Data Collection, Analyses and Organization

     The 21 Web Tables data and analyses that follow support the numbers used in Imperial Japan's World War Two. . . The casualty numbers shown throughout this book are for the most part approximations due to the dearth of precise counts from official government sources. Most of the territories and countries invaded by Japan made little attempt to collect detailed statistical data on their victims since their governments and infrastructure were too badly disrupted by the war and occupation. Following the war, East Asia was confronted with the huge task of recovery, and some areas were engulfed in civil wars and wars of independence. Estimates and calculations have been made during and since the war by governments, observers, and writers. Countries like the U.S. and Japan provide the most reliable data on their losses, but inconsistencies still exist, including a wide range of estimated deaths from the atomic bombs.

     The process has been to collect available statistical data, cross checking it with anecdotal and other evidence, making adjustments and sometimes new calculations to arrive at estimates. An effort has been made to locate as many credible estimates as possible for the wide range of war situations and to average or use numbers that are in the middle of a range of these estimates. In some cases, illustrative rough calculations are made to estimate well-documented atrocities and other tragic occurrences for which there are no other estimates. For example, estimates were made for the number of rape victims and Kempeitai torture victims. Aggregated with other numbers these provide some idea of the magnitude of the war's casualties.

     The statistics of aggregation means that the sum is more accurate than its parts, i.e., in the estimates of the individual subparts will in some cases be too high and others too low, but their total is probably not too far from reality. Thus I believe the total deaths and total numbers of severely affected that are tabulated here give a good indication of the scale of the war's human devastation. The annual distribution of these casualties is less accurate but still provides a useful measure of the yearly toll for analyses purposes.

     The sources of statistical data represent a wide range of major organizations and authorities on World War II (1931-1945) in China, Southeast Asia, and the Pacific. The tables provide the sources of the data in the form of the last name of the source author, the date of publication, and the page on which the data can be found. The reader of Imperial Japan's World War Two 1931-1945 need only refer to the book's Bibliography to find the full description of the source.

     At the end of the war, the Japanese had two weeks before the Allies took control and were under orders from senior Imperial government officials to destroy any records that contained information about their victims. Allied bombing also destroyed documents. Those that were taken by the United States after the war were returned to the Japanese government before many pages had been copied, which has, for the most part, kept them closed to researchers. Japanese estimates of casualties they caused, that are available, are often far lower than those of Western news personnel, POWs, embassy officials, and other witnesses to the war in the theater.

     Korean casualties are included under the tabulation of Allied casualties because they opposed Japan's annexation in 1910 and suffered severely from it in many ways. The Japanese acquired Taiwan after they defeated the Chinese in the 1894-5 war. Taiwan did not appear to have suffered as much under Japanese control and therefore that nation is not considered part of the Allied victims for this study.

     Japan's war and occupation policies undermined the already-fragile Asian life-sustaining economies that supported over one third of the world's population. The scale of famine and disease death caused by World War II was not out of line with Asian history, as examples will show. From 1876 to 1879, a drought--caused famine in Northern China is estimated to have taken five million lives. North China experienced drought again in 1892-94, which caused an estimated million deaths. In 1920-21, another caused 500,000 deaths. Droughts in Bengal, India, 1899-1900 caused deaths of over a million. As bad as these famine numbers were, millions of lives were saved because of peacetime relief efforts not available under wartime conditions. The numbers of deaths due to the war's disruptions is probably underestimated in this work, considering the huge populations at risk and susceptible to famines over a four to 14-year period. [1]

     A considerable part of the war and Japanese occupation in China and East Asia went unobserved by the West. Many examples of death and terror probably went unreported, or are in Asian language studies not much pursued by Westerners because of translation challenges. While these factors suggest that the numbers in this book may be too low, the tendency for war observers to overestimate casualties may balances things out. For this reason, I believe the casualty numbers provided here reasonably represents the dimensions of the human 14-year tragedy. However, as noted in Imperial Japan's World War Two. . .more research is necessary particularly with respect to those dying after the war from war causes.

List of all 21 Tables with further explanation when required.

     These notes will augment the notes provided in the table pages themselves. All discussion refers primarily to my reasoning and calculations. In all tables the casualties occurring on and after December 7, 1941 are included in the 1942 statistics to simplify the multi year statistical display. Table pages often include more than one set of statistical tables. For the full reference based on the last names and publication data shown in the Tables' sources, refer to the parent book Imperial Japan's World War Two 1931-1945 Bibliography. Tight space permits only the major sources to be listed, but often there are more. The following acronyms are sometimes used because of space constraints:

  • "J" Japan or Japanese
  • "C" China or Chinese
  • "WG" indicates that I made the estimate or contributed to it.
  • "A-P" Asian-Pacific War, 1931-1945.
  • "KIA" Fighting men killed in action.
  • "WIA" Fighting men wounded in action.
  • "~" Substitute for the word approximation.
  • A comprehensive index to all Web Tables can be found at the: Web Tables Listing

    Table 1: The Scale of World War II, 1931-1945

    Table 2:

          World War II Deaths, 1931-1945

          Asian-Pacific Death Share, 1931-1945

    Table 3: Allied Deaths in Asia and the Pacific by State

    The approach was to locate as many sources as could be found of documented and anecdotal evidence to average or make judgments as to the most accurate approximation of the number of dead for each subgroup or state. There are usually many sources but because of space limitations one or two of the most important are listed. For China, Rummel's average of 11 estimates of Chinese total military and civilian deaths (14 million) for 1937-1945 is used and the 1931-33 toll is added. The Communist and Nationalist military deaths from all causes are entered and subtracted to arrive at the first order of the civilian toll. Rummel's numbers do not fully deal with war and Japanese occupation caused famine and disease deaths, and tend to include those for Chinese civil war and Nationalist mismanagement deaths. Adjustments are made. The famine estimates from Web Table 8 are added and Chinese civil war and Nationalist mismanagement deaths, also from that table, are subtracted. This provides the total estimate for Chinese war toll attributable to Imperial Japan's aggression from 1931-1945. The following web Table 4 corroborates the total thought the process of building up a more detailed estimate. There are probably millions more who perished prematurely after 1945 from war causes that are not included in this total.

    The Philippine toll involves two sources that provide estimates of a million deaths. However, most other available evidence suggested a figure about half this number when all war violence, Japanese massacres including the sacking of Manila and anti-guerilla actions, plus the war's disruptive effects on the economy are accounted for. While Luzon bore the brunt of the Japanese invasion, other islands also paid a price. Among them were Leyte, Panay, Cebu, Mindoro and Mindanao.

    High Korean numbers reflect forced or low wage labor at the toughest and more dangerous jobs, e.g., mining, and neglect in Korea, Manchuria, and Japan. It includes losses among Koreans conscripted into the Imperial Army or into civilian support organizations, including comfort women sex slaves.

    Table 4: Allied Deaths by Cause

    In the China count, an adjustment is made to reduce the possibility for double counting the dead in the first eight line items and the three general category line items that followed. The sum of the three (93,000 atrocities + Germ and gas warfare + Civilians bombed) are multiplied by 65% and then added to the preceding eight to arrive at the Subtotal, Civilians from Violence (adjusted) line. Thirty five percent of the deaths are assumed to have already been included in the preceding eight, leaving 65 percent as attributable to the large number of remaining 1931 to 1945 battles and war crimes. This adjustment is based on considerable analysis of all the wartime combat and atrocity accounts occurring over the 14 years of war and occupation.

    The Chinese Refugee death estimate is based on considerable anecdotal evidence that establishes that a large number of the elderly, very young, weak, pregnant, and infirm succumbed to the rigors and privation of refugee flight and homelessness. That particularly vulnerable group made up at least 30 percent of the refugees. It is thus assumed that at least 10 percent of the approximately 35 million refugees lost their lives as shown. The impact on the villages and cities the refugees past through and competed for shelter, food and medicine must have also paid a price in lives from shortages. The distribution of Chinese lives lost in the two periods, 1931-1941 and 1942-1945, where it was not clear, is judged based on historical evidence.

    This Table arrives at the total Chinese death toll through a more detailed accounting of causes, and corroborates the total arrived at in Web Table 3.

    Table 5:

          Allied Deaths by Cause (Another Display)

    Simply a more easily readable version of the Table 4 analysis.

          Allied Deaths by Cause Summary

    Arranges the Table 4 data by major Asian and Western group and major death categories. This provides the reader with a quick reference to the number of Allied dead by major category and location.

    Table 6:

          Allied World War II Global Deaths, 1931-1945, Summary

          World War I Deaths, 1914-1918

          Allied POW Deaths Under the Japanese

    Table 7:

          Allied Severely Affected Casualties (Surviving)[2]

    This attempts to analyze and estimate the scale of suffering within the invaded states. The author believes the estimates are conservative and the real numbers of humanity who suffered severely are probably much larger. The total could easily double when the impact on related families and communities of these casualties are considered.

          Death Estimate Range

    Provides the reader with examples of the range of macro numbers found for nations and groups and the aggregate corresponding numbers I arrived at.

    Table 8:

          Chinese Famine Victims of Japan's Invasion

    This lists the major Chinese famines caused by the war and Japanese economic policies and natural famines that could not be relieved because of the war. The analysis assigns responsibility between Japanese aggression and the Chinese Nationalists' mismanagement effects. Under the "Number Savable" column, 100% indicates Japan's war was the primary cause. Smaller percentages indicate natural famines and the percent that possibly could have been saved in peacetime from studies of prewar famines. "Japanese Controlled" column indicates the famine region invaded or the percentage Japan controlled during the time of the natural famine. "Japan's Fault" and "Chinese Fault" percentages are assigned. Though Japan occupied less than half a particular region, it is often assigned greater responsibility because its war disrupted relief. For example, in the 1942-43 Honan "Famine Province", the 3 million dead X 30% "Savable" = .9 million X 70% judged "Japan's Fault" = "Japan's Responsibility" of ~630,000. In turn the 30% "China's Fault" = "Chinese Responsibility" of ~270,000.

    Particularly useful in the analyses are maps of Chinese population density and percentage of cultivated land where the Imperial Army operated or threatened, and where famines took place (Dear & Foot eds. 1995: 213 & 230).

          Chinese Civil War & Mismanagement Deaths, 1937-1945

    Primarily attributable to China's internal disarray. These circumstances and the reason for them are explored in the Imperial Japan's World War Two..., Chapter 11.

          Chinese Opium Addiction from Japanese Sales

    Table 9:

          Forced Civilian Laborers and Deaths

          Population Under Japanese Occupation

    Shows the population numbers by nations occupied over time and expected still to be occupied if the war had not ended when it did.

    Table 10: Rape Victims; Torture Victims; Child Victims

    Shows the calculations undertaken to arrive at a rough estimate of the number of people directly affected from well-documented occurrences, but for which estimates could not be found. The calculated numbers of rape and torture victims are probably conservatively low. But by being low they reduce the chance of double counting with other major categories listed in Table 7.

    Table 11: Japanese Casualties

    The reader's attention is drawn to the wide range of uncertainty as to Hiroshima and Nagasaki death totals. Note that the deaths caused by U.S. strategic bombing, including the A-bomb, represent only approximately two percent of all Allied civilian loss of life resulting from Japan's aggression.

    Table 12: War Deaths Per Year and Week

    There is ample data to show the casualty rate was increasing as the war progressed from 1931 to 1945. This made it possible to show the approximate number of annual deaths. Allocations by year are made as follows: The total Chinese civilian and military losses are distributed about equally from 1931-1933 and 1937-1941. From 1942 to 1945 the more extensive data makes the greater difference by year clearer.

    The U.S. military records provide a breakdown of U.S. combat fatalities for all services by year. These percentages are used as the best approximation of Allied annual losses not otherwise locatable. Available Chinese Nationalist combat data by year made possible the rough allocation of both the deaths in the Chinese Nationalist and Communist forces.

    Asian war-caused famine victims are distributed using the dates during which they occurred. Most Allied Asian civilian losses from other conflict causes are distributed evenly by year. The exception is the civilian losses from the Japanese 1944 China Ichigo Campaign and the U.S. 1945 retaking of Manila and its sacking by the Japanese. The civilian losses, in each case equal to those at Hiroshima, are entered in the years they took place. On the Japanese side, the military losses per year were not available but are assumed to grow proportionally to those of the Allied side. Japanese civilian deaths occurred in large numbers only after mid 1944 and primarily in 1945 and are so entered. This reflects primarily the civilian loss during the battle for Okinawa and the bombing of the Japanese Home Islands. They exclude the 1945 Atomic bomb deaths in order to show the death rate prior to the bomb being used.

    The Allied Severely Affected surviving casualties [2] by year can be approximated for the Chinese by multiplying total yearly Chinese death figures by a factor of 6 from 1937 to 1945 and the total Allied deaths by 3.4 from 1942 to 1945 (From Table 7). The Japanese Badly Harmed factor from 1942 to 1944 averages 4; however the ratio in 1945 is about 9 because of the impact of the battle for Okinawa and the U.S. bombing of Japanese cities. (From Table 11)

    Deaths per week calculation are based on 52 weeks per war year from mid 1931 through 1944. In 1945, however, the number of major periods of deaths are assumed to have differed. The majority of military deaths including those from wounds among all belligerents took place from January 1 through the end of August, 1945. Therefore 8 X 4.35 (weeks per average month) = 35 week divisor. The majority of Japanese civilian deaths from Okinawa, the bombing of the home islands, plus other war related deaths took place from March through the end of August, therefore over six months or 26 weeks. Allied civilians are assumed to have died through December 1945 because of the lingering effects of war-caused privation and thus their 1945 deaths are divided by 52 weeks.

    Table 13: War Deaths per Year Footnotes

    Provide the bulk of the data, sources, and calculations that distributed the casualties over the 14-year war period shown in Table 12.

    Table 14: Japanese and Allied Death Ratios

    Provide the length of major battles and the death numbers and ratios of the Japanese and Americans.

    Table: 15

          Japanese Deaths in the Pacific Ratios

    Display the high rate and percentage of Japanese deaths in the Pacific island combat. Here, withdrawal was usually not possible and maneuver was constrained; on average 87 percent gave their lives for the emperor and nation, most fighting to the death but some committing suicide.

          Other Ratios

    These are useful in making comparisons and calculations.

    Table 16:

          Manchuria/Jehol, and Shanghai Battles, 1931-6

    Tables 16-18 show the military and civilians dead statistics from major Japanese and Chinese army clashes that are displayed with the calculations and sources. Ratios from Table 15 are used to break out approximate Chinese and Japanese deaths when only total casualty (dead + wounded) numbers could be found.

          Sanghai, Yangtze Delta Battles and Nanking, Aug. 1937-Jan.'38

          Hankow/Wuhan Region Battle, June-Oct.1938

    Table 17:

          Changsha, Three Major Battles,1939-1942

    Represents my rougher calculations of the Chinese civilian and military deaths based on more diverse data and analyses from other battle statistics.

          "Kill all, Loot All, Burn All", 1937-1945

          Chinese Counter Offensive, Nov. 1939-March, 1940

          Japanese Ichigo Offensive, May-Nov. 1944

    Includes the fourth battle for Changsha.

    Table 18:

          Allies in Burma, 1942-45

          Chinese Military Casualties, 1931-1945

          Major Chinese Resolute Actions & Victories Against Japanese

    Table 19:

          Japanese Military Deaths, 1931-1945

          Japanese Army Troop Dispositions

    Table 20:

          Projection of All Casualties Post Aug. 1945, W/O A-Bomb Use

    This provides the step-by-step projected death estimate calculation for each added month of war from August through December 1945. This shows the bill for each added month the war might have lasted had it not ended when it did. "Japanese Troop Numbers Faced" in August 1945 are taken from Table 19. Estimates are made of the number of "Battle Months", percent of "Japanese Fighting to Death", and "Ratio of Japanese to Allied Deaths" based on Table 14 and 15 data. From this the "Resulting Japanese" and "Allied (military) Dead" for the total battle months per offensive are calculated and summed as shown under "Scheduled Offensives Total" and "U.S. Invasion of Kyushu (Operation Olympic) Total".

          Projected Lives Spared by A-bomb From the Above Data

    This displays the second step of the estimating process. Scheduled "Offensive Total" and "U.S. Invasion of Kyushu Total" are brought forward and divided by the expected months of combat, yielding the projected monthly rate of death. The invasion of Kyushu deaths start in December because of the October typhoon is assumed to have delayed the invasion date from November to December 1945. The "Allied Civilian" number is taken from Table 12, "Total Allied Civilian" count in 1945 and divided by 12 months to establish a projected monthly rate. The "Japanese Civilian" number from the same Table "Japanese Civilian Pre A-bomb 1945" count is divided by the last six months Japanese large civilian death rate period to arrive at their projected monthly death rate. Finally, the four major death rate projections are summed by month with a cumulatively total for every added month of war.

    Table 21:

          U.S. Military Estimates of Casualties to Invade all of Japan

          U.S. Estimates for Invasion Deaths in First 30 Days

    Both provide the estimates from various sources for the U.S. human toll to invade the Japanese Home Islands. It is important to note that these casualties are projected based on a June 1945 intelligence assessment of Japanese troop strength. By August the estimated number of troops to be encountered on Kyushu, the first Japanese Island to be invaded, had doubled. I could find no estimates of expected Japanese losses.

    Any estimate of U.S. military deaths can easily be multiplied by 20 to 100 times to arrive at an estimate of Japanese who would have been killed based on the American to Japanese death ratios in the Philippines and Burma (Tables 14-15). The fact that the defenders would include civilian militia armed with the crudest of weapons and used in a suicidal role would have assured a civilian massacre. The massiveness of U.S. naval gun bombardment and wave after wave of destruction-laden B-29 bombers, along with the rest of the U.S. massive airpower fleets, would have pulverized military and civilians alike opposing the American landing. Assuming a hypothetical 50,000 American combat deaths, the Japanese share could easily have reached well over 1,000,000 killed.


    1. Cornell, 1976, 170, for all famine dates. Other sources provide similar dates and casualty levels.

    2. A reminder: the Allied Severely Affected include the wounded, maimed, raped, tortured, victims of biological war, forced/slave labor, refugees, war orphans, and widows, Japanese-caused opium addicts, victims of severe war-caused malnutrition and diseases, POWs, internees, and most Asian military conscripts less the dead and wounded. This does not account for the family relations and communities traumatized by these casualties. Nor does it account for the communities over run by desperate refugees seeking help when little was available. The Japanese Badly Harmed included basically the wounded, maimed, refugees and homeless, war orphans and widows, the malnourished, and men conscripted into the armed forces less the dead and wounded.

    A comprehensive index to all Web Tables can be found at the: Web Tables Listing

    Available at bookstores worldwide, on the web or from the publisher.

         Barnes & Noble and Borders bookstores.
         Toll Free Telephone: 1-888-999-6778

    Hardcover, 254 pages

    Return to Main Directory

    Copyright © 2006 Werner Gruhl