Mathwords logoMathwords

Alabama Paradox — Definition, Formula & Example

The Alabama paradox is what happens when a legislature gains one seat, populations stay the same, and yet some state ends up with fewer seats. It is a known flaw of Hamilton's method of apportionment, also called the largest remainder method.

An apportionment method exhibits the Alabama paradox if there exist fixed populations p1,…,pnp_1, \dots, p_n and a house size hh such that some state ii satisfies ai(h+1)<ai(h)a_i(h+1) < a_i(h). Here ai(h)a_i(h) is the number of seats the method gives state ii when there are hh seats in total.

Key Formula

qi=pid=pi hP,d=Phq_i = \frac{p_i}{d} = \frac{p_i \, h}{P}, \qquad d = \frac{P}{h}
Where:
  • qiq_i = Standard quota of state i, its exact fair share of seats
  • pip_i = Population of state i
  • PP = Total population of all states
  • hh = Total number of seats in the house
  • dd = Standard divisor, the number of people per seat

Worked Example

Problem: Three states have populations A = 6,000, B = 6,000, and C = 2,000, for a total of 14,000. Use Hamilton's method to apportion 10 seats, then 11 seats. Hamilton's method gives each state the whole-number part of its quota, then hands the leftover seats to the largest fractional parts. Does any state lose a seat when the house grows?
Step 1: With 10 seats, compute each quota. Rounding down gives 4 + 4 + 1 = 9 seats, so 1 seat is left over. C has the largest fractional part, about 0.43 compared with 0.29 for A and B, so C gets the extra seat. The apportionment is A = 4, B = 4, C = 2.
qA=qB=6000⋅1014000=307≈4.29,qC=2000⋅1014000=107≈1.43q_A = q_B = \frac{6000 \cdot 10}{14000} = \frac{30}{7} \approx 4.29, \qquad q_C = \frac{2000 \cdot 10}{14000} = \frac{10}{7} \approx 1.43
Step 2: With 11 seats, rounding down again gives 9 seats, so 2 seats are left over. Now A and B have the largest fractional parts, about 0.71 compared with 0.57 for C, so A and B each get one. The apportionment is A = 5, B = 5, C = 1.
qA=qB=6000⋅1114000=337≈4.71,qC=2000⋅1114000=117≈1.57q_A = q_B = \frac{6000 \cdot 11}{14000} = \frac{33}{7} \approx 4.71, \qquad q_C = \frac{2000 \cdot 11}{14000} = \frac{11}{7} \approx 1.57
Step 3: Here is why C lost its seat. Adding one seat raises each quota by that state's share of the population, pi/Pp_i / P, so larger states' quotas rise faster. A's quota rose by 3/7 but C's rose by only 1/7. As a result, A's fractional part climbed from 2/7 to 5/7 and passed C's, which grew from 3/7 to 4/7.
ΔqA=600014000=37,ΔqC=200014000=17\Delta q_A = \frac{6000}{14000} = \frac{3}{7}, \qquad \Delta q_C = \frac{2000}{14000} = \frac{1}{7}
Answer: Yes. State C drops from 2 seats to 1 seat when the house grows from 10 to 11 seats. This is the Alabama paradox.

Why It Matters

The paradox is named after the 1880 census. Calculations showed Alabama would get 8 seats in a 299-seat House but only 7 in a 300-seat House. It helped push Congress away from Hamilton's method. Since 1941 the US House has been apportioned with the Huntington–Hill method, a divisor method that cannot produce the Alabama paradox. The Balinski–Young theorem, studied in college courses on voting and fair division, shows that no apportionment method can avoid every paradox and also always stay within quota.

Common Mistakes

Mistake: Assuming that adding seats can only help or leave every state unchanged.
Correction: Under Hamilton's method, the leftover seats depend on how fractional parts compare. Those comparisons can reorder when the house size changes, so a state can lose a seat. Always recompute the whole apportionment for the new house size.

Related Terms

  • Borda Count — Another fairness method with its own counterintuitive paradoxes
  • Condorcet Winner — Candidate who beats every rival head to head, yet some voting methods fail to elect
  • Fair Division — Apportionment is a fairness problem in dividing seats