Stata Analysis for Bank Risk-taking Dissertation
Budget: £20 – £250 GBP
I am writing to ask whether you could help with the Stata analysis for my dissertation.
My dissertation examines whether the PRA’s climate-related supervision under SS3/19 influenced the risk-taking behaviour of UK banks. More specifically, I am testing whether banks with stronger ESG profiles responded differently after the introduction of this supervisory framework.
The analysis should be carried out in Stata as a bank-year panel study using publicly listed UK commercial and investment banks over 2015–2023, subject to data availability.
The dependent variable should be NPLTL as the proxy for bank risk-taking. The main explanatory variables are:
PostPRA (1 for 2019–2023, 0 otherwise)
ESG
L_ESG (one-year lag of ESG)
PostPRA × L_ESG
The key hypothesis is that the coefficient on the interaction term should be negative, meaning that after SS3/19, banks with higher ESG scores are expected to show lower risk-taking relative to their peers.
The control variables are:
MV
NIM
LTDTA
CASHTA
The preferred model should include both bank fixed effects and year fixed effects.
If possible, could you also:
check merges carefully by bank and year
identify duplicates or unmatched observations
ensure one observation per bank-year
set the panel correctly before creating lagged variables
provide a short summary of missing values across the main variables
For the results, I would ideally need:
a baseline model with PostPRA, L_ESG and the interaction
a model adding the controls
the full model with bank and year fixed effects
a reduced-control model if the full version causes a major drop in observations
Please also provide descriptive statistics, a correlation matrix, and a brief interpretation of the results, especially the interaction term, even if it is insignificant.
If feasible, I would also appreciate a few robustness checks, such as winsorisation and testing whether results remain similar when some controls are removed.
The final deliverables I would need are the cleaned dataset, do-file, regression tables, summary statistics, and a short note explaining the final sample, cleaning decisions, and main conclusion. Nothing more than 40 pound please
My dissertation examines whether the PRA’s climate-related supervision under SS3/19 influenced the risk-taking behaviour of UK banks. More specifically, I am testing whether banks with stronger ESG profiles responded differently after the introduction of this supervisory framework.
The analysis should be carried out in Stata as a bank-year panel study using publicly listed UK commercial and investment banks over 2015–2023, subject to data availability.
The dependent variable should be NPLTL as the proxy for bank risk-taking. The main explanatory variables are:
PostPRA (1 for 2019–2023, 0 otherwise)
ESG
L_ESG (one-year lag of ESG)
PostPRA × L_ESG
The key hypothesis is that the coefficient on the interaction term should be negative, meaning that after SS3/19, banks with higher ESG scores are expected to show lower risk-taking relative to their peers.
The control variables are:
MV
NIM
LTDTA
CASHTA
The preferred model should include both bank fixed effects and year fixed effects.
If possible, could you also:
check merges carefully by bank and year
identify duplicates or unmatched observations
ensure one observation per bank-year
set the panel correctly before creating lagged variables
provide a short summary of missing values across the main variables
For the results, I would ideally need:
a baseline model with PostPRA, L_ESG and the interaction
a model adding the controls
the full model with bank and year fixed effects
a reduced-control model if the full version causes a major drop in observations
Please also provide descriptive statistics, a correlation matrix, and a brief interpretation of the results, especially the interaction term, even if it is insignificant.
If feasible, I would also appreciate a few robustness checks, such as winsorisation and testing whether results remain similar when some controls are removed.
The final deliverables I would need are the cleaned dataset, do-file, regression tables, summary statistics, and a short note explaining the final sample, cleaning decisions, and main conclusion. Nothing more than 40 pound please