The Cost of Access: Understanding Price Efficiency of ADRs

As more investors embrace direct indexing to personalize portfolios, many wealth managers are advising clients on expanding strategies into developed markets outside the U.S. Direct indexing involves buying and selling individual securities, which would typically require trading non-U.S. equities, but American depositary receipts (ADRs) offer an alternative to the administrative and cost overhead of foreign stock markets. 

ADRs are certificates issued by U.S. banks representing shares in non-U.S. companies. They allow U.S. investors to trade international firms’ stock in USD through U.S. exchanges, eliminating the need to access overseas markets directly. Although arbitrage typically keeps ADR prices aligned with their underlying locally listed securities, discrepancies can still occur due to: 

• Market liquidity: Lower trading volume in ADRs can widen bid-ask spreads. 
• Trading hours: Differences between U.S. and foreign market hours may hinder price discovery; arbitrage is most effective when the market for the underlying security is open during U.S. market hours. 
• Fees and costs: Custodian fees, conversion costs and regulation-related expenses can lead to divergence. 
• ADR level: Level-one ADRs trade over the counter with limited regulatory requirements, while level-two and -three ADRs trade on exchanges and follow stricter rules, leading to greater transparency and liquidity.¹

To help wealth managers better understand this pricing divergence, we analyzed the broad universe of developed-market ADRs to investigate how ADR prices compared to their underlying overseas securities in the last four years. This universe is comprised of level-one, -two or -three ADRs² for the constituents of the MSCI EAFE Index. For each ADR and its underlying security in this universe, we first converted the price of the underlying security into USD using the applicable exchange rate³ and adjusted for the ADR ratio.⁴ We then calculated the price divergence as the percentage difference between each ADR and its underlying stock in USD. This divergence could reflect price inefficiencies, market dislocation or structural factors. 

To obtain a comprehensive picture and eliminate idiosyncratic discrepancies such as stock splits, we calculated this difference on the last trading day of each month from January 2021 through March 2025. Our ADR universe contained an average of 715 ADRs per day during the analysis period. Within that universe, we also examined the potential investable universe, proxied by the MSCI EAFE Expanded ADR Index. The average number of EAFE Expanded ADR Index constituents is 360. 

What’s the cost? 

For any investors using ADRs, the premium (or sometimes discount) they would have to pay, relative to the local shares, is important. We plot below the distribution of the price divergence, calculated for all assets in our study periods, across all months.⁵

MSCI monthly data from January 2020 to March 2025.

It’s immediately reassuring that the average price divergence between ADRs and local shares is 0,  both for the broad universe and the MSCI EAFE Expanded ADR Index.⁶ On average, there has not been a persistent premium to pay to use ADRs. There was nevertheless a difference between the underlying local security and the ADR. The standard deviation of the price divergence for the broad universe was 3.26%, while for the MSCI EAFE Expanded ADR Index, it was 1.81%. 

The MSCI EAFE Expanded ADR Index is constructed by first applying a one-month frequency-of-trading screen on the ADR universe. The remaining securities are sorted in descending order of 12-month annual-traded-value ratio (ATVR), and the cumulative coverage of the free-float-adjusted market capitalization is calculated for each security. Securities ranked below the cumulative 95% of free-float-adjusted market-capitalization coverage are excluded from the index.

MSCI monthly data from January 2020 through March 2025.

We next examined the mean and standard deviation of the price divergence between the different ADR levels. We grouped the level-two and -three ADRs together, as they are both exchange-traded. The mean pricing divergence of both our groupings is again close to the zero. The key driver of the price divergence is instead the ATVR screen — the average (cap-weighted) ATVR for is 0.17 for the MSCI Expanded EAFE Index and 0.09 for the broad ADR universe. The higher liquidity of the expanded index helps reduce the size of the price divergence.  

Price-divergence patterns over time and by country 

We have established that, on average, the historical price divergence was 0, but what were the pattern and magnitude of variation for individual ADRs? Was there a positive or negative bias? Did some ADRs always price over their local securities while others priced under? 

Stability over time 

We calculated the average pricing divergence for each month to assess whether there was any long-term trend or seasonality. We also compared the fifth and 95^th percentile of the pricing divergence for the broad ADR index as well as the MSCI EAFE Expanded ADR Index.

MSCI monthly data from January 2020 through March 2025.

The divergence is quite dispersed for the middle quintiles. For example, of the 19.8% of securities in the MSCI Expanded EAFE Index with a divergence in the third quintile in the first month, only 5% still had a divergence in the third quintile for the second month. However, 7.5% of assets for that same index with a divergence in the first quintile in the first month still had a divergence in the first quintile in the second month. 8.3% of assets with a divergence in the fifth quintile in the first month were still in the fifth quintile in the second month. 

Even though the overall distribution of price divergence is normal, if an ADR had a very high or very low price divergence, it was more likely to persist. This persistence was weaker within the MSCI Expanded EAFE ADR Index compared to the broad universe of ADRs, however, and was also weaker for assets with a larger market cap. 

Pricing-divergence variation by country

In response to growing client demand for country-specific exposures, wealth managers may limit ADR selection by region. To understand the historical range of price divergence, we plotted its fifth and 95th percentiles by country in the chart below (along with the total number of ADRs in each grouping). 

MSCI monthly data is from January 2020 through March 2025.

In the chart above, the 50^th percentile fluctuates around 0, with the largest spike of 2% occurring in November 2022. There appear to be no long-term cycles, suggesting that an investor looking to use ADRs for international exposure should not be overconcerned with timing the trades — on average, the pricing divergence is close to 0 over time. 

The distribution of the pricing divergence was consistently larger for the broad ADR universe, however, averaging about +/- 4%, compared to under 2% for constituents of the MSCI EAFE Expanded ADR Index.  

To further understand the stability of price divergence, we examine how it changed month to month. For each month, we grouped the price divergence into quintiles and observed how a given ADR’s divergence compared to the previous period. We aim to understand how much the first month’s divergence influences the following month — to what extent divergences persist. This is averaged over the 62 months in our analysis. We again compare the stability of the broad ADR universe with that of the MSCI EAFE Expanded ADR Index. 

MSCI monthly data from January 2020 through March 2025.

The range of historical price divergence (from the fifth to 95th percentile) varied notably by country — with New Zealand, Australia and Israel showing the greatest variation — but was consistently smaller for the MSCI EAFE Expanded ADR Index during the analysis period. For instance, Australia’s ADRs showed a divergence range of about ±4.75% in the broad universe, compared to only ±2% for those in the index. Median divergence across countries was near 0, with the U.K. highest at 34 basis points (bps) and New Zealand lowest at -60 bps. 

The size of a country’s ADR universe didn’t directly drive pricing divergence. For example, Japan, with an average of 210 ADRs in the broad universe for our period of study, still showed significant variation between the broader universe and the subset of ADRs included in the index.  

Controlling the cost of access

ADRs promise convenience for U.S. investors: They can be traded through their usual brokers in USD on familiar exchanges and often have simpler regulatory burdens. On average, ADRs price on par with the underlying local share. Though the pricing can diverge from par at times, on average, it has typically reverted back to 0 quickly.  

We’ve shown that, across all countries, the average pricing divergence ranged between +/- 4% for the broad ADR universe and +/- 2% for the MSCI EAFE Expanded ADR Index. The pricing divergence is stable across countries, but for the broad ADR universe, the metric experiences a much broader range than when restricted to the constituents of the MSCI EAFE Expanded ADR Index. Our analysis is therefore supportive of the use of ADRs by wealth managers who are seeking to personalize global portfolios through direct indexing. The narrower, and less persistent divergence between the ADR and its underlying security illustrates the value of the frequency-of-trading and ATVR filters used in constructing the MSCI EAFE Expanded ADR Index.