Backtest Period
Markets Traded
Commodities
Maximum Drawdown
Period of Rebalancing
Monthly
Return (Annual)
Sharpe Ratio
Standard Deviation (Annual)
Original paper
SSRN-id920496.pdf209.7KB
Abstract
Gold shows the characteristics of a zero-beta asset. It has approximately the same mean return as a Treasury Bill and bears no market risk. Silver also bears no market risk but has returns inferior to Treasury Bills. Both gold and silver show evidence of inflation-hedging ability, with the case being much stronger for gold. The prices of both metals are cointegrated with consumer prices, showing additional evidence of hedging ability.
Keywords:Â gold, silver, inflation
Trading rules
- Implement the Fed model for determining market entry and exit points.
- Compare stocks’ forward earnings yield (E/P) to the yield on 10-year government bonds monthly.
- Take a long position in gold under the following conditions:
- The market is undervalued by at least 15% (E/P is higher than bond yield).
- The E/P to bond yield ratio reaches or exceeds 1.15.
- Adjust the portfolio's composition every month.
Python code
Backtrader
import backtrader as bt
import datetime
class FedModel(bt.Strategy):
params = (
('rebalance_interval', 21),
)
def __init__(self):
self.earnings_yield = self.data0.close / self.data0.earning
self.bond_yield = self.data1.close
def next(self):
if len(self.data0) % self.params.rebalance_interval == 0:
if self.earnings_yield > self.bond_yield:
ratio = self.earnings_yield / self.bond_yield
if ratio >= 1.15:
self.order_target_percent(self.data0, target=1.0)
else:
self.order_target_percent(self.data0, target=0.0)
if __name__ == '__main__':
cerebro = bt.Cerebro()
stock_data = bt.feeds.YahooFinanceData(
dataname='SPY',
fromdate=datetime.datetime(2000, 1, 1),
todate=datetime.datetime(2021, 1, 1),
plot=False
)
cerebro.adddata(stock_data)
bond_data = bt.feeds.YahooFinanceData(
dataname='IEF',
fromdate=datetime.datetime(2000, 1, 1),
todate=datetime.datetime(2021, 1, 1),
plot=False
)
cerebro.adddata(bond_data)
cerebro.addstrategy(FedModel)
initial_cash = 100000
cerebro.broker.setcash(initial_cash)
cerebro.run()
final_value = cerebro.broker.getvalue()
print(f'Initial Cash: {initial_cash}, Final Value: {final_value}')Make sure to replace the placeholder stock and bond data sources with your desired data sources containing earnings and bond yield information. This code sample assumes earnings information is available in the earning line of the stock data feed.