Taal: probing the judgment of scientists and the power of bureaucrats

My wife is Batangueño; some of her relatives are evacuees. So I have a personal interest in the human drama around the Taal volcano eruptions.

Let me paint the scenario as it played out:

Threatened by eruptions which began Jan. 12, many residents evacuate on their own. Others stay longer, to save their livelihoods. The Philippine Institute of Volcanology and Seismology (Philvolcs) warns against imminent, more powerful eruptions, declares a 14-kilometer danger zone, and recommends its complete evacuation. Invoking their police powers, local government units (LGUs) forcibly evacuate people and stop residents from returning to salvage whatever they could of their livelihoods. Other officials allow a four-hour window. More than a week after the first eruption, a stern order comes from the Department of the Interior and Local Government (DILG) about Tagaytay businesses: stay closed or your business permits will be revoked. National officials also talk about permanent relocation. On Jan. 26, Philvolcs downgraded the Taal alert level from 4 to 3. Many evacuees return to their homes to pick up the pieces and resume their lives.

Areas around Taal have been under threat of eruption for centuries. Batangueños, more than anyone else, feel the threat in their guts. This conscious decision to live near a volcano is a badge of courage. That they have flourished is a tribute to their resilience. This right to decide what and how much risk to take should be respected. In the current context, this is how it could be done:

Philvolcs’ knowledge and judgment guide government decisions. Philvolcs knows the mechanisms of volcanic behavior. It knows how specific volcanoes the world over have behaved in the past. It knows the details of Taal’s previously recorded eruptions. Its instruments tell it how magma is flowing underneath the Taal geological formation. But it does not know how soon a deadly eruption will actually occur. It knows the various threats (base surge, toxic gases, ashfall, pyroclastic flows, lava flow, tsunami, landslides, etc.), but it does not know which of these will occur and to what extent. It has judgments, based on its knowledge.

The problem is the uncertainty in its judgments. The solution is to put numbers to this uncertainty: How imminent? How powerful? How risky? Initially, Philvolcs gave no number to the probability of a deadly eruption. A few days after the eruptions subsided, its spokesperson said, “we need to look at the probabilities again.” This means it did have probability estimates, but it kept it to itself. On Jan. 23, Philvolcs finally announced its estimate: 30%.

This 30% presumably refers to the probability of an eruption within “hours or days” that can kill people within the 14-km danger zone. To be far more useful, Philvolcs should word its risk estimates roughly as follows: “As of (say) 6 a.m. today, we estimate the probability within the next (say) 10 days of a Taal eruption that can kill people within (say) 14 km is (say) 30%. Inside the danger zone, the risk of getting caught in a deadly eruption is around (say) 0.15% for every hour inside.”

The very useful last sentence above assumes a period of 10 days and that the probability is evenly distributed over those 240 hours. Philvolcs can use its updated probability parameters whenever it issues its daily bulletins.

The 0.15% is calculated as follows: The chances of a deadly eruption not occurring within 10 days is 70%; the chances of one not occurring on a particular one-hour period among the 240 hours/10 days is 99.85% (for the math-literate: 0.9985 raised to 240 is 0.70); the chances of one occurring on that particular hour is 0.15% — roughly one in 673. These are by no means precise, but they are the best we have. Even in wars, staying in a battle zone for an hour is far less risky than staying there for a month. Senator Bato calls such a risk “one-time, big-time.”

Consider a one-time retrieval of animals to bring them to a safe zone. If this takes four hours, the probability of getting caught by a deadly eruption quadruples: 0.6% or 1 in 167. Residents and LGU officials who wanted to retrieve their animals must have instinctively felt that the risk was worth taking. Had higher officials allowed it, and provided the means to do it quickly, many animals could have been retrieved and part of the people’s livelihoods salvaged.

A lot of homes could be saved if ash is swept from roofs regularly to prevent collapse. Assume a two-hour cleaning operation every five days. This also totals four hours of risk exposure, like the one-time animal retrieval operation. Residents who thought the risk was worth taking should have been supported by LGUs by providing transport, face masks, ladders and brooms, and risk-reduction personnel could stand by just outside the danger zone in case a quick exit was necessary.

Doing both operations doubles the risk further to 1.2%. But not for individuals, if a different set of people did the second operation. For instance, it was wrong to assign the same police personnel in the danger zone throughout the eruptions. Rotation would have reduced the risks for individuals.

It was painful to watch coverage of residents pleading to bureaucrats and police to let them clean their roofs and save their animals. With the benefit of hindsight, we now know for a fact that no deadly eruption occurred during those particular hours. Given the prior odds of 99.85% against a deadly eruption, that was entirely expected. The roofs would have been cleaned, and the animals retrieved. Nobody would have died.

On forced evacuations: the following argument can hopefully convince LGUs why they should let households make the final decision.

Economists use the term moral hazard for this problem. Evacuees bring extra money and resources to LGUs; the more evacuees, the more aid LGUs receive. This was highlighted by the state of emergency declared in Cavite, so the neighboring province can also use its emergency funds and receive aid for evacuees.

When LGUs force residents to evacuate against their will, then receive millions in aid as a result, a moral hazard arises. The hazard is that genuine concern may at some point become a callous way of getting more aid. No such hazard arises in voluntary evacuations.

The DILG order to the Tagaytay mayor that businesses should stay closed was clearly overreacting. For businesses within the danger zone, the order was superfluous. The lock-down policy already applied to all. For struggling businesses outside the danger zone, the order would have been a death sentence. Tagaytay City is the crown jewel of Southern Tagalog tourism. Big land developers — President Duterte sometimes derisively calls them the Rich — are moving in. A string of bankruptcies would have created new opportunities for the Rich to muscle in.

President Duterte won partly for his promise to end the rule of “imperial Manila.” He should realize that the DILG order for Tagaytay businesses to stay closed smacks of this Imperial Manila Syndrome.

For centuries, Batangueños did not give up their livelihoods for fear of a volcano. Now and in the future, they will not surrender to it without a fight. If the government ties their hands with lock-downs and closure orders, they will surely become dependent evacuees requiring huge infusions of aid and vulnerable victims of Manila-based land speculators.

Roberto Verzola has an academic background in economics and engineering. He runs a non-profit organization promoting renewable energy, and he is a senior fellow of Action for Economic Reforms.