Why are "pseudo firms" useful as experimental entities?

Because they are extremely flexible entities, and we observe the values of both assets and liabilities by using traded securities. Like with any parametric structural model of the firm, we can also change the characteristics of pseudo firms, such as

1. Leverage
2. Default probability
3. Type and riskiness of assets
4. Losses given default

Moreover, it is possible to run “what if” experiments by changing the characteristics of the pseudo firms under various scenarios of interest, and then look at the outcome directly in the data. For instance, Culp, Nozawa, and Veronesi “Option-Based Credit Spreads” run three “what if” experiments

1. What is the bias in average credit spreads from infrequent credit rating updates?
2. What is the impact of idiosyncratic asset uncertainty on credit spreads?
3. What is the source and the size of tail risk of banks that make commercial loans? How does tail risk depend on discount rate shocks? How does it affect the default correlation across banks?

The applications of the framework are endless, and bounded only by researcher’s imagination