Bid-ask spread is an important trading cost, which has a great impact on investment decisions. Amuid and Mendelson (1986) (hereafter A&M) propose that assets with larger spreads are held for longer holding periods and assets with larger spreads yield a higher average return. These two propositions are called spread clientele effect hypothesis and liquidity premium hypothesis, respectively. This paper provides an empirical evidence for A&M's propositions via the equity markets in Taiwan. The results show that both propositions are supported in the GreTai Securities Market. However, the spread clientele effect exists only in the TSEC. The inconsistency may be attributed to the quotation behavior and the minimum tick system.