How Many Freshman Financial Aid Letters Are Generated Every Year? Six-and-a-half million.

The answer is missing from the web, so let's pull it ourselves, using the IPEDS data tables

New Financial Aid Letter Issuance

Below I’ve pulled from IPEDS the number of financial aid letters produced each year:

First, a quick history - generating financial aid letters was an artisanal process, generally performed by the Dean of Students, until the 1965 Higher Education Act:

The image of the dean sitting by himself in a dark room awarding $100 scholarships to “fine young men” was probably not too far from a true picture of how aid was awarded before the government identified the concept of ‘need’ as a priority… financial aid was the responsibility of the admissions office and/or the dean of students’ office until the mid- to late 1960s Crafting a Class, p. 175

The 1965 Higher Education Act introduced three programs: Educational Opportunity Grants (EOGs), Subsidized Student Loans, and College Work Study. These three programs quickly became major revenue sources, and spurred the development in the late 1960s of what is now a mission-critical piece of university infrastructure: the financial aid office.

Financial aid offices were soon tasked with developing ways to award aid consistently to thousands of students: initially they relied heavily on lookup tables, but in the early 1990s they began to harness algorithms. Algorithmic aid user manuals are now published online and describe some of the popular algorithmic packaging rules, like awarding financial aid around pre-defined GPA cutoffs. The below text is from an Ellucian user manual that describes how its software can be used to implement algorithmic financial aid:

>  award $5000 to the student if their EFC is <10,000 and their GPA is at least 4.0, 
$2500 if their EFC is <10,000 and GPA is at least 3.5, and $1500 if their EFC 
is <10,000 and their GPA is at least 3.0. 
Subsequent sequences are evaluated only if the prior sequences fail.

Algorithmic aid is fascinating, especially because very little has been published around the inner-workings of algorithmic financial aid, like:

• How much algorithmic diversity is there?
• What do the fine-tuning processes look like?
• Are steps being taken to audit the algorithims and make them “more fair” every year?
• How do financial aid algorithims compete with each other?

But before we explore those bigger questions, a fun first step is just counting: how many financial aid award letters do the algorithims generate every year?

As I write this, none of the major AI products has a real answer:

• Perplexity:

  > While there isn't a precise figure available... 
  

• ChatGPT

  > The exact number of financial aid letters generated each year isn't widely 
  reported, but it's likely in the millions, given the size of higher education 
  enrollment in the U.S. alone. The number of financial aid letters generated 
  annually likely exceeds tens of millions.
  

• Claude

  > I don't have access to precise data on the number of financial aid letters 
  generated annually. 
  

This creates a fun opportunity to dig into IPEDS and generate the data. Because we are using IPEDS, this analysis is limited to undergraduate students, and does not count graduate student financial aid letters.

The formula is straightforward: we first need to know how many acceptance letters are generated each year, and following that, we then need to know what percentage of accepted students recieve any aid. Then we just multiply:

Formula

1. Total number of acceptances, multiplied by
2. % of students recieving any aid

The two IPEDS variables we need reside in two separate IPEDS tables:

Voila! By stringing together 40 CSV files we now have a clearer picture of the financial aid landscape:

A few observations:

1. Going back as far as the data allows, to 2001, over 88 million financial aid letters have been produced.
2. The bulk of those 88 million aid letters have been generated algorithimically.
3. We know relatively little about how these algorithims work.

This is incredibly energizing! The in’s and out’s of algorithimic aid have largely escaped inquiry, but I have so many questions!

Caveats:

Our above analysis above is limited by the ANYAIDP and ADMSSN variables to first-time full-time freshmen enrollments, which means that 6.5m financial aid letters is part of the answer, but not a complete answer to: “how many financial aid letters are generated each year, including new aid letters + sophomore aid letters + junior aid letters + senior aid letters + part-time enrollment letters?“

26m (=6.5m x 4yrs) would be too crude of an guesstimate because there is attrition: IPEDS 4yr national graduation rate is 46.5% and IPEDS 6yr national graduation rate is 64.6%. To build an ideal estimate, we would need to build a curve estimating separation after each year. “First-time, full-time” enrollments also excludes many online enrollments, which IPEDS reported in 2021 at 4.4m.