I do not usually watch documentaries. When I am on a device, I am a news article junkie. When I am doing chores, I really enjoy listening to podcasts. But for various reasons, I never enjoyed watching documentaries. Of course, when my son was born in 2019, I quickly discovered that the main task in parenting a newborn is convincing them that it is in their best interest to sleep. One of the only ways to do it at times is to hold them while walking around until they fall asleep and then to sit on the couch with them sleeping on your shoulder. This made me appreciate documentaries more than I ever have before.
During Christmas 2019, I remember watching a series of short Vox documentaries on Netflix called “Explained.” Many of the episodes were quite whimsical and funny, but one, in particular, struck me: “Pandemics Explained.” Pandemics were never a subject I spent much time thinking about. Yet, I found the episode so impressive that I ended up showing it again to my wife and a friend a few days later. The show (in only 20 minutes) went through a short history of 20th-century pandemics (mostly the Spanish flu) and some recent pandemic near misses (SARS in 2002). My main takeaway was: “I really hope we do not have a global pandemic anytime soon because we are not prepared for this.” Little did I know that we were only weeks away from the global spread of COVID-19, which would utterly change our lives and our world for a long time to come.
In retrospect, a pandemic was a risk hiding in plain sight: history has been full of pandemics, and yet we were woefully unprepared for COVID. Our hospitals, economy, government agencies, safety net: none of them were not prepared for the havoc that COVID has wrought in the past year. Yet it was not unpredictable: we have seen global pandemics before, and we should have known that another one was coming eventually. Nonetheless, we didn’t prepare for one.
We, as human beings, are terrible at understanding risk. Even though the risk is often quantifiable, our bias towards optimism pushes us to avoid confronting small but significant chances at adverse outcomes. We know stocks naturally fluctuate up and down, still we underestimate the potential that we will be the one holding the bag on the downturn. We understand that all forested landscapes inevitably burn on a set pattern, yet we assume that our neighborhood will not have any wildfires that impact us. Actuaries create tables that quantify our likelihood of dying or becoming disabled in any given year, but we often don’t buy life or long-term disability insurance.
One upside to getting burned by these risks: we sometimes change our behavior due to experiencing a bad outcome (or those we love experiencing one). It is a terrible way to learn, but often it is the only way we do. The most competent countries in fighting COVID have been Pacific Rim countries like Taiwan, Hong Kong, and Singapore. For them, SARS was not a distant news story but a terrifying disease that managed to infect and kill enough people to wake them up to pandemic risk. In the US, after COVID, we can hope that we will more realistically anticipate the likelihood of pandemics and do what we can to mitigate those risks.
For instance, it seems sensible to invest in public health to “stand ready” for future pandemics. The US supports our military even in peacetime in the name of “readiness” if we face an unexpected attack. We should take a similar approach with public health:
We can invest in our healthcare system's capacity to handle surges, something that private hospitals don’t have an incentive to do.
We can create government prizes that would buy vaccines patents for diseases and viruses (like Ebola) that carry a public health risk. We do not expect these vaccines to be in demand outside of a pandemic, so the market often under-incentivizes these vaccines being created.
We can mitigate the risks of zoonotic transmission posed by modern agriculture by pricing risk into business models that rely on factory farming.
We can modernize the FDA to include “wartime” provisions that would force the agency to act more quickly and be less obstructive during a pandemic outbreak. The rigorous processes designed for reviewing drugs that may make a heart attack 10% less likely SHOULD NOT be the same as processes during a pandemic. Next time, we should prepare to approve tests, antiviral therapies, and vaccines much faster.
COVID has hopefully taught us that it is worth investing in low-cost measures to mitigate future pandemics before they grow out of control. We have seen the immense human cost once they are out of control.
The Even Bigger Problem of Uncertainty
However, I would argue that as bad as our capacity to judge risk is, it is not nearly as dangerous to public policy as uncertainty. These two ideas are often conflated, but I think it is helpful to cite Frank Knight’s work that attempts to distinguish them (emphasis is mine):
Risk is present when future events occur with measurable probability
Uncertainty is present when the likelihood of future events is indefinite or incalculable
Risk is quantifiable because it relies on comparisons with past actions, as when actuaries quantify the risk of an individual dying in any given year. Uncertainty involves almost entirely unprecedented threats, and thus, we have no idea what risk we face. When we deal with risk, we can expect most people to misjudge risk. Still, at least the possibility that those who rigorously study risk can create tools to keep us safe. Think of the scientists whose years of extensive study of coronaviruses helped us act more quickly to develop tests and vaccines for COVID:
Problems of uncertainty are particularly “wicked” in a way that can be confounding to specialized experts. This is because we cannot use past data to predict the future. We can create theoretical models of the future, but they do a poor job of holding up in the real world. This makes it incredibly difficult to identify where the trade-offs for any particular issue lie: we are forced to combine incomplete data and historical analogies to guess the future. A dollar spent on mitigating an uncertain risk is a dollar we cannot spend on known risks! And it is actually quite scary to think of how many problems involve both a high level of uncertainty and the potential for immense harm. Here is a non-exhaustive list:
Asteroids
You may laugh at me for starting with this problem, but I think it is a fascinating example of uncertainty. We know that asteroids can strike our earth: they have in the past and will again. But our data is so limited that we have minimal idea of how often this phenomenon occurs and how likely it is to be cataclysmic. It is theoretically possible for all life on earth to be wiped out by an asteroid; Hollywood was so taken with this dramatic possibility in the late 90s that they made two movies about it in one year (only for the subject to quickly pass out of our consciousness). We also know the human race has yet to be wiped out by one, and we seem to have been on this planet for quite some time.
Nature tried to quantify an asteroid's risk and estimated a 1 in 10,000 chance that a large asteroid would hit the earth during the 21st century. But our limited data create a great deal of uncertainty around that number. How much spending is justified on a problem with a small chance of cataclysm and a large chance of mattering zero? Would it matter whether one country (probably the US) pays for efforts to mitigate this risk while the rest of the world has a free ride? Congress seems to think this matters a great deal! How would we ever create international co-operation to fund such an abstract risk? It looks like something we should actually discuss, not just dismiss out of hand.
Nuclear Weapons
We know that for the past 75 years, we have possessed weapons powerful enough to wipe out life on earth. However, we also understand that these weapons have only been used twice in history (and they were much less potent than today’s weapons). There is a whole theory of nuclear deterrence, which says that nuclear weapons make countries in possession of them less likely to fight each other. This has historical merit: only once (India and Pakistan) has there ever been a war between two countries possessing nuclear weapons, and that war was quite limited.
Is this a sign that deterrence works? Or a sign have we merely gotten lucky for 75 years? Could a miscalculation (like we almost had in the mid-1980s) doom us in the future? Could a weapon getting into the hands of a rogue state or terrorist group send our world reeling? I don’t personally have a great deal of faith in people who seem 100% confident one way or the other. There is a huge range of possible outcomes here: from good (we continue to see nuclear weapons deter war) to bad (cities being taken out by rogue terrorists) to cataclysmic (a significant amount of civilization is wiped out by nuclear war between states). How do we understand the tradeoffs of a policy issue with so little historical data? It is wickedly hard to quantify these risks! And probably should give us pause that this is not really a topic of debate in this post-cold-war era. We should be doing our best to think through this problem!
Political Violence
This is where we start to get into more controversial territory: looking around the world, it is easy to see that political systems are fragile. But most Americans (specifically white Americans) have generally thought that the risk to our political system was relatively low. But recent events have caused many to question whether we should be more “uncertain” about the ability of democratic institutions to hold up.
As we have seen recently, partisans increasingly justify political violence to achieve desired political outcomes. This should not shock us: research before the 2020 election found that 36% of Republicans and 33% of Democrats thought that using violence to achieve political goals was “a little” justified. Over 40% of both sides thought electoral defeat in an election was a reasonable justification for political violence. As I have written before, any level of political violence is a threat to democracy. One of the best ways to understand democracy is as a “substitute for war.” Democracy enables us to channel our anger or dislike for others into voting against them, rather than trying to harm them. Once you start to legitimize violence as an alternative to democracy, the whole system can quickly descend into chaos.
Of course, one can find individual cases where violence did achieve good ends: the American civil war did end the institution of slavery. However, suppose you survey the history of civil wars and violent revolutions across human history. In that case, it becomes much harder to have a lot of certainty that contemporary violence would be similarly effective. We should be willing to sacrifice some of our own ideological goals to keep the system of democracy working. Maybe our post-pandemic political world will involve a return to normalcy, with less political violence. That may prove true, but I would not be “certain” about it.
Debt and Deficit
In recent years, the US has spent far more money than we ever have in our history outside of mobilization for war. While a pandemic is undoubtedly a good reason to justify this (a pandemic is very similar to a war), we were running up huge government debts before COVID and almost certainly will after COVID. It is important to say that because of monetary policy, countries running up debts do not have the disastrous consequences that individuals running up debts have. Most recent debt crises involve entities that do not have their own currency (Greece is in the Eurozone, Puerto Rico is an American territory). Not only does the US have the advantage of printing its own money, but the US also maintains a unique position as a kind of “reserve currency” for the world, and thus, even after huge deficit spending, demand for federal treasury bonds is as healthy as ever. The normal warning signs that economists look for, inflation and interest rates on debt, have not gone up. So there is no sign that we are plunging into a hole we cannot dig ourselves out of.
For this reason, many on the left (and even some on the center-right) now take it as an article of faith that deficits don’t matter as much as we thought and that governments should spend as much as they need to. The extreme articulation of this can be found in Modern Monetary Theory, which is increasingly popular on the progressive left. And they have a point: there is an opportunity cost to governments not spending money when interest rates are low! But there are three reasons to be skeptical of taking this argument too far:
We know that deficits and debt do matter; almost all economists still agree there could come a tipping point where debt will become an issue. Even if the US never will face a Greece in 2009 style crisis because we have more power over our currency, history has many examples of inflation becoming a severe issue. The problem is, as economist Noah Smith puts it, we have no idea when that point will come. The “fiscal cliff” is somewhere in the distance, but it could be 5 miles away or 500.
Suppose we start to approach a debt crisis. In that case, Congress will need to act quickly to mitigate that crisis's potentially terrible effects. For those who have not been paying attention, Congress has been a quite dysfunctional institution for some time. We should not assume that they will suddenly “come to Jesus” when facing a debt crisis.
Even if we never reach a catastrophic crisis point, we can put ourselves in a position where our future fiscal flexibility is constrained because of the obligations to debt, interest payments, and entitlements we have built up through the years. Our children and grandchildren may not have the same capacity for new and necessary government spending; instead, they will be hamstrung by our choices to borrow money.
I think it is fair to say that there is no absolutely correct way to respond to this problem. But I think it would be unwise to throw out all limiting principles when it comes to debt. I do not have space to go into what I think the best approach is here, but I think a good, level-headed approach is laid out in what conservative scholar Yuval Levin says in his recent article:
If we are concerned with debt, we should be more attuned to the issue of long-term entitlement funding (mostly medicare) rather than short-term emergency funding.
We need to continually ask the same question that individuals should ask when taking on debt: will this spending serve as an investment that can pay off in the future as economic growth?
Spending on things like physical infrastructure and keeping kids out of poverty likely does this (again, in my opinion), as it helps build up our tangible and human capital. But giving upper-income individuals $1400 checks in the middle of a pandemic should be subject to more scrutiny (I will be writing about this more later this week). Even if constraints don’t matter as much as we think, they still exist, and thus priorities are still important.
Climate Change:
Now we are entirely into controversial territory! From studying the scientific debate, I think there is actually a great deal of certainty that two things are happening:
1) The world is getting warmer
2) Human economic life is contributing a great deal to that warming.
I understand some readers will think differently (please email me if this is a conversation you want to have), but I will not spend time reasoning through these points here. HOWEVER, certainty about those two points still leaves a great deal of uncertainty around two other matters that are vital to this conversation:
3) How damaging will the effects of climate change be?
4) What are the best strategies for fighting those effects?
Most of our inferences about how global warming will impact human civilization come from mathematical climate models. Using math to simulate a complex system is, unsurprisingly, a wicked problem that experts are doing their best to make educated guesses about. Anyone who pays attention to economic models knows that we cannot expect them to be 100% accurate. We should have a great deal of uncertainty: we shouldn’t dismiss the idea that climate change could be a cataclysmic event for humanity. There is also the potential that climate change is a severe challenge that creates bad outcomes well short of cataclysm. This uncertainty renders policy-making really difficult.
On the one hand, doing only the bare minimum to avoid climate change risks being an even worse example of unpreparedness than the pandemic. The potential outcomes of warming 4-5°C make COVID look like a cakewalk. Governments would be justified in bombing their geopolitical rivals’ coal factories if we knew with certainty the outcomes of this warming were this dire. Of course, this is not happening because most policy-makers either do not yet believe or find it convenient to think that the worst consequences will not happen. But as time goes on, and if things worsen, more and more drastic solutions might be proposed.
But we also cannot know that we will see the worst-case scenario. In the 1960 and 1970s, there was a widespread belief that overpopulation would become a defining problem of the latter half of the 20th century. Prophets of doom like William Vogt and Paul Erlich believed humanity needed to take drastic measures to avoid the mass catastrophe:
China’s one child policy, which has had brutal consequences on women and racial minorities, was primarily implemented in response to overpopulation concerns. So was the forced sterilization of lower-class people in India, which American NGOs played a crucial role in. As it turns out, overpopulation was a non-issue: technology to improve food production meant that yield outpaced the growing population. Economic prosperity reduced women's incentives to have more than two kids, and the world population did not grow as fast as predicted. Now we are at the point where many countries are more concerned with shrinking populations! Now, just because alarmists were wrong on overpopulation does not mean they are faulty on climate. Personally, I do not believe the climate issue will go away without concerted action that will involve some short-term pain. But this historical example should pause to those who feel confident we are headed towards the worst-case scenario and sure about the best policy response. Those who preached the dangers of overpopulation then were just as passionate as those who are sure of the risks of climate change today.
Predicting how well we can mitigate climate change is also uncertain. The pace of improvement in renewable energy has been phenomenal over the last 10 years. By some measures, solar power is now the cheapest form of power across the board (even before subsidies and carbon taxes). But carbon capture technology, which was highly touted 15 years ago, has mostly proven to be a dud. I tend towards techno-optimism on this front: I am more persuaded that widespread efforts to incentivize markets to push for de-carbonization will achieve climate goals without catastrophic consequences. But I cannot be certain that is what will happen!
Conclusion
You may not agree with me on all these areas of risk and uncertainty, which is expected! But I hope you will agree with my broader point:
We should probably be more concerned with precise thinking about them than we are.
We should be careful to maintain a level of epistemic humility about them.
To fail to do so risks genuine peril!