UNIT 9: RESEARCH IN BEHAVIORAL ECONOMICS
1. Introduction
2. Measuring preferences
3. Usefulness of preferences measures
4. Preferences and financial and economic outcomes
• Micro-evidence on risk tolerance
• Risk tolerance, growth and entrepreneurship
• Risk tolerance and fertility (children)
• Trust and corporate cash holding
5. Determinants of preferences
• Shocks and preferences
• Shocks and preferences
1. INTRODUCTION
“In economics, preferences are evident in theoretically optimal choices or real (behavioural) choices when people
decide between alternatives.” Preferences are present in every decision making.
1. Preferences can be related to the perception of risk: Risk preferences
2. Preferences can be related to trust, reciprocity, fairness, altruism: Social preferences
3. Preferences can be related to time discounting: Time preferences. People have different perceptions of time:
people are ore present bias than other. We show how present bias or time inconsistency can be very important in
nudges to make better decisions.
Preferences are of fundamental importance for human behaviour. But for a certain period of time preferences has been
shown as innate meaning that preferences were assumed as something that is given. Everyone has a level of preferences
that are exogenous, and we do not know where they come from. So, for a long time we were not able to investigate the
determinants of preferences, the drivers of preferences, their effect in economic outcomes (financial or economic
development. Etc). One of the main reasons, it is that we did not have access to data bases. They were used experiments
and incentives measures that were costly for cross-country studies.
This started to change with the appearance of survey measures and the World Value Surveys where they were asked
questions about people’s preferences.
Risk preferences have been shown to affect various economic and financial outcomes from financial investments to job
and education choices, marriage decisions and health behaviours etc…
2. MEASURING PREFERENCES
How can we measure preferences?
2.1 SURVEY MEASURES
Survey measures: They are based usually on self-declared levels of preferences (e.g., How do you see yourself? Are you
generally a person who is fully willing to take risks?…etc.). They are cost-effective and easier to obtain compared to
incentivised measures
Example of measure of risk preference from German Socio-Economic Panel (SOEP). Here we have a categorical value.
Basically, you have the scale, and every individual tries to pick the level of his/her risk aversion. This measure is self-
declared risk perception. It can be seen if this risk tolerance changes by sector (for example), maybe you are willing to
take more risks if you are driving and less if you are at work. In this way, Dohmen and Falk produce preferences
measures that are widely use in research nowadays, they do duplicate some of the word value surveys for the measure
of risk tolerance…etc.
, Example of their surveys:
For instance, the Intertemporal choice sequence using staircase method for patients: How long are you willing to wait? Social
preference measures: As positive reciprocity measures.
When having categorical variables, we can see the correlations meaning how these variables move together (it is not
really about causality). We can do studies about correlations, but it is really difficult to do causal analysis as we need the
proper tools to isolate the effects. Because if we have two variables and say: the hedge causes risk tolerance or trust
for example, but how do you say that? Because the are other variables can affect trust at the same time. Imagine you
have education and risk tolerance, and you want to try to study the casual relationship between these two variables. If
education has a casual on risk tolerance, you cannot say that with only correlation. And, even if you find a relationship,
it cannot be casual because so many other variables can have an effect on risk tolerance and on the education at the
same time (background, age). So, in order to talk about causality, you really need to choose the suitable methods.
If you are in a macrolevel, meaning you sue macrolevel variables, on a country-level for instance, you talk about
correlation. So, it is very difficult to infer in causality except if you have individual characteristics.
2.2 INCENTIVISED MEASURES
Incentivised measures: We talk about incentivised measures for e.g., with the experiments as the one we conducted in
class. Meaning you do an experiment, and you pay for example the participant. In this way, they are incentivized usually
1. Introduction
2. Measuring preferences
3. Usefulness of preferences measures
4. Preferences and financial and economic outcomes
• Micro-evidence on risk tolerance
• Risk tolerance, growth and entrepreneurship
• Risk tolerance and fertility (children)
• Trust and corporate cash holding
5. Determinants of preferences
• Shocks and preferences
• Shocks and preferences
1. INTRODUCTION
“In economics, preferences are evident in theoretically optimal choices or real (behavioural) choices when people
decide between alternatives.” Preferences are present in every decision making.
1. Preferences can be related to the perception of risk: Risk preferences
2. Preferences can be related to trust, reciprocity, fairness, altruism: Social preferences
3. Preferences can be related to time discounting: Time preferences. People have different perceptions of time:
people are ore present bias than other. We show how present bias or time inconsistency can be very important in
nudges to make better decisions.
Preferences are of fundamental importance for human behaviour. But for a certain period of time preferences has been
shown as innate meaning that preferences were assumed as something that is given. Everyone has a level of preferences
that are exogenous, and we do not know where they come from. So, for a long time we were not able to investigate the
determinants of preferences, the drivers of preferences, their effect in economic outcomes (financial or economic
development. Etc). One of the main reasons, it is that we did not have access to data bases. They were used experiments
and incentives measures that were costly for cross-country studies.
This started to change with the appearance of survey measures and the World Value Surveys where they were asked
questions about people’s preferences.
Risk preferences have been shown to affect various economic and financial outcomes from financial investments to job
and education choices, marriage decisions and health behaviours etc…
2. MEASURING PREFERENCES
How can we measure preferences?
2.1 SURVEY MEASURES
Survey measures: They are based usually on self-declared levels of preferences (e.g., How do you see yourself? Are you
generally a person who is fully willing to take risks?…etc.). They are cost-effective and easier to obtain compared to
incentivised measures
Example of measure of risk preference from German Socio-Economic Panel (SOEP). Here we have a categorical value.
Basically, you have the scale, and every individual tries to pick the level of his/her risk aversion. This measure is self-
declared risk perception. It can be seen if this risk tolerance changes by sector (for example), maybe you are willing to
take more risks if you are driving and less if you are at work. In this way, Dohmen and Falk produce preferences
measures that are widely use in research nowadays, they do duplicate some of the word value surveys for the measure
of risk tolerance…etc.
, Example of their surveys:
For instance, the Intertemporal choice sequence using staircase method for patients: How long are you willing to wait? Social
preference measures: As positive reciprocity measures.
When having categorical variables, we can see the correlations meaning how these variables move together (it is not
really about causality). We can do studies about correlations, but it is really difficult to do causal analysis as we need the
proper tools to isolate the effects. Because if we have two variables and say: the hedge causes risk tolerance or trust
for example, but how do you say that? Because the are other variables can affect trust at the same time. Imagine you
have education and risk tolerance, and you want to try to study the casual relationship between these two variables. If
education has a casual on risk tolerance, you cannot say that with only correlation. And, even if you find a relationship,
it cannot be casual because so many other variables can have an effect on risk tolerance and on the education at the
same time (background, age). So, in order to talk about causality, you really need to choose the suitable methods.
If you are in a macrolevel, meaning you sue macrolevel variables, on a country-level for instance, you talk about
correlation. So, it is very difficult to infer in causality except if you have individual characteristics.
2.2 INCENTIVISED MEASURES
Incentivised measures: We talk about incentivised measures for e.g., with the experiments as the one we conducted in
class. Meaning you do an experiment, and you pay for example the participant. In this way, they are incentivized usually
