The framework used to endogenise technology growth by Acemoglu, Aghion, Bursztyn, and Hemous (2012), hereafter AABH, allows the existence of unstable equilibria and does not provide a rationale for specifying which equilibrium should apply when more than one exists. This paper: (i) suggests a rationale for choosing one corner solution used in AABH that constitutes a lower bound for the subsidy or tax required to direct clean research; (ii) argues against use of the other corner solution; and (iii) provides an alternative equilibrium that constitutes an upper bound to the policy required. The alternative methods can produce substantially different results when the elasticity of substitution between clean and dirty inputs is high.