Food Stocks Are Crashing — Is This Finally the Bottom?
Examining Tyson & General Mills: what 52-week lows do (and don’t) tell you about when to buy.
Some investors are buying because “food isn’t a fad.” Others are waiting. I ran the data.
Here is what 52-week lows actually mean in a down-trend — and the conditions that must show up before buying makes sense. The rule is not “buy lows” but “buy only what fits your portfolio.”
Motivation. This study is motivated by three objectives:
Current market context. Several large, mature food companies have recently traded at or near fresh 52‑week lows, including TSN and GIS which hit new lows today. Are these new 52-week lows a sign that it is time to buy?
Signal design problem. Many stocks do not make a single clean bottom; they may produce clusters of consecutive 52‑week lows while continuing to drift downward. We therefore test alternative cool‑down rules that attempt to create more meaningful, non‑redundant entry signals.
Trend regime risk. A 52‑week low behaves differently depending on whether the stock is in a temporary dislocation (mean‑reverting regime) or in a multi‑year structural drawdown. We explicitly examine how a long down‑trend regime degrades the effectiveness of a 52‑week‑low entry strategy.
Background — Food stocks at fresh lows
At the time of writing, multiple widely‑followed food manufacturers and protein processors — including names such as TSN (Tyson Foods), GIS (General Mills), CAG (Conagra), CPB (Campbell Soup), and KHC (Kraft Heinz) — have traded at or near 52‑week lows. TSN and GIS, the companies studied in this memo, hit new 52-week lows today.
All five of these companies are seasoned with long operating histories, stable brands, and established dividends. Intuition might suggest that such lows present attractive entry points for long‑horizon investors seeking defensive exposure at a discount.
However, price alone does not reveal whether the low represents an exhausted washout or merely one waypoint in a continuing deterioration process.
This study tests entry rules – purchase at 52-week lows regardless of long-term trends.
Methods
A basic “52-week low” rule says: buy when the stock hits the lowest price seen in the past year. The problem is that when a stock is in a steady decline, it can keep setting new 52-week lows day after day. Those repeated lows are not fresh opportunities — they are just the same drop continuing. Counting each of them as a new buy signal would give a false picture of how well the rule works.
The evaluation of whether purchase at 52-week lows can lead to long term profitability should account for different trading rules which restrict purchases during steady downturns. Three rules are considered.
Rule 1 — No cool‑down (naïve baseline). Every 52‑week low counts as a distinct entry, even if the stock is falling in a straight line.
Rule 2 — Fixed calendar cool‑down. After a low signal, we require a minimum number of trading days (e.g. 20 or 63) before another low can register. This suppresses dense clusters but remains agnostic to price behavior.
Rule 3 — Logical (price‑based) cool‑down. After a low signal, no new low is recorded until the price first rebounds sufficiently (e.g. +10%) — meaning the prior down‑cycle is likely “completed” before evaluating a fresh low as a new episode.
This structure allows us to separate (i) the mechanical fact that the price printed a new low, from (ii) the interpretive question of whether the new low represents a new information regime or just repetition inside an ongoing decline.
Results
Where stocks sit in a 5-year time frame
Using adjusted prices over the most recent five years:
TSN is at roughly the 7th percentile
→ lower than about 93% of all closes in five years
→ a deep multi-year low, not just a local dipGIS is at roughly the 1st percentile
→ lower than about 99% of all closes in five years
→ an extreme long-horizon low, effectively a worst-case zone
Neither stock is making a one-off 52-week low — they are sitting in the bottom decile or worse of a five-year distribution. That is the statistical fingerprint of a persistent down-trend regime, not a short-lived mean-reversion setup.
Profits for the three purchase at 52-week low rules
Six month free:
https://bernsteinbook1958.substack.com/cea31403
50 percent off annual membership ($30 total.)
