We explore the welfare costs of the loss of animal life in a consequentialist total utilitarian framework that incorporates the intrinsic value of living beings through their sentience. Moral philosophy and neuroscience define sentience as the capacity for valenced experience. Observable performance on cognitive tests represents a conservative lower bound on sentience under plausible neuroscientific assumptions. We estimate the relationship between cognition and neuron counts to cardinalize sentience. We collect population and neuron data for a wide range of animal species, and aggregate them into a single welfare measure. Across a wide range of parameters, the aggregate sentience losses of animals in the past half century outstrip the sentience gains to humans from population growth by at least 10,000-fold. Breaking even in welfare terms requires assigning a correspondingly small welfare weight to animals. Including the benefits of economic growth to humans does little to change these overall welfare losses.

In my view macro welfare comparisons are most meaningful when gains from trade are possible, which of course holds across humans, and also holds across humans and (some) dogs.  I believe we should act to boost the welfare of non-human animals at the margin, but cannot make the larger comparisons across the totals.  We are, inevitably, on the side of the humans and we are not obliged, as welfarists, to commit collective suicide to replenish the biodiversity of the earth.

The title of the paper is the somewhat misleading “The Intrinsic Valuation of Biodiversity Loss.”  The authors are