How are behaviors of gold and silver exchange-traded funds (ETF) similar and different? To investigate we consider iShares Silver Trust (SLV) versus SPDR Gold Shares (GLD). Using daily returns for SLV, GLD and the S&P 500 Index (SP500) during late April 2006 (limited by SLV) through early September 2020, we find that:
The following table summarizes daily performance statistics for GLD and SLV over the full sample period. Results show that SLV underperforms GLD over time and is more volatile. Correlations with SP500 indicate that SLV is somewhat more equity-like than GLD, perhaps due to wider industrial application.
For perspective, we look at a scatter plot.
The following scatter plot relates contemporaneous daily returns for SLV (vertical axis) versus GLD (horizontal axis) over available sample periods. Based on a linear relationship:
- The R-squared statistic is 0.65, meaning that daily GLD price movements explain 65% of contemporaneous SLV price movements.
- SLV has a slightly negative daily alpha relative to GLD, confirming its underperformance.
- SLV has daily beta relative to GLD of 1.39, confirming its greater volatility.
Is the GLD-SLV relationship consistent over time?
The next chart tracks rolling 252-day R-squared statistic for contemporaneous daily SLV versus GLD returns over the sample period. Explanatory power varies over time, with an R-squared range of 0.53-0.84.
Does either asset lead the other?
The final chart summarizes correlations between SLV daily returns and GLD daily returns over the full sample periods for lead-lag relationships ranging from SLV return leads GLD return by 10 trading days (-10) to GLD return leads SLV return by 10 trading days (10). Day 0 is the contemporaneous relationship. All correlations other than the contemporaneous one appear to be noise. In other words, the two series move together day by day, with neither leading the other.
In summary, evidence from simple tests over the available sample period indicates that GLD and SLV are similar assets, with the former outperforming the latter over time.
Cautions regarding findings include:
- The available sample period is short in terms of number of independent 252-day intervals for the rolling correlation and, especially, in terms of variety of financial market conditions.
- Analyses are in-sample.
See also “Survey of Research on Silver, Platinum and Palladium as Investments” and “Comparing Precious Metals as Safe Havens”.
