Prop 13 polls missed final margin by eight points

California voters have repeatedly rejected or altered legislative tax initiatives in ways that surveys failed to predict. Jon Coupal, president of the Howard Jarvis Taxpayers Association, argues that polling data for fiscal measures consistently undercounts support for taxpayer interests.

The most recent discrepancy occurred with Proposition 5, a legislative effort to lower the vote threshold for certain local bonds from two-thirds to 55%. A Public Policy Institute of California poll released just before the election found likely voters split evenly, with 49 percent supporting the measure and 50 percent opposing it. The final tally showed the proposition failing by a margin of over 55 percent to 45 percent, a gap of approximately five percentage points from the projected split.

A similar pattern emerged in 1996 with Proposition 218, a tax reform measure also sponsored by the HJTA. In the weeks leading up to the November 5 election, surveys indicated the measure was failing badly. The final Field Poll released prior to the vote projected only 36 percent support among likely voters. On election day, the measure passed with 56.55 percent of the vote, a swing of more than 20 points from the last major survey.

The earliest case study in this trend dates to 1978 and Proposition 13. After Howard Jarvis collected over 1.2 million signatures in early 1978, initial polls showed a divided electorate. Support stood at approximately 25 percent, opposition at 27 percent, and nearly half of all voters remained undecided. As spring arrived and homeowners received property tax bills, support rose sharply. Even the last Field Poll taken before the June 6 primary underestimated the final outcome, predicting passage by 57 to 60 percent. The actual result was 64.8 percent in favor to 35.2 percent against.

Coupal notes that while respected pollsters like J. Ann Selzer of the Des Moines Register have faced high-profile misses, such as the 2024 Iowa survey, the issue with state ballot measures is systemic. He contends that in a progressive state like California, polling on conservative fiscal causes faces specific challenges. Voters who do not identify as taxpayer activists often fail to respond to surveys or prefer to conceal their fiscal conservatism.

The reliability of polling has come under broader scrutiny in recent election cycles. The Wedgewood Polls scandal highlighted how fabricated data can skew aggregated results. That firm admitted its data was entirely fabricated, yet its results were included in major aggregators like The New York Times, FiftyPlusOne, and VoteHub. The National News Desk reported that Wedgewood had no history, formal credentials, or real data behind most of its releases.

Coupal does not argue that polling is useless. He describes it as indispensable for measuring voter sentiment and testing arguments. However, he suggests the discrepancies reveal a behavioral gap. Voters may feel comfortable expressing moderate or progressive views in private surveys but vote for tax relief when the ballot is in hand.

The Howard Jarvis Taxpayers Association is currently active in California politics, monitoring upcoming fiscal measures. The organization plans to use historical polling data to inform its messaging for the next cycle. Coupal stated that the only poll that counts is the one conducted on election day.