Solving the steady-state

[nikhilwoodruff]

Filing this to keep track of my attempts to move the best working parametrisation of OG-UK from the default OG-Core model, to the full UK model with fitted tax functions.

Version: 0.0.1

• [x] Default OG-Core model (solves)
• [x] As above, DEP, not age-specific, start year = 2019 (solves)
• [x] As above, plus oguk_default_parameters.json (solves)
• [x] As above, with OpenFisca-UK linear tax function calibration
• [x] As above, with GS tax function calibration
• [ ] As above, with DEP tax function calibration (fails to solve)

@jdebacker @rickecon @jpycroft @MaxGhenis any ideas on "middle-ground" points between 3 and 4? Or I could first reforming the microsim model to make the tax functions effectively flat or otherwise simplified?

[jdebacker]

@nikhilwoodruff for a middle ground, I'd recommend fitting the linear tax rates first (before the GS or DEP functions that allow for progressive rates).

[nikhilwoodruff]

Thanks @jdebacker - updated with the success! I'll see how the TPI looks.

[nikhilwoodruff]

OK, some results and a few questions:

1. Here's what I get when just running SS functions, for the default reform which raises the basic rate of income tax from 20% to 30%. How does this compare with what we'd expect?

Variable	Baseline	Reform	% Change (or pp diff)
GDP ($\bar{Y}$)	0.689195	0.678031	-1.61978
Consumption ($\bar{C}$)	0.600977	0.606212	0.871164
Capital Stock ($\bar{K}$)	2.3755	2.37737	0.0785319
Labor ($\bar{L}$)	0.353969	0.345041	-2.52229
Real interest rate ($\bar{r}$)	0.0488827	0.0472211	-3.39929
Wage rate ($\bar{w}$)	1.26558	1.2773	0.925861

2. The TPI takes a while to run: if there are issues, are there ways we can diagnose/test them just from SS runs?

[nikhilwoodruff]

Now runs without issue after #50.