aboutsummaryrefslogtreecommitdiffstats
path: root/python/notebooks/swaption_scenarios.ipynb
diff options
context:
space:
mode:
Diffstat (limited to 'python/notebooks/swaption_scenarios.ipynb')
-rw-r--r--python/notebooks/swaption_scenarios.ipynb225
1 files changed, 85 insertions, 140 deletions
diff --git a/python/notebooks/swaption_scenarios.ipynb b/python/notebooks/swaption_scenarios.ipynb
index e632d922..6399d8e4 100644
--- a/python/notebooks/swaption_scenarios.ipynb
+++ b/python/notebooks/swaption_scenarios.ipynb
@@ -18,6 +18,8 @@
"from scipy.interpolate import SmoothBivariateSpline\n",
"from utils.db import dbconn, dbengine\n",
"from risk.swaptions import get_swaption_portfolio\n",
+ "from scipy.optimize import brentq\n",
+ "from pandas.tseries.offsets import BDay\n",
"\n",
"conn = dbconn('dawndb')\n",
"dawn_engine = dbengine('dawndb')\n",
@@ -32,50 +34,45 @@
"metadata": {},
"outputs": [],
"source": [
+ "############# Current portfolio one day PNL/Delta scenario\n",
"portf = get_swaption_portfolio(datetime.date.today() - pd.offsets.BDay(), conn, source_list=['GS'])\n",
"\n",
- "hedges = pd.read_sql_query(\"SELECT security_id as redcode, maturity, notional, folder FROM list_cds_positions_by_strat(%s) \"\n",
+ "hedges = pd.read_sql_query(\"SELECT security_desc, notional FROM list_cds_positions_by_strat(%s) \"\n",
" \"WHERE folder in ('IGOPTDEL', 'HYOPTDEL')\",\n",
" conn, params=(datetime.date.today(),))\n",
"\n",
"for i, r in hedges.iterrows():\n",
- " strategy = r.pop(\"folder\")\n",
- " trade_index = CreditIndex(**r, value_date=datetime.date.today() - pd.offsets.BDay())\n",
- " trade_index.mark()\n",
- " portf.add_trade(trade_index, (strategy, i))\n",
+ " portf.add_trade(CreditIndex(r['security_desc'].split(\" \")[1],\n",
+ " r['security_desc'].split(\" \")[3][1:],\n",
+ " '5yr', value_date=datetime.date.today() - pd.offsets.BDay(),\n",
+ " notional = r['notional']), ('delta', i))\n",
"\n",
"vol_surface = {}\n",
"for trade in portf.swaptions:\n",
- " k = (trade.index.index_type, trade.index.series, trade.option_type)\n",
- " if k not in vol_surface:\n",
- " vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
- " value_date=datetime.date.today(), interp_method = \"bivariate_linear\")\n",
- " vol_surface[k] = vs[vs.list('GS', option_type=trade.option_type)[-1]]\n",
+ " vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
+ " value_date=datetime.date.today(), interp_method = \"bivariate_linear\")\n",
+ " vol_surface[(trade.index.index_type, trade.index.series, trade.option_type)] = vs[vs.list(source='GS', option_type=trade.option_type)[-1]]\n",
"\n",
"#Set original_pv as of yesterday's EOD levels, don't reset PV after this time\n",
+ "portf.mark(interp_method=\"bivariate_linear\", source_list=['GS'])\n",
"portf.reset_pv()\n",
"\n",
"#set ref to today's levels\n",
"portf.value_date = datetime.date.today()\n",
- "portf.mark(interp_method=\"bivariate_linear\", source_list=['GS'])"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
+ "portf.mark(interp_method=\"bivariate_linear\", source_list=['GS'])\n",
+ "\n",
"spread_shock = np.round(np.arange(-.1, .1, .01), 4)\n",
- "scens = run_portfolio_scenarios(portf, [datetime.datetime.now()], params=['pnl', 'hy_equiv', 'sigma'],\n",
+ "scens = run_portfolio_scenarios(portf, [datetime.datetime.now()], params=['pnl', 'hy_equiv'],\n",
" spread_shock=spread_shock,\n",
" vol_shock=[0],\n",
+ " corr_shock=[0],\n",
" vol_surface=vol_surface)\n",
"pnl = scens.xs('pnl', level = 2, axis=1).sum(axis=1)\n",
"hy_equiv = scens.xs('hy_equiv', level = 2, axis=1).sum(axis=1)\n",
"\n",
"ig = CreditIndex('IG', 32, '5yr')\n",
"ig.mark()\n",
+ "\n",
"pnl.index = pnl.index.set_levels((1+pnl.index.get_level_values('spread_shock')) * ig.spread, level = 'spread_shock')\n",
"hy_equiv.index = pnl.index"
]
@@ -86,16 +83,8 @@
"metadata": {},
"outputs": [],
"source": [
- "pnl.reset_index([\"date\", \"vol_shock\"], drop=True).to_frame(\"pnl\").plot()"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "hy_equiv.reset_index([\"date\", \"vol_shock\"], drop=True).to_frame(\"hy_equiv\").plot()"
+ "pnl, hy_equiv\n",
+ "#plot_trade_scenarios(portf)"
]
},
{
@@ -104,7 +93,40 @@
"metadata": {},
"outputs": [],
"source": [
- "pnl, hy_equiv"
+ "#breakeven calc\n",
+ "index = 'HY'\n",
+ "series = 32\n",
+ "value_date = datetime.date.today()\n",
+ "option_delta = CreditIndex(index, series, '5yr')\n",
+ "#option_delta.spread = 56.5\n",
+ "option_delta.price = 106.0\n",
+ "option1 = BlackSwaption(option_delta, datetime.date(2019, 12, 16), 106.5, option_type=\"receiver\") \n",
+ "option1.sigma = .35\n",
+ "option1.notional = 100_000_000 \n",
+ "option_delta.notional = -option1.delta * option1.notional\n",
+ "portf = Portfolio([option1, option_delta], trade_ids=['opt1', 'delta'])\n",
+ "portf.value_date = value_date\n",
+ "portf.reset_pv()\n",
+ "portf.value_date = value_date + BDay(1)\n",
+ "orig_ref = portf.ref\n",
+ "\n",
+ "def get_pnl(portf, x):\n",
+ " portf.ref = x\n",
+ " portf.sigma = float(vs[surface_id](self.T, np.log(self.moneyness)))\n",
+ " return portf.pnl\n",
+ "\n",
+ "if index == 'IG':\n",
+ " widening = brentq(lambda x: get_pnl(portf, x), portf.ref, portf.ref + 10)\n",
+ " portf.ref = orig_ref\n",
+ " tightening = brentq(lambda x: get_pnl(portf, x), portf.ref-10, portf.ref)\n",
+ " portf.ref = orig_ref\n",
+ "else: \n",
+ " widening = brentq(lambda x: get_pnl(portf, x), portf.ref-3, portf.ref)\n",
+ " portf.ref = orig_ref\n",
+ " tightening = brentq(lambda x: get_pnl(portf, x), portf.ref, portf.ref + 3)\n",
+ " portf.ref = orig_ref\n",
+ "\n",
+ "tightening, orig_ref, widening"
]
},
{
@@ -117,16 +139,16 @@
"index = 'IG'\n",
"series = 32\n",
"option_delta = CreditIndex(index, series, '5yr') \n",
- "option_delta.spread = 60\n",
- "option1 = BlackSwaption(option_delta, datetime.date(2019, 9, 17), 90, option_type=\"payer\") \n",
- "option2 = BlackSwaption(option_delta, datetime.date(2019, 11, 19), 90, option_type=\"payer\") \n",
- "option1.sigma = .6\n",
- "option2.sigma = .58\n",
+ "option_delta.spread = 55\n",
+ "option1 = BlackSwaption(option_delta, datetime.date(2019, 10, 16), 50, option_type=\"receiver\") \n",
+ "option2 = BlackSwaption(option_delta, datetime.date(2019, 10, 16), 77.5, option_type=\"payer\") \n",
+ "option1.sigma = .4\n",
+ "option2.sigma = .63\n",
"option1.notional = 100_000_000 \n",
"option2.notional = 100_000_000 \n",
"option1.direction = 'Long' \n",
"option2.direction = 'Short' \n",
- "option_delta.notional = option1.delta * option1.notional + option2.delta * option2.notional\n",
+ "option_delta.notional = -option1.delta * option1.notional - option2.delta * option2.notional\n",
"portf = Portfolio([option1, option2, option_delta], trade_ids=['opt1', 'opt2', 'delta'])"
]
},
@@ -135,6 +157,13 @@
"execution_count": null,
"metadata": {},
"outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
"source": [
"date_range = pd.bdate_range(portf.value_date, portf.value_date + pd.tseries.offsets.BDay(30), freq='3B')\n",
"vol_shock = np.arange(-.15, .31, 0.01)\n",
@@ -143,7 +172,7 @@
"for trade in portf.swaptions:\n",
" vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
" value_date=portf.value_date, interp_method = \"bivariate_linear\")\n",
- " vol_surface[(trade.index.index_type, trade.index.series)] = vs[vs.list(option_type='payer')[-1]]\n",
+ " vol_surface[(trade.index.index_type, trade.index.series, trade.option_type)] = vs[vs.list(source='GS', option_type=trade.option_type)[-1]]\n",
"\n",
"df = run_portfolio_scenarios(portf, date_range, params=[\"pnl\"],\n",
" spread_shock = spread_shock,\n",
@@ -171,7 +200,7 @@
" for trade in portf.swaptions:\n",
" vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
" value_date=portf.value_date, interp_method = \"bivariate_linear\")\n",
- " vol_surface[(trade.index.index_type, trade.index.series)] = vs[vs.list(option_type='payer')[-1]]\n",
+ " vol_surface[(trade.index.index_type, trade.index.series, trade.option_type)] = vs[vs.list(option_type=trade.option_type)[-1]]\n",
" \n",
" df = run_portfolio_scenarios(portf, date_range, params=[\"pnl\",\"delta\"],\n",
" spread_shock = spread_shock,\n",
@@ -188,11 +217,16 @@
" else:\n",
" df['spread'] = portf.indices[0].spread * (1 + df.spread_shock)\n",
" df = df.set_index(['date', 'spread', 'vol_shock'])\n",
- " sort_order = [True, True]\n",
- " \n",
+ " sort_order = [True, False]\n",
+ " \n",
+ " #If the multilevels index contains strategy drop it\n",
+ " if df.columns.nlevels == 3: \n",
+ " df.columns = df.columns.droplevel(level=0)\n",
" pnl = df.xs('pnl', axis=1, level=1).sum(axis=1)\n",
" for trade_id, t in portf.items():\n",
" if isinstance(t, BlackSwaption):\n",
+ " if len(trade_id) == 2:\n",
+ " trade_id = trade_id[1]\n",
" df[(trade_id, 'delta')] *= -t.notional \n",
" delta = df.xs('delta', axis=1, level=1).sum(axis=1).xs(0, level='vol_shock')\n",
" delta += sum([x.notional * -1 if x.direction == 'Buyer' else 1 for x in portf.indices])\n",
@@ -223,7 +257,7 @@
" for trade in portf.swaptions:\n",
" vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
" value_date=portf.value_date, interp_method = \"bivariate_linear\")\n",
- " vol_surface[(trade.index.index_type, trade.index.series)] = vs[vs.list(option_type='payer')[-1]]\n",
+ " vol_surface[(trade.index.index_type, trade.index.series, trade.option_type)] = vs[vs.list(option_type=trade.option_type)[-1]]\n",
"\n",
" df = run_portfolio_scenarios(portf, date_range, params=[\"pnl\"],\n",
" spread_shock = spread_shock,\n",
@@ -239,16 +273,6 @@
"metadata": {},
"outputs": [],
"source": [
- "plot_trade_scenarios(portf)\n",
- "plot_trade_scenarios(portf, -.15, .5, vol_time_roll=False)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
"#Dec Jan 2017 Trade\n",
"option_delta = CreditIndex.from_tradeid(864)\n",
"option1 = BlackSwaption.from_tradeid(3, option_delta)\n",
@@ -336,100 +360,21 @@
"execution_count": null,
"metadata": {},
"outputs": [],
- "source": [
- "#Set up Portfolio\n",
- "from risk.swaptions import get_swaption_portfolio\n",
- "from risk.tranches import get_tranche_portfolio\n",
- "rundate = datetime.date(2019,6,21)\n",
- "\n",
- "portf = get_swaption_portfolio(rundate, conn)\n",
- "\n",
- "#index positions\n",
- "df = pd.read_sql_query(\"SELECT * from list_cds_positions_by_strat(%s)\",\n",
- " dawn_engine, params=(rundate,))\n",
- "df = df[df.folder.str.contains(\"OPT\")]\n",
- "for t in df.itertuples(index=False):\n",
- " portf.add_trade(CreditIndex(redcode=t.security_id, maturity=t.maturity, notional=t.notional),\n",
- " (t.folder, t.security_desc))\n",
- "\n",
- "portf.value_date = rundate\n",
- "portf.mark(interp_method=\"bivariate_linear\")\n",
- "portf.reset_pv()\n",
- "\n",
- "#------------------------Calc Scenarios\n",
- "vol_surface = {}\n",
- "for trade in portf.swaptions:\n",
- " vs = BlackSwaptionVolSurface(trade.index.index_type, trade.index.series, \n",
- " value_date=rundate, interp_method = \"bivariate_linear\")\n",
- " vol_surface[(trade.index.index_type, trade.index.series)] = vs[vs.list(option_type='payer')[-1]]\n",
- "vol_shock = [0]\n",
- "corr_shock = [0]\n",
- "spread_shock = np.round(np.arange(-.2, 1, .05), 3)\n",
- "scens = run_portfolio_scenarios(portf, [pd.Timestamp(rundate)], params=['pnl', 'delta'],\n",
- " spread_shock=spread_shock,\n",
- " vol_shock=vol_shock,\n",
- " corr_shock=[0],\n",
- " vol_surface=vol_surface)\n",
- "\n",
- "pnl = scens.xs('pnl', axis=1, level=2)\n",
- "pnl = pnl.xs(0, level='vol_shock')\n",
- "\n",
- "scenarios = (pnl.\n",
- " reset_index(level=['date'], drop=True).\n",
- " groupby(level=0, axis=1).sum())\n",
- "\n",
- "options = ['HYOPTDEL', 'HYPAYER', 'HYREC', 'IGOPTDEL', 'IGPAYER', 'IGREC']\n",
- "scenarios['options'] = scenarios[set(scenarios.columns).intersection(options)].sum(axis=1)"
- ]
+ "source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
- "source": [
- "#Now say that it widens by X percentage, then rebalance, then do the shock again\n",
- "rundate = datetime.date(2019,6,21)\n",
- "x = 1.2\n",
- "for t in portf.swaptions:\n",
- " t.index.spread *= x\n",
- " vs = vol_surface[(t.index.index_type, t.index.series)]\n",
- " t.sigma = max(0.2, float(vs(t.T, math.log(t.moneyness))))\n",
- "for t in portf.indices:\n",
- " t.spread *= x\n",
- "pnl = portf.pnl\n",
- "\n",
- "analytics.init_ontr(value_date=rundate)\n",
- "rebal = analytics._ontr()\n",
- "rebal.notional = portf.hy_equiv\n",
- "rebal.direction = 'Seller'\n",
- "\n",
- "rebal.spread *= x\n",
- "portf.add_trade(rebal, ('rebalance', 'HYOPTDEL'))\n",
- "portf.reset_pv()\n",
- "\n",
- "swaptions_scens = portf.swaptions[0].shock(params=['pnl', 'pv'],\n",
- " spread_shock=spread_shock,\n",
- " vol_shock=vol_shock,\n",
- " vol_surface=vol_surface)\n",
- "\n",
- "#------------------------Calc Scenarios\n",
- "scens = run_portfolio_scenarios(portf, [pd.Timestamp(rundate)], params=['pnl', 'pv'],\n",
- " spread_shock=spread_shock,\n",
- " vol_shock=vol_shock,\n",
- " corr_shock=[0],\n",
- " vol_surface=vol_surface)\n",
- "\n",
- "pnl = scens.xs('pnl', axis=1, level=2)\n",
- "pnl = pnl.xs(0, level='vol_shock')\n",
- "\n",
- "scenarios = (pnl.\n",
- " reset_index(level=['date'], drop=True).\n",
- " groupby(level=0, axis=1).sum())\n",
- "\n",
- "options = ['HYOPTDEL', 'HYPAYER', 'HYREC', 'IGOPTDEL', 'IGPAYER', 'IGREC', 'rebalance']\n",
- "scenarios['options'] = scenarios[set(scenarios.columns).intersection(options)].sum(axis=1)"
- ]
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": []
}
],
"metadata": {
generated by cgit v1.2.3-70-g09d2 (git 2.52.0) at 2025-12-21 22:27:54 +0000