XLU — XLU | S&P 500 Utilities Sector ETF 13F holdings and portfolio analysis

Directly following the baseline means accepting a concentrated, one-sector utility portfolio with modest absolute return and limited benchmark competitiveness. The baseline artifact says sectorWeights are 100.01% Utilities, with concentration of 40.03% in the top five, 58.8% in the top ten, and 83.45% in the top twenty. The largest disclosed names are NEE at 12.72%, CEG at 8.39%, SO at 7.3%, DUK at 6.93%, and AEP at 4.69%. Performance-wise, the baseline produced 6.54% annualized return, 87.19% total return, beta of 0.60, negative alpha of -0.43, and a -36.56% max drawdown. So the exposure is clear: lower-beta utilities concentration, but still meaningful drawdown risk and persistent underperformance versus SPY.

The recent baseline periods that best show the trade-off are the strong but defensive months in early 2025 and the weaker relative months in spring and summer 2025. Positive evidence includes 2025-01 with optimizedReturn -0.14% versus benchmarkReturn -2.79%, 2025-02 with -1.02% versus -6.48%, and 2025-06 with 4.27% versus 2.69%, showing the portfolio can cushion market weakness and occasionally lead. But the cost of the utility tilt is visible in 2025-04, when the baseline returned 3.45% versus SPY’s 6.28%, and 2025-07, when it fell -2.01% versus SPY’s +2.05%, a -4.07 point excess shortfall. Turnover stayed fairly modest in those months at 0.82 in January, 1.58 in February, 0.99 in April, and 0.81 in July, so the trade-off was mostly exposure-driven rather than caused by extreme implementation churn.

Before accepting the baseline, a user should inspect three things: concentration, drawdown behavior, and implementation burden. Concentration is high with top5 at 40.03% and top10 at 58.8%, so reviewing whether NEE 12.72%, CEG 8.39%, SO 7.3%, and DUK 6.93% are acceptable single-name exposures is essential. Drawdown behavior also matters because despite beta of only 0.60, the backtest still shows a -36.56% max drawdown and 439 recovery days. Finally, implementation is not trivial: the baseline risk notes say the backtest required 3,480 trades, while totalEstimatedCost was 0.4086, so a user should verify whether real-world slippage, taxes, and account constraints could be worse than the estimate.