1. IntroductionThe canonical model

1. Introduction
The canonical model of portfolio choice (Markowitz, 1952), predicts first that households
will hold a positive share in risky assets and second that managing risk will require a
higher diversification of an investment portfolio optimizing its risk-return characteristics.
Previous empirical studies based on microeconomic data have verified that a great
fraction of investors hold very few risky stocks (Haliassos, 1995). A larger theoretical
model has been motivated by these results.
Recently, there has been greater attention given to the lack of financial skills as a driver
of poor financial behavior. The output measure of the financial decisions has most often
been a lack of savings (Hilgert and Beverly, 2003; Lusardi, 2007), while recent research
studies focus on stock market participation (Rooij et al., 2011), over indebtedness (Lusardi
and Tufano, 2009), and mortgage delinquency (Gerardi et al., 2013).

In recent years, the economic literature has uncovered two features of investor
behavior at odds with standard portfolio theory; the lack of portfolio diversification
and proximity investment, contrary to what standard theory would require, investors
hold highly undiversified portfolios made up of a limited number of stocks (Barber and
Odean, 2000) and they mostly select these stocks on the basis of geographical or
professional proximity to them (Coval and Moskowitz, 2001).
Also many recent papers have examined whether the financial illiteracy reflects
fundamental mispricing of the asset, over-optimistic expectations about future price
appreciation, or the buyer’s failure to correctly anticipate consumption risk, in the
specific context of financial plans for housing. The empirical results indicate that
financially illiterate households pay more for a given house than do more sophisticated
buyers since the transaction price reflects the bargaining power of the buyer and the
seller, and depends on the state of the credit market, the state of the housing market
and the information sets of the buyer and the seller (Turnbull and Sirmans, 1993;
Harding et al., 2003).
In the last papers, several authors have also associated financial literacy with the
portfolio diversification (e.g. Kimball). Contrary to what is required by the standard
theory, investors own a greater undiversified portfolios which are composed of a small
number of different assets (Goetzmann and Kumar, 2005). In fact, most investors hold
nearly all their wealth in local stocks (French et al., 1991; Goetzman et al., 2004),
concentrate their portfolio in the equity market (Barber et al., 2003) and choose these
stocks generally on the basis of geographical or professional proximity (Coval and
Moskowitz, 1999; Curcuru et al., 2005).
Therefore, rational and informed investors are supposed to have diversified
portfolios, irrespective of their levels of risk aversion. In the framework of traditional
finance theory, important contributions have been made to explain why a rational
investor may own a less diversified portfolio than what might be considered optimal.
Simultaneously, a group of alternative explanations draws interest to the fact that
the thoughts about diversification vary systematically among groups of agents with
different characteristics (Shiller, 2002). Portfolio diversification may vary according
to differences of the investors’ age (DaSilva and Giannikos, 2004), competence, wealth
(Bertaut, 1998; Kumar, 2005), trading experience (Nicolasi et al., 2004), occupation
(Christiansen et al., 2005).
In this paper, we try to determine the extent to which differences in the investors’
diversification behavior can be justified by differences in their financial literacy. We also
control socioeconomic differences and behavioral differences among groups of investors
and consider three distinct aspects of financial literacy: specific financial knowledge, the
investors’ educational level (used as a proxy for their ability to use gathered information)
and the sources of information commonly used by investors as the basis for their
financial choices. Our results provide some potentially useful evidence on the investors’
behavior in emerging financial markets. Most studies focus on well-developed financial
markets and very little is known about the investor’s profile, motivations and behavior
less-developed markets. By using data from a survey of Tunisian individual investors,
we contribute to bridging this gap.
The remainder of this paper is structured as follows: we discuss the previous
studies and measures of financial literacy used in our work. The section after
that presents the model, discusses the data, and describes the construction of
variables and the estimation method used. The final section of the paper reports the
empirical results.