Legacies of Wartime Destruction: Lessons from the Tokyo Firebombing
By Daniel M. Smith
The wars in Ukraine and Gaza have filled newsfeeds with haunting images of destroyed communities and tragic stories of civilians who have been exposed to brutal wartime violence (Figure 1).
Two years since the start of the war in Ukraine, tens of thousands of civilians have been killed, millions have been displaced, and numerous towns and villages have been completely destroyed. In one recent survey, about half of the people in eastern and southern Ukraine reported that they had lost their homes to damage or destruction. In Gaza, the death toll now exceeds 30,000 people, and experts estimate that somewhere between a third and a half of all buildings have been damaged or destroyed.
Sadly, humankind’s most violent and destructive behavior has been repeated throughout history. But we can also try to learn something from history––if not how to prevent war from starting in the first place, at least whether and how communities might recover from its damaging effects, and how long the process of recovery might take after the fighting ends.
A rapidly expanding literature in historical political economy (HPE) broadly takes up this challenge, with a focus on understanding the various long-term effects of war (and other forms of political violence) on the communities affected by it (e.g., Charnysh and Finkel 2017; Lupu and Peisakhin 2017; Rozenas, Schutte, and Zhukov 2017; Walden and Zhukov, 2020; Wang 2021).
Exposure to wartime violence and destruction––whether in civil wars or interstate wars––is a complex and bundled treatment that can leave dramatic and enduring legacies (positive or negative) affecting individuals, families, communities, and local institutions (e.g., Collier et al., 2003; Wood, 2008; Bellows and Miguel, 2009; Balcells, 2012; Arjona, 2014; Bauer et al., 2014; Gilligan, Pasquale, and Samii, 2014; Dell and Querubin, 2018). These include (1) psychological changes such as post-traumatic stress disorder (PTSD); (2) activation of collective-action behaviors; (3) social disruption due to displacement and erosion of social networks; (4) economic costs of destruction and reconstruction; (5) transformation of formal and informal institutions; and (6) environmental degradation and contamination of natural resources, among other outcomes.
The Socioeconomic Effects of Wartime Destruction
Several studies attempt to understand the long-term socioeconomic effects of one particular form of wartime violence: destruction caused by bombing campaigns. But the findings are mixed.
An early and influential study by Davis and Weinstein (2002) looks at the effects of Allied bombings of Japan in World War II on the population sizes of its cities. They find that it only took about two decades for city-level populations to recover to their prewar trends.
Building on this finding, Brakman, Garretsen, and Schramm (2004) consider the case of Germany––which was also bombed by Allied forces in multiple cities. The authors find that the pattern of population recovery in West Germany was similar to that of Japan, but diverged in East Germany. This could be due, the authors argue, to the centrally planned economy in East Germany, which constrained the reconstruction process for housing.
Other studies examine the effects of US bombing campaigns in the Vietnam War in subsequent development in Southeast Asia. Miguel and Roland (2011), for example, find no evidence of long-term effects of US bombings on population density, local poverty rates, consumption levels, infrastructure, or literacy in Vietnam.
Yamada and Yamada (2021) consider the effects of bombings on population density and economic development (measured by nighttime luminosity) in villages in Laos, and find no evidence of long-term effects in southern Laos four decades after the bombings. However, they but do find negative effects in northern Laos.
Finally, Lin (2022) looks at the legacies of the bombing campaigns in Cambodia. She finds that long-term community-level economic development was suppressed––not because of the direct destruction caused by the bombings, but rather because unexploded bombs in high-fertility land areas left these areas unsafe to farm.
The Case of the Tokyo Firebombing
In a new study, Masataka Harada, Gaku Ito, and I investigate the long-term socioeconomic consequences of wartime destruction using the case of the 1945 firebombing of Tokyo (Harada, Ito, and Smith, 2024).
While most historical and popular accounts of World War II have focused on the atomic bombings of Hiroshima and Nagasaki, the US’s deadliest bombing operation in Japan took place in Tokyo over the course of several air raid campaigns between March and August 1945. These air raids destroyed the homes of approximately 2.5 million residents, killed about 100,000 people, and injured roughly the same number of people (Fedman and Karacas, 2012).
The air raids left varying degrees of damage across neighborhoods of Tokyo, due to the indiscriminate bombing strategy of the US Army Air Forces (USAAF) under Major General Curtis LeMay, the wood-and-paper materials used to construct homes, and winds that spread the resultant fires caused by the napalm-like bombs. Community fire brigades had little chance of containing the raging infernos that swept through the city (e.g., Moore, 2018).
As illustrated in the top cover image for this post (taken by the USAAF after the war), the firebombing left some neighborhoods largely intact, while other neighborhoods, even in the immediate vicinity, were completely razed to the ground. We use the conditionally exogenous variation in damages caused by the firebombing (controlling for known treatment assignment mechanisms like aiming points and other factors) to investigate whether the amount of damage sustained in the air raids affected the long-term socioeconomic development of communities at the neighborhood level.
As Figure 2 illustrates stylistically, variation in damages caused by the air raids could theoretically have resulted in diverging patterns in the postwar redevelopment of Tokyo’s neighborhoods. While neighborhoods that sustained minimal damage could have experienced relatively minimal disruption to the existing urban environment and social ties between residents, neighborhoods that were more heavily damaged were potentially exposed to various forms of disruption, and these immediate postwar differences could have produced enduring legacies.
We use aerial photos taken after the war to measure the level of damage sustained across neighborhoods, and then examine the relationship between the level of damage and contemporary measures of social capital and socioeconomic outcomes: the presence of authorized neighborhood associations (a key indicator of the strength of local social capital), and census-based data on the unemployment rate, occupational class composition, residential stability, and education levels of residents.
Our findings indicate that, even decades after the end of the war, the most heavily damaged neighborhoods were less likely to have registered an authorized neighborhood association, and these neighborhoods also exhibited slightly worse socioeconomic outcomes. Although the effect sizes are modest (and vary across years for some outcomes), it is nevertheless remarkable that such inequalities persisted even decades after the war, and in a city that experienced a general recovery of its population and civil society activity (Davis and Weinstein, 2002; Kage, 2010), rapid economic growth, and dramatic transformations of its urban space.
Broader Implications?
This kind of HPE research is important, since it may help us understand how and when communities that are currently being devastated by war––whether in Ukraine, Gaza, or elsewhere in the world––might recover in the aftermath of destruction.
But caution is warranted, as many of the downstream consequences of the bombing cases scholars have examined thus far have idiosyncratic factors bundled together with the destruction, as well as potentially crucial differences in postwar policies. An example of the former is the fact that bombs in Cambodia were less likely to detonate in more fertile soil, thereby “reversing the fortunes” of different areas in the country’s postwar economic development (Lin, 2022). An example of the latter is the divergent housing policies of West and East Germany (Brakman, Garretsen, and Schramm, 2004), which were not caused by the destruction itself, but nevertheless seem to have affected post-bombing outcomes.
Tokyo’s experience might not be replicated in other contexts, but might still offer some lessons for postwar public policies. One reason why the firebombing of Tokyo had such lasting effects could be because no government program was established to help or compensate air raid victims, and public remembrance of the air raids was overshadowed by the bombings of Hiroshima and Nagasaki (Karacas, 2010). Reconstruction of Tokyo following the war was also notoriously slow to begin (Hein, Diefendorf, and Ishida, 2003), leaving an opportunity for the seeds of inequality to take root and compound over time.
Going forward, it is important for HPE scholars working on these questions to continue to ask, as a recent review of the historical persistence literature does, “under what conditions does historical persistence take place, and under what conditions does it not?” (Acharya, Blackwell, and Sen 2024). If we can learn enough from the variety of observed outcomes following the horrors of wartime destruction experienced in the past, we can perhaps identify ways to mitigate these legacy effects from taking hold in the future.