Using CAPE Regimes with Momentum and RSI for Equity Entries
Article Strategy library · Author: QuantConnect
Summary
This example combines the cyclically adjusted price-to-earnings ratio (CAPE) with MACD and RSI to trade equities. In a high-CAPE regime, it considers buying when RSI is below 70 and selling when RSI rises above 70; in a new-low CAPE regime, it considers buying below RSI 30 and selling above 30. Actual orders are further gated by MACD direction: positive for purchases and negative for liquidations. The example uses SPY by default and includes commented alternatives for historical stock groups.
CAPE is treated as monthly custom data, with a new low detected against a rolling set of four observations. The algorithm runs from 1998 to mid-2014, while the referenced CAPE dataset is described as extending from 1990 through 2014. No performance statistics are provided. The code’s regime logic and order conditions may not fully match its stated goals, and its narrow asset selection, data availability, timing assumptions, and fixed cash checks limit what can be inferred from the example.
Key ideas
- CAPE levels and recent lows are used to distinguish overvalued and undervalued market conditions.
- RSI thresholds suggest entry or exit conditions, while MACD sign gates the actual order actions.
- The sample trades SPY by default and checks for CAPE lows across four observations.
- The document provides an implementation example but no reported returns or risk statistics.
Tags
Full text
# BubbleAlgorithm
# BubbleAlgorithm
Strategy example algorithm using CAPE - a bubble indicator dataset saved in dropbox. CAPE is based on a macroeconomic indicator(CAPE Ratio), we are looking for entry/exit points for momentum stocks CAPE data: January 1990 - December 2014 Goals: Capitalize in overvalued markets by generating returns with momentum and selling before the crash Capitalize in undervalued markets by purchasing stocks at bottom of trough
## Source (Apache-2.0)
```python
# QUANTCONNECT.COM - Democratizing Finance, Empowering Individuals.
# Lean Algorithmic Trading Engine v2.0. Copyright 2014 QuantConnect Corporation.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from AlgorithmImports import *
### <summary>
### Strategy example algorithm using CAPE - a bubble indicator dataset saved in dropbox. CAPE is based on a macroeconomic indicator(CAPE Ratio),
### we are looking for entry/exit points for momentum stocks CAPE data: January 1990 - December 2014
### Goals:
### Capitalize in overvalued markets by generating returns with momentum and selling before the crash
### Capitalize in undervalued markets by purchasing stocks at bottom of trough
### </summary>
### <meta name="tag" content="strategy example" />
### <meta name="tag" content="custom data" />
class BubbleAlgorithm(QCAlgorithm):
def initialize(self):
self.set_cash(100000)
self.set_start_date(1998,1,1)
self.set_end_date(2014,6,1)
self._symbols = []
self._macd_dic, self._rsi_dic = {},{}
self._new_low, self._curr_cape = None, None
self._counter, self._counter2 = 0, 0
self._c, self._c_copy = np.empty([4]), np.empty([4])
self._symbols.append("SPY")
# add CAPE data
self.add_data(Cape, "CAPE")
# # Present Social Media Stocks:
# self._symbols.append("FB"), self._symbols.append("LNKD"),self._symbols.append("GRPN"), self._symbols.append("TWTR")
# self.set_start_date(2011, 1, 1)
# self.set_end_date(2014, 12, 1)
# # 2008 Financials
# self._symbols.append("C"), self._symbols.append("AIG"), self._symbols.append("BAC"), self._symbols.append("HBOS")
# self.set_start_date(2003, 1, 1)
# self.set_end_date(2011, 1, 1)
# # 2000 Dot.com
# self._symbols.append("IPET"), self._symbols.append("WBVN"), self._symbols.append("GCTY")
# self.set_start_date(1998, 1, 1)
# self.set_end_date(2000, 1, 1)
for stock in self._symbols:
self.add_security(SecurityType.EQUITY, stock, Resolution.MINUTE)
self._macd = self.macd(stock, 12, 26, 9, MovingAverageType.EXPONENTIAL, Resolution.DAILY)
self._macd_dic[stock] = self._macd
self._rsi = self.rsi(stock, 14, MovingAverageType.EXPONENTIAL, Resolution.DAILY)
self._rsi_dic[stock] = self._rsi
# Trying to find if current Cape is the lowest Cape in three months to indicate selling period
def on_data(self, data):
if self._curr_cape and self._new_low is not None:
try:
# Bubble territory
if self._curr_cape > 20 and self._new_low == False:
for stock in self._symbols:
# Order stock based on MACD
# During market hours, stock is trading, and sufficient cash
if self.securities[stock].holdings.quantity == 0 and self._rsi_dic[stock].current.value < 70 \
and self.securities[stock].price != 0 \
and self.portfolio.cash > self.securities[stock].price * 100 \
and self.time.hour == 9 and self.time.minute == 31:
self.buy_stock(stock)
# Utilize RSI for overbought territories and liquidate that stock
if self._rsi_dic[stock].current.value > 70 and self.securities[stock].holdings.quantity > 0 \
and self.time.hour == 9 and self.time.minute == 31:
self.sell_stock(stock)
# Undervalued territory
elif self._new_low:
for stock in self._symbols:
# Sell stock based on MACD
if self.securities[stock].holdings.quantity > 0 and self._rsi_dic[stock].current.value > 30 \
and self.time.hour == 9 and self.time.minute == 31:
self.sell_stock(stock)
# Utilize RSI and MACD to understand oversold territories
elif self.securities[stock].holdings.quantity == 0 and self._rsi_dic[stock].current.value < 30 \
and self.securities[stock].price != 0 and self.portfolio.cash > self.securities[stock].price * 100 \
and self.time.hour == 9 and self.time.minute == 31:
self.buy_stock(stock)
# Cape Ratio is missing from original data
# Most recent cape data is most likely to be missing
elif self._curr_cape == 0:
self.debug("Exiting due to no CAPE!")
self.quit("CAPE ratio not supplied in data, exiting.")
except:
# Do nothing
return None
if not data.contains_key("CAPE"): return
self._new_low = False
# Adds first four Cape Ratios to array c
self._curr_cape = data["CAPE"].cape
if self._counter < 4:
self._c[self._counter] = self._curr_cape
self._counter +=1
# Replaces oldest Cape with current Cape
# Checks to see if current Cape is lowest in the previous quarter
# Indicating a sell off
else:
self._c_copy = self._c
self._c_copy = np.sort(self._c_copy)
if self._c_copy[0] > self._curr_cape:
self._new_low = True
self._c[self._counter2] = self._curr_cape
self._counter2 += 1
if self._counter2 == 4: self._counter2 = 0
self.debug("Current Cape: " + str(self._curr_cape) + " on " + str(self.time))
if self._new_low:
self.debug("New Low has been hit on " + str(self.time))
# Buy this symbol
def buy_stock(self,symbol):
s = self.securities[symbol].holdings
if self._macd_dic[symbol].current.value>0:
self.set_holdings(symbol, 1)
self.debug("Purchasing: " + str(symbol) + " MACD: " + str(self._macd_dic[symbol]) + " RSI: " + str(self._rsi_dic[symbol])
+ " Price: " + str(round(self.securities[symbol].price, 2)) + " Quantity: " + str(s.quantity))
# Sell this symbol
def sell_stock(self,symbol):
s = self.securities[symbol].holdings
if s.quantity > 0 and self._macd_dic[symbol].current.value < 0:
self.liquidate(symbol)
self.debug("Selling: " + str(symbol) + " at sell MACD: " + str(self._macd_dic[symbol]) + " RSI: " + str(self._rsi_dic[symbol])
+ " Price: " + str(round(self.securities[symbol].price, 2)) + " Profit from sale: " + str(s.last_trade_profit))
# CAPE Ratio for SP500 PE Ratio for avg inflation adjusted earnings for previous ten years Custom Data from DropBox
# Original Data from: http://www.econ.yale.edu/~shiller/data.htm
class Cape(PythonData):
# Return the URL string source of the file. This will be converted to a stream
# <param name="config">Configuration object</param>
# <param name="date">Date of this source file</param>
# <param name="is_live_mode">true if we're in live mode, false for backtesting mode</param>
# <returns>String URL of source file.</returns>
def get_source(self, config, date, is_live_mode):
# Remember to add the "?dl=1" for dropbox links
return SubscriptionDataSource("https://www.dropbox.com/s/ggt6blmib54q36e/CAPE.csv?dl=1", SubscriptionTransportMedium.REMOTE_FILE)
''' Reader Method : using set of arguments we specify read out type. Enumerate until
the end of the data stream or file. E.g. Read CSV file line by line and convert into data types. '''
# <returns>BaseData type set by Subscription Method.</returns>
# <param name="config">Config.</param>
# <param name="line">Line.</param>
# <param name="date">Date.</param>
# <param name="is_live_mode">true if we're in live mode, false for backtesting mode</param>
def reader(self, config, line, date, is_live_mode):
if not (line.strip() and line[0].isdigit()): return None
# New Nifty object
index = Cape()
index.symbol = config.symbol
try:
# Example File Format:
# Date | Price | Div | Earning | CPI | FractionalDate | Interest Rate | RealPrice | RealDiv | RealEarnings | CAPE
# 2014.06 1947.09 37.38 103.12 238.343 2014.37 2.6 1923.95 36.94 101.89 25.55
data = line.split(',')
# Dates must be in the format YYYY-MM-DD. If your data source does not have this format, you must use
# DateTime.parse_exact() and explicit declare the format your data source has.
index.time = datetime.strptime(data[0], "%Y-%m")
index["Cape"] = float(data[10])
index.value = data[10]
except ValueError:
# Do nothing
return None
return index
```Shown in full with attribution under the source's licence. Licence: Apache-2.0
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.