by KB, WG, CS Group Alphabet Soup
The dataset that our group is analyzing is Volume 25 (totaling 283 pages), which contains the press conferences that Henry Morgenthau Jr. held between November 4, 1943 and June 29, 1944. During the period of these press conferences, the United States was heavily involved in World War II, and the Treasury was responsible for national fundraising efforts for war operations. Some examples of the national fundraising strategies include selling war bonds, and increasing income tax. By analyzing the most frequently used terms, and when those words were most used in consideration to date throughout Volume 25, we hoped to learn more about the concerns, problems, solutions and public opinions that would add to our understanding of the plans the Treasury conceived in response to funding World War II operations.
All the pages of Volume 25, sans the index, were transcribed by our group using a 2-step process. Screenshots or photos of each page were first transcribed using the Optical Character Recognition (OCR) function in Google Image search or the Android camera app. The text extracted were subsequently cleaned up manually by comparing with the original text. To obtain the most frequently occurring topics related to national fundraising, our group fed the transcribed text into Voyant in order to obtain a list of most frequently occurring terms. Unsurprisingly, there were a number of fundraising, expenditure or currency-related terms that were on the list. We went through the list again and extracted the most frequently occurring currency or expenditure-related terms, namely, “gold”, “silver”, “tax”, “loan” and “bond”. With these results, we produced two data visualizations. The first data visualization shows a comparison of the frequency of these 5 terms within Volume 25. By mapping these five terms to the dates that were mentioned in the press conferences, we were able to produce another data visualization of the frequency of these 5 terms between November 1943 and June 1944.
Based on our data visualization, we found out that “gold” and “silver” were actually mentioned in the context of international financing, specifically in April 1944, in relation to the creation of the International Monetary Fund. The other terms like “tax”, “bonds” and “loans” were mentioned fairly consistently across this 8-month duration in relation to national financing. Our data visualizations can help researchers pinpoint specific time periods to conduct further research on to obtain greater understanding of how the United States financed World War II (WG).
After World War I, the United States (U.S) adopted an isolationist policy which was a policy of “disarmament and nonintervention” (Gopnik, et. al., 2024). However, as World War II began with Germany’s invasion of Poland in 1939, the U.S gradually began to move away from its isolationism policy. Key events like the Japanese attack on Pearl Harbor on December 7, 1941, played a significant role in the change in her political stance (Gopnik, et. al., 2024). Eventually, in 1943, the U.S became militarily strong enough to be able to participate in “large-scale offensive operations” (Gopnik, et. al., 2024). In order to fund these war operations, it was necessary for the United States to raise money using various economic strategies like selling war saving bonds to citizens, issuing loans to industries that supported war production, and pegging interest rates at low levels (Richardson, 2013) (WG). Another way to acquire funds for the war was to increase the national income tax. The American population paying taxes increased from 3% in 1939 to 30% in 1943. With this money, Morgenthau intended to pay for at least fifty percent of the war debt acknowledging that soldiers did not want to come home from fighting a war to then pay for it (November 4, 1943 Press Conference, page 23) (KB). According to the Institute of Economics and Peace report (2011), the United States had “one of the most significant short term increases in economic growth in the history of the U.S. economy” during the war years of 1941 - 1945, due to “government spending” and “declines in consumption and investment” (IEP, 2011, p. 7) (WG).
During a majority of the Franklin D. Roosevelt administration, from January 1934 to July 1945, Henry Morgenthau Jr. served as secretary of the Treasury. Before playing an important role in government and politics, Morgenthau had quit university to become a farmer. It was his close friendship with Roosevelt that led him to the role as head of the Farm Credit Administration and then, as secretary of the Treasury a year later (Ostrow, 1994) (KB).
Due to this lasting relationship, he was a known close advisor to the president. Morgenthau is credited with establishing and supporting many well-known and historically significant programs during his almost 10 years in this role including the War Bond Program, Works Progress Administration as part of the New Deal and the War Refugee Board in 1944 (Richardson, 2013). Unique to funding the second World War, Morgenthau set up a system that would allow the Federal Reserve to provide support for Treasury borrowing, to ensure financial stability, and allow the Federal Reserve to buy unsold bonds by the public at a specific rate, therefore increasing financial support of the war in a strategic manner (U.S. Department of Treasury, n.d.) (KB).
Our group chose to work with Volume 25, covering November 4, 1943 to June 29, 1944. We chose to focus on the subject of money and its many forms referenced in the press conference transcripts. We also chose to only present terms that were related to the national funding efforts of World War II. Due to the span and international impact of the war, we thought that concentrating on the United States’ funding plans and actions would contribute to research regarding American economic health and currency development during this time period (KB).
To begin the cleaning process, we broke the volume into three equal parts. Excluding the first twelve pages of the index, each section came out to be 95 pages per person. Because the document could not easily be converted to searchable text to be fed into OpenRefine we decided to extract the text and place it into a shared Google Doc (CS). Optical Character Recognition (OCR) software available on Adobe PDF was unable to extract the text from the images to a high degree of accuracy, so we tried different methods in order to extract the text (WG):
Each of these methods did not yield perfect results, so manual correction to formatting, spelling, and other noticeable errors was necessary. Though we encountered issues with each method, we all found them preferable and less time-consuming as compared to hand transcribing each page (CS).
After deliberating about what method would be best to support the data computation and transformation, our group decided to emphasize frequently occurring terms that provide a general insight into the data within Volume 25. We gathered these terms by inputting raw data from our individual data sets into the program, Voyant. After compiling our lists, we cross referenced our individual lists and added context by including the dates of the press conferences that these terms appeared in, i.e. November 1943 to June 1944.
We had originally considered using OpenRefine, but found that given the assignment window and working together remotely we could not find a solution as to what fields we would focus on. We considered transformation through significant phrases. We transcribed our sections of the press releases using the varying methods listed aboveredFrom here, we individually input our transcriptions into Voyant, a text analysis tool. We were able to choose a list of terms most used. Our only stipulation at this point was it had to be currency related.
We had originally considered using OpenRefine, but found that given the assignment window and working together remotely we could not find a solution as to what fields we would focus on. We considered transformation through significant phrases. We transcribed our sections of the press releases using the varying methods listed above.
From here, we individually input our transcriptions into Voyant, a text analysis tool. We were able to choose a list of terms most used. Our only stipulation at this point was it had to be currency related.
Soon after completing this phase, we agreed that we should narrow down our focus even more (KB). We selected 5 key terms that were the most closely related to the war funding efforts, i.e., “gold”, “silver”, “tax”, “loan” and “bond”. We correlated the frequency of these terms to the dates of the press conferences and created data visualizations for them (Figure 1) (WG).
To create visuals to chart our findings we pulled out various terms (variables) and plot them on graphs. Figure 2 on the right represents frequency of the top 5 terms relating to finance found throughout the volume. After feeding the sections of the volume into Voyant, Gold, Silver, Loan, Tax, and Bond appeared as the most frequent single variables. Plural forms of the words were combined with singular forms to better tally frequency. Gold and Tax were the most frequent variables throughout the volume, but to better examine the context of how these variables were being used frequency in relation to press conference date was also graphed. By creating Figure 2 on the left a narrative showing the rise and fall of variable frequency emerges. Gold appears with the most frequency during the April 1944 press conferences but is absent from those in November, December and March. Conversely, taxes which represent the most frequent variable in the data set appear more evenly throughout the conferences. Several tools to create the visuals were explored but ultimately pulling out the selected variables and plugging them into a Canva template yielded the most straightforward and eye-appealing results (CS).
While carrying out the data computation and manipulation for this paper, we have ensured that our processes and decisions made are in line with the Society of American Archivists’ Code of Ethics. The SAA Code of Ethics states that decisions made by archivists should be made “mindfully… to ensure the preservation, authenticity, diversity, and lasting cultural and historical value of materials” (SAA, 2024). We have also carefully documented each step of our workflow, in order to ensure full transparency. This is in-line with the Code of Ethics which states that “all collections-related policy decisions, including preservation treatments, descriptive work, processing activities, and access guidelines” should be carefully documented (SAA, 2024). Finally, during the transcription process, we have ensured a high level of intellectual integrity and authenticity by “not willfully alter[ing], manipulat[ing], or destroy[ing] data or records to conceal facts or distort evidence” (SAA, 2024) (WG).
Through our data computation and manipulation, researchers will be able to pinpoint more easily what keywords were major factors within Volume 25 in regard to national policies and programs related to funding the war – which can serve as access points to further research. America had been fighting in the war for two to three years during these transcripts and the data within provides insight into the stabilization of gold, history of tax and bond developments and national loans during this pivotal time in American history.
In looking more closely at the terms that can loosely describe the context of these transcripts we realized that the term gold was used mostly in an international sense regarding the amount that different countries could contribute to paying for the war and reimbursements for agreements made between nations. So, although gold was our second most overall used term, below tax, it doesn’t necessarily support our parameters of looking only at national funding. We concluded that this could possibly be due to the fact that gold was the standard of which almost all countries fixed their value to, in a system known as The Gold Standard (World Gold Council, n.d.). There were no restrictions placed on the import and export of gold. That changed after the second world war, however, when the Bretton Woods System became the standard. Henry Morgenthau Jr. played a role in that historic economic change as well.
The other terms that were most frequently used within our volume were tied to the loans, bonds and tax programs the U.S put in place to help fund the war. Lastly, silver was used also in an international context, but with the understanding that the context focused on whether the U.S would use silver in addition to gold and if it was going to be used in future stabilization of currency.
In reflecting on the project, we realized that in the future more time allotted for data discovery and increased group work would have allowed for more detailed computation and manipulation of the data. Much time was spent on transcribing the digitized microfilm which we felt could have been used to comb through the transcripts for a more in-depth analysis. Additionally, once we started looking at the project through an experimental lens, we were able to find a workflow that was more attainable (KB).