def headline_metrics(baseline: Microsimulation,
reformed: Microsimulation) -> dict:
"""Compute headline society-wide metrics.
:param baseline: Baseline simulation.
:type baseline: Microsimulation
:param reformed: Reform simulation.
:type reformed: Microsimulation
:return: Dictionary with net_cost, poverty_change, winner_share,
loser_share, and gini_change.
:rtype: dict
"""
new_income = reformed.calc("equiv_household_net_income", map_to="person")
old_income = baseline.calc("equiv_household_net_income", map_to="person")
gain = new_income - old_income
net_cost = (reformed.calc("net_income").sum() -
baseline.calc("net_income").sum())
poverty_change = pct_change(
baseline.calc("in_poverty_bhc", map_to="person").mean(),
reformed.calc("in_poverty_bhc", map_to="person").mean(),
)
winner_share = (gain > 1).mean()
loser_share = (gain < -1).mean()
gini_change = pct_change(old_income.gini(), new_income.gini())
return dict(
net_cost=gbp(net_cost),
net_cost_numeric=(net_cost),
poverty_change=float(poverty_change),
winner_share=float(winner_share),
loser_share=float(loser_share),
gini_change=float(gini_change),
)
def compute_reform():
params = request.json
reform = Microsimulation(create_reform(params))
new_income = reform.calc("equiv_household_net_income", map_to="person")
old_income = baseline.calc("equiv_household_net_income", map_to="person")
gain = new_income - old_income
net_cost = reform.calc("net_income").sum() - baseline.calc("net_income").sum()
decile_plot = px.bar(gain.groupby(old_income.percentile_rank()).mean()).update_layout(
title="Income effect by percentile",
xaxis_title="Equivalised disposable income percentile",
yaxis_title="Average income effect",
yaxis_tickprefix="£",
width=800,
height=600,
template="plotly_white",
showlegend=False
).to_json()
top_1_pct_share_effect = gain[old_income.percentile_rank() == 100].mean()
top_10_pct_share_effect = gain[old_income.decile_rank() == 10].mean()
median_effect = new_income.median() - old_income.median()
return {"net_cost": gbp(net_cost), "decile_plot": json.loads(decile_plot), "1pct": top_1_pct_share_effect, "10pct": top_10_pct_share_effect, "median": median_effect}
def hover_label(component: str, amount: float, is_pop: bool) -> str:
"""Create a label for an individual point in a waterfall hovercard.
:param component: Name of the component, e.g. "Tax revenues".
:type component: str
:param amount: Name of the component, e.g. "Tax revenues".
:type amount: float
:param is_pop: Whether the component is population- vs. household-level.
:type is_pop: bool
:return: Label for hovercard.
:rtype: str
"""
# Reset household net income label to match the headline.
if component == "Your net income":
component = "Your annual net income"
# Net impact bars should match the title.
res = component
# Flip the amount for labeling population benefits and household taxes.
if component in ["Benefit outlays", "Your taxes"]:
amount *= -1
# Round population estimates, not individual.
abs_amount = round(abs(amount))
abs_amount_display = gbp(abs_amount) if is_pop else f"£{abs_amount:,}"
# Branch logic, starting with no change.
# More special handling of the net impact to match the title.
if amount == 0:
if component == "Net impact":
return "Reform has no budgetary impact"
return res + " would not change"
if amount > 0:
if component == "Net impact": # Population.
return "Reform produces " + abs_amount_display + " net surplus"
return res + " would rise by " + abs_amount_display
if amount < 0:
if component == "Net impact": # Population.
return "Reform produces " + abs_amount_display + " net cost"
return res + " would fall by " + abs_amount_display
def test_rounding():
from rdbl import gbp
assert gbp(231e9) == "£231bn"