The strategy buys NYSE, AMEX, and NASDAQ stocks on analyst days, identified via 8-K filings or press releases, holding for 20 days with value-weighted portfolios.
ASSET CLASS: stocks | REGION: United States | FREQUENCY:
Daily | MARKET: equities | KEYWORD: Analyst
I. STRATEGY IN A NUTSHELL
Trades NYSE, AMEX, and NASDAQ stocks around analyst days, going long on the stock on the event day and holding for 20 days. Portfolios are value-weighted to capture potential abnormal returns from market attention.
II. ECONOMIC RATIONALE
Analyst days convey predominantly positive firm information, leading to market underreaction. Persistent abnormal returns and improved firm metrics post-event create exploitable opportunities, as investors gradually adjust to credible disclosures.
III. SOURCE PAPER
Analyst Days, Stock Prices, and Firm Performance [Click to Open PDF]
Di (Andrew) Wu — University of Michigan, Stephen M. Ross School of Business; Amir Yaron — University of Pennsylvania – Wharton School of Business; Bank of Israel; National Bureau of Economic Research (NBER).
<Abstract>
We construct a comprehensive dataset of 3,890 analyst days, which are firm-hosted gatherings where information is disclosed to equity analysts and institutional investors. We demonstrate that firms holding these events have significantly higher abnormal returns after these events, despite the Regulation Fair Disclosure requirement that such information be simultaneously disclosed to the public. A buy-and-hold strategy that holds these stocks for 20 days earns a market-adjusted return of 1.6%, and a similar calendar-time portfolio has a one-month, four-factor alpha of 1.8%. We find no evidence of mean reversion or change in risk exposure after analyst days, and abnormal returns remain significantly positive for up to six months. We classify analyst days into four major types—product announcement, review of results, discussion of strategy, and technology and markets—according to the textual content of their announcements, and we show that product- and market-related analyst days earn significantly higher returns than events reviewing past financial results. Finally, firms holding analyst days have significantly higher revenue growth, earnings per share, and dividend yields up to two years after these events. Analyst coverage, earning estimates, and price targets also increase, and these estimates have lower dispersion. Our results thus suggest that firms use analyst days to convey positive incremental information that has not been incorporated in their stock prices, and market participants significantly underreact to this information
IV. BACKTEST PERFORMANCE
| Annualised Return | 18.3% |
| Volatility | 24.82% |
| Beta | 0.355 |
| Sharpe Ratio | 0.59 |
| Sortino Ratio | -0.005 |
| Maximum Drawdown | N/A |
| Win Rate | 59% |
V. FULL PYTHON CODE
from AlgorithmImports import *
#endregion
class AnalystDays(QCAlgorithm):
def Initialize(self):
self.SetStartDate(2010, 1, 1)
self.SetCash(100_000)
self.equally_weighted_flag: bool = False # False - VW; True - EW
self.analyst_days: Dict[str, List[datetime.date]] = {}
self.selected: List[Fundamental] = []
self.tickers: List[str] = []
self.market_cap:dict[Symbol, float] = {}
market: Symbol = self.AddEquity('SPY', Resolution.Daily).Symbol
self.spy_future_symbol: Symbol = market
self.opened_long_positions_period: Dict[Symbol, int] = {} # opened long stock postions with holding period
self.holding_period: int = 20 # n days of holding period
self.leverage: int = 10
csv_string_file: str = self.Download('data.quantpedia.com/backtesting_data/economic/analyst_days.csv')
lines: List[str] = csv_string_file.split('\r\n')
header: List[str] = lines[0].split(';')
for ticker in header[1:]: # skip first column == date
self.tickers.append(ticker)
for line in lines[1:]: # skip header line
if line == '':
continue
line: List[str] = line.split(';')
# convert string to date
str_date: str = line[0]
date: datetime.date = datetime.strptime(str_date, '%d.%m.%Y').date()
for index in range(1, len(line)): # skip date as a first value of the row
# retrive ticker from list created based on csv header
ticker: str = self.tickers[index - 1]
if ticker not in self.analyst_days:
self.analyst_days[ticker] = []
# if analyst_day_flag == 1, then this company had analyst days
analyst_day_flag: str = line[index]
if analyst_day_flag == '1':
# company had analyst days
self.analyst_days[ticker].append(date)
self.selection_flag: bool = False
self.UniverseSettings.Leverage = self.leverage
self.UniverseSettings.Resolution = Resolution.Minute
self.AddUniverse(self.FundamentalSelectionFunction)
self.settings.daily_precise_end_time = False
self.Settings.MinimumOrderMarginPortfolioPercentage = 0.
self.Schedule.On(self.DateRules.MonthStart(market), self.TimeRules.BeforeMarketClose(market, 0), self.Selection)
self.Schedule.On(self.DateRules.EveryDay(market), self.TimeRules.BeforeMarketClose(market, 1), self.ManageTrade)
def OnSecuritiesChanged(self, changes: SecurityChanges) -> None:
for security in changes.AddedSecurities:
security.SetFeeModel(CustomFeeModel())
def FundamentalSelectionFunction(self, fundamental: List[Fundamental]) -> List[Symbol]:
# rebalance monthly
if not self.selection_flag:
return Universe.Unchanged
self.selection_flag = False
# select S&P500 stocks from fundamental based on ticker
self.selected = [
x for x in fundamental
if x.MarketCap != 0
and x.Symbol.Value in self.analyst_days
]
if self.equally_weighted_flag:
self.selected_values = list(map(lambda stock: stock.Symbol, self.selected))
return self.selected_values
else:
self.market_cap = { stock.Symbol : stock.MarketCap for stock in self.selected }
return list(self.market_cap.keys())
def ManageTrade(self):
# liquidate opened symbols
symbols_to_remove: List[Symbol] = []
rebalance_flag: bool = False
for symbol in self.opened_long_positions_period:
holding_period_remaining: int = self.opened_long_positions_period[symbol]
if holding_period_remaining == 0:
# remove stock from holdings
symbols_to_remove.append(symbol)
else:
# decrement remaining holding period
self.opened_long_positions_period[symbol] -= 1
for symbol in symbols_to_remove:
self.Liquidate(symbol)
del self.opened_long_positions_period[symbol]
rebalance_flag = True # rebalance is required
today: datetime.date = self.Time.date()
long: List[Symbol] = [] # storing symbols of stocks of those companies that had analyst day
# check if selected companies had analyst day
for stock in self.selected:
symbol: Symbol = stock.Symbol
ticker: str = symbol.Value
# make sure selected stock has analyst days data
if ticker not in self.analyst_days:
continue
if today in self.analyst_days[ticker]:
# store stock in opened postions selection
self.opened_long_positions_period[symbol] = self.holding_period
rebalance_flag = True # rebalance is required
# rebalance is required
if rebalance_flag:
# rebalance whole active selection
if self.equally_weighted_flag:
n: int = len(self.opened_long_positions_period)
if n != 0:
# rebalance whole trade selection
for symbol in self.opened_long_positions_period:
price = self.Securities[symbol].Price
if price != 0:
self.SetHoldings(symbol, 1/n)
else:
total_cap: float = sum([self.market_cap[x] for x in self.opened_long_positions_period])
for symbol in self.opened_long_positions_period:
if symbol in self.market_cap:
self.SetHoldings(symbol, self.market_cap[symbol] / total_cap)
def Selection(self) -> None:
self.selection_flag = True
# Custom fee model.
class CustomFeeModel(FeeModel):
def GetOrderFee(self, parameters: OrderFeeParameters) -> OrderFee:
fee: float = parameters.Security.Price * parameters.Order.AbsoluteQuantity * 0.00005
return OrderFee(CashAmount(fee, "USD"))VI. Backtest Performance
