FEAT: Add multi-year expected loss selection example.

faslr/index/index_grid.py

@@ -0,0 +1,116 @@
+import numpy as np
+import pandas as pd
+
+trend = 0.03425
+
+years = [
+    2002,
+    2003,
+    2004,
+    2005,
+    2006,
+    2007,
+    2008
+]
+
+rate_changes = [
+    0,
+    .05,
+    .075,
+    .15,
+    .10,
+    -.2,
+    -.2
+]
+
+tort_changes = [
+    0,
+    0,
+    0,
+    0,
+    - (1 - .67 / .75),
+    -.25,
+    0
+]
+
+rate_factors = [x + 1 for x in rate_changes]
+tort_factors = [x + 1.0 for x in tort_changes]
+
+d = {}
+
+for year in years:
+    d[str(year)] = [1.03425 ** (year - x) for x in years]
+
+df_idx = pd.DataFrame(data=d, index=years)
+
+
+def rate_level(
+        base_yr: int,
+        years: list
+) -> dict:
+    idx = years.index(base_yr)
+    res = {}
+    for x in range(len(years)):
+        if years[x] < base_yr:
+            adj = np.array(rate_factors[x+1:idx + 1]).prod() ** - 1
+        else:
+            adj = np.array(rate_factors[idx + 1:x + 1]).prod()
+        res[str(years[x])] = adj
+    return res
+
+d = []
+for year in years:
+    d += [rate_level(base_yr=year, years=years)]
+
+df_rl = pd.DataFrame(data=d, index=years)
+
+
+
+def tort_level(
+        base_yr: int,
+        years: list
+) -> dict:
+    idx = years.index(base_yr)
+    res = {}
+    for x in range(len(years)):
+        if years[x] < base_yr:
+            adj = np.array(tort_factors[x+1:idx + 1]).prod() ** - 1
+        else:
+            adj = np.array(tort_factors[idx + 1:x + 1]).prod()
+        res[str(years[x])] = adj
+    return res
+
+d = []
+for year in years:
+    d += [tort_level(base_yr=year, years=years)]
+
+df_tort = pd.DataFrame(data=d, index=years)
+
+df_loss = df_idx * df_tort
+df_prem = df_rl
+
+df_loss['claims'] = [
+    48953,
+    47404,
+    77662,
+    78497,
+    65239,
+    62960,
+    61262
+]
+
+df_prem['premium'] = [
+    61183,
+    69175,
+    99322,
+    138151,
+    107578,
+    62438,
+    47797
+]
+
+d = {}
+for year in years:
+    d[str(year)] = (df_loss[str(year)] * df_loss['claims']) / (df_prem[str(year)] * df_prem['premium'])
+
+df_res = pd.DataFrame(data=d, index=years)