“该策略使用标普500数据，通过VIX和GARCH预测计算净情绪波动指数（NEVI）。当NEVI低于第2分位时做多，高于第8分位时做空，持仓周期为30天。”
资产类别：差价合约（CFDs）、交易所交易基金（ETFs）、基金、期货 | 地区：美国 | 频率：每日 | 市场：股票 | 关键词：净值、情绪、波动性、指数
I. 策略概述
该策略使用标普500的历史数据估算GARCH(1,1)模型系数并预测30天的波动率。将这些预测的平均值定义为平均GARCH预测（AGF），并通过计算VIX与AGF的差值得到净情绪波动指数（NEVI）。将NEVI值划分为十分位，并将其作为动量指标。当NEVI低于第2分位时，投资者建立多头头寸；当其高于第8分位时，建立空头头寸。信号发出后，头寸持有30天，旨在利用波动性驱动的市场动量效应。
II. 策略合理性
VIX可以分解为两部分：理性市场波动性和情绪波动性。从VIX中扣除GARCH预测的情绪波动性，是比VIX更纯粹的市场情绪指标。
III. 论文来源
Volatility Based Sentiment Indicators for Timing the Markets [点击浏览原文]
- 作者：Cacia, Tzvetkov
- 所属机构：瑞典隆德大学经济与管理学院
<摘要>
VIX是由芝加哥期权交易所发布的著名隐含波动率估算指标。本文评估了VIX作为情绪指标的能力，以及其在短期投资策略中发出的信号。研究证明并讨论了基于VIX的策略（通常称为“逆向”策略）如何实现比市场更高的回报。我们还提出了一种从VIX派生的更纯粹的情绪指标，能够提供更准确的市场时机信号。我们称之为“净情绪波动指数”（NEVI）。研究表明，该指标具有显著的统计关系和有趣的属性。
IV. 回测表现
| 年化收益率 | 12.65% |
| 波动率 | 5.92% |
| Beta | -0.25 |
| 夏普比率 | 1.46 |
| 索提诺比率 | N/A |
| 最大回撤 | N/A |
| 胜率 | 57% |
V. 完整python代码
from AlgorithmImports import *
#endregion
# https://quantpedia.com/strategies/net-emotional-volatility-index-effect/
#
# Historical S&P 500 index data are used to estimate the coefficients in the GARCH (1,1) model. Once the GARCH coefficients are obtained, for each day, the investor forecasts the volatility
# for the next 30 days, reiterating the forecast, which means that every forecast becomes the last observation for the next forecast. The average of these values is called the Average GARCH
# Forecast (AGF). The investor then calculates the Net Emotional Volatility Index (NEVI) as the difference between the VIX and AGF. The historical NEVI index is then sorted into deciles, and
# the current NEVI is used as a “Momentum” indicator – the investor opens long on a position when the indicator goes below the 2nd decile and short on positions when it goes above the 8th decile.
# The position is held 30 days from the day of the last long or short signal.
import numpy as np
from arch import arch_model
class NetEmotionalVolatilityIndexEffect(QCAlgorithm):
def Initialize(self):
self.SetStartDate(2008, 1, 1)
self.SetCash(100000)
self.period:int = 21
self.holding_period:int = 30
self.SetWarmUp(self.period + self.holding_period)
self.market:Symbol = self.AddEquity("SPY", Resolution.Daily).Symbol
self.vix:Symbol = self.AddData(QuandlVix, "CBOE/VIX", Resolution.Daily).Symbol # starts in 2004
self.days_held:int = 0
self.market_data:RollingWindow = RollingWindow[float](self.period)
self.volatility:RollingWindow = RollingWindow[float](self.period)
self.nevi_index:list[float] = []
def OnData(self, data):
# market data are present in the algorithm
if self.market in data and data[self.market]:
market_price:float = data[self.market].Value
if market_price != 0:
# store market price data
self.market_data.Add(market_price)
if self.market_data.IsReady:
# calcualte market daily returns
market_prices:np.ndarray = np.array([x for x in self.market_data][::-1])
market_returns:np.ndarray = (market_prices[:-1] / market_prices[1:] - 1) * 100
# forecast market volatility
exp_market_volatility:float = self.garch_vol_predict_volatility(market_returns, 1, 1, 30)
self.volatility.Add(exp_market_volatility)
if self.volatility.IsReady:
agf:float = np.mean([x for x in self.volatility])
# vix data are present in the algorithm
if self.vix in data and data[self.vix]:
vix:float = data[self.vix].Value
self.Plot("Expected Market Volatility", "Expected Market Volatility", exp_market_volatility)
self.Plot("VIX", "VIX", vix)
if vix != 0:
nevi:float = vix - agf
self.nevi_index.append(nevi)
# wait until at least 30 data points is present for NEVI
if len(self.nevi_index) >= self.period and not self.Portfolio[self.market].Invested and not self.IsWarmingUp:
low_decile:float = np.percentile(self.nevi_index[:-1], 20)
high_decile:float = np.percentile(self.nevi_index[:-1], 80)
recent_nevi:float = self.nevi_index[-1]
if recent_nevi > high_decile:
self.SetHoldings(self.market, -1)
self.days_held = 0
elif recent_nevi < low_decile:
self.SetHoldings(self.market, 1)
self.days_held = 0
if self.Portfolio[self.market].Invested:
self.days_held += 1
if self.days_held == self.holding_period:
self.Liquidate(self.market)
self.days_held = 0
def garch_vol_predict_volatility(self, returns, p, q, h) -> float:
am = arch_model(returns, mean='Zero', vol='GARCH', p=p, q=q)
res = am.fit(disp='off')
forecasts = res.forecast(horizon=h)
return np.sqrt(forecasts.variance.values[-1,:][0] * 252)
class QuandlVix(PythonQuandl):
def __init__(self):
self.ValueColumnName = "close"
