Hanna, S. D. (2002). Research esoterica. Financial Counseling and
Planning , 13(1), iii-vi.
Sherman D. Hanna 1
Readers of this journal
are interested in topics such as why people get into credit problems,
how should people invest for retirement and other goals, and why do
people make bad financial decisions. Writers for this journal are
mostly academics who get rewarded for publishing esoteric research. In
order to make research more understandable, some basic ideas of
research are presented in this article.
Theoretical approaches can
be prescriptive (normative) or descriptive (Hanna, 1989). Those trained
in finance tend to focus on prescriptive theories, such as, what is the
optimal portfolio? Those trained in other social sciences tend to focus
on descriptive theories, although it is more appropriate to refer to
these theories as
having a positive science approach. The goals of positive social
are to describe, explain, predict, and perhaps to control behavior.
The most important
prescriptive theoretical approach in economics and finance is expected
utility theory. Although the expected utility model has been criticized
as not adequately describing human behavior, it is the best
prescriptive model available (Schoemaker, 1982). In finance, there are
various levels of prescriptive approaches, some explicitly using
expected utility theory (e.g., Hanna & Chen, 1997) and many using
approaches such as the Capital Asset Pricing Model (CAPM) which
are based on expected utility theory (e.g., Bodie & Merton, 1998,
13). This journal has published approximately 20 purely prescriptive
i.e., articles that derive a prescriptive rule but do not examine how
actually behave. In addition to articles explicitly using expected
theory for optimal portfolios and optimal credit use (Fan, Chang &
1992), some articles use cost-benefit analyses that are consistent with
utility analysis, e.g., Hatcher (2000) or any level of risk aversion,
Kish and Hogan (2000).
This journal has also
published at least 30 articles with an explicit discussion of a
prescriptive theory, e.g., the Life Cycle Model, and also including an
empirical analysis of household behavior. Topics include retirement
adequacy (Hatcher, 1997; Yuh, Montalto & Hanna, 1998), emergency
fund adequacy (e.g., Chang, Hanna & Fan, 1997), and debt (e.g.,
Chen & Finke, 1996). The advantage of specifying optimal or
efficient behavior and then analyzing actual behavior is that
problems can be more clearly identified and implications for consumer
and public policies can be discussed.
Use of economic theory for
describing and explaining behavior has limitations, so a number of
theories have been proposed to better explain how people actually
behave. Before dismissing economic theories, though, it is important to
recent developments. Unfortunately, most discussion of the Life Cycle
Model in articles in this journal end with the state of the model as of
in Ferber (1973). There have been many extensions of the theory in the
past 30 years, and all authors proposing to discuss the Life Cycle
should at least start with a recent review such as Browning and Lusardi
(1996). As they note (p. 1798), there is a problem with terminology, as
“life-cycle model” to refer to the original simplistic Modigliani
but others use the same term to “… refer to any model in which agents
forward looking intertemporal consumption problems that may allow for,
bequests, habits and liquidity constraints.”
model has been used in at least five articles analyzing household
behavior (e.g., Parrotta & Johnson, 1998). One reason for the low
number of articles using this model since 1990 is that the large
national datasets readily available to researchers do not include the
“throughput” variables that are a crucial part of the model.
Behavioral economics or
finance concepts have been mentioned in a number of articles, e.g.,
Xiao (1995). Rabin (1998) has a review of some of the basic literature
in this field. Typically, as with Xiao (1995), there is a discussion of
the limitations of standard economic theory and some discussion of
behavioral economics. Such research can be useful in helping to
understand how people behave, but
it is not as clear what the research implies about what experts should
people to do.
It is interesting to
discover the relationship between two variables. For instance, does age
affect the likelihood that a college student will take out an
educational loan? In Cha and Weagley (2002, p. 66), Table 2 shows that
the mean age of borrowers was significantly lower than the mean age of
non-borrowers. With all bivariate relationships, the researcher should
also consider the possibility that
other factors have caused the relationship, so multivariate analysis
controls for other variables is needed.
Researchers can analyze the
effect of many variables at the same time on the dependent variable,
e.g., whether a college student took out an educational loan. In Cha
and Weagley (2002, p. 67), Table 3 shows that age did not have a
significant effect on whether a college student took out a loan, and
also, among those who took out loans, age did not have a significant
effect on the amount borrowed. In all multivariate analyses, the
appropriate interpretation is, for instance in this example, all other
things equal, age did not have a significant effect on whether a
student took out a loan.
The Significance of
In Table 3 of Cha and
Weagley (2002, p. 67), results are listed as significant or not
significant at the 0.05, 0.01, and 0.001 levels. These significance
levels indicate that the relationship between a particular independent
(or “predictor”) variable and the dependent variable of interest might
occur by chance 5%, 1%, or 0.1% of the time, respectively.
One way to get an
intuitive understanding of significance tests is to think about tossing
a coin. If one suspected that a coin was weighted to come up heads all
of the time, how many times in a row would it have to come up heads on
tosses before you were confident that it was actually biased toward
heads? If the coin came up heads four times in a row, you might get
suspicious, given that there is
only one chance in 16 of four heads in a row. In terms of the
conventions of statistical significance tests, the significance level
is 0.065 or 6.5%. The most common standard of significance tests is
0.05 or 5%, simply because higher probability levels are not very
convincing. There is a case to be made for even stricter standards, not
only because of various problems in the gathering of household data,
but also because researchers are investigating the effects of many
variables. Imagine a room of 30 people all tossing coins, and one
person flips heads five times in a row. Does that person have a
coin weighted to come up heads all of the time? Even though there is
a 3% chance of that result by one person, there is a much higher chance
one person among the 30 people will have that result. We could be more
confident of the result if we insisted on p < 0.01, in this simple
example, having seven heads in a row, and even that would not give us
complete confidence, just a basis for further investigation.
Miscellaneous Research Issues
What Variables Should
One interesting issue
facing researchers is which variables to include in analyses. Large
national datasets such as the Survey of Consumer Finances have many
hundreds of variables. A theoretical framework may only have
predictions about a few variables. Often researchers will include other
variables because they are of interest for policy, marketing purposes,
or educational programming. In bivariate analysis, a variable might
have significant relationships with the dependent variable of interest
because the variable is related to income or other economic variables,
though then in multivariate analysis, the effect might be reduced
because the effects of income and wealth are being controlled. Race and
ethnic identification are often of interest for various reasons, for
to answer questions about discrimination.
Race and Ethnic
Fan and Burton (2002)
included a question about race/ethnic identification in the survey they
designed. The survey questions, available in a link on the journal web
page, include the term Caucasian . Even though
previous articles in this journal have used that term to refer to those
identifying themselves as white, I would urge future authors to avoid
the term and those designing surveys in the future to avoid the
term, as it is an archaic term that has no scientific basis. It is true
that there is controversy about many designations for racial and ethnic
but researchers working with U. S. data should follow the practices of
major U. S. government-sponsored surveys. For instance, the Survey of
Consumer Finances asks about race in the following way: “Which of these
categories do you feel best describe you: white, black or
African-American, Hispanic, Asian, Native American, or another race?”
Interviewers are instructed to code all that apply (Board of Governors
of the Federal Reserve System, 2000). The United States Census first
asks whether each person identifies himself/herself as
“Spanish/Hispanic/Latino” and allows for more detailed categories such
as Mexican-American (U. S. Census, 2000). The Census then asks each
person to mark one or more races, with White, Black/African
Researchers are limited
by what is included in datasets, and by statistical limitations. For
the Survey of Consumer Finances, one serious limitation is that the
dataset released to researchers does not identify subcategories in the
“Other races” category, so that it is impossible to distinguish Native
Americans/American Indians from those identifying themselves as Asian
or Pacific Islander. Even researchers working with datasets listing
many racial categories have to keep in mind that it does not make
statistical sense to include a category that includes relatively few
people in the particular sample. In multivariate analysis controlling
for income, for instance, one might ask how many elderly high income
Hispanics are in the sample? If the number is low, in some cases the
researcher might appropriately decide to combine several categories
into one “other” category to compare to White non-Hispanics.
Volpe, Kotel and Chen
(2002) report on a survey of investment literacy. I am not sure that
investors need to know all of the answers to be successful over the
long run – one of the strongest arguments for index funds is that the
investor can make one decision and then not think about investing for
many years. However, their sample is of online investors, some of whom
do online trading, so clearly there
is some need for knowledge of investing details. Knowledge tends to
increase with age, and is higher for males than females. Presumably
males have more experience than females of the same age. This type of
research can be helpful in designing educational programs.
Hogarth and Hilgert
(2002) examine consumers with high-rate home loans. An important
question is why such consumers have high interest rates. Obviously, a
poor credit rating can be one cause of having to accept high rates,
though the authors find that factors not necessarily related to credit
risk have an influence.
Cha and Weagley (2002)
also look at credit, namely taking out student loans. It makes as much
sense to borrow for education as for a home, so this is certainly one
type of credit use where one should not assume that the decision of
borrow is necessarily bad. However, Fan and Burton (2002) investigate
college student perceptions of what are status goods, which might
influence credit use to the extent that students borrow to purchase
goods they feel will enhance their status. Cavanagh and Sharpe (2002)
investigate the impact of debt on contributing to retirement accounts,
and find that installment debt had an effect on participation.
Howton, Howton and Olson
report on research that has very practical implications for investors -
you are going to invest in an Initial Public Offering (IPO) which stock
should you prefer? Obviously IPOs have been in the news for a number of
years (e.g., Casey, 2002). Howton et al. (2002) do not offer a
definitive explanation as to why NASDAQ IPOsS are better for investors
than NYSE IPOs, but the result is both an intriguing anomaly that
investors might be able to exploit and a topic for further research.
Zielonka (2002) analyzes
perceptions of financial analysts, and has nice review of objective
research about what really does influence stock prices. Originally his
manuscript included “Polish” in the title, but I felt that a similar
survey and analysis of analysts in the United States and other
countries might have similar results. It would be interesting for
someone to design a similar survey for analysts in the United States.
Bodie, Z. & Merton,
R. C. (1998). Finance. Prentice Hall.
Board of Governors of
the Federal Reserve System. (2000). Codebook for the 1998 Survey
of Consumer Finances. Washington, DC: Federal Reserve. [Available:] URL:
Browning, M. &
Lusardi, A. (1996). Household saving: Micro theories and micro facts. Journal
of Economic Literature, 34(4), 1797-1855.
Casey, M. (2002; September 30).
Dot-com IPO pricing baffles economists. The Wall Street Journal,
240(64), p. A2.
Cavanagh, J. A. &
Sharpe, D. L. (2002). The impact of debt levels on participation in and
level of discretionary retirement savings. Financial Counseling
and Planning , 13 (1), 47-59.
Cha, K.-W. &
Weagley, R. O. (2002). Higher education borrowing. Financial
Counseling and Planning , 13(1), 61-73.
Chang, Y. R., Hanna, S.
& Fan, J. X. (1997). Emergency fund levels: Is household behavior
rational?, Financial Counseling and Planning, 8(1),
Chen, P. & Finke,
M. S. (1996). Negative net worth and the life cycle hypothesis . Financial
Counseling and Planning, 7, 87-96.
Fan, J. X. &
Burton, J. R. (2002). Students’ perception of status-conveying goods. Financial
Counseling and Planning, 13(1), 35-46.
Fan, X. J., Chang, Y.
R. & Hanna, S. (1992). Optimal credit use with uncertain income. Financial
Counseling and Planning, 3, 125-133.
Ferber, R. (1973).
Consumer economics. A survey, Journal of Economic Literature,
11 (4), 1303-1342.
Hanna, S. (1989).
Optimization for family resource management. Proceedings of the
Southeastern Regional Association for Family Economics-Home Management
Conference, 4-16. Available URL:
Hanna, S. & Chen,
P. (1997). Subjective and objective risk tolerance: Implications for
optimal portfolios. Financial Counseling and Planning, 8(2),
Hatcher, C. B. (1997).
model of desired wealth at retirement, Financial Counseling and
Planning , 8(1), 57-64.
Hatcher, C. B. (2000).
households establish emergency funds? Financial Counseling and
Planning , 11(2), 77-83
Hogarth, J. M. &
Hilgert, M. A. (2002). A profile of consumers with high-rate home
loans. Financial Counseling and Planning, 13(1),
Howton, S. D., Howton,
W. & Olson, G. T. (2002). The role of exchange listing in the
initial and aftermarket performance of IPOs. Financial Counseling
and Planning , 13(1), 75-85.
Kish, R. J. &
Hogan, K. M. (2000).Small stocks for the long run. Financial
Counseling and Planning , 11(2), 21-31.
Parrotta, J. L. &
Johnson, P. J. (1998). The impact of financial attitudes and knowledge
on financial management and satisfaction of recently married
individuals. Financial Counseling and Planning, 9(2),
Rabin, M. (1998).
Psychology and economics. Journal of Economic Literature, 36(1),
Schoemaker, P. J. H.
(1982). The expected utility model: Its variants, purposes, evidence
and limitations. Journal of Economic Literature, 20(2),
U. S. Census (2000).
Census 2000 Questionnaire. U. S. Department of Commerce. Retrieved
October 6, 2002. URL: http://www.census.gov/prod/
Volpe, R. P., Kotel, J.
E. & Chen, H. (2002). A survey of investment literacy among online
investors. Financial Counseling and Planning, 13(1),
Xiao, J. J. (1995).
Patterns of household financial asset ownership. Financial
Counseling and Planning , 6, 99-106.
Yuh, Y., Montalto, C.
P. & Hanna, S. (1998). Are Americans prepared for retirement? Financial
Counseling and Planning, 9 (1), 1-12.
Zielonka, P. (2002). How financial analysts perceive
macroeconomic, political news and technical analysis signals. Financial
Counseling and Planning , 13(1), 87-95.
1. . Sherman D. Hanna, Professor, Consumer
Sciences Department, The Ohio State University, 1787 Neil Ave.,
Columbus, OH 43210-1295. Phone: 614-292-4584. E-mail: