what in Simon's work is dubbed (Reference Simon, McGuire and Radner1972, Reference Simon and Latsis1976) a procedural rationality. In this view, when a decision is procedurally rational it means that it is the outcome of an appropriate deliberation, or as dual-system proponents would say reasoned and effortful process (e.g., Stanovich & West, Reference Stanovich and West2000, see also Stanovich, West, & Toplak, Reference Stanovich, West and Toplak2011). The stress is put on the choice of the proper reasoning rules regardless of the quality of data the decision was taken upon. Hence, the economically irrational decision, which was effortfully reasoned, should be considered procedurally rational. What follows, increase in the cognitive effort should enhance the procedural rationality and in turn should reduce the framing effect (e.g., McElroy & Seta, Reference McElroy and Seta2003; Simon, Fagley, & Halleran, Reference Simon, Fagley and Halleran2004). However, when the Asian disease scenario was labelled as “medical” as opposed to “statistical,” the framing effect was increased although the decision took longer and was more effortful (Igou & Bless, Reference Igou and Bless2007). This example shows that procedurally rational process may produce the biased outcome when the very core of the problem is wrongly understood because of its biased formulation.

The framing effect has been described in the prospect theory (Kahneman & Tversky, Reference Kahneman and Tversky1979). In this framework, a decision is a sequential process in which the methods of achieving a goal are elicited, then edited and finally evaluated. Each of the stages may be judged against some rationality criteria and found to be faulty in many ways. Prospect theory separates the editing phase from the evaluation phase. In the editing phase the representation of a problem is created: The prospects are coded against a reference point (i.e., the outcome is framed). The prospects become gains or losses and then as such are subject to evaluation. Prospect theory states that the value (or utility) carriers are the changes of utility, rather than the value of utility itself. This implies that the prospects should be treated as deviations from a reference point rather than as absolute values, and this opens the possibilities for many biases stemming from the fact that reference points may be the status quo but also an aspiration level (e.g., an employee who gets the bonus of 1,000 USD would consider it a gain, while they would consider it a loss if they expected the bonus of 2,000 USD).

The list of potential culprits – biases related to the editing – is quite long. The least complex of them is the anchoring effect (Tversky & Kahneman, Reference Tversky and Kahneman1974). It consists in imposing the reference point, which in turn produces the biased value of a prospect. Among others and more complex one can find the translation effect (Abelson & Levi, Reference Abelson, Levi, Lindzey and Aronson1985), the endowment effect (Kahneman, Knetsch, & Thaler, Reference Kahneman, Knetsch and Thaler1991), the disposition effect (Dacey & Zielonka, Reference Dacey and Zielonka2008), and the certainty effect (Kahneman & Tversky, Reference Kahneman and Tversky1979).

Most of the arguments of Bermúdez about the rationality of the framing effect refer to the subsequent evaluation phase. Here, people tend to violate normative models, for example, by differently treating gains and losses (losses appear larger than gains) and distort probabilities (sure outcomes are weighted as more important than the uncertain ones). Bermúdez claims these distortions are rational. There is a possibility that the decision maker can be involved in “rational” evaluation of the prospects that are inherently faulty because of being the product of a biased editing.

A procedurally rational agent acts to maximize chances to achieve a goal, regardless of whether the goal was set rationally in the editing phase. As we argue, the goal setting itself is biased by a frame of a problem, and so is the entire decision. Bermúdez argues, and we agree, that loss-avoidance and gain-approach seem to be universally correct and beneficial. The problem with the framing effect, however, is that it is irrational because the same prospects are not only treated differentially, but also the same prospects are malleable – they change when framed differentially on the basis of individual's aspiration levels or imposed reference points.

We argue that the way we set a reference point against which outcomes are compared is temporarily unstable, subject to external manipulation (not only by gain/loss frames, see also anchoring effect, see Tversky & Kahneman, Reference Tversky and Kahneman1974), and thus unreliable to the extent that the decisions are unpredictable. Hence, displaying framing effect is merely procedurally rational consequence of biased editing phase. We conclude that Bermúdez is presenting the framing effect as rational, simply because the evaluation phase is correct. As we argue, there is another bias involved at an earlier stage, and the whole process of decision making may still be irrational. Bermúdez simply moved the source of the bias elsewhere.