Financial Literacy & Stock Market Participation in Different Generations 🔗
-
In the introduction, I still miss a paragraph about the costs (and other consequences) of reduced stock market participation, and a sentence about why fin. literacy should influence stock market participation
-
The first paragraph is still too much like a report rather than a study (“This research question has been chosen…")
- Instead, you should provide a motivation from the literature and supported by anecdotes about why the effect of fin. literacy on SMP is different for people from different generation
- In places throughout the thesis, this writing style still persists
-
It is not really clear what the lightning flashes in the figures mean, and the arrows cannot point two ways in a DAG (Acyclical)
-
In the literature, I think the selection process can be omitted.
- I would also like to see the hypotheses development in a separate section, while briefly recapitulating the arguments from the other sections.
- Probably, a logical order is also: financial literacy, stock market participation, heterogeneous effects according to generation, (hypothesis development)
-
Methodology:
-
You should repeat the hypotheses and give more context surrounding the hypotheses on pp. 22-23
-
You can also use linear probability models instead of Probit/Logit models (easier with interaction effects)
-
You can use interaction effects between fin. lit. proxies and the age groups to find whether the relationship changes
-
Empirical analysis still quite messy, many analyses not related to what we’ve discuss in last meeting, or to hypotheses
Some Questions 🔗
-
Should I add control variables in the mediation analysis?
-
Answer: yes!
-
Do I add the controls one by one rather than including all control directly to see changes in the main coefficients?
-
Answer: You can include them directly, but also run the analysis with a few control variables excluded that have the most missing values to investigate sensitivity.
-
Do you think it is necessary to compare the survey (from 2018) to an older version of the survey (2009) to check if things changed over time?
-
Answer: I guess not, but it depends on how precisely you should formulate your theory
- I think your theory is about generation Z, rather than an age difference per se, so in that case, not.
-
On the other hand, you can do a placebo test using the 2008 or other surveys. Since there is no generation Z, your theory should predict there should be no difference between generations in the fin. literacy -> SMP relationship.