CCCS study shows that Korean Americans contributed as much as — if not significantly more — $200 million in federal campaign contributions since the Federal Election Commission (FEC) kept records starting in 1977. Since 2000, the total contributions when adjusted for inflation is $167 million.

The methodology applied in the research had to make trade-offs between accuracy over coverage. Therefore a significant number of contributors may have been deliberately overlooked to avoid potential selection bias and sampling errors.
In the 2020 election, contributions from all groups, including Vietnamese and Asian Indians spiked exponentially

Key Highlights:

Korean American contributors are overwhelmingly represented by white collar professionals, especially attorneys, physicians and corporate executives who collectively make up more than 60% of identified professions in the top 10.

Korean Americans on average contribute twice as much as the average FEC donations, $222 versus $100 (Pew Research Center data).
Four of the top ten cities for Korean American contributions are from one state, California: Los Angeles, San Francisco, Los Altos Hills, and Irvine. Nearly ⅓ of all contributions in the top ten cities are from Korean Americans with California residences.

Korean American Contributions to Federal Election Campaign from 2000-2022

The first edition of Money Matters: Korean American Political Contributions from 1994 to 2000 (“2000 Report”), was published over twenty years ago by the Center for Civic Culture Studies (“Center”). It was among the first systematic efforts to filter and analyze federal campaign contribution data based on a specific ethnicity, the Korean Americans. The results of this updated study is that money continues to matter in politics and even more today with the exponential growth in online fundraising by networked digital platforms like ActBlue and WinRed.

Money in politics raises questions about the current state and future of American democracy and especially electoral politics. This report is more modest as the research is more narrowly tailored to parsing political contributions data from a small ethnic group, the Korean Americans. Ethnicity research about the civic participation of a small minority in a country where most are still considered to be immigrants is challenging but necessary. Korean Americans have a relatively shorter history in American history compared to other ethnic groups, yet they also have a compelling place in modern American history much like the Vietnamese Americans, as both countries are linked inextricably by wars where nearly a combined number of 90,000 Americans lost their lives. Future reports on Vietnamese Americans and especially a rising and politically active origin group like Indian Americans will add to this evolving and dynamic story of Asian Americans in America.

While ethnic origin is a useful tool for social sciences and political research, the more compelling story is how that “melting pot” of peoples from around the world continue to engage in the most vibrant civic marketplace of ideas and action in one of the most dynamic democracies in the world.

Since the Federal Election Campaign (“FEC”) database does not identify contributors’ ethnicities as separate data points, Center researchers adopted an innovative approach involving ethnic surnames with other statistically relevant details of contribution records to identify Korean Americans from the pool of over 95 million FEC database records. As described in further detail below in the methodology section, the validated surnames are based on original research by Professor Diane Lauderdale of the University of Chicago and Bert Kestenbaum, Office of the Chief Actuary of the Social Security Administration.

This report had to make difficult and necessary trade-offs between accuracy and coverage as part of the filtering process. By choosing to produce a more accurate report with narrower results, we had to accept that the study would produce a lower count in total contributions, and in some cases significantly so. It was important to avoid the risk of selection bias and other sampling errors as the Romanization of ethnic names is often subject to the vagaries of the immigration and naturalization bureaucratic process. The Korean surname Kim is fairly straightforward in its Romanized spelling of 김 in the Korean Hangul language. But there are myriad variations in the Romanized version, such as Ghim, Kym, Keem, Keam, Gym, etc. that are not part of the validated list of surnames. The surname Lee (and its Romanized variations) is the second most common surname in Korea, representing approximately 14.7% of the population. But Lee could not be part of the validated surname list that was used to filter the data because it is also among the most common surnames in China and even in the United States where it ranks 21st among popular last names and its inclusion would fatally skew the data.

Therefore, while the filtered data below may be reliably accurate, it also significantly undercounts by statistical design the total sum of potential political contributions and the number of Korean Americans donors.

We also discovered that the numbers do not always add up. With individual contribution records approaching 100 million lines and increasing daily and consisting of over half a dozen variables, everything from name, address, amount, date, address, etc., reporting errors by candidates and committees are to be expected. The www.opensecrets.org website sponsored by the Center for Responsive Politics founded in 1983 by former U.S. Senators Frank Church (D-Idaho) and Hugh Scott (R-Pa.) which pioneered campaign contribution research which says it best:

WHY DON'T THE NUMBERS ADD UP?

Sometimes it's hard to make apple-to-apple comparisons across some of the pages in a candidate's profile. Here's why: Summary numbers for federal officeholders - specifically "Total Raised and Spent" and "PAC/Individual Split" - are based on summary reports filed by the candidates with the Federal Election Commission….All other numbers in these profiles for state and federal officeholders are derived from contribution records disclosed in campaign finance reports from the FEC or state campaign finance agencies that reach applicable itemization thresholds…There is also a time lag in posting the information. While summary numbers are reported almost immediately by the FEC — and listed quickly on OpenSecrets — processing and analyzing the detailed records takes much longer. For that reason, summary numbers are usually higher (and more current) than the numbers based on detailed records. Data for state officeholders may be subject to longer time lags, as OpenSecrets aggregates this information from many sources and formats.

The Center is particularly grateful to Jae Ku, Ph.D., formerly with the Johns Hopkins School of Advanced International Studies and the U.S. Korea Institute, for research based on the original report by the Center for Civic Culture Studies on the underlying methodology used in this report.

Korean American Political Contributions: 2000 – 2022

The U.S. is home to the largest Korean diaspora community in the world. It is the fifth-largest Asian American subgroup after the Chinese Americans, Filipino Americans, Asian Indian American, and Vietnamese American communities but account for only a little over half a percent of the US population, or just under 2 million out of 331 million (2020 census).

The first Korean settlement is believed to have been established by Korean workers in early 1905 in the citrus orchards in what is known today as Riverside, California. While that camp did not survive, it did mark the beginning of a remarkable growth since then of the largest Korean diaspora outside of Korea in Los Angeles and major pockets of Korean Americans in New York City, Chicago, Northern Virginia, and now increasingly in major metropolitans cities in the Sun Belt such as Atlanta and Dallas.

The 2000 Report that covered the years from 1994 to 2000 identified 13,808 records of Korean American contributions totaling more than $10 million, a modest yet surprisingly large amount back in 2000 before the explosive growth of money in politics since then. The data used in this research is based exclusively on downloads from the FEC website’s public database. Additional data was also retrieved from the Harvard Dataverse for comparative purposes, an open source web application open to all researchers from any discipline, both inside and outside of the Harvard community.

In this follow up report twenty years later, Korean Americans are estimated to have contributed as much as $200 million or more (adjusted for inflation and sampling coverage) since the FEC launched its contributions database in 1977.

As in the 2000 Report, Korean Americans continue to give more generously compared to the U.S. average in terms of average political donation per contributor.. Korean Americans on average contribute $222 while the average U.S. donation is less than $100. Korean American contributors are geographically diverse hailing from all fifty states including the U.S. territories of Puerto Rico and Guam. Their giving is also ideologically diverse spanning the entire political spectrum from progressives to conservatives and all political stripes in between. Over 5,000 different campaign committees from ActBlue to WinRed and all varieties in between received contributions from Korean Americans in varying amounts.

Below are some of the key findings regarding political contributions to federal campaigns by Korean American individuals according to FEC data and the applied methodology.

1.

First, the threshold question: What does it mean to be a Korean-American? Or for that matter, an American?

This is a fairly narrow study of the political campaign contributions by a small ethnic minority, the Korean-Americans. Yet the question of what it means to be a hyphenated American or for that matter, an American, as deceptively simple as it may seem, is a question that has been debated since the arrival of the first among many ethnic groups that came to these shores and answered in countless ways and often controversial over the course of our nation’s history.

Almost any discussion regarding hyphenated Americans these days engenders passionate reactions about identity politics whether described as Asian Americans, LatinX or the Patriotic Front, and exploited often by politicians. Even Francis Fukuyama in his 2018 book, Identity, which was spurred by the divisive election in 2016 of Donald Trump, could not resist describing some of the populist inspired movements as the politics of resentment in the book’s subtitle: The Demand for Dignity and the Politics of Resentment.

But as Fukuyama masterfully does in his book, which this report modestly hopes it has tried to do, is to speak to the passionate desire by all Americans demanding some measure of human dignity. Critical to this is the continuing debate about what it means to be an American: does the term "American" have a singular or plural definition? Is there a common culture or bank of commonly shared values that unite Americans of widely different backgrounds? Have the disagreements, sometimes fierce and violent, coexisted with a rough consensus on values and some concept of the “American dream”?

The motto of the Great Seal of the United States, E pluribus unum – "From many, one" – and the Seal’s sheaf of arrows suggest the coexistence of many peoples under a unified citizenship based on shared ideals that are worth protecting and fighting for. Being an American means in varying degrees sharing a national culture that is often shaped and enriched by race, ethnicity and religions, but above all by the basic principles laid out in the U.S. Constitution.

There is no question that a strong and sustainable civic culture requires a commitment to a shared set of values and beliefs, something that is most easily understood when an immigrant takes an oath to become an American citizen: a professed allegiance to the principles of the U.S. Constitution, adoption of a shared culture by demonstrating the ability to read, write, and speak English, and some basic understanding of American history.

It is clearly beyond the scope of this report to address the deeper issues raised by that simple question. This study was begun many years ago by what its authors thought at the time was an easy inquiry that turned into a fascinating bot not surprising discovery. By typing in one of the most common Korean surname, “Kim,” in the FEC search database back in 2000, we got a glimpse, a privileged one if you will, into the political contribution patterns of what was and still is a small ethnic group awakened into political consciousness by the events of the LA riots in 1993 or “Saigu” (Korean for “4/29”). When riots were finally quelled six days later, more than 50 people would be dead and 3,000 businesses destroyed or looted, nearly half of which were Korean-owned. Damages totaled about $1 billion. That seminal event and especially the random but clearly targeted attacks against Asian Americans in recent years make this report relevant.

In 1989, Professor Michael Walzer, the noted political theorist, attempted to answer this question at the Morgan Lecture at Dickinson College in Carlisle, Pennsylvania, titled appropriately, “What does it mean to be an ‘American’?” His answer is a hopeful one but difficult to achieve.

If the manyness of America is cultural, its oneness is political, and it may be the case that the men and women who are free from non-American cultures [their ethnic origins] will commit themselves more fully to the American political system. (Brackets added).

As mentioned, the idea for this research was started over 20 years ago when one of its researchers typed in the ubiquitous surname “Kim” into the FEC database. What he discovered was remarkably moving and hopeful. Thousands of Korean Americans from nearly all 50 states with varied occupations such as business owner, retiree, student, physician, investment banker, even housewife, had taken the time to think about a candidate and made a conscious decision to part with her hard-earned money that would not even be tax deductible. Apart from voting, nothing speaks more clearly about committing to a political system than this simple act.

2.

When adjusted for inflation and other sampling factors, Korean Americans have contributed as much as $167 million to federal campaigns between 2000-2020 made by 69,272 contributors in over 800,000 contributions identified as Korean Americans based on the surname sampling methodology. CCCS estimates that the total amount may be significantly higher if data could be collected starting in 1977 (when records were first collected by the FEC but not digitized in searchable database) and includes all possible Romanized variations of Korean surnames.

In a preview of future research on other ethnic groups to be released in the coming months, preliminary analysis comparing the top surnames among three Asian American groups — “Kim” for Korean Americans, “Nguyen” for Vietnamese Americans and Patel” for Indian Americans — found the same pronounced donation spikes in 2020 from all three groups. This is consistent with the national contribution trend shown below.

*Again, please note that the numbers do not always add up for reasons best explained here.

3.

Korean American contributors are geographically diverse, hailing from all fifty states and the U.S. territories of Puerto Rico and Guam. Candidates, both incumbents and challengers, received contributions from a broad spectrum of Korean Americans from all 50 states, including nearly 80 contributions from the one state, Vermont, which had no contributions back in the 2000 Report. Although the geographical diversity this time is even more striking in 2022, as with the 2000 Report, most Korean Americans call California and New York their home, but now increasingly live in the sunbelt states of Texas and Georgia and the mid-atlantic but essentially the southern state of Virginia.. More than 40% of Korean American contributions are from California and New York, and almost 80% of contributions were made by the Top 10 States.

4.

Korean Americans follow the general trend of the average U.S. contributor of making increasingly more contributions to candidates and parties over the last 20 years.

According to a Pew Research Center report on US Elections and Voters based on studies by the American National Election Studies (“ANES”), Americans are now more likely to contribute to political candidates and parties “with the share of adults who say they have donated directly to candidates doubling since 1992.”5 And the share of Americans who say they have donated to an individual running for public office in 2016 has doubled, increasing from 6% in 1992 to 12% in 2016, with an exponential spike in the 2020 election.

Korean American donations since 2000 similarly grew exponentially with a pronounced spike in 2020. In the absence of opinion surveys, we can only speculate that enthusiasm both for and against Donald Trump may have driven that spike. If perception is reality in electoral politics, the more that issues are perceived through a political lens in an increasingly polarized political atmosphere, the more likely they are involved with the varieties of political participation, including political contributions.

When President Donald Trump first met with North Korea’s Kim Jong Un in 2018 in Singapore after a series of provocative missile launches in 2017 by North Korea and the revelations of “Love Letters” between the two, the resulting media coverage could only have fed the perception of President Trump as someone willing to take risks, perhaps recklessly to some people, to solve a problem that continues to roil the Korean peninsula now nearing 70 years since the end of a Korean War in 1953.

5.

The digital revolution in the money chase; ActBlue (Democrats) vs. WinRed (Republicans).

The real story of political contributions especially in the last few years leading up to the 2020 congressional and presidential election may in fact be the digital revolution in small donor network platforms that launched fundraising juggernaut for both the Democrats with ActBlue and the Republicans with WinRed. Both ActBlue and WinRed are digital platforms that simplify the process of making political contributions often with just a single click of the mouse.T

The Republicans platform was launched in 2014 nearly ten years after ActBlue and ramped up by the Republican party and Trump administration in 2019. As the animated chart shows below, WinRed reached $1 billion in much shorter time than ActBlue which had started earlier in 2004, propelled by the Trump presidency and the pandemic which severely limited the traditional meet and greet high dollar fundraisers. We were unable to find WinRed’s total raise during the years since 2014, but ActBlue to date reports raising over $11 billion, most of which were raised in recent years.

The Center believes that both the potential and the nature of online fundraising platforms have significantly influenced both the tone and tenor of online fundraising pitches that are increasingly shrill and even dangerous to our democracy. Most Americans believe in the principle that under the free exercise of speech clause guaranteed by the First Amendment of the U.S. Constitution and upheld by the Supreme Court, any American should have the right to exercise that freedom by supporting particular candidates and issues with their hard-earned money. Social media in particular has become easier to exploit various schisms in American culture and politics. There is, for example, growing unease regarding a multiracial democracy in which the U.S. will move ever closer to a non-white majority. Some of this may be fueling the attacks on democratic institutions (including the January 6 riot) and a perception that democratic norms are weakening in America and around the world. While both Democrats and Republicans are concerned about the state of American democracy, neither hesitate to point the finger at each other when looking for blame. Finding common ground in this charged atmosphere, even between ethnic minorities, appears increasingly difficult.

That standoff is now fueling excesses in fundraising pitches by both sides and forcing candidates — even the moderate ones — to adopt ever more vitriolic and incendiary language. Even between two Asian American candidates in a newly drawn district in California pitting incumbent Michelle Steele, a Korean American candidate, against a Vietnamese American challenger, Jay Chen, charges of red-baiting have erupted over photo-shopped campaign fliers. Positive messages and any talk about compromise and bipartisanship just do not rile up the base for either side. How America grapples with how much “freedom” we should allow while ensuring that all parts of society have access to these technologies, and whether social media platforms like Facebook should be the ones policing national information security, is a critical issue going forward.

This “brave new world” of political fundraising going forward will now increasingly leave behind the old-fashioned political fundraising mixers in posh locales of Washington, D.C.’s Georgetown or Mclean in Northern Virginia, where you can press the flesh and kick the tires, so to speak, of potential candidates and incumbents. The power of the digital revolution wrought by the network effect of the internet that has upended and disrupted so much of modern life and business has now taken over politics.

Those from corporate boardrooms and the private equity community gave in aggregate amounts that rank among the top campaign contributions in the country. The largest aggregate yearly contributions on record from the past two decades by a Korean American individual is over $1.6 million.

6.

*Again, please note that the numbers do not always add up for reasons best explained here.

The average wealth among billionaires doubled in just one year from 2020 to 2021 and, predictably, so has their political contributions, especially in the years after the Citizens United Supreme Court decision that opened the floodgates to unlimited amounts. Three of them—Donald Trump, Michael Bloomberg and Tom Steyer—even ran for president. A fourth, Howard Schultz, flirted with the idea. Collectively, more than 200 billionaires spent $2.3 billion in 2020 alone. While Korean Americans may not be at these levels, it would be interesting to see how fast and how much groups like Korean Americans, Indian Americans and other ethnic minorities will quickly catch up in the relentless money chase. It is likely just a matter of time.

7.

Do Korean Americans donate to Korean American candidates? Contrary to public perception, Korean American contributors do not give in overwhelming numbers exclusively to Korean American candidates. The usual suspects like the national presidential committees get the lion’s share of campaign funds from Korean Americans just as they do nationally. This just shows that ethnicity, while useful as a research tool, is not a political tool for raising money. If money mattered and was the sole reason for getting certain members of congress election, the representation in the U.S. Congress by four Korean Americans represents a remarkable and justifiably proud achievement for the Korean American community but not because Korean Americans gave overwhelming campaign contributions. To their credit, the Korean American members of Congress got elected because they ran on the issues that mattered to a majority of voters in their district. Nevertheless, collectively, the current members with Korean heritage, three of whom were born in Korea to immigrant parents, raised over $3 million from Korean Americans.

Korean Americans In Congress

Young Kim - CA 39th District

Andy Kim - NJ 3rd District

Michelle Steel - CA 48th District

Marilyn Strickland - WA 10th District

These amounts from the Korean American community are actually a small percentage of the total amounts raised by each of the candidates for their races. Rep. Andy Kim raised over $18 million since 2017 when he first ran for Congress; Rep. Young Kim raised over $15 million over her career; Rep. Michelle Steel raised close to $11 million; and Rep. Marilyn Strickland raised over $3 million. In other words, ethnic identification only gets you so far in the political money chase with your own ethnic identity, although they are still significant.

While money matters and it will continue to play a significant role in electoral politics, at the end of the day, electoral politics will continue to be about representing constituents on issues that matter most to them.

8.

Korean Americans are also ideologically diverse; money matters, but so do issues, like the rest of America. Korean Americans contributed to nearly 5,000 different campaign committees and political action committees spanning the ideological spectrum, from leading recipients ActBlue, the democratic fundraising platform, and WinRed, the Republican fundraising platform, to pro-Trump and even an anti-Trump PAC. Below is only a small sample of the top committees and PACs that received contributions from Korean Americans.

Methodology

In order to analyze any contribution patterns for ethnic origin, a sample of a particular ethnic donor group, in our case Korean Americans, needed to be extracted from the Federal Election Commission (FEC) data of individual contributions. Without any given demographic information in the FEC data, this study used surnames to infer ethnicity and a surname sampling list was applied to the original FEC data to create a resulting ethnic Korean sample dataset. Surname sampling creates a list of surnames related to a target ethnicity (or group) that has been validated with a dataset containing surname, race, and ethnicity.

Surname sampling could have been combined with geocoding to use known geographic concentrations of a target ethnicity like Los Angeles which has a high concentration of Koreans to inform the probability of an individual being of that target ethnicity, but given the lack of publicly available data that includes surnames, zip codes, and Korean ethnicity at a national scale, this was not pursued. Another possible enhancement is the use of given names or occupational information to identify those of a target ethnicity [1], but these techniques are limited by the researchers’ ability to include all possible outcomes that could indicate that target ethnicity [2], and would otherwise introduce too much selection bias.

For a population with no known identifiable links between surname and social status such as ethnic Koreans, a surname sampling technique should provide a sample that is representative of the entire population of individuals with Korean surnames [3]. However, individuals of Korean ethnicity who have non-Korean surnames (e.g. Korean adoptees, spouses of non-Korean individuals, and children of multiethnic or multiracial households) would not be discovered through surname sampling. The lack of demographic and other identifying variables in the FEC data limits this study’s ability to include such individuals. This study relies on surname sampling to produce a dataset of Koreans with ethnic Korean surnames that is a representative sample of both Koreans with and without ethnic Korean surnames.

In order to create and validate a surname sampling list, researchers derive a surname list from a source dataset that includes surnames and ethnic/racial information and then validate the surname list against the source dataset and often against another validation dataset, containing surname and ethnic/racial information. The geography of the derivation source dataset and the validation dataset(s) impact the efficacy of a surname list. When a surname list is validated, it is tested for accuracy and coverage, but these statistics will deviate from the performance of the original validated list if the list is applied to a dataset covering a different geographic region. For example, a list derived from a statewide dataset will lose accuracy if applied to a nationwide dataset [4]. Because FEC data contains national data, a surname list derived and validated using a national-level dataset is required for this study.

Many surnames used by individuals of Korean ethnicity are not specific to Korean ethnicity (e.g. Lee, Park, and Yu). Building an ethnic surname list demands a trade-off between coverage and accuracy. If a list were to include all surnames used by Koreans, this list would also include several names not unique to Koreans, and when applied to the dataset, the results would include a large number of non-Koreans. On the other hand, if a list were to select only names with a high correlation to Korean ethnicity (and low correlation to other ethnic groups), this list would be small, but when applied to the dataset, the smaller resulting dataset would be more accurate and only include a small number of non-Koreans [5]. In order to identify political contribution patterns specific to ethnic Koreans in America and to limit the impact of non-Korean cases on the findings, this study prioritized accuracy over coverage.

There are several studies on Asian surname sampling, but the pool of research on Korean surname sampling is much smaller. Due to a dearth of publicly available national-level datasets that include name and ethnicity information, this study relies on a previously created and twice-validated Korean surname list. Diane Lauderdale and Bert Kestenbaum have created an unconditional Korean surname list for application to datasets without race information available by using the Social Security Administration’s (SSA) Master Beneficiary Record (MBR) of individuals entitled to social security benefits or enrolled in Medicare to derive ethnic Korean surname lists [6], and they validated the lists against the SSA record of social security card applicants against a sample file of the 1990 US Census [7].

Because the overwhelming majority of Korean-Americans immigrated to the US after the Immigration and Nationality Act of 1965 [8], the age limit of the source files suggests that most of the Koreans in the source file will be predominantly first generation immigrants with ethnic Korean surnames or maiden names [9].

However, ethnic Koreans with non-ethnic surnames or maiden names would not have a high likelihood of being included in any surname sampling technique. In order for a surname to meet the strength of association threshold used by surname sampling techniques, there would need to be a large proportion of individuals with a specific non-ethnic Korean surname who are of Korean ethnicity. It is unlikely that a surname sampling technique would accurately identify someone of Korean ethnicity without a Korean surname because the pool of non-Korean surnames is incredibly large. Additionally, the cross-validation of Lauderdale and Kestenbaum’s surname lists against the more recent census file and the lack of socio-demographic differences across Korean surnames suggests that this list is not age-specific and can be considered a representative subset of the Korean ethnic population [10].

Additionally, the national-level data used and validated in their study presents a good fit for application to the FEC data. To increase sampling coverage, researchers also devised surname sampling techniques that include geographic data coupled to ethnicity (geocoding) and given name ethnic origins, but for this study, Lauderdale and Kestenbaum’s surname sampling technique best fit the study parameters while offering external validity statistics indicating the degree to which the sampling results can be generalized to the Korean-American population and the ability to optimize accuracy against coverage.

Lauderdale and Kestenbaum statistically evaluate their surname list using the metrics of sensitivity (SE) and positive predictive value (PPV) to analyze their lists’ coverage and accuracy, respectively. The SE is the proportion of the number of positive cases in the control dataset accurately identified as positive by the test to the number of all positive cases in the control dataset. A high SE indicates that the surname list has high coverage with few false negatives, and a low SE indicates that the surname list has low coverage and many false negatives. The PPV is the number of positive cases in the control dataset accurately identified as positive by the test to the number of all cases (both accurately and inaccurately) identified as positive by the test. A high PPV implies the testing method has high accuracy and results in few false positives, while a low PPV suggests the testing method is inaccurate with a large number of false positives.

Lauderdale and Kestenbaum developed predictive (N = 285 names) and strongly predictive (N = 110 names) unconditional Korean surname lists. In order to be included on these surname lists, each surname associated with an individual of Korean ethnicity needed to meet specific thresholds for their strength of association with Korean ethnicity [11]. Both lists were more accurate against the 2000 Census sample file than the SSA source file [12], and the strongly predictive list has an SE range of 0.36 to 0.43 and a PPV range of 0.83 to 0.88. Although the application of the strongly predictive list identifies fewer ethnic Koreans from the validation datasets than would the predictive list, it also identifies fewer non-ethnic Koreans. The strength of this study’s analysis depends on identifying political contribution traits specific to Korean-Americans, and in order meet those ends, the importance of having a resulting dataset that is comprised of a large majority of ethnic Koreans outweighs the need to maximize the size of the resulting dataset at the cost of including more non-ethnic Koreans. Thus, this study analyzed the application of Lauderdale and Kestenbaum’s strongly predictive unconditional list to the FEC data.

[1] Wong, Palaniappan, and Lauderdale validated a sampling technique using ethnic given names in conjunction with ethnic surnames, but for Koreans this lowered the accuracy (positive predictive value) at only a small increase in coverage (sensitivity).

[2] Wong, Palaniappan, and Lauderdale validated a sampling technique using ethnic given names in conjunction with ethnic surnames, but for Koreans this lowered the accuracy (positive predictive value) at only a small increase in coverage (sensitivity).

[3] Jibum Kim et al., “Surname Sampling: Reevaluating Kim Sampling in Korea and the United States,” Field Methods 26, no. 1 (February 2014): 87-104, accessed March 5, 2014, http://fmx.sagepub.com/content/26/1/87.

[4] Wong, Palaniappan, and Lauderdale, 543.

[5] Surname sampling techniques do not address the challenge of identifying ethnic Koreans who have changed their surname through marriage or who have been adopted to non-ethnic Korean families.

[6] For application datasets with race information available, they used the SSA’s record of social security card applicants who were born before 1941 to create a conditional Korean surname list.

[7] Diane S. Lauderdale and Bert Kestenbaum, “Asian American Ethnic Identification by Surname,” Population Research & Policy Review 19, no. 3 (June 2000): 283-300, accessed February 19, 2014, http://www.jstor.org/stable/40230271.

[8] Eui-young Yu and Peter Choe, “Korean Population in the United States as Reflected in the Year 2000 US Census,” Amerasia Journal 29, no. 3 (2003-2004): 1-2, accessed April 10, 2014, http://www.metapress.com/content/b3vg03788h44v552/.

[9] Lauderdale and Kestenbaum substituted maiden names for the married surnames of female individuals in their datasets.

[10] Jibum Kim et al., “Surname Sampling: Reevaluating Kim Sampling in Korea and the United States,” Field Methods 26, no. 1 (February 2014): 87-104, accessed March 5, 2014, http://fmx.sagepub.com/content/26/1/87.

[11] For inclusion on the predictive surname list Lauderdale and Kestenbaum required that a surname be associated with the target ethnicity in at least 50 percent of all cases and with any other specific ethnicity in less than 50 percent of all cases. For inclusion on the strongly predictive surname list, the strength of association with the target ethnicity was increased to 75 percent.

[12] Unconditional Predictive Korean Surname List: Source Dataset (SE = 0.52, PPV = 0.74), 1990 Census Sample (SE = 0.54, PPV = 0.81). Unconditional Strongly Predictive Korean Surname List: Source Dataset (SE = 0.36, PPV = 0.83), 1990 Census Sample (SE = 0.43, PPV = 0.88). Lauderdale and Kestenbaum, 290-2.

FEC Data

To examine Korean-American political contribution patterns, this study uses FEC data from the Itemized Records (Individual/Candidate Contributions), Committee Master Information, and Candidate Master Information files from 2000-2022 (N = 95,000,000 entries). These files were downloaded from the FEC FTP server on Jun 30, 2022. The Itemized Records list every individual contribution to political candidates and political action committees (PAC), organized by election year; The Committee Master Information and Candidate Master Information files contain every political committee and candidate, respectively, registered with the FEC for each election year.

The Itemized Records include individual contributor information such as first and last names, city of residence, state of residence, zip code, and occupation, but demographic information like race, ethnicity, gender or age are not collected, Contribution information includes the amount of the receipt, the date of receipt, and the unique FEC Identification Number of the recipient committee. The Committee Master Information lists all unique FEC committee ID numbers used for each election year, along with committee information such as name, type, etc.

This study linked all individual and committee files, by election year, via the FEC committee ID number in Microsoft Office Access 2019. The individual and committee files were linked by matching the FEC ID number of the recipient committee in the Itemized Records to the corresponding FEC ID number in the Committee Master Information files. Thus, the Committee Master Information and Candidate Master Information files during this time period were linked by matching the unique FEC Identification Number of the candidate from the candidate files to the Supporting Candidate variable of the committee files.

Using Microsoft Access 2019, Lauderdale and Kestenbaum’s strongly predictive unconditional Korean surname list was applied to the linked FEC individual, committee, and candidate files, and the resulting dataset included all individual contributors with surnames matching those on the list and their corresponding donation information. Only whole surname matches were included. Thus, individuals with nonconventional surnames, such as hyphenated surnames or more than one surname, are not represented in the sample [13]. The application of the Lauderdale-Kestenbaum strongly predictive unconditional Korean surname list to the FEC dataset produced a resulting dataset of contribution records from Korean Americans.

[13] It is unclear how many Koreans have non-conventional surnames. Previous research suggests that 6.4% of all women in the total US population have a surname that differs from their spouse, but this number reaches 15% for Asian alone women. Additionally, the most common non-conventional surname option chosen by women in the general population is keeping their maiden name. Gretchen E. Gooding and Rose M. Kreider, “Women’s Marital Naming Choices in a Nationally Representative Sample,” Journal of Family Issues 31, no. 5 (May 2010): 681-701, accessed March 6, 2014, http://jfi.sagepub.com/content/31/5/681.

Data Analysis

The large size of the original dataset prevents verification of every entry for accuracy. Geographically, Itemized Record entries that did not list a two-character US state or territory abbreviation were examined by zip code, and the corresponding city and state information were entered in the blank fields when applicable. For analyses across multiple years, contribution amounts were adjusted for inflation by contribution year using the Consumer Price Index for All Urban Consumers (CPI-U) [14].

The Itemized Records also included refunds (listed as negative contribution amounts), but due to the large size of the dataset and entry inconsistencies between contribution and refund amounts, the refunds and the matching original contributions were not removed from the dataset. Additionally, when aggregating contribution amounts that have been adjusted for inflation, refund and original contribution dollar amounts net zero (except in cases when the refund and original contribution years do not match). Any mismatches between contribution and refund years lead to a small overestimation of aggregate dollar amounts because a late refund adjusted for inflation is smaller than the original contribution adjustment. While refunds alter aggregate contribution counts and dollar amounts, the number of refunds given in the resulting dataset was small (N = 213, 1.01% of sample) compared to the total number of cases in the original and resulting datasets. Further, any overestimation of dollar amounts constitutes a small percentage (i.e. the difference in CPI-U across years) of the proportion of refunds made in a different year from the original contribution. It is unlikely these refunds would have a noticeable impact on the results.

As a collection of several decades of records, the resulting FEC dataset contains inconsistencies across years. In particular, candidates’ official names can differ by election cycles or by the type of office sought. This study manually examined variations of candidate names against FEC candidate records and candidate identification numbers and merged variations into one candidate name when appropriate. This study analyzes candidates by their aggregate amounts of receipts (adjusted for inflation), including contributions to non-principal and non-authorized committees that contain the candidate’s name in the committee name and list only that candidate as the recipient of transferred funds for the relevant election year.

This study also performed analyses of the type of office of sought (election type) and political parties. The amount of contributions to each election type is analyzed by election year and includes contributions to committees that are not principal campaign committees but do contribute to specific election types. This study examines the political parties of recipient candidates and committees. However, not all contributions were made to committees with an affiliated candidate or party.

Despite efforts to verify the dataset for accuracy, due to the size of the dataset and recording methods of the FEC it is likely that human error remains, especially when the data is initially entered by candidates and committees. The path from individual contribution to entry in the FEC records is often indirect–while contributions made online will be recorded in the FEC files as is, contributions made in person or by phone require an individual other than the donor to transcribe or record information. Thus, it is expected that the FEC files are subject to human error, such as misspellings and illegible handwriting.

[14] US Bureau of Labor Statistics, Table 4. Consumer Price Index for All Urban Consumers (CPI-U): Selected areas, all items index - 2022 M08 Results (bls.gov)

Center for Civic Culture Studies, Inc.
Sean Woo, President
Sam Choi, Director
Annie Woo, Assistant

Korean American Political Contributions Project Staff

Sam Choi, Director of Operations
Ifunanya Nwogbaga, Advisor, Ph.D. Candidate, Department of Biophysics, Johns Hopkins University
Daanyaal Khan, Advisor, M.S. Candidate, Information Systems Engineering Candidate, Johns Hopkins University

The Center is a private nonpartisan independent research organization dedicated to the study of the norms and networks of civic and political engagement among and within a variety of civic groups in the United States. Subjects currently under research include campaign finance and elections, healthcare, education, and emerging technologies.

Disclaimer:

In using outside resources, especially data, in preparing and producing this report, Center makes no representations with respect to the accuracy or completeness of data and/or sources. Center only used data submitted to and compiled by the Federal Election Commission (FEC), including disambiguated FEC campaign contribution data available through the Harvard Dataverse. Any and all opinions expressed herein reflect only those of the Center and official representatives and are not intended to reflect an official position. Any views expressed in this report are solely those of Center and are not intended to reflect the views of any other individual and/or their organizations nor are their listing either an implicit or explicit endorsement by them or their organizations.