US Decision Making: Formosa or Luzon?
One of the trickiest problems facing US strategic planners during World War II was what to do as the two main strategic drives across the Pacific approached the Japanese Home Islands. One was commanded by Admiral Chester Nimitz (Central) while the other was commanded by General Douglas MacArthur (Southwest). As the two drives entered the Western Pacific, the choice facing the US was whether to invade either the island of Formosa (now known as Taiwan) or the island of Luzon, part of the Philippines. The decision was eventually made by the US Joint Chiefs of Staff (JCS), who, by agreement of the US-British Combined Chiefs of Staff, were responsible for the conduct of the war in the Pacific. It was made after a long debate with discussions involving senior commanders, taking into account a number of factors, not least the current state of logistical support available within the Pacific Theatre as well as Allied operations in other theatres, especially Europe. It also considered the plan that had been made after the Pearl Harbor disaster that befell the US Pacific Fleet, rendering obsolete the inter-war plans for conducting hostilities with Japan. The plan was not written in stone, no plan should be, but it helped guide the discussions that took place for about eighteen months over the relative priorities of both Luzon and Formosa as objectives in the Allied drive towards the Japanese Home Islands.
The plan recognised that to end the war in the Pacific, the Allies may well have to invade the Japanese Home Islands. A prerequisite to such an operation would be a sustained, intensive strategic bombardment, combined with aerial, naval and submarine resources in order to cut Japan off the resource-rich territories it had acquired early in the war. Early on, the JCS believed that the best way to accomplish this bombardment was to establish extensive airfields in eastern China. In order to do this, it would be necessary to seize at least one major port on the southern Chinese coast, to replace the poor overland and aerial routes from India and Burma, the then main routes for men and material moving into China. To secure such a port on the Chinese coast and simultaneously cut Japan's lines of communication (LOC) meant that the Allies would have to seize control of the South China Sea. This in turn, the JCS realised, would require the seizure and development of a large number of sizeable air, sea and logistical bases in a strategic triangle formed by the southern Chinese coast, Luzon and Formosa. To move safely into this triangle, the JCS realised, would mean the Allies would have to secure air bases in the central and southern Philippines from which to neutralise Japanese air power on Luzon. They might also have to develop bases in the central and southern Philippines with which to mount amphibious attacks against Luzon, Formosa and the southern Chinese coast. Therefore, in accordance with this plan, the Allies had advanced westward towards this strategic triangle along two axes of advance, the Southwest Pacific Area (MacArthur) which had driven up through New Guinea to Morotai Island and the Central Pacific Area (Nimitz) which had driven through the Gilbert, Marshall, Mariana and Palau Islands.
Formosa: Relative Importance
After studying various plans for the Allied entry into the strategic triangle, the JCS concluded that Formosa was possibly the single most important objective in the target area. The island possessed many advantages and was located in such a strategically important position that many planners thought that the Allies would have to occupy it no matter what other operations they might conduct in the Western Pacific. Until the Allies seized Formosa, the Allies would have great difficulty in establishing a sea-based supply route to China. Plus, Allied air and naval forces could cut Japanese LOC to the south much more effectively from Formosa than they could from either Luzon or the southern Chinese coast alone. In addition, the USAAF's new B-29 bombers could carry heavier payloads against Japan from Formosa than they could from the more distant Luzon. Many considered Formosa such a valuable strategic objective that a great deal of effort was expended in looking at the possibility of bypassing the Philippines entirely, in favour of a direct assault against the island. This proposal generated a great deal of discussion, which increased and decreased in intensity during 1943 and 1944, despite the fact that the overall plan called for the seizure of bases in either the central or southern Philippines as a prerequisite for moving on to the Chinese coast-Formosa-Luzon triangle. Discussion in and between the War and Navy Departments found both internally divided. Admiral Ernest J. King, Chief of Naval Operations and US Navy representative in the JCS was a major advocate of plans to bypass the Philippines. Other US Navy commanders, such as Nimitz, favoured at least a partial re-occupation before striking at Formosa as they feared it would be impossible to establish and secure Allied LOC to Formosa without being able to base Allied air power in the Philippines to neutralise Japanese air power on Luzon.
General George C. Marshall, Chief of Staff and US Army representative on the JCS was relatively inactive in the debate until late 1944, but at one point seemed to favour bypassing both the Philippines and Formosa for a direct assault on Kyushu, one of the Japanese Home Islands. Some officers high in the Army councils, including Lt Gen Joseph T McNarney (Deputy Chief of Staff), strongly advocated bypassing the Philippines in favour of Formosa while General Henry H Arnold, USAAF representative on the JCS, also seems to have been in favour of bypassing the Philippines during much of 1943 and 1944. Other US Army officers, such as Chief Logistician Lt Gen Brehon B. Somervell, were in favour of taking the entire Philippines before making a more elsewhere, including Formosa. Naturally, General MacArthur was strongly in favour of liberating the Philippines and was supported by many other officers in the Pacific.
In March 1944, the JCS had directed MacArthur to be ready to move against the southern Philippines before the end of the year and to plan for an assault against Luzon during February 1945. At the same time, they ordered Nimitz to prepare plans for an assault against Formosa, again during February 1945. These orders, which left unclear the relative priority of Luzon and Formosa, seemed to settle the question of liberating the Philippines but in mid-June, the JCS reopened the question of whether or not to bypass the archipelago. Developments in the war at a strategic level between mid-March and mid-June 1944 tended to support those planners who wanted to bypass the Philippines. The Allies received intelligence that the Japanese were reinforcing many of their Western outposts, including Formosa. Thus, the longer the Allies delayed an attack on Formosa, the greater the eventual cost would be. Some Army planners thought that the Allies could strike Formosa in November 1944 if the JCS decided to bypass the Philippines immediately. In addition, the JCS were worried about the possibility of a collapse in Chinese resistance, the only way to avert such a development being the early seizure of Formosa and a port on the Chinese coast without taking any intermediary operations in the Philippines. Feeling the need to in some way accelerate the pace of operations in the Pacific following the Normandy invasion, The JCS approached Nimitz and MacArthur on 13 June 1944 to consider the possibility of bypassing all the objectives already selected in the Western Pacific, which included both the Philippines and Formosa. Neither commander was particularly happy with this and reiterated that any advance beyond Palaus-Morotai line would need to be supported by airbases in either the central or southern Philippines. Such an assertion was backed up by additional staff work, which highlighted the need to seize parts of the Philippines before advancing on either Formosa or Luzon …. directly advancing on Japan was at this stage, a non-starter. Both commanders again emphasised the importance of firmly establishing bases in the southern or central Philippines before either Luzon or Formosa could be attacked when they met at Pearl Harbor with President Roosevelt in late July 1944. No major planning decisions were made at these conferences and the debate over whether to go for Luzon or Formosa rumbled on for the next couple of months. Only the decision to seize bases in central or southern Philippines before moving on elsewhere was confirmed.
Luzon: The Different Views
Of all those advocating that the Allies should secure Luzon before any further advances towards Japan, the most vociferous was General MacArthur. He believed that Luzon was a more valuable strategic objective than Formosa, in direct opposition to the JCS, and thought that the Allies would have to occupy the whole of the Philippines archipelago in order to sever Japan's LOC to the south. It would also provide valuable air and logistical support to any invasion of Formosa however, he also suggested that taking Luzon could very well nullify the necessity of striking at Formosa and be cheaper in terms of time, men and material. He also considered that bypassing part of the Philippines may well cause additional hardship to the civilian population plus liberating the entire archipelago was a national obligation and political necessity – bypassing any of the Philippines would have international political implications later on. Of course, just as MacArthur was the strongest proponent for striking at Luzon, Admiral King was the strongest advocate for striking at Formosa. He believed that attacking Luzon before Formosa would delay more important operations in the north, plus taking Formosa first would facilitate operations against Luzon. In addition, the Allies could not secure a port on the Chinese coast (and therefore better support the Chinese Nationalists) until they had seized Formosa. Finally, he suggested that, if the Allies bypass Formosa, then the principal objective in the western Pacific for the Allies should be Japan itself, not Luzon.
MacArthur believed the plans to bypass Luzon were Navy-inspired but the fact is that both the War and Navy Departments were as internally split over the question of whether to bypass the Philippines as they were over the Luzon vs Formosa debate. For example, until at least mid-September 1944, General Marshall had favoured the Formosa-first strategy and like Admiral King, had considered that Japan itself should be the next objective (rather than Luzon) if Formosa was to be bypassed. Most Army members of the JCS' subordinate committees held a similar view and consistently pressed for an early decision favouring Formosa, alongside army air force planners who expressed an interest in Formosa to site B-29 bases. On top of that, Admiral Nimitz went on record until late September that he favoured a Formosa-first strategy. However, it seems that his staff were less-than enthusiastic about the plan and it seems that Nimitz himself was growing more lukewarm towards the idea of attacking the island as time went on. Nimitz had been somewhat at odds with King over bypassing the Philippines archipelago and it might have ben that his support for the Formosa-first strategy at least partly stemmed from his deference to King's judgement. A possible indication of this is that Nimitz's staff were developing plans to attack Okinawa as a substitute for Formosa well before such an operation became a serious consideration in Washington.
Next in line in the US Navy's chain of command was Admiral William F Halsey, who commanded the Third Fleet (and until 15 June 1944, the South Pacific Area) and was extremely opposed to any Formosa-first strategy – he wanted to seize Luzon and bypass Formosa to strike at Okinawa. Further down the chain of command, most other senior Army and Navy commanders in the Pacific favoured a Luzon-first strategy, bypassing Formosa. This included Lt Gen Robert C Richardson (US Army), Lt Gen George C Kenney (US Army Air Force) and Vice Adm Thomas C Kinkaid (US Navy). However, by this stage, the only member of the JCS to now support a Luzon-first strategy to now support a Luzon-first strategy was Adm William D Leahy, the President's Chief of Staff. It is worth noting then, at this point, most of the ranking Army and Navy commanders in the Pacific (with the possible exception of Nimitz), the officers who were responsible for executing or supporting operations, opposed attacking Formosa. The majority favoured a series of operations capturing first Luzon, then a attack either on Okinawa or striking directly at Japan itself. This contracted with most of the high-level planners in Washington (the main exceptions being Leahy and Somervell) who thought the Formosa-first strategy was better strategically. This, along with several other factors, brought the Luzon vs Formosa debate to a climax in late September 1944.
The first factor to bring this debate to a climax was a major change to the target date set for the assault on the Philippines. Until mid-September, MacArthur's plans had called for an initial assault on 15 November 1944 on southeast Mindanao, followed by the main attack to take place on 20 December 1944 on the island of Leyte. On 15 September, MacArthur cancelled the preliminary operation against Mindanao (with the approval of the JCS), favouring a direct move on 20 October 1944 from the Palaus-Morotai line onto Leyte. Not long after this change of schedule, MacArthur informed the JCS that he would be able to push on from Leyte and attack Luzon itself on 20 December 1944, two months earlier than the date set out in the existing plan. Although this new schedule could well allow the Allies to conduct an attack on Formosa on the date already scheduled, MacArthur argued that the seizure of Luzon would render the assault on Formosa unnecessary. The new schedule had much to recommend it to the JCS. The sequence Leyte (20 October) – Luzon (20 December) – Formosa (possibly, 20 February) would allow the Allies to maintain a steady pressure on the Japanese, but should the Allies drop Luzon, it would allow the Japanese to realign their defences against the next operation in the sequence, during the pause. Plus, dropping Luzon would not allow the Allies to advance the operation against Formosa – logistical problems would make the mounting of an invasion of Formosa unlikely before late February anyway. Meanwhile, the Formosa-first strategy was losing favour. Plans developed in Washington called for the entire island to be captured before US forces would land on the coast of China to seize the port of Amoy, but Nimitz's latest plans called for a simultaneous strike against both Formosa and the Amoy area, with the intention of only occupying a part of the island initially, with the remainder being occupied after Amoy had been secured.
Army planners soon decided that Nimitz's plan had major drawbacks. The Japanese were hardly likely to just sit back and allow the Allies to operate freely from southern Formosa. There would be counterattacks from the remaining Japanese forces in the north, possibly with reinforcements from the Chinese mainland. In addition, in past campaigns, the Japanese had found it difficult to move against the newly taken islands due to the distances from the next major base, but that wouldn't be the case with Allied bridgeheads on southern Formosa and around Amoy. Also, the Allies simply did not have enough air and naval power to neutralise all the Japanese airfields in range of southern Formosa and Amoy, as well as try and cut off the flow of reinforcements. With these drawbacks in mind, Army planners believed an attack on Formosa would be impractical and lead to a protracted, costly campaign to secure both Formosa and a large portion of the Chinese mainland. A campaign of such magnitude would inevitably delay the overall advance on Japan and be a major drain on (limited) Allied resources. In further studies of the manpower needed for such an operation, the use of more recent Army intelligence estimates of Japanese strength in the Formosa-Amoy region indicated that the Allied invasion force would need many more combat units that initially estimated by Nimitz’s staff. This therefore would necessitate an expansion of the combat support and combat service support troops backing them up. Planners looked at ways to expand the service support forces available. One suggestion was to advance the target date to attack Formosa and take service forces from the Southwest Pacific Area. However, MacArthur’s command was already short of service troops and taking any away might jeopardise the success of the Leyte operation and would certainly immobilise MacArthur’s forces in the central Philippines area until long after Nimitz had secured the Formosa-Amoy region. Although the Formosa-Amoy and Luzon operation would require roughly the same number of combat troops in the assault phase, MacArthur could count upon hundreds of thousands of Filipinos to augment both his combat and service forces. There would be no similar source of friendly manpower on Formosa.
By mid-September 1944, there were so few service units left in the USA that the only way Army planners could see to solve the manpower shortage was to await redeployment from Europe. Both they and the Joint Logistics Committee estimated that Nimitz would be able to launch the Formosa-Amoy operation by the beginning of March 1945 if the war in Europe ended by the beginning of November 1944. Even then, many believed the preparations would only be complete by the beginning of March 1945 if the JCS immediately cancelled the Luzon operation to allow for an early, sustained and uninterrupted build-up of resources for Nimitz’s campaign. On the other hand, many were convinced that MacArthur would be able to advance on Luzon regardless of what happened in Europe. Army planners were less optimistic than they had been about an early end to the war in Europe (probably prompted by the apparent Germany recovery after the disaster at Falaise, leading to the failure of Operation Market Garden) and reasoned that it would be unsound to make plans based around a German collapse by the beginning of November 1944. In addition, Army planners believed that the Formosa-Amoy campaign would tie up so many combat troops, ships, landing craft and aircraft that any campaign to take Luzon would not be able to commence until November 1945, with any further step taken towards Japan, such as the invasion of Okinawa, similarly delayed. A Luzon-first strategy would also be safer logistically, as any LOC to Luzon is far shorter and easier to protect than any to Formosa, especially if Luzon remained in Japanese hands. On top of that, Admiral Leahy believed that while the Formosa-first strategy might ultimately hasten the end of the war in the Pacific, the Luzon-first strategy might be cheaper in terms of men and resources. By mid-September, he and many Army planners favoured the longer course at cheaper cost option, while MacArthur still regarded the Formosa-first strategy as the longer course at more cost strategy. General Marshall also began to have misgivings about the cost although he remained convinced the Formosa-first strategy was the sounder strategically. Admiral Nimitz expressed no firm opinion on the relative costs of the campaigns but commenting in a ‘roundabout’ way, thought that an attack on Luzon after Formosa need not significantly delay the course of the Pacific War. If Formosa came first, he argued, MacArthur’s campaign on Luzon would probably be easier and less costly. Admiral King also thought that the Formosa-first strategy would save time and be less costly over the long term.
Meanwhile, events in September elsewhere were to have a profound effect. While the high-level discussions continued apace in Washington, the Allied position in China was deteriorating. General Joseph L Stillwell, Commander of the Allied Forces in China, Burma and India and Allied Chief of Staff to Generalissimo Chiang Kai-Shek reported to the JCS that Japanese offensives in eastern and south-eastern China were overrunning the last airbases of the US Fourteenth Air Force that could effectively be used to support an offensive against either Luzon or Formosa. This upset the plans of the USAAF which were looking to expand the airbases in Eastern and south-eastern China to allow B-29s to stage from there so they could attack targets in Japan, Manchuria, Formosa, and Korea, as well as much of the tactical air support needed for the actual invasion of Japan. The airbases in eastern China now looked to be irretrievably lost and the Allies could not afford to divert the manpower necessary to recapture them. The requirement to seize and develop a port on the Chinese coast therefore lost much of its urgency as there was no longer a need to have one to facilitate the development of the airbases and therefore the necessity to capture Formosa, to facilitate the capture of a port on the Chinese coast also lost much of its urgency. This heralded a major shift in the thinking of the Navy planners forcing them to reconsider the Formosa-Amoy strategy. Without the seizure of a mainland port, Formosa by itself lacked the anchorages, port facilities and logistical bases to support a large fleet and large-scale operations in the western Pacific. If the need to seize a mainland port had evaporated, was there any need to seize Formosa? The Army planners thought not. The Army Air Force planners though were more hesitant. There was no question that the B-29s would operate more effectively against Japan from northern Formosa than they could from Luzon, the Mariana Islands or western China but would be equally effective from southern Formosa as they would be from the other basing areas. Indeed, Saipan and Tinian (in the Marianas) were closer to Tokyo than Nimitz’s proposed base area in southern Formosa, plus they were secure from Japanese air attack. Even northern Luzon, some 200 miles further from Tokyo than southern Formosa had certain advantages – it had more room for B-29 airbases and was safe from air attack. Even assuming Nimitz could achieve the optimistic target date for the invasion of Formosa (1 March 1945), B-29s would not be able to begin operations until the late spring or early summer and the Army Air Force was planning on commencing B-29 operations from the Marianas before the end of 1944. Finally, there were political considerations. MacArthur’s warning that it would be disastrous for American prestige to bypass any part of the Philippines had merit and Admiral Leahy shared his point of view. Given his close proximity to President Roosevelt, it can be assumed that his colleagues in the JCS gave Leahy’s opinion considerable weight.
By the end of September 1944, the weight of military opinion had shifted away from favouring a Formosa-Amoy first strategy to favouring a Luzon first strategy, and thereby bypassing Formosa, ignoring the ports on the Chinese coast and moving on to Okinawa. The only major figure who still favoured going for Formosa-Amoy was Admiral King. He therefore took a negative tack and pointed out possible problems with the Luzon strategy in a broader sense. He argued that the Luzon campaign, as MacArthur had planned it, could tie up all the Pacific Fleet’s fast carrier task forces for at least six weeks while they protected the initial beachhead, provided close air support to the ground forces, escort reinforcement and resupply convoys as well as neutralising Japanese air power in the wider region. To tie up these mobile assets for so long was impractical and there fore unacceptable to the US Navy. MacArthur counter this by arguing to the JCS that the only requirement he would have for carrier air power after the initial assault would be a few escort carriers to remain on station to provide close air support while the engineers prepared an airfield to host land-based air power. In addition, only the initial assault convoys would be routed through the dangerous waters north of Luzon and require protection from the fast carrier task forces – later convoys would come through the Central Philippines and be protected by land-based aircraft from Mindoro Island, south of Luzon. As a follow-up, he argued that the fast carrier task forces would be tied to a single region much longer if Formosa-Amoy were attacked, especially if Luzon remained in Japanese hands. Not only did this take out much of the wind from King’s sails but Nimitz withdrew what little support he still gave to the Formosa-Amoy strategy after concluding that sufficient ground forces would not be available to execute the operation in the timescales needed. At the end of September 1944, he gave his backing to the Luzon strategy, proposing that the Formosa strategy be dropped, at least temporarily. He also presented a plan for a series of rolling operations to maintain continuous pressure against the Japanese and advance Allied forces towards the Japanese Home Islands.
King accepted Nimitz’s recommendations with one last reservation. He felt that the dangers involved in routing the assault convoys into the area between Formosa and Luzon were of such magnitude that the approval for such action should come directly from the JCS and raised similar objections over the fast carrier task forces operating in the same restricted waters. The other three members of the JCS agreed to leave a decision on these matters to Nimitz and MacArthur, something which King finally accepted. With King’s change of position, the JCS finally achieved the unanimity that taking major strategic decisions required. On 3 October 1944 they directed General MacArthur to invade Luzon on or around the 20 December 1944 and Admiral Nimitz to attack both Iwo Jima and Okinawa on the dates planned. Nimitz would provide fast carrier task forces to provide air cover for the invasion of Luzon, while MacArthur would provide as much air cover as he could from Luzon for the assault on Okinawa. Both commanders would coordinate their efforts with the B-29 units then operating in the Pacific and from India, along with General Stilwell and the Fourteenth Air Force in China.
Interestingly, the JCS did not formerly cancel the Formosa operation. Instead, they left the final decision over the seizure of the island in abeyance, and the question of the seizure of Formosa as a wartime operation never came up again around the higher echelons in Washington DC. The JSC had not arrived at the decision to bypass Formosa, attack Luzon, and then effectively substitute Okinawa for Formosa easily. At the start, they had believed that seizing Formosa and a port on the Chinese coast, essentially bypassing Luzon, was the best course of action in the Pacific. Eventually, they had realised that the Allies could not assemble the forces and material needed to undertake the strategy until after the war in Europe had been concluded. However, they could not delay operations in the Pacific until Germany had been defeated, something which could be (and was) many months away. Logistical considerations alone then, dictated that the JCS change strategy to one favouring Luzon, although other factors, both military and political certainly played a role in influencing the formulation of strategy in the western Pacific. The directive sent out on 3 October 1944 ended months of uncertainty for Allied forces across the Pacific Theatre …. Luzon was to be the next objective; Formosa would be bypassed. The entire Philippines archipelago would be recaptured in a series of linked operations.
Miranda, J. (2022) 'Operation Causeway: The Planned Invasion of Formosa, 1944' in World at War, Issue 83 (April-May 2022), pp. 7-22.
Smith, Robert. (1960) ‘Luzon versus Formosa' in Greenfield, Kent. (1960) Command Decisions, CMH Publication 70-7, GPO S/N 008-029-00482-8, Chapter 21, pp. 461-477, available at https://history.army.mil/books/70-7_21.htm, as of 2nd October 2022.
Wikipedia. (2022) 'United States Color-Coded War Plans', last edited 2nd August 2022, located at https://en.wikipedia.org/wiki/United_States_color-coded_war_plans, as of 15th August 2022.
Wikipedia. (2022) 'War Plan Orange', last edited 4th August 2022, located at https://en.wikipedia.org/wiki/War_Plan_Orange, as of 15th August 2022.
The map of Formosa showing POW camps was sourced from https://ww2db.com/images/5458685d745ad.jpg and is listed as being in the Public Domain under US copyright law.
'Map of the Empire of Japan in 1914', URL: https://nzhistory.govt.nz/media/photo/map-empire-japan-1914, (Ministry for Culture and Heritage), updated 14th August 2014. The map is freely available as a PDF download.
Topographical map showing principal installations including air facilities on Formosa (Taiwan) sourced from https://www.history.navy.mil/research/library/online-reading-room/title-list-alphabetically/n/northern-formosa-pescadores.html, as of 7th February 2023.
The map of Formosa (Taiwan) in 1917 is listed as having an expired copyright under the copyright laws of both the People’s Republic of China and the Republic of China. It was sourced from here: https://commons.wikimedia.org/wiki/File:Map_of_Formosa_%28Taiwan%29_in_1917_from-_STANFORD_%281917%29_p19_PLATE_2._FUKIEN_%2814803676033%29_%28cropped%29.jpg.
Satellite image of Taiwan (Formosa): Geography of Taiwan. (2023, January 25). In Wikipedia. https://en.wikipedia.org/wiki/Geography_of_Taiwan.
The map of Kinmen, Xiamen and the surrounding area in November 1958 is a cropped version of a larger map, prepared by the US Army Corps of Engineers. As a work of the US Federal Government, it is considered to be in the Public Domain.