Fix: Remove join fan-out bug in agg_monthly_loans

[bmiller-dh]

Problem

The Risk Analytics Reporting Dashboard is showing inflated loan values this month. The agg_monthly_loans model contains a join fan-out bug that's causing aggregated loan amounts to be multiplied incorrectly.

Root Cause

The model joins the aggregated monthly_originations CTE back to the loans table using only loan_type_name:

left join loans
    on orig.loan_type_name = loans.loan_type_name

Since monthly_originations is aggregated by month and loan_type, but the loans table contains multiple individual loan records per loan_type, this creates a Cartesian product. Each aggregated row gets multiplied by the number of loans matching that loan_type.

Example: If you have 100 loans of type "Mortgage" and the aggregated row shows $5M originated, the join multiplies this by 100, showing $500M instead.

Solution

Remove the problematic left join loans clause. The customer_id field that was added in the previous commit is not needed for the aggregation and was causing the data quality issue.

Impact

• Fixes the inflated amount_originated values in the Risk Analytics Reporting Dashboard
• Restores accurate monthly loan origination reporting
• Removes unnecessary join operation, improving query performance

Testing

After merge, please run a dbt test to verify the aggregated values match expected ranges for previous months.