| This paper empirically studies the impact of Sino-Japanese War on human capital in China.Since September 18th,1931,also known as“9.18”incident,Japanese military personnel invaded China.This war ended in 1945,lasted almost14 years.During the war,a large number of civilian displaced,many school aged children were unable to access to normal school.The tragedy was aggravated by Japanese bacterial warfare,epidemics outbroke everywhere.Existing literature have estimated the loss caused by Japanese invasion,yet less of them explain the correlation between the loss and the war.What makes it difficult is that historical archives during the war cannot be documented thoroughly and accurately.Accordingly,this paper firstly utilizes nationwide population census microdata to examine the impact of Sino-Japanese War on education level of the affected cohorts and their offspring.Moreover,by using all kinds of historical epidemic data and bacterial warfare sites,this paper examines the causal relationship between Japanese bacterial warfare and epidemics at the same time.Also,this paper discusses the effect of bacterial warfare on birth during the war.The main content is as below:Firstly,the paper studies the direct effect of war on education human capital.School aged children who exposed to the war were unable to school,would produce inefficiently,thus resulting in low human capital stock.However,less studies explain how much of the effect was.This paper treat Japan surrender in 1945 as a natural experiment,which differentiates the affected children and the unaffected children.Having perused anti-Japanese map and local gazettes,the occupied times are collected.By using 1%population census data that covers almost 2 million observations,and difference-in-differences method,this paper examines the effect of war on the education level of children who exposed to the war.The results show that the children exposed to the war had at least less 0.2 school years,compared to those not exposed to the war.At least 1 out of 100 children would become illiterate because of the war.Heterogeneity analyses show that the effect on women was larger than men,the probability of accomplishing junior middle school is the lowest.In mechanism analyses,the paper finds that education infrastructures,such as school buildings,were the primary channels in which war affected education.However,this paper also finds that intergenerational transmission of parent education and migration of higher-educated people are unlikely the channels.Secondly,this paper studies the lasting effect of war on education human capital.Parents’education has been shown to affect the education level of their offspring,yet the mechanism is not explained clear.By using parent-child matching data in population census,this paper examines whether the education level of the individuals whose parents experienced the Sino-Japanese war was affected.The results show that the individuals whose parents were exposed to the war had 0.1lower school years than those whose parents were not exposed to the war.According to“Quantity-Quality”tradeoff hypothesis in fertility decision,parents will consider between the number of children and human capital investments in children.Those parents who experienced war will lower the expected return to education investment in children,resulting in having more children and decreasing education investments in children.The finding of this paper confirms this hypothesis.Specifically,parents who experienced war had more children and married earlier.Also,this paper finds that the children exposed to war had lower employment probability in labor market when they became adult.This implies that lower income is likely to one of channels which war affected education of next generation.Therefore,there is a lasting effect of war on education of next generation,the possible channels are lower return to education investment in children and lower income of parents.Thirdly,this paper discusses the causal relationship between Japanese bacterial warfare,epidemics,and loss of birth during the wartime.Due to the incompleteness of historical archives documented during the war,the reason why the death and casualties of civilian were severe was less discussed.Existing historical literature reveal that Japanese military personnel had used mass destructed biological weapons on the battlefields.Of these weapons,the bacterial weapons are transmissible and concealable,such as cholera,plague,and typhoid weapons.However,less literature empirically discuss the relationship between bacterial warfare,epidemics,and loss of birth.This paper utilizes epidemics data of almost70 years to examine the causal relationship between bacterial warfare and epidemics at the same time.The results show that cholera warfare increased the probability of cholera outbreaks in 250km of cholera warfare center,and the navigable rivers were one of possible transmissible channels.Plague and typhoid warfare increased the probability of corresponding epidemic outbreaks in 25km,and 50km of warfare center,respectively.Despite the different transmissible spread,the nearest the region approaches the bacterial warfare,the larger is the probability of epidemic outbreaks.In addition,this paper finds that 100km closer to the bacterial warfare will decrease cohort birth growth rate by 8.4 percentage points.Moreover,in the nearest 50km,the cohort birth growth rate decreased by 30 percentage points.When the distance is further away,the effect is becoming lower,and remains not significant until350km further.These findings confirm the negative effect of bacterial warfare on epidemic spread and birth loss.Different from other literature that study the effect of Sino-Japanese War on human capital,this paper focuses on the causal analysis.On the basis of exploring and compiling historical data by historians,this paper uses more thorough and more accurate data to examine the effect of war.The empirical findings complement the existing case studies,and enforce the credibility of conclusions.Due to the findings above,policy implications are proposed as follows:First,proper education policy intervention should be given to children exposed to war.Second,the next generation whose parents experienced war should be treated properly by policy intervention.Those children who experienced war would have more children when they become adult,invest less education in children,thus resulting a vicious cycle.Third,in the process of postwar restore,the policy intervention should be different for different groups.Since the discrepancy of effect existed as this paper shows,such as gender,urban or rural areas,targeted policies would maximize the welfare at the social level.Fourth,epidemic outbreaks should be emphasized during wartime since it threatens health human capital.Biological weapons,such as bacterial weapon,would infect a large number of civilian because of weapons’transmissibility and concealability.Fifth,these findings apply to not only postwar policy intervention,but also policy intervention after other disasters,such as human capital loss caused by pandemic,natural disasters.The innovations in this paper are:First,in terms of research content,the micro-data used in this paper cover two million observations,and the time when Chinese cities occupied by the Japanese army and the prevalence of various major plagues in the past 70 years were manually collected.This paper uses the 1%census microdata provided by IPUMS in 1982,1990 and 2000.This study utilizes population data with millions of observations and covers a wide area,which improves the reliability of this study.By consulting a large number of authoritative materials and various local gazetteers,this paper has sorted out the years when the counties were occupied by the Japanese army and the year when the war ended.Finally,on the basis of the newly published monographs on epidemics by historians,the data of various major epidemics in the past 70 years are collected.Second,for the method,this paper devoted to the causal inference of the impact of war on human capital.The historical event of the end of the Anti-Japanese War in 1945 was used as a natural experiment to examine the educational level of the war-affected cohorts and whether the educational level of the next generation of them was affected.For the impact of bacterial warfare on the spread of epidemics,this paper adopts the counterfactual estimation method proposed by Dell and Olken（2020）to obtain the spread range of epidemic around the center of bacterial warfare,so as to verify the causal relationship between bacterial warfare and the epidemic at the same time.Third,according to the findings,this paper verifies the hypothesis about the"quality-quantity"trade-off when parents make fertility decisions.There is more or less consensus on this hypothesis,because there are too many factors affecting fertility and education.This paper tests this hypothesis by observing whether there is a simultaneous decrease in education level and an increase in fertility rate through the unexpected shock of war.And by validating the theory,this paper enriches the literature on the relationship between fertility and education.The shortcomings of this paper are:First,due to limited data,the individuals studied in this paper are from the survivors from the war,it is impossible to obtain a completely accurate estimated effect.The 1%census microdata provided by IPUMS only surveyed survivors of the war,and some populations more affected by the war may have died earlier,the bias may be more.Fortunately,the estimated value in this paper is downward,that is,the actual effect is greater than the effect estimated in this paper,therefore the main conclusion of this paper would not be affected.This paper also attempts to make up for this shortcoming through more robustness tests.Second,for the impact of war on healthy human capital,this paper only examines the impact on the spread of the epidemic and the birth cohort growth rate.Indeed,there are many variables to measure health,therefore how to select better quality indicators will also be the core of follow-up research.For example,more detailed historical data should be explored,such as the number of people who died from the epidemics can be measured as a proxy for mortality.Therefore,this paper cannot discuss further with limited data,but such issues should be discussed in future research. |
