Still Money for Nothing? Two Decades of Empirical Evaluation of Conservation Investments

Alex Caruana , Joseph W Bull, Paul J Ferraro, Hannah S. Wauchope, Alec P Christie, Julia P. G. Jones

Twenty years ago, the landmark paper “Money for Nothing?” argued that biodiversity conservation relied too little on empirical evidence. It called for more evaluations of conservation effectiveness based on explicit counterfactuals, comparing observed outcomes with those that would likely have occurred in the absence of intervention. To assess progress towards this goal, we systematically reviewed the study designs used to evaluate one of the most widely implemented conservation interventions: protected areas. Across 614 studies published over the past two decades, half still relied on simple Before-After or Control-Impact designs that do not reliably support causal inferences, although their use has declined in recent years. The other half used more formal causal identification strategies, most commonly conditioning strategies that control for observed confounders. However, most of these studies lacked pre-protection outcome data, limiting their ability to address unobserved confounders. Because causal claims depend on causal assumptions, it is notable that few studies stated these assumptions explicitly, let alone interrogated their plausibility. Although the design of conservation impact evaluations has advanced substantially, much remains to be improved. Combining causal inference methods with expanding data streams from remote sensing and biodiversity monitoring offers a major opportunity to strengthen the evidence base for conservation.