Dividend Adjustment Anomaly in Alpaca's Data

Alpaca Market Data
1

Hi Alpaca community,

I’ve noticed a possible discrepancy with the dividend-adjusted prices from Alpaca for the stock $PFE (Pfizer). It appears the dividend amount might be subtracted twice.

Context:
PFE has a recent dividend ex-date of 2023-07-26 with a value of $0.41. For the date 2023-07-25, the expected adjusted close should be $37.21 (raw close) minus the dividend of $0.41, resulting in $36.80. However, Alpaca’s data indicates an adjusted close of $36.38. This is essentially subtracting the dividend twice. This pattern appears to be consistent for dates prior to 2023-07-25.

Raw Data from Alpaca:

symbol | timestamp                 | open  | high  | low    | close | volume      | trade_count | vwap      
-------|---------------------------|-------|-------|--------|-------|-------------|-------------|-----------
PFE    | 2023-07-24 04:00:00+00:00 | 37.41 | 37.80 | 37.370 | 37.51 | 19,414,710  | 123,538     | 37.596311 
PFE    | 2023-07-25 04:00:00+00:00 | 37.40 | 37.72 | 37.045 | 37.06 | 21,862,764  | 120,078     | 37.284072
PFE    | 2023-07-26 04:00:00+00:00 | 36.87 | 37.345| 36.760 | 37.21 | 20,983,650  | 123,832     | 37.143291
PFE    | 2023-07-27 04:00:00+00:00 | 36.85 | 37.01 | 36.18  | 36.20 | 25,346,304  | 146,050     | 36.413980

Adjusted Data from Alpaca:

symbol | timestamp                 | open  | high  | low    | close | volume      | trade_count | vwap      
-------|---------------------------|-------|-------|--------|-------|-------------|-------------|-----------
PFE    | 2023-07-24 04:00:00+00:00 | 36.58 | 36.96 | 36.54  | 36.67 | 19,414,710  | 123,538     | 36.760000 
PFE    | 2023-07-25 04:00:00+00:00 | 36.57 | 36.88 | 36.22  | 36.23 | 21,862,764  | 120,078     | 36.450000
PFE    | 2023-07-26 04:00:00+00:00 | 36.05 | 36.51 | 35.94  | 36.38 | 20,983,650  | 123,832     | 36.320000
PFE    | 2023-07-27 04:00:00+00:00 | 36.85 | 37.01 | 36.18  | 36.20 | 25,346,304  | 146,050     | 36.413980

Data from Yahoo Finance:

Date       | Open       | High      | Low       | Close     | Adj Close  | Volume    
-----------|------------|-----------|-----------|-----------|------------|-------------
2023-07-24 | 37.410000  | 37.799999 | 37.369999 | 37.509998 | 37.096695  | 19,414,700
2023-07-25 | 37.400002  | 37.720001 | 37.049999 | 37.060001 | 36.651657  | 21,771,000
2023-07-26 | 36.869999  | 37.349998 | 36.759998 | 37.209999 | 36.799999  | 20,983,700
2023-07-27 | 36.849998  | 37.009998 | 36.180000 | 36.200001 | 36.200001  | 25,346,300

Has anyone else observed similar discrepancies? Insights or clarifications would be much appreciated!

Thank you!

Edit: the data was fetched using a paper trade account.
Code used:

  request_params = StockBarsRequest(
      symbol_or_symbols=symbol,
      timeframe=timeframe,
      start=datetime.fromisoformat(self._validate_datetime(start)),
      end=datetime.fromisoformat(self._validate_datetime(end)),
      adjustment=adjustment,
  )
  df = self.client.get_stock_bars(request_params).df