Original paper
Abstract
Theory has linked price momentum with price reversals (Barberis, Shleifer, and Vishny (1998), Daniel, Hirshleifer, and Subrahmanyam (1998), and Hong and Stein (1999)). The models generally rely on behavioral descriptions of irrational investors who push prices beyond their fundamental value thus leading to an inevitable price reversal. While significant empirical evidence has shown the presence of momentum in global equity returns, there have been no large-scale global studies of the subsequent long-term price reversals. We study returns from twenty-three developed countries categorized into the regions of North America, Europe, Japan, and Asia, over 1993-2014 and find evidence supporting the global presence of long-term price reversal. The positive return differential between loser stocks over the past three years and winner stocks over the past three years is economically and statistically significant. Results from independent double sorts and from Fama-MacBeth regressions show that long-term reversals remains significant after controlling for size, book-to-market equity, and momentum.
Keywords:Â Return predictability, overreaction, long-term reversals, market efficiency, cross-section of returns, international asset pricing
Trading rules
- Investment universe: 26 countries from emerging markets.
- Evaluate and rank on a monthly basis using previous five-year returns.
- Top quartile: Late-stage outperformers (LW).
- Last quartile: Late-stage underperformers (LL).
- Within each subgroup, classify nations by their half-year (6-month) momentum.
- Long positions: Top 50% of best 6-month momentum in LL group.
- Short positions: Bottom 50% of worst 6-month momentum in LW group.
- Equal weighting for each country.
- Hold positions for 6 months.
- Monthly portfolio rebalance (1/6th of the portfolio each month).
Python code
Backtrader
import backtrader as bt
class EmergingMarketStrategy(bt.Strategy):
def __init__(self):
self.month_counter = 0
self.rebalance_period = 6
def next(self):
# Check if it's time to rebalance the portfolio
self.month_counter += 1
if self.month_counter % self.rebalance_period != 0:
return
# Calculate 60-month performance
past_60_month_perf = [d.close[0] / d.close[-60] for d in self.datas]
# Calculate 6-month momentum
six_month_momentum = [d.close[0] / d.close[-6] for d in self.datas]
# Sort countries by 60-month performance and create late-stage groups
sorted_indices = sorted(range(len(past_60_month_perf)), key=lambda k: past_60_month_perf[k])
lw_group = sorted_indices[:len(self.datas) // 4]
ll_group = sorted_indices[-len(self.datas) // 4:]
# Rank countries in each subgroup by 6-month momentum
lw_ranked = sorted(lw_group, key=lambda k: six_month_momentum[k], reverse=True)
ll_ranked = sorted(ll_group, key=lambda k: six_month_momentum[k], reverse=True)
# Set target positions
for i, data in enumerate(self.datas):
target_weight = 0
# Set long positions for top 50% of best 6-month momentum in LL group
if i in ll_ranked[:len(ll_group) // 2]:
target_weight = 1 / (len(ll_group) // 2)
# Set short positions for bottom 50% of worst 6-month momentum in LW group
if i in lw_ranked[-len(lw_group) // 2:]:
target_weight = -1 / (len(lw_group) // 2)
# Rebalance the portfolio
self.order_target_percent(data, target_weight)Remember to add your data feeds and set up a Cerebro instance to execute this strategy in Backtrader. The code provided above assumes that you have already prepared the data feeds for the 26 emerging market countries.