WHAT DECIDE THE ENDOWMENT AND ANNUAL GIVING SIZE? — B ASED ON THE EMPIRICAL ANALYSIS OF UNIVERSITIES IN AMERICA XINWEI LIU
WHAT DECIDE THE ENDOWMENTAND ANNUAL GIVING SIZE?
—BASED ON THE EMPIRICAL ANALYSIS OF UNIVERSITIES IN
AMERICA
XINWEI LIU
Abstract
This Paper, picking the university data as research samples,
takes advantages of the Multiple Linear Regression Model and
uses the empirical method to testify whether performances of
universities in the current year have impact on the size of
Endowment Assets and Annual Giving next year. The regression
results signify: Better performances of universities in
current year significantly increase the total number of
Endowment Assets and Annual Giving next year by sending to
donors and public signals that universities with better
performances have good research potential and high-quality
education potential and are deserved to be donated.
【Key Words 】 University Performances Endowment Annual
Giving
Introduction
Everyone wants to stand first in line, first in the hearts
of the country and first in the standings. The pursuit of
Number One is surely an important thing in sports or games.
Universities, however, standing in the first line is not so
important as standing among top ones. As the development of
society, 21st century shows no lessening of interest among
researchers, donors, government supports in using various
criteria to assess the performance of research institutes.
As always, University Endowment and Annual Giving in
America play an important role in the development of
universities, not only for public universities but also for
private universities. First, endowment is one of the main
sources of financial income of almost all the universities in
America since university is a kind of special non-profit
organization whose main goal is not making money. The fact is
that Endowments, Federal Support and Tuition Income consist of
more than 90% of every university’s income. Second, endowments
lay the foundation of research achievements of universities.
Given the fact that abundant financial power is the footstone
of the beginning research studies, even starting a research
will not come true without the guarantee of financial power.
Consequently, maintaining the Endowment size is the key factor
to the long-term success of universities.
Recently, Endowments issues attract more attention not
only because of the importance itself but also because of the
increased gap between top universities such as Ivy League and
other universities. Since 1992, the gap has been increased:
top universities (top 25) are richer while others fall behind.
(Josh Lerner and Jialan Wang, 2008) Why do top universities
such as Ivy League become richer? It’s not only because of
outstanding breakthroughs or research achievements they made
every year but also because of other factors such as National
Academy Membership, Faculty Awards, Doctorates Awarded,
Postdoctoral Appointees and the high average SAT Scores,
mirroring their high research and education qualities. These
measurements reflect the quality of a university from
different area. National Academy Membership represents the
excellent scholars and their researches. Faculty Awards is an
element that can reflect the faculty reputation and
achievements of a university. Doctorate Awarded and
Postdoctoral Appointees evaluates the ability to cultivate
potential researchers in the future. Average SAT Score imply
the quality of students entering the university. All these
factors reflect the quality of a university from different
aspect. Since these measurements are used to measure the
performance of University and the performances are criteria
for endowment. There does exist a necessity to find relations
between endowment size and measurements of performance, thus
providing evidence and methods for universities to keep long-
term successes.
Literature Review
When talked about the Endowments and measurements of
universities, many scholars have showed interests in some
aspects. Hoover and Farrell in 2007 found that top research
institutions such as Harvard and Princeton have used their
deep pocket to increase admissions and reduce effective
tuition rates, thus broadening the excess for elite education.
Gifts to Annual Giving allow universities to provide
extraordinary opportunities for learning and discovery; to
extend the financial aid program to more students who needs
it; and to help meet the emerging needs and challenges. The
funds of Annual Giving provide universities the flexibility to
undertake critical new initiatives such as developing the new
science research program and attracting world-renewed faculty
and so on so forth.
On the contrary, poor financial endowment will endanger the
development of a university. As Carlson said in 2007,
Antioch’s only 36 million endowments made financial
insecurity, leading to decaying facilities and declining
students enrollment. Some scholars focused on superior
investment returns of endowments compared with other
institutional investors (Lerner, Schoar and Wongsungwai,
2007). Josh Lerner and Jialan Wang, 2008 found that much of
growth in endowment size has been driven by investment
performance. What’s more, while investment in some assets have
been successful, there have also been some loss, like the $350
million investment loss that Harvard experienced in Sowood
Capital Management (Karmin and Zucker, 2007).
Most universities measure themselves on a wide range of
dimensions that they believe important for them to make
decisions. Moreover, no single indicator or study shows what
research institutes have done, can do and will do. To improve
the quality of education and research of a research institute,
its students, faculty, staff and supporters have to put the
power together to achieve the improvement among the best
universities. Since it is a fact that both private and public
universities live on the resources generated from tuition,
government support and endowment, the critical resource for
universities is the income generated from Endowment or Annual
Giving. Although Endowment attracts many scholars’ interests,
the real relation between endowment and performance
measurements of university still remains uncovered.
According to the review above, I made my hypothesis of this
paper:
H1: Performances of Universities are signals for
University donors.
H2: Performances of Universities have positive impacts on
the Size of Endowment Assets.
H3: Performances of Universities have positive impacts on
the Size of Annual Giving.
My research, in some extent, provides answers to unanswered
questions as to Endowment and Annual Giving. By finding
relationships between long-term strength “Endowment” and
measurements of university performance and between short-term
indicator “Annual Giving” and measurements of university
performance, my research have contributions to enrich the
university study. From university’s perspective, my research
gives basic considerations of the ways to improve the quality
of universities and to attract more endowment. From donors’
perspective, my research lays a foundation for making
endowment decision in terms of the performances of
universities. Not only for keeping the long-term success of
universities but also for making endowment decisions of
investors, my research provides some guidance.
Research Design
1. Sample Selection
All the Sample data of this paper come from The Center for
Measuring University Performance, samples including 771
universities and colleges in United States. In order to be
more accurate, I selected 771 universities’ data from 2005 to
2011. I choose the data of Endowment Assets, Annual Giving,
Total Research Expenses, Average SAT Scores, Faculty Rewards,
National Academy Membership, Postdoctoral Appointees and
Doctorates Awarded from 2004 to 2011 as samples. Each item has
the number of 5391 data from 2004 to 2011. Given the fact that
I focus on the relationship between the performances of
university in t year and the Endowment and Annual Giving Size
in t+1 year, data of the dependent variables such as Endowment
Assets and Annual Giving are selected from 2005 to 2011. Other
independent variables are chosen from 2004 to 2010.
According to the experience of regression and the basic
background of these university data, I made the following
standards as criteria to adjust data.
(1)Since a certain number of universities didn’t have
Endowment Assets or Annual Giving and some didn’t have
Research Expenses, I delete all the vacant data that may
affect the regression results.
(2)Moreover, I eliminated useless data and extreme-value data
to minimize the regression error.
After the adjusted work, the total number of each item changed
from 5390 to 4823 for each item.
2. Data Description Statistics
Table 1 shows the Description Statistics of the sample
data before elimination. The total Observed Sample is 4823. I
pick the following items as performance measurements of
universities and specific data description is in Table 1.
Table 1. Date Description
Mean Median Min Max Observed
Value
EndAti,t+170374 133971 182 36556284 4823
AnulGi,t+131257 10202 147 709423 4823
TRexpi,t71761 4057 0 1997252 4823
SATavgi,t1119 1105 735 1525 4823
FRewdi,t3.589 1 0 116 4823
NAMembi,t5.7536 0 0 348 4823
PostDAi,t77.5796 0 0 5827 4823
DoctAi,t73.4047 6 0 911 4823
3. Empirical Modeling Design
(1). Empirical Modeling
lnEndAti,t+1=α0+α1lnTRexpi,t+α2lnSATavgi,t+α3lnFRewdi,t+α4lnNAMembi,t+α5lnPostDAi,t+α6lnDoctAi,t(1)
In the Model (1):
The EndAti,t+1 represents the Endowment Assets in dollar of
i university in t+1 year; TRexpi,tis Total Research Expense in
dollar of i university in t year; SATavgi,t is Average SAT Score
of i university in t year; FRewdi,t is Faculty Rewards of i
university in t year; NAMembi,t is the number of National
Academic Membership of i university in t year;PostDAi,t is the
number of Post Doctor Awarded of i university in t year;
DoctAi,t is the number of Doctor Awarded i university in t year.
In this Model, I focus on figuring out the relationship
between the performances measurements of universities in t year
and the Total number of Endowment Assets in t+1 year. Is the
size of Endowment Assets in t+1 year affected by how universities
performed in t year. If the Coefficients of each variable are
positive, it means that the good performance of universities
in t year will obviously increase the Total Size of Endowment
Assets in t+1 year, or vise versa. Moreover, the Coefficients, α1,
α2, α3, α4, α5, α6, mean how the dependent variable changes
with the change of six independent variables.
lnAnulGi,t+1=β0+β1lnTRexpi,t+β2lnSATavgi,t+β3lnFRewdi,t+β4lnNAMembi,t+β5lnPostDAi,t+β6lnDoctAi,t(2)
In the Model (2):
The AnulGi,t+1is Annual Giving in dollar of i university in
t+1 year. TRexpi,tis Total Research Expense in dollar of i
university in t year; SATavgi,t is Average SAT Score of i
university in t year; FRewdi,t is Faculty Rewards of i
university in t year; NAMembi,t is the number of National
Academic Membership of i university in t year;PostDAi,t is the
number of Post Doctor Awarded of i university in t year;
DoctAi,t is the number of Doctor Awarded i university in t year.
In this Model, I focus on the relations between the Annual
Giving in dollars of universities in t+1 year and the
performance measurements of universities in t year. If the
Coefficients of each Independent variables are obviously
positive, it means that good performances for universities in t
year will lead to better impressions to donors, thus
increasing the total number of Annual Giving in t+1 year, or
vise versa. Like the meaning of coefficients in Model (1),
coefficients of six independent variables, β1, β2, β3, β4, β5,
β6, explain how the Annual Giving changes with the changes of
these six independent variables.
(2). Analysis of Regression Results
Based on the Model (1), I use R-software to do the
regression work. The results of regression are listed in Table
2.
Table 2: Effects of Performance Measurements on Endowment
Assets
Estimate
(αn)
Std.
Error
T Value P Value Sig.
α0-44.02586 1.44655 -30.435 <2e-16 ***
lnTRexpi,t0.02263 0.01445 1.566 0.117548
lnSATavgi,t7.81373 0.20542 38.037 <2e-16 ***
lnFRewdi,t0.14956 0.04528 3.303 0.000982 ***
lnNAMembi,t0.13143 0.03780 3.477 0.000522 ***
lnPostDAi,t0.04093 0.02803 1.460 0.144413
lnDoctAi,t0.15224 0.01925 7.908 5.35e-15 ***
R- Square 0.7539
Adjusted
R-Square
0.7528
P – Value <2.2e-16 ***
Notes: “*”, “**”, “***” means significance in 10%, 5%, and 1%
levels respectively.
From the results above, four coefficients of four
independent variables are obviously positive, which means: the
increase of Average SAT Scores, the increase of Faculty
Rewarded, the increase of National Academic Membership and the
increase of Doctor Awarded of a university in one year will
lead to an increase of Endowment Assets next year. That is,
good performance significantly increase the Size of Endowment
Assets by sending good signals to donors that the potential
good research ability of a university and the potential high
education quality of a university. Moreover, the P Value of
variables lnSATavgi,t, lnFRewdi,t, lnNAMembi,t and lnDoctAi,t
are quite small with three stars “***”, meaning that their
effects are extremely significant under the 1% significant
level.
On the contrary, the P-Value of the lnTRexpi,t and
lnPostDAi,tare not significant, meaning the changes of Total
Research Expenses and Number of Post Doctor Awarded in one
year have no obvious impacts on the Endowment Assets next
year. The R-Square Value and Adjusted R-Square Value are
0.7539 and 0.7528, respectively, meaning that changes of all
the six independent variables are completely enough to explain
the changes of dependent variable: Endowment Assets. The
explained power is 75%. The P-Value of the whole Model (1) is
“< 2.2e-16” which is quite small, pointing out the Model I
designed is quite appropriate and the degree of fitting is
quite powerful.
Using same method to do the regression, the results of
Model (2) regression are listed in Table 3.
Table 3: Effects of Performance Measurements on Annual Giving
Estimate
(βn)
Std.
Error
T Value P Value Sig.
β0-21.19810 1.27240 -16.660 <2e-16 ***
lnTRexpi,t0.06432 0.01271 5.062 4.72e-07 ***
lnSATavgi,t4.15510 0.18070 22.995 <2e-16 ***
lnFRewdi,t0.22635 0.03983 5.683 1.62e-08 ***
lnNAMembi,t0.09407 0.03324 2.830 0.00472 **
lnPostDAi,t0.12573 0.02465 5.101 3.8e-07 ***
lnDoctAi,t0.07862 0.01693 4.643 3.76e-06 ***
R- Square 0.7362
Adjusted
R-Square
0.7418
P – Value <2.2e-16 ***
Notes: “*”, “**”, “***” means significance in 10%, 5%, and 1%
levels respectively.
According to Table 3, unlike the regression results in
Model (1), six coefficients of all the independent variables
are positive, signifying that the increase of each item in t
year will lead to the increase of Annual Giving in t+1 year.
Given the fact that the smaller the P Value, the more
significance of the regression results. According to the P
Value in Table 3, Total Research Expenses, SAT Average Scores,
Faculty Rewards, Post Doctor Awarded and Doctor Awarded are
all significant under the 1% significance level, while the
National Academic Membership is less significant under 5%
significance level than others. The results point out that the
improvement or better performances of a university in one year
will obviously have positive impact on the increase of the
Annual Giving next year since good performances send donors
signals that one university is deserved to be given money for
research and education.
Once a university expends its research by increasing its
Total Research Expenses, enhances its students’ quality with
higher Average SAT Scores, improves its’ research by adding
the National Academic Membership and makes advance in future
research by cultivating more Post Doctors and Doctors, it
sends signals to donors, attracts their attentions and finally
increases the Endowment and Annual Giving, thus verifying the
hypothesis H1, H2 and H3 I proposed in the beginning.
Robustness Test
In order to testify the robustness of the regression
results I listed above, I certified my results in other two
areas.
1. Add two performance measurements into Independent Variables
In both Model (1) and Model (2), I added another two
independent variables in the original six measurements:
Federal Research Expenditures and National Merit and
Achievement Scholars. After the same regression process, the
conclusions remain unchanged.
2. Group by Ranking
Given the fact that the original samples include different
level of universities, I classified those data into three
groups: top 100, 100-200 and universities ranking after 200.
Clarifying different levels, I did the same regression
process, the conclusion remain unchanged.
Conclusion
My paper, picking the university data as research samples,
taking advantages of the Multiple Linear Regression Model and
using the empirical method testified whether performances of
universities in the current year have impact on the size of
Endowment Assets and Annual Giving next year. The regression
results signify: Better performances of universities in
current year significantly increase the total number of
Endowment Assets and Annual Giving next year by sending to
donors and public signals that universities with better
performances have good research potential and high-quality
education potential.
The findings of my paper not only ensure the relations
between performances and Endowment /Annual Giving with
statistic evidence, but also enrich the university researches,
hence enlightening universities to find bright way for their
further development and for maintaining their long-term
successes, especially for top universities.
References:
1. Larry L. Lesloe, Sheila Slaughter, Liang Zhang, “How do
Revenue Variations Affect Expenditures Within U.S. Research
Universities? ”, Research in Higher Education, (Sep. 2012),
pp.614-639
2. Josh Lerner and Jialan Wang, “Secrets of The Academy: The
Drivers of University Endowment Success”, National Bureau of
Economic Research, (2008);
3. Stephen G. Dimmock, “ Background Risk and University
Endowment Funds”, The Review of Economics and Statistics,
(Aug. 2012), pp.789-799;
4. Donald L. Basch, “Changes in the Endowment Spending of
Private Colleges in the Early 1990s,” The Journal of Higher
Education, (May. 1999), pp.278-308;
5. Harold A. Davidson, “Investing College Endowment Funds: A
Comparison of Internal and External Management,” (Feb.1971),
pp. 69-74