Thursday, July 10, 2014
In April, 2013, I wrote a post about The Aransas Project v. Shaw, a case involving water management agencies, whooping cranes, and the Endangered Species Act. The defendant water agencies had issued permits for water withdrawals upstream of important whooping crane habitat. According to the plaintiffs, the combination of those permits and the 2008-09 drought reduced freshwater inflows to the estuary where whooping cranes feed, leading to food shortages that killed off large numbers of whooping cranes. Those deaths, the plaintiffs argued, were unpermitted "takes" and were prohibited by section 9 of the ESA.
A federal district court agreed with this theory, but on June 30, the Fifth Circuit reversed. It determined that the district court had failed to apply a "proximate cause" analysis to the take claims. Because the district court failed to apply the correct legal standard, the Fifth Circuit found that it owed no deference to the district court's factual findings. Considering the record de novo, the Fifth Circuit concluded that the water agencies' activities were not the proximate cause of the whooping cranes' deaths.
The reasoning that led the Fifth Circuit to its de novo standard is, to say the least, interesting. In fact, the district court did recite the proximate cause standard, and it did purport to apply it. It just applied it in a way that the Fifth Circuit found overly simplistic. That could be a basis for reversal--appellate courts do review factfinders' factual conclusions for abuses of discretion--but it doesn't seem like a basis for de novo appellate review.
There's also a whiff of hypocrisy in the Fifth Circuit's reasoning. It concluded that the district court had been too simplistic because it had failed to engage with the complexities and contingencies of the alleged causal chain. Perhaps that's a fair critique (for an argument that the district court made a mess of the case, see the comments on my earlier post); I am not familiar enough with the factual record of the case to know.
But the Fifth Circuit's substitute reasoning is also simplistic. Its core conclusion is that "[c]ontingencies concerning permittees' and others' water use, the forces of nature, and the availability of certain foods to whooping cranes demonstrate that only a fortuitous confluence of adverse factors caused the unexpected 2008-09 die-off found by the district court." That conclusion contains a big logical leap, which I don't think is remedied elsewhere in the opinion: it ignores the reality that uncertainty is often a matter of degree. Just identifying some uncertainty within a chain of causation does not mean that the ultimate outcome could fairly be described as "fortuitous." Instead, one must ask how contingent the links in the causal chain were, how much other contributing causes might have added, and how long the causal chain was.
To put it in simple mathematical terms, suppose that event A has a 90% chance of causing event B, which has a 90% chance of causing event C, which has a 90% chance of causing event D. There is uncertainty at every stage of this causal chain, yet the odds of event A leading to event D still are just under 66%. We'd probably be comfortable calling event A the proximate cause of event D. On the other hand, if each event has only a 30% chance of causing the next event, then the odds of event A leading to event D are less than 1%. Contingencies are present in both causal chains, and the chains themselves are the same length. But the causal relationships are drastically different.
Or consider historical examples. One hundred years ago, Gavrilo Princip fired two shots that set in motion a series of events culminating in the slaughter of World War I. In hindsight, history often looks falsely inevitable, yet I suspect most historians would agree that the links between those shots and the horrors of trench warfare in northeastern France were too many, and too contingent, to identify Princip as the proximate cause of the Battle of the Somme. But historians probably will identify the 9/11 attacks as the proximate cause of the U.S. invasion of Afghanistan, even though a contingent causal chain linked those events as well.
The point, again, is that it's not enough to just name uncertainties or other contributing causes, though that's a logical first step in the analysis. We also need to think about how much they matter. And the Fifth Circuit didn't really do that. If it had, the results might have been different. The relationship between permits that allow water withdrawals and actual water withdrawals is pretty direct. So too is the relationship between upstream water withdrawals and downstream reductions in flow. The occurence of a severe drought in 2008-09 was, of course, a chance event, but there's very little chance of Texas avoiding severe droughts throughout the entire duration of a water use permit. The last links in the causal chain--reduced inflows allegedly causing ecological effects to ripple up the food chain, causing whooping cranes to die off--are probably the most uncertain, but reduced inflows damaging an estuary's food chain is pretty plausible, and just labeling that outcome contingent ought to have been the start, not the end, of the analysis.
So how much does the case mean for ESA litigation? The answer, I suspect, is not a whole lot. Even if courts were committed to embracing the gory details of uncertainty analysis, winning ESA section 9 cases would still be a challenge, for the evidentiary burdens plaintiffs face would be substantial. For that reason, plaintiffs don't often try. This latest decision will likely just reinforce that reality.
- Dave Owen