Original paper

SSRN-id2492082.pdf800.0KB

Abstract

We assess the properties of currency value strategies based on real exchange rates in a cross-sectional portfolio setting. We find that real exchange rates predict currency excess returns, but in a way that is inconsistent with the notion of currency value because a high valuation level forecasts a relative appreciation of the foreign currency going forward. However, adjusting real exchange rates for two key country-specific fundamentals -- productivity and the quality of export goods -- generates a measure of currency value which still predicts currency excess returns and is consistent with the spot exchange rate reverting toward fundamental value.

Keywords: Currency value, macro fundamentals, real exchange rate, predictability

Trading rules

Investment scope: 10-20 foreign currencies.
Use to the most recent OECD PPP statistics to calculate the month's fair value of each currency relative to the USD.
Adjust the prior month's fair PPP value by incorporating monthly CPI and currency rate fluctuations.
Take long positions in the trio of currencies deemed most underpriced (smallest PPP fair value).
Adopt short positions in the three currencies deemed most overpriced (largest PPP fair value).
Allocate any uninvested funds as collateral in overnight interest rates.
Adjust portfolio positions either on a monthly or quarterly basis.

Python code

Backtrader

import backtrader as bt

class PPPCurrencyStrategy(bt.Strategy):
    params = (
        ('num_currencies', 10),
        ('long_count', 3),
        ('short_count', 3),
        ('rebalance_freq', 30),  # Rebalance every 30 days (monthly) by default
    )

    def __init__(self):
        self.oecd_ppp_data = {}  # Load your OECD PPP data here
        self.cpi_data = {}  # Load your monthly CPI data here
        self.currency_pairs = []  # Load your currency datafeeds here
        self.rebalance_days = 0

    def next(self):
        self.rebalance_days += 1
        if self.rebalance_days % self.params.rebalance_freq == 0:
            self.rebalance()

    def rebalance(self):
        ppp_values = {}
        for pair, data in self.currency_pairs:
            ppp_values[pair] = self.calculate_ppp_value(pair, data)

        sorted_pairs = sorted(ppp_values.items(), key=lambda x: x[1])

        long_pairs = sorted_pairs[:self.params.long_count]
        short_pairs = sorted_pairs[-self.params.short_count:]

        for pair, data in long_pairs:
            self.order_target_percent(data, target=1 / self.params.long_count)

        for pair, data in short_pairs:
            self.order_target_percent(data, target=-1 / self.params.short_count)

        for pair, data in self.currency_pairs:
            if pair not in long_pairs and pair not in short_pairs:
                self.order_target_percent(data, target=0)

    def calculate_ppp_value(self, pair, data):
        # Calculate PPP fair value based on OECD PPP, CPI, and exchange rate data
        pass

# Create a Backtrader Cerebro instance
cerebro = bt.Cerebro()

# Add the datafeeds and the strategy to the Cerebro instance
# cerebro.adddata(datafeed)
cerebro.addstrategy(PPPCurrencyStrategy)

# Run the backtest
results = cerebro.run()

# Plot the results (optional)
cerebro.plot()

Please note that this code is a basic structure for your requirements. You will need to load your own OECD PPP data, CPI data, and currency datafeeds, and implement the calculate_ppp_value method to calculate the PPP fair value based on the provided data.

Currency Value

Original paper

Abstract

Trading rules

Python code

Backtrader