Direct answer: Year-end S&P 500 price targets from major banks have been systematically inaccurate when measured at the start of the year. They tend to cluster near current prices plus a historical average return, miss turning points by wide margins, and are subject to mid-year revision. Academic research shows short-term stock returns are close to unpredictable, while longer-horizon returns (5 to 10 years) have some predictability from valuation ratios.
Forecast Scorecard: Stock Market Predictions
Key Takeaways
- Year-end S&P 500 targets from large bank strategists have tracked actual outcomes with low correlation; the spread across forecasters often understates realized uncertainty.
- Valuation ratios such as CAPE (Shiller P/E) have negative correlations with subsequent 10-year returns, but no predictive power at 1-year horizons.
- Earnings growth and GDP growth forecasts are not reliably correlated with stock returns because valuation changes dominate at most horizons.
- Forecast consensus provides less information than the distribution of forecasts; when forecasters cluster tightly, the realized outcome often falls outside the consensus range.
The Annual Target Record
Each December, major banks and brokerage houses publish year-end S&P 500 targets for the coming year. Retrospective studies show that these targets explain very little of the variance in actual returns. Targets are revised mid-year, often toward recent price action, which improves apparent year-end accuracy in a purely mechanical way. Individual strategists who dramatically diverge from consensus are rarely correct and face career risk for being wrong in isolation.
What Actually Drives Returns
Stock returns have two components: earnings growth and multiple change (P/E expansion or contraction). At 1-year horizons, multiple change tends to dominate and is difficult to predict because it depends on sentiment, liquidity conditions, and macroeconomic surprises. At 10-year horizons, starting valuation has meaningful predictive power because P/E ratios are stationary over long periods.
CAPE and Long-Horizon Predictability
The cyclically adjusted P/E ratio (CAPE or Shiller P/E) divides price by 10-year average real earnings. Research by Robert Shiller and subsequent authors finds CAPE negatively correlates with subsequent 10-year returns: high starting CAPE tends to produce lower long-run returns. However, CAPE has poor 1-year predictive power and has remained elevated by historical standards in US markets for extended periods without correction.
Forecaster Herding and Tail Risk
Professional forecasters face incentives to cluster near consensus, limiting their willingness to forecast extreme outcomes. This herding means that actual tail events (25%+ drawdowns, 30%+ up years) are underrepresented in the distribution of published forecasts. Investors who take professional targets as reflecting the realistic return distribution are implicitly understating tail risk.
Frequently Asked Questions
Why are year-end S&P 500 targets so inaccurate?
Markets are driven by surprises relative to what is already priced in. If consensus expects 10% earnings growth, stocks do not rise 10% unless they actually grow more than expected. Year-end targets are usually anchored near historical average returns applied to current prices, which misses both strong and weak years. The variance of annual returns is also very high: plus or minus 15 percentage points around the expected value is common.
Is the stock market predictable at any horizon?
Short-term returns (days to months) are close to unpredictable in the sense that no simple rule consistently outperforms. Over 5 to 10-year horizons, starting valuation ratios (CAPE, price-to-book, earnings yield) have shown negative correlations with subsequent returns in many studies, suggesting some mean reversion in valuations. This long-horizon predictability is too noisy to time markets but useful for long-run asset allocation decisions.
Should I follow bank S&P 500 targets?
Year-end price targets are best understood as scenario anchors rather than reliable forecasts. Their value is in the underlying assumptions they reflect (earnings growth, interest rate paths, sector views) rather than the specific year-end number. Examining the range across forecasters and the implicit assumptions is more informative than treating any single target as a prediction.