Impact of Tunisian Political and COVID-19 Crisis on Asset Allocation:

Traditional Theory of Portfolio Selection Versus Behavioral Theory

YASMINA JABER

Department of Finance and Accounting,

Higher Institute of Management Gabes University,

Jilani Habib street, Gabès 6002,

TUNISIA

effectively manage the portfolio. Within this framework, various strategies have been implemented such as

modern portfolio theory (MPT) and behavioral portfolio theory (BPT). We concentrate on portfolio

optimization for two alternative approaches: the MVT and the BPT. This study aims to compare portfolios

generated by these two approaches during political and COVID-19 crisis periods using data from the Tunisian

stock market exchange for the period 2009 –2022. The results show that in the case of a higher degree of risk

aversion induced by investors’ BPT, all the stock is located at the top right of the mean-variance frontier.

However, during the crisis, the portfolios selected by rational investors were not systematically selected by

irrational investors, even if the optimal portfolio of BPT coincides with the Markowitz efficiency frontier. The

results indicate that the crisis induces simultaneously an increase in risk and a sharp decrease in the portfolio

return of individuals who follow the mean-variance theory of Markowitz.

Investing in the financial market is a means used to

build wealth. Successful investing relies on

knowledge of financial industry developments,

investor behavior, and portfolio management

techniques. Multitudes of strategies have been

implemented over time to effectively manage the

portfolio. In this context, we cite the mean-variance

theory (MVT), developed by [1], which was

revolutionary in the world of portfolio management.

that the distribution of returns is normal as

mentioned in [9], [10], [11], [12], [13]. To

overcome this shortcoming, [14], was the first

researcher to use probability as an alternative risk

measure to that proposed by Markowitz which takes

recent developments, the theory of Markowitz and

that of [20], have become the central research

hypotheses. This gives rise to a new flow of

literature that attempts to compare the asset

allocation generated by the BPT model with that

generated by the MVT model. There are two

opposite ways of literature. The first path is that of

[21], [22], [23], [24], [25], [20], [26], who compared

the efficiency frontier of [1], with the efficiency

frontier of [20], and has shown that generally, these

portfolios, a mix of bonds and lottery tickets, on

expected wealth, desire for security, and level of

aspiration. This was asserted by [26], who proved

that the optimal portfolio of a normal investor, who

also considers the three dimensions of benefits, is

lower than the optimal portfolio of a rational

investor who ignores expressive and emotional

benefits. Behavioral investors are willing to give up

a certain portion of the expected return to gain

expressive and emotional benefits. This is precisely

theory of optimal portfolio diversification varies

from one investor to another, depending on the

investor's attitude towards risk, while the optimal

portfolio in terms of BPT varies from investor to

investor not only because of the different levels of

risk tolerance but also because of the different

wants, needs, biases, habits, preferences, and

emotions of these investors.

The second path is that of [27], [28], [29], [30],

as well as, [19], who have shown that certain

assets are normally distributed. While the normality

hypothesis is often accepted in the literature but not

verified in real markets as shown in [34], [19], [35],

[36]. Whereas, [37], [38], have shown that the

optimization of the mean-variance portfolio and the

The paper is organized as follows: Section 2 is a

To construct the sample for our study, we first

We start our empirical part by calculating the

We find that, during the four periods studied, the

values of Kurtosis were greater than three. This

means that the distributions of returns on securities

are sharper than the distribution of the normal

distribution. As a result, the distributions of returns

on the securities studied are sharper with thick tails

on the left and the right. Consequently, the

securities in the portfolio is greater than thirty to

achieve the optimal level of diversification.

However, after 16 years, it has shown that the

portfolio is as diversified as possible by one that

contains several securities. The results show that

their number is greater than 120. [44], have shown

by a study carried out on 40,000 investors from

1991 to 1996, that the average number of securities

that make up a portfolio is 4. For that, the principle

enunciated by [1], is far from being applied in the

number of securities. First, we randomly chose,

among the 43 available stocks, the number of assets

in each portfolio where n can take 2, 3 up to 43.

Subsequently, we calculate for each value of n the

average variance of 10,000 randomly constructed

portfolios made up of n stocks of equal weights.

The results of diversification for the four sub-

periods are presented in Figure 1.

We note that for the three sub-periods the variance

of the portfolio which contains all the securities

diversified the portfolio, the lower its risk exposure.

This conclusion was affirmed by Table 2.

B. Efron in 1979 from our series of daily returns.

RJ is given by

initial wealth, his portfolio. Therefore, the first

investors, we consider several configurations for α

4.1 Impact of the Political and COVID-19

Crisis on the Optimization of Markowitz

Portfolios

To study the effect of the Tunisian revolution and

COVID-19 crisis on the portfolios of investors

according to the mean-variance model of [1], we

covid-19 crisis). In Figure 2, we present the 100000

Pandemic on Markowitz's portfolio optimization

During the period of the revolution and Covid-19,

we find that these periods of disruption induce both

an increase in risk and a decrease in the expected

return on the 100,000 portfolios of [1]. Thus, the

lowest negative return of the portfolios can be

observed in times of crisis. We also observe that the

efficiency frontier for the two crisis periods is below

According to Table 3 and Table 4 the number of the

optimal portfolio constructed according to the

model of Shefrin and Statman (2000) during the

four sub-periods, from 1,200,000 draws made, is

respectively equal to 814970, 320330, and 701680

The same result is noted for the period of the

COVID crisis.

We also note, for all the sub-periods, that the

given level of profitability, are the least numerous.

And as a result, their set of secure wallets is hugely