Chinese American Depositary Receipts: The Valuation Conundrum
Suuronen, Sampo (2018)
Suuronen, Sampo
2018
Kuvaus
Opinnäytetyö kokotekstinä PDF-muodossa.
Tiivistelmä
Over the past three decades, China has witnessed significant economic development, sparking the interest of international and U.S. investors. However, due to capital and foreign ownership restrictions, individual investors are unable to invest directly into the Chinese companies. For this purpose, investment vehicle called Chinese American Depositary Receipts (“ADRs”) were created. However, by 2012 one third of the Chinese ADRs were revealed to have fraudulent accounting practices and/or little to no actual business operations. This resulted in significant stock price declines and shareholder wealth was lost in tens of billions. Companies that had legitimate business operations were associated with the fraudulent ones leading their stock to be significantly undervalued relative to their western peers.
Thus, this thesis examines relative valuation and its accuracy for the Chinese ADRs. Furthermore, the paper determines whether the accuracy of the analysis was impacted by the 2012 fraud crisis and if the potential valuation discrepancies can be used in an investment strategy. In addition, two asset pricing models are used to test if they are able to capture the stock return variations. First asset pricing model is the Fama and French (1993) three factor model, followed by the Carhart’s (1997) four factor model. The sample data include 446 Chinese companies listed in the U.S. main markets or in the over-the-counter (“OTC”) markets from January 2000 to December 2016. The sample is examined in three different periods, before 2012, after 2012 and the whole period. Furthermore, the analysis is divided into ten industries to sort out if the relative valuation has especially robust or weak observations for a specific industry.
The paper yields comprehensive results through broad comparable analysis using six (6) different valuation ratios. The smallest forecasting errors are attained consistently by using Price to Earnings (P/E) and Price to Book (P/B) ratios. Furthermore, utilities are priced with the highest accuracy within the ten industries and through the different subperiods. Surprisingly, the Price to Sales (P/S) ratio yields largest errors for the comparable analysis, which is in contrast to the previous studies. Overall, the Chinese ADRs price based multiples have continuously declined, while at the same time the valuation errors have declined as well. In addition, the asset pricing models perform remarkably well for the sample data and for the period. The model alphas are twenty-one (21) times out of the twenty-five (25) indistinguishable from zero for the three and four factor models. Thus, the models are able to capture stock return variations from majority of the portfolios. Small to mid-size portfolios that have high book values relative to their market values are not captured by the models. The added fourth factor, momentum, has a marginal impact on the results and is therefore a redundant factor.
Thus, this thesis examines relative valuation and its accuracy for the Chinese ADRs. Furthermore, the paper determines whether the accuracy of the analysis was impacted by the 2012 fraud crisis and if the potential valuation discrepancies can be used in an investment strategy. In addition, two asset pricing models are used to test if they are able to capture the stock return variations. First asset pricing model is the Fama and French (1993) three factor model, followed by the Carhart’s (1997) four factor model. The sample data include 446 Chinese companies listed in the U.S. main markets or in the over-the-counter (“OTC”) markets from January 2000 to December 2016. The sample is examined in three different periods, before 2012, after 2012 and the whole period. Furthermore, the analysis is divided into ten industries to sort out if the relative valuation has especially robust or weak observations for a specific industry.
The paper yields comprehensive results through broad comparable analysis using six (6) different valuation ratios. The smallest forecasting errors are attained consistently by using Price to Earnings (P/E) and Price to Book (P/B) ratios. Furthermore, utilities are priced with the highest accuracy within the ten industries and through the different subperiods. Surprisingly, the Price to Sales (P/S) ratio yields largest errors for the comparable analysis, which is in contrast to the previous studies. Overall, the Chinese ADRs price based multiples have continuously declined, while at the same time the valuation errors have declined as well. In addition, the asset pricing models perform remarkably well for the sample data and for the period. The model alphas are twenty-one (21) times out of the twenty-five (25) indistinguishable from zero for the three and four factor models. Thus, the models are able to capture stock return variations from majority of the portfolios. Small to mid-size portfolios that have high book values relative to their market values are not captured by the models. The added fourth factor, momentum, has a marginal impact on the results and is therefore a redundant factor.