Original paper
Abstract
Using an event study approach at the stock level, we examine the effect of North Atlantic hurricanes on U.S. stock returns over the period January 1990 to December 2014. We document a substantial economic impact of hurricanes on the aggregate market: an accumulated loss of 0.522% (6.264% annualized) to 30 days post-landfall. For Industry Portfolios encompassing the entire U.S. stock market, we document several patterns. Many manufacturing-based and consumer-focused Industry Portfolios exhibit consistently negative abnormal returns over the event window from formation of the hurricane to 30 days post-landfall. Although the market tends to incorporate the foreseen loss related to the hurricane before landfall, there is no recovery post-landfall and the losses continue to accumulate. One industry in particular reacts differently, namely Gold. Gold-related stocks experience substantial positive abnormal returns during these extreme weather events, which suggests that this commodity acts as an endogenous safe asset during the market disruption caused by such meteorological phenomenon. We also show that degradation in liquidity of Gold stocks' is of the same magnitude as other industries which fairly rule out a flight to liquidity.
Keywords:Â Event study, extreme weather events, gold, hurricanes, market efficiency, safe assets
Trading rules
- Hold gold-based instruments (be it ETFs, futures, or CFDs) for the months of September and November.
- For the remaining months, keep holdings in cash.
Python code
Backtrader
import backtrader as bt
class SeasonalGoldStrategy(bt.Strategy):
def __init__(self):
self.gold = self.datas[0]
def next(self):
current_month = self.datetime.date(0).month
if current_month in (9, 11): # September and November
if self.getposition(self.gold).size == 0:
self.buy(data=self.gold)
else:
if self.getposition(self.gold).size > 0:
self.sell(data=self.gold)
cerebro = bt.Cerebro()
data = bt.feeds.YourGoldDataFeed(dataname='your_gold_data')
cerebro.adddata(data)
cerebro.addstrategy(SeasonalGoldStrategy)
cerebro.run()Replace YourGoldDataFeed with the appropriate data feed for gold (ETF, futures, or CFD) and your_gold_data with the actual data source.