外汇价值 v2 – 实际汇率变动

“该策略交易39种货币对，根据5年实际汇率变化做多被低估的货币，做空被高估的货币，形成等权重、每月再平衡的投资组合。”

资产类别: ETF | 地区: 美国 | 周期: 每日 | 市场: 股票 | 关键词: VIX, ETN

I. 策略概要

该策略交易39种货币对美元，使用5年实际汇率变化作为货币价值的指标。实际汇率使用名义汇率以及本地和美国消费品价格水平的比率计算。每个月，货币从最被低估（最低5年回报）到最被高估（最高5年回报）进行排名。投资者做多前20%（最被低估的），做空后20%（最被高估的），形成一个等权重的投资组合，每月进行再平衡以利用估值差异。

II. 策略合理性

实际汇率是货币市场中的一项基本指标，反映了国家之间的相对价值。理论认为，长期来看，实际汇率的差异应该会收敛。实际汇率低于1的货币被认为是“被低估的”，而高于1的货币则被认为是“被高估的”。然而，均衡实际汇率可能偏离1，因此使用（对数）实际汇率的五年累计变化来定义“价值”更为实际，这为评估货币随时间的估值提供了一种稳健的方法。

III. 来源论文

外汇市场中的价值策略是否有利可图？[点击查看论文]

• 艾哈迈德·拉扎，奥塔哥大学 – 会计与金融系。

<摘要>

本文研究了基于四种不同货币估值指标的价值策略的盈利能力。实际汇率水平策略在长达1个月的区间内产生最大的超额回报，而实际汇率变化在1-12个月的区间内产生最大的超额回报。基于购买力平价和巨无霸指数的策略表现不佳。这些回报无法用经济状态变量或货币风险因素解释，并且大于估计的交易成本。基于所有四种估值方法的综合策略产生更高的平均回报，但由于投资组合集中度，波动性也更高。

IV. 回测表现

年化回报	9.57%
波动率	9.51%
β值	-0.114
夏普比率	1.01
索提诺比率	-0.375
最大回撤	N/A
胜率	41%

V. 完整的 Python 代码

from AlgorithmImports import *
import data_tools
from typing import List, Dict
class FXValuev2RealExchangeRateChanges(QCAlgorithm):
    def Initialize(self):
        self.SetStartDate(2000, 1, 1)
        self.SetCash(100000)
        
        self.data:Dict[str, RollingWindow[float]] = {}
        
        self.leverage:int = 3
        self.period:int = 12 * 5            # five years of monthly values
        self.selected_currencies:int = 3    # select these many of forex futures for long and short each rebalance
        
        # currency future symbol and PPP yearly symbol
        self.symbols:Dict[str, str] = {
            "CME_AD1" : "AUS_PPP", # Australian Dollar Futures, Continuous Contract #1
            "CME_BP1" : "GBR_PPP", # British Pound Futures, Continuous Contract #1
            "CME_CD1" : "CAN_PPP", # Canadian Dollar Futures, Continuous Contract #1
            "CME_EC1" : "DEU_PPP", # Euro FX Futures, Continuous Contract #1
            "CME_JY1" : "JPN_PPP", # Japanese Yen Futures, Continuous Contract #1
            "CME_NE1" : "NZL_PPP", # New Zealand Dollar Futures, Continuous Contract #1
            "CME_SF1" : "CHE_PPP"  # Swiss Franc Futures, Continuous Contract #1
        }
        
        self.last_ppp:Dict[str, Union[None, str]] = {
            "AUS_PPP" : None,
            "GBR_PPP" : None,
            "CAN_PPP" : None,
            "DEU_PPP" : None,
            "JPN_PPP" : None,
            "NZL_PPP" : None,
            "CHE_PPP" : None,
            "USA_PPP" : None
        }
        
        for symbol, ppp_symbol in self.symbols.items():
            data = self.AddData(data_tools.QuantpediaFutures, symbol, Resolution.Daily)
            data.SetFeeModel(data_tools.CustomFeeModel())
            data.SetLeverage(self.leverage)
            
            # PPP data
            self.AddData(data_tools.PPPData, ppp_symbol, Resolution.Daily)
            self.data[symbol] = RollingWindow[float](self.period)
            
        self.AddData(data_tools.PPPData, 'USA_PPP', Resolution.Daily).Symbol
        
        self.Settings.MinimumOrderMarginPortfolioPercentage = 0.
        self.selection_flag:bool = False
        self.Schedule.On(self.DateRules.MonthStart('CME_AD1'), self.TimeRules.At(0, 0), self.Selection)
    def OnData(self, data):
        ppp_last_update_date:Dict[str, datetime.date] = data_tools.PPPData.get_last_update_date()
        future_last_update_date:Dict[str, datetime.date] = data_tools.QuantpediaFutures.get_last_update_date()
        # store PPP values, when they are available
        for ppp_symbol, _ in self.last_ppp.items():
            if ppp_symbol in data and data[ppp_symbol]:
                # only last PPP for each country is stored
                self.last_ppp[ppp_symbol] = data[ppp_symbol].Value
        
        if not self.selection_flag:
            return
        self.selection_flag = False
        
        symbols_to_delete:List[str] = []
        # store monthly data
        for symbol, ppp_symbol in self.symbols.items():
            # data is still coming
            if self.Securities[symbol].GetLastData() and self.Time.date() > future_last_update_date[symbol] \
                or self.Securities[ppp_symbol].GetLastData() and self.Time.date() > ppp_last_update_date[ppp_symbol]:
                symbols_to_delete.append(symbol)
                continue
            # check if all data are ready for calculation
            if symbol in data and data[symbol] and self.last_ppp[ppp_symbol] and self.last_ppp['USA_PPP']:
                # forex future price * (forex future PPP / USA PPP)
                self.data[symbol].Add(data[symbol].Value * (self.last_ppp[ppp_symbol] / self.last_ppp['USA_PPP']))
        
        if len(symbols_to_delete) != 0:
            [self.symbols.pop(symbol) for symbol in symbols_to_delete]
        # rebalance            
        years_return:Dict[Symbol, float] = {}
        
        for symbol, _ in self.symbols.items():
            if not self.data[symbol].IsReady:
                continue
            
            # calculate years return
            values = [x for x in self.data[symbol]]
            years_return[symbol] = (values[0] - values[-1]) / values[-1]
        
        if len(years_return) == 0:
            return
        
        # sort forex futures by years return
        sorted_by_years_return:List[Symbol] = [x[0] for x in sorted(years_return.items(), key=lambda item: item[1])]
        
        # long self.selected_currencies with the lowest past years return
        long:List[Symbol] = sorted_by_years_return[:self.selected_currencies]
        
        # short self.selected_currencies with the highest past years return
        short:List[Symbol] = sorted_by_years_return[-self.selected_currencies:]
        
        # trade execution
        invested:List[str] = [x.Key.Value for x in self.Portfolio if x.Value.Invested]
        for symbol in invested:
            if symbol not in long + short:
                self.Liquidate(symbol)
        for i, portfolio in enumerate([long, short]):
            for symbol in portfolio:
                if symbol in data and data[symbol]:
                    self.SetHoldings(symbol, ((-1) ** i) / self.selected_currencies)
    def Selection(self):
        self.selection_flag = True