Which Characteristics or Experiences Influence Interpersonal Trust? Comments and Criticism of the Paper "Who trusts others" by Alesina and La Ferrara
Term Paper (Advanced seminar) 2015 11 Pages
3. Data and descriptive statistics
4. Econometric evidence
5. Main findings / conclusions
6. Criticism and further research
7. Own conclusions
In the following I will present the paper “Who trusts others?” from the authors Alberto Alesina and Eliana La Ferrara. Afterwards I will comment on the paper, criticise some points of it and finally I will give proposals for further research.
The structure of this term-paper is as follows: At first I will give a general introduction into the theoretical basics of this paper. This part will explain why it is important to analyse interpersonal trust and it will contain information about the main research hypothesis of the two authors. The second part is about the data and descriptive statistics of the reference paper. This section will contain information about the most important variables used for the regressions of the paper “Who trusts others?” and it will present the first conclusions about the data. The third segment of this paper will expose the econometric evidence. Here I will give a summary about every table used in the reference paper and explain the most important findings of the regressions. After that there will follow a very short segment where I will summarise the main results of the reference paper. In the last part I will give a comment about the reference paper and will add some criticism. This section will end with proposals for further research based on the current economic perspectives. The paper will end with a short summery of my comments and research proposals.
The reference paper “Who trusts others?” written by Alberto Alesina and Eliana La Ferrara analyses the question which characteristics or experiences influence interpersonal trust. But why is this question important?
Trust is one aspect of “social capital” which is defined by Putnam (1999) as “features of social life, networks, norms, trust that enable participants to act together more effectively to pursue shared objectives.” Trust is very important in economic activities. It is reducing transaction costs and stimulates the efficiency of governments. Therefore the authors have formulated five hypotheses about what determines trust.
First, trust may be dependent on moral beliefs or ones culture. If that is true the individual’s education or religion are playing a big role in building trust towards others. Second, past experiences may affect trust. If an individual was treated well in the past, it is more likely that he/she will trust others in the future. Third, trust may be based on the similarity of people. An individual is more likely to trust a person of the same family, race or ethnic group. From this it follows that in a homogeneous community trust is higher than in a heterogeneous community. Fourth, a longer interaction among people may support trust. When people cooperate frequently it is possible to retaliate in the next period if the other person deviates from cooperation. Fifth, better legal institutions lead to more trust. A community with good legal institutions will have less criminals. Hence a person living in this community has less incentives not to trust.
The factors influencing trust that can be derived from these hypotheses are:
- Individual culture, tradition, and religion
- How long an individual has lived in a community with stable composition
- Recent personal history of misfortune
- The perception of being part of a discriminated group
- Several characteristics of the composition of one’s community
3. Data and descriptive statistics
In the following I will describe the data collection of my reference paper. The main data source of the paper “Who trusts others?” is the General Social Survey (GSS) from 1974 to 1994. This survey interviews several people from a nationally representative sample every year.
The most important question for the authors was the following: “Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people?” If the interviewed person answered “most people can be trusted”, he was categorised as trusting and got the value 1. If the person answered “you can’t be too careful” or “it depends”, he was categorised as non-trusting and got the value 0. Therefore the dependent variable is trust and is a dummy with the values 0 or 1. Another variable the authors want to explain is trust in other institutions, like financial institutions, the army, congress, medical doctors, etc.
The other variables they use from the GSS in their analysis are individual characteristics like education, marital status, income and the recent history of the individual (a previous divorce can be caused by trust issues).
The community variables are income inequality and racial and ethnic fragmentation. The racial fragmentation index they use represents the probability that two random individuals belong to the same race. Their ethnic fragmentation index represents the probability that two random individuals have the same origin / ethnic. The community variables shall analyse in which extend the community is homogenous / heterogeneous.
The descriptive statistics of the paper analyse the states of America and their institutions in terms of high and low trust states and institutions.
High trusting states are in the North / North West like North Dakota, Minnesota or Montana. In terms of racial / ethnic fragmentation and income equality they are very homogeneous. Low trusting states are in the South / South East like Mississippi, Delaware or Alabama. These states stick out by their high rate of heterogeneity in terms of racial / ethnic mixing and income inequality.
The statistics of trust in institutions show that impersonal institutions are more trusted than personal institutions. The institutions that are trusted the most are “medicine” and the “scientific community”, the ones that are trusted the lowest are “organized labour” and “congress”.
4. Econometric evidence
In the next part I will give a summary of the econometric evidence of my reference paper. I will focus on table two from the Alesina / Ferrara paper, because in my opinion it contains the most important information about trust. For this reason table two will be the only table displayed in this paper. This table deals with information about the individual’s characteristics. The findings about the community variables I will summarise more quickly.
illustration not visible in this excerpt
In the first column it can be noted that the age is positively correlated with trust, but it has a declining rate. That means older people trust more, but the effect is decreasing. The fact whether the respondent is married or not is not significant and has no influence on trust. Hence people who trust more are not more likely to get married. The next variables in column one is female and black. Both have been discriminated in the history and both have a significantly negative influence on trusting others. Income and education are both significantly positive correlated with trust. That means if one had a longer education and got a better paid job he is more likely to trust another person.
Column two describes the influence of a past trauma on trust. This includes not only a trauma of the individual but also traumata of close relatives. By adding these variables, the variables of column one stay mostly the same. The data says a past trauma is significant and has a negative influence on trust. In addition to that the authors say that if the trauma is older it has less influence on trust (they are quickly forgotten) and a trauma of financial misfortune decreases trust the most.
In column three religious beliefs are added to the table but they are all insignificant. The authors mention, that because of the missing influence of religious beliefs the American “melting pot” works. From this it follows that social interactions between people are more important than religion.
The next table from the reference paper contains information about the size of the respondent’s place, about the median household income, racial fragmentation and ethnic fragmentation. Here the authors are trying to estimate which of the variables has the strongest relationship with trust. In table three only the median household income and racial fragmentation are significant. It shows that a high variation of income and racial fragmentation have a negative influence on trust, but between this two the negative influence of racial fragmentation is higher than the one of income inequality.
One of the hypotheses mentioned in section one is “a longer interaction among people may support trust”. The idea of the authors is that if a person is living longer in the same place the trust to other people in the community rises. A second assumption is that if the people living in the community of the respondent change frequently, his trust to these people decreases. These assumptions are based on the idea that people who are not interacting in a frequent way, lose the ability to punish the other person for bad behaviour.