Two methods have been proposed; one by Scholes and Williams (1977) and the other by Dimson (1979) to estimate beta correctly when there is infrequent trading. In the aggregated co-efficient method, Scholes and Williams (1977) and Dimson (1979) suggested a nonsynchronous adjustment that required summation of coefficients pertaining to the lagged, coincident and leading market return variables that provide an unbiased estimate of the stock’s systematic risk. These co-efficients are estimated from multiple regressions of stock returns on lagged, coincident and leading market returns.
The alternative method for solving the problem of infrequent or nonsynchronous trading is suggested by Marsh (1979). In the alternative method betas are estimated on the basis of variable rather than fixed length periods and each period is defined as the time between two adjacent recorded trades. That is why this method is known as ‘trade-to-trade’ method, and does not involve making the adjustment in the systematic risk. The market rate of return represented by an appropriate index and stock’s rate of return are calculated over the same period. These paired observations are used to estimate the value of beta.